biology join - Coweeta LTER - University of Georgia

Articles 

Long-Term Ecological Research 

in a Human-Dominated World 

G. PhiliP RobeRtson, scott l. collins, DaviD R. FosteR, nicholas bRokaw, huGh w. Ducklow, 

teD l. GRaGson, coRinna GRies, stePhen k. hamilton, a. DaviD mcGuiRe, John c. mooRe, 

emily h. stanley, RobeRt b. waiDe, anD maRk w. williams 

The US Long Term Ecological Research (LTER) Network enters its fourth decade with a distinguished record of achievement in ecological science. 

The value of long-term observations and experiments has never been more important for testing ecological theory and for addressing 

today’s most difficult environmental challenges. The network’s potential for tackling emergent continent-scale questions such as cryosphere loss 

and landscape change is becoming increasingly apparent on the basis of a capacity to combine long-term observations and experimental results 

with new observatory-based measurements, to study socioecological systems, to advance the use of environmental cyberinfrastructure, to promote 

environmental science literacy, and to engage with decisionmakers in framing major directions for research. The long-term context of network 

science, from understanding the past to forecasting the future, provides a valuable perspective for helping to solve many of the crucial environmental 

problems facing society today. 

Keywords: coupled natural–human systems, cyberinfrastructure, environmental observatories, environmental education, socioecological systems 

The US Long Term Ecological Research (LTER) Network 

was started in 1980 to provide sites for ecologists to 

address questions that require long periods of study in 

order to be resolved. Hypothesis-driven research conducted 

over an extended period is a hallmark of the LTER Network 

today and the foundation of its scientific contributions 

(see Callahan 1984, Franklin et al. 1990, Hobbie et al. 2003). 

At 26 sites (figure 1), ecologists conduct synthetic and crosssite 

research that builds on site-based data, experiments, and 

models across diverse regions. 

Historically, studies at LTER Network sites have addressed 

long-term questions not easily addressed in short-term 

funding cycles: How do populations change in response to 

long-term environmental forcings such as landscape and 

climate change? How do these changes affect biodiversity 

and trophic interactions and, in turn, primary productivity, 

element cycles, and other ecosystem processes? What are 

the lags in ecosystem responses to and the legacies of past 

human and natural disturbances? What precipitates ecological 

tipping points, and are such changes predictable? 

These questions are broadly applicable to all ecosystems, 

and as the value of addressing them became clear during 

the first 20 years of the program, the LTER Network grew to 

include additional biomes and ecosystem types, to encompass 

broader regional scales of inquiry, and to incorporate 

human-dominated systems in its research. Today’s network 

of forest, grassland, desert, freshwater, coastal, and other 

ecosystems spans a broad geographic range of both climate 

and human impact. Climates within the network range from 

polar to tropical and from maritime to continental, with 

correspondingly diverse biotic assemblages. Human influences 

among sites range from no intentional disturbance to 

intensive management for agricultural, rangeland, forestry, 

and urban outcomes. 

Creation of the network thus substantially altered the 

range of research sites used by US ecologists. Although 

sites in national parks, national forests, agricultural experiment 

stations, and biological field stations have historically 

provided a rich context for asking long-term ecological 

questions, most questions have been addressed in an ad hoc 

manner. The LTER Network provides an explicit opportunity 

to document ecological changes and to simultaneously 

address long-term questions across a broad array of ecosystems. 

Documenting these changes provides an important 

opportunity to ask mechanistic questions about the causes 

and consequences of change, an additional hallmark of 

LTER: place-based long-term experimentation (Knapp et al. 

2012 [in this issue]). 

The ability to link results at one site to findings at another 

allows the exploration of questions at broader geographic 

BioScience 62: 342–353. ISSN 0006-3568, electronic ISSN 1525-3244. © 2012 by American Institute of Biological Sciences. All rights reserved. Request 

permission to photocopy or reproduce article content at the University of California Press’s Rights and Permissions Web site at www.ucpressjournals.com/ 

reprintinfo.asp. doi:10.1525/bio.2012.62.4.6 

342 BioScience • April 2012 / Vol. 62 No. 4 www.biosciencemag.org

SBC 

CCE 

BNZ 

0 10 km 

ARC 

MCR 

AND 

CAP 

NWT 

JRN 

SEV 

SGS 

CDR 

KNZ 

NTL 

KBS 

CWT 

BES 

FCE 

scales in order to explore both regional patterns and controls, 

as well as to explore the degree of connectivity among 

disparate parts of regional and continental landscapes (sensu 

Peters et al. 2008). By the end of the network’s third decade, 

the number of cross-site studies had ballooned (figure 2; 

Johnson et al. 2010). The following five examples demonstrate 

this development: 

(1) In the Long-Term Intersite Decomposition Experiment, 

the decomposition rates of leaf litter and roots were 

measured in a 10-year reciprocal-transplant experiment 

among 21 long-term sites in seven biomes. The results 

showed that relatively simple models can predict decomposition 

rates on the basis of litter quality and regional 

climate (Gholz et al. 2000) but that the rate of nitrogen 

release from leaf litter is largely independent of climate 

(Parton et al. 2007). Nitrogen release instead depends on 

the initial tissue nitrogen concentrations and mass, except 

in arid environments in which exposure to large amounts 

of ultraviolet radiation overrides the influence of nitrogen 

content. 

(2) In the Lotic Intersite Nitrogen Experiment (LINX), a set 

of comparative studies of nitrogen dynamics were conducted 

HBR 

PIE 

HFR 

GCE 

Articles 

in 70 headwater streams from 

across North America on the basis 

of collaborations that grew out 

of the LTER Network and later 

included other sites. Using coordinated 

whole-stream nitrogen-15 

isotope-addition experiments, 

LINX studies demonstrated the 

importance of headwater streams 

for maintaining downstream 

water quality (Peterson et al. 

2001, Helton et al. 2011), quantified 

their sensitivity to excess 

nitrate loading (Mulholland et al. 

2008), and clarified their role as 

sources of the potent greenhouse 

gas nitrous oxide (Beaulieu et al. 

2011). LINX research has now 

expanded to include nitrogen 

cycling in large rivers and wetlands 

in addition to streams. 

(3) In a study of very longterm 

records of lake ice initiated 

at LTER Network sites and then 

expanded to other Northern 

Hemisphere locations, Magnuson 

and colleagues (2000) exposed a 

trend of shorter and more variable 

durations of ice cover over 

the past century. These trends 

of reduced ice cover offered an 

integrated, long-term indication 

of a warming climate over broad 

geographic regions. 

(4) A working group convened at the National Center 

for Ecological Analysis and Synthesis to examine the relationship 

between plant productivity and diversity at LTER 

Network and other sites (Waide et al. 1999) led to a 

meta-analysis of over 170 studies of species richness and 

productivity (Mittelbach et al. 2001), which changed the 

prevailing view that species richness peaks at intermediate 

productivities. Building on this result, a more recent multisite 

international experiment that included LTER Network 

sites showed that species richness per se cannot be used to 

predict productivity, except in reconstructed communities 

(Adler et al. 2011). 

(5) An analysis of more than 900 species responses from 

34 nitrogen-fertilization experiments across long-term sites 

(Cleland et al. 2008) showed that trait-neutral and traitbased 

mechanisms operated simultaneously to influence 

diversity loss as net primary production increased with 

fertilization (Suding et al. 2005). Although soil-buffering 

capacity modulated some responses (Clark et al. 2007), 

low abundance was consistently an important driver of 

species loss across ecosystems, and both trait-based and 

species-specific responses were also evident (Pennings et al. 

www.biosciencemag.org April 2012 / Vol. 62 No. 4 • BioScience 343 

VCR 

PAL 

0 2000 km 

MCM 

Figure 1. Map of the 26 Long Term Ecological Research Network sites on an ecoregion 

map from Olson and colleagues (2001). Descriptions of the sites can be found at 

www.lternet.edu. Abbreviation: km, kilometers. For the site-name abbreviations, 

see Knapp and colleagues’ (2012) table 1 (in this issue, on p. 379). 

LUQ

Articles 

1992 

BNZ 

AND 

HBR 

2003 

AND 

PAL 

SGS 

FCE 

HFR 

ARC 

VCR LUO 

VCR 

CWT 

NTL 

KNZ 

HFR 

LUQ 

CDR 

JRN 

HBR 

PAL 

CWT 

ARC 

SEV 

JRN 

NTL 

NWT 

MCM 

SEV 

NWT 

KBS 

KNZ 

KBS 

SGS 

GCE 

BNZ 

PIE 

CAP 

BES 

Figure 2. Evolution of cross-site research in the Long Term 

Ecological Research Network in the decade prior to 2003, 

quantified by joint intersite publications (lines between 

sites), recalculated from the data in Johnson and colleagues 

(2010). For the site-name abbreviations, see Knapp and 

colleagues’ (2012) table 1 (in this issue, on p. 379). 

2005, Cleland et al. 2011). The results from these syntheses 

demonstrated that rarity, species identity, functional traits, 

and physical environment all contributed to changes in 

plant-community composition in response to soil nitrogen 

availability. 

Clearly, cross-site syntheses add substantial value to 

highly mechanistic site-based analyses and provide the 

opportunity to develop and test ecological theories that can 

lead to broader ecological knowledge. 

The new emergence of environmental observatories— 

including the National Ecological Observatory Network 

(NEON; Keller et al. 2008); the Oceans Observatory 

Initiative (Isern and Clark 2003); the Global Lake Ecological 

Observatory Network (Kratz et al. 2006); and others, 

including open-source networks such as NutNet (Adler 

et al. 2011)—underscores a growing scientific appreciation 

for the power of cross-site observations (Carpenter 2008, 

Robertson 2008). The LTER Network was not designed 

to be a single integrated observatory in which each 

site employs standardized instrumentation and capacity, 

as does NEON, nor are network sites optimally located to 

capture environmental trends at continental scales. Rather, 

the LTER Network was designed to provide key places for 

long-term, biome-specific observations and experimentation, 

where investigations can reveal the underlying causes 

and future consequences of patterns detected by distributed 

observatories that include LTER Network sites. And 

by increasingly engaging with diverse stakeholders—land 

managers, policymakers, and decisionmakers at all levels— 

LTER Network scientists can ensure that their inquiries 

are relevant to addressing societal concerns (Driscoll et al. 

2012 [in this issue]). 

LTER Network sites also play a unique role in science 

education at all levels. The sites are closely associated with 

institutions of higher learning, and graduate, undergraduate, 

and postdoctoral scholarship at these sites serves to 

advance ecology, as well as to introduce undergraduates 

to field research and graduate and postdoctoral scientists 

to the value of distributed research networks (e.g., Kane 

et al. 2008). Researchers at the sites are also actively engaged 

with local communities and with state and national agencies 

and boards, through which they can advance a range 

of informal education approaches, including professional 

advancement and environmental training for community 

leaders. The resulting synergies have included contributions 

to kindergarten through twelfth grade (K–12) science education 

through professional-development activities for science 

teachers (e.g., McKnight 2010), broadening and strengthening 

of local and state science curricula, and pedagogical contributions 

to the development of an environmental literacy 

movement, as is described below. These diverse educational 

efforts are increasingly providing new avenues for improving 

the quality and relevance of LTER science (Driscoll et al. 

2012). 

As the LTER Network enters its fourth decade, it is poised 

to contribute substantively to helping society respond to 

the ever-growing challenges of environmental sustainability, 

including climate-change mitigation and adaptation. How 

will the network help meet these challenges? In this article, 

we describe the vision of that future that emerged through 

a multiyear process of engagement across and beyond the 

entire LTER community, its National Science Foundation 

(NSF) associates, and colleagues from many other programs, 

observatories, and agencies. 

We first lay out the place of LTER in a world increasingly 

subject to human influence. We describe a common 

framework for addressing important questions and examples 

of three overarching research themes that can best be 

addressed with a network of sites: landscape vulnerability 

and resilience to global change, cryosphere loss, and coastalzone 

climate change. We then describe LTER contributions 

toward building environmental science literacy among K–12 

students, undergraduates, and those who work with broader 

audiences, as well as the cyberinfrastructure demands of 

long-term ecological science and the network’s approach to 

meeting these needs. 


LTER in a human-dominated world 

Thirty years of LTER Network research have yielded valuable 

knowledge about ecosystem change in response to 

both natural and human influences. Changes ranging from 

climate alteration to species introductions and to land- and 

water-use decisions have far-reaching impacts on ecosystem 

function, community structure, and population and evolutionary 

dynamics, which in turn strongly affect the critical 

ecosystem services on which we all depend. Ecological 

research seeks to test theory and to provide the empirical 

knowledge needed to forecast change and to devise effective 

management and policy responses. And theory and 

knowledge increasingly cross the boundary between natural 

and human systems, effectively linking science with policy 

(Liu et al. 2008, Driscoll et al. 2012). 

A framework for exploring coupled natural–human systems over 

the long term. Several recent studies have shown how couplings 

between human and natural systems exhibit nonlinear 

dynamics across space, time, and organizational scales 

and have revealed complexities that cannot be disentangled 

by ecological or social research alone. The importance of 

understanding these dynamics cannot be underestimated: 

Without understanding the couplings between natural and 

human systems, workable policy solutions to some of the 

most recalcitrant environmental problems of today, which 

range from degraded water quality to biodiversity loss to 

climate-change vulnerability, will remain difficult to design 

and even more difficult to achieve. 

At individual LTER sites, research on the couplings 

between natural and human systems has a rich history, 

ranging from inherently coupled working lands (row crop 

systems, timber plantations, grazing lands, coastal fisheries) 

to urban and exurban areas and sites in which direct 

human impact ceased decades ago but in which its legacies 

continue to condition ecosystem patterns and processes. 

In fact, no LTER Network site is uncoupled from human 

influence: The network’s most remote Arctic and Antarctic 

sites are also affected by human decisions and behaviors, 

although humans are far away and their effects mostly 

unintentional. As a whole, the LTER Network provides a 

broad range of sites with differing intensities of human 

influence, degrees of intent, and levels of connectedness 

(Peters et al. 2008). 

The integration of social and ecological research within 

the context of LTER Network sites and scientists (Redman 

et al. 2004), coupled with a rich ecological information base 

for the sites, is a promising new research frontier for LTER. 

The network has responded to this challenge by adopting as 

an organizing framework a common model that provides 

a standardized terminology and generalized structure to 

facilitate investigations of a wide variety of questions. The 

press–pulse dynamics (PPD) model (Collins et al. 2011) 

provides a comparative framework to integrate the biophysical 

and social sciences through an understanding of how 

human decisionmaking and behaviors interact with natural 

Articles 

processes to affect the structure, function, and dynamics of 

ecosystems and the services they provide to people. 

The PPD model (see Collins et al. 2011, but cf. figure 3) 

is iterative, with linkages and feedbacks between biophysical 

and social domains (in figure 3, environmental and human 

systems, respectively). Model linkages are mediated by the 

biophysical system’s delivery of ecosystem services and by 

the perception of these services by the social system. Model 

feedbacks are mediated by how services change human 

outcomes, perceptions, and behaviors that in turn affect the 

biophysical systems and their capacity to deliver subsequent 

services. Behavioral changes in the PPD model range from 

shifts in consumer preferences to environmental and energy 

policies and reproduction and migration rates. Such changes 

deliver to the biophysical system short-term pulses, such as 

nutrient inputs, fires, and management interventions, as well 

as long-term presses, such as atmospheric carbon dioxide 

loading, climate change, nitrogen deposition, and sea-level 

rise. In time, presses, pulses, and pulse–press interactions 

affect community structure and ecosystem function (Smith 

et al. 2009), eventually changing the delivery of ecosystem 

services such as the provision of food and fiber, pest and disease 

suppression, soil fertility, greenhouse gas stabilization, 

and clean water. 

There are many potential socioecological questions that 

could be asked across a network of long-term sites. Building 

on a long history of prior research, LTER Network sites and 

scientists have identified several environmental challenges 

that represent critical issues for science and society that 

the network seems particularly well positioned to address 

today, including (a) landscape vulnerability and resilience 

to climate and land-use change, (b) the consequences of 

cryosphere loss and changes in associated services that 

range from urban water supply to rural livelihoods, and 

(c) coastal-zone climate change as it interacts with rising 

sea levels and coastal population change. For each of these 

Exogenous 

forces 

Coupled human-ecosystem interactions 

Human 

systems 

Ecosystem 

services 

Actions/Policy 

Ecosystem boundary 

Environmental 

systems 

Exogenous 

forces 

Figure 3. A press–pulse dynamics (PPD) framework, 

simplified for use by K–12 learners. The complete PPD 

model with more comprehensive linkages and feedbacks is 

available in Collins and colleagues (2011). 

www.biosciencemag.org April 2012 / Vol. 62 No. 4 • BioScience 345

Articles 

challenges, a comprehensive socioecological framework is 

required for them to be addressed effectively; each is best 

addressed with long-term observations and experiments in 

multiple locations; and for each, a subset of network sites 

in partnership with other networks and observatories could 

provide a core set of locations at which questions could be 

effectively addressed. 

Future scenarios: Examining landscape vulnerability and resilience 

to global change. Science to help us understand, anticipate, 

and adapt to global change, including land-use and climate 

change, is becoming an ever more pressing need. How will 

global change alter the future of regional socioecological 

systems, and how and why do regional systems differ in 

vulnerability, resilience, and adaptability to change? These 

questions cannot be addressed by discipline-bound thinking 

but, rather, require new approaches that also incorporate 

broad-scale comparative investigations of diverse systems. 

One such approach is that of scenario studies (e.g., Baker 

et al. 2004, Thompson et al. 2012 [in this issue]), which provide 

a framework for addressing socioecological questions 

by crafting and evaluating suites of plausible scenarios that 

follow from current and historical trajectories. By examining 

multiple visions of the future that reflect a range of assumptions 

about land and water use, the burden of prediction is 

lifted, and comparisons among contrasting scenarios can be 

used to understand the dynamics of complex systems. New 

insights come from the examination of the perceived bounds 

of plausibility and from the discovery of commonalities 

across scenarios. Indeed, intrinsic vulnerabilities and robust 

management strategies are often identified when patterns 

recur across disparate scenarios. 

Depictions of future scenarios are often articulated by 

regional stakeholders, including residents, policymakers, 

and social and ecological scientists, in order to illustrate 

major strategic choices (Hoag et al. 2005). These qualitative 

scenarios can be an end in themselves, or they may 

lead to quantitative simulations of future landscape change. 

This is frequently an iterative process whereby the narratives 

inform and are in turn informed by integrated spatial 

models of socioecological change that might include, for 

example, agent-based models that link land-use change, 

econometric, and ecosystem process models (Evans and 

Kelly 2004). At its best, fundamental site-based science 

underpins the development of the scenario-to-simulation 

framework, the creation of which is itself a form of scientific 

synthesis. This approach for coupling qualitative and 

quantitative scenarios has informed prescient planning and 

policy decisions and has generated a rich set of fundamental 

research questions. 

For example, researchers at the Harvard Forest LTER site 

have begun a statewide scenario-studies project to examine 

the future of Massachusetts’s forests. Their work began with a 

landscape-simulation study to examine the relative influence 

of 50 more years of the current trends in forest conversion, 

timber harvest, and climate change in terms of their effects 

on forest carbon storage and tree species composition 

(Thompson et al. 2011). This work was rooted in 20 years 

of ecological research at Harvard Forest. Researchers then 

convened a group of around 12 stakeholders, including 

natural-resource managers and decisionmakers from state 

government, representatives from conservation nongovernmental 

organizations, and academics from multiple 

disciplines. They asked this group to chart three alternative 

futures of their choosing to compare with the current 

trends that had already been modeled. Through spirited 

discussion, the group settled on (a) a “free-market future,” 

characterized by a rollback in environmental regulations and 

incentives for new business; (b) a “resource-limited future,” 

characterized by high energy prices, a resurgence of agriculture, 

and a strong demand for woody-biomass energy; and 

(c) a “green-investment future,” characterized by government 

incentives for conservation, green energy, and land-use 

planning. Through an iterative process with stakeholders, 

the researchers were able to describe the types, distribution, 

and intensity of land uses under each of the scenarios. Each 

of the scenarios is now being integrated into a simulation 

framework, which will superimpose the land-use scenarios 

onto a common template of climate-change and ecological 

dynamics. The goal is to examine the aggregate and interactive 

effects of land use within each scenario, as well as to 

make comparisons across the scenarios. Clearly, none of 

the scenarios will manifest exactly as they were described; 

nonetheless, by examining multiple potential pathways, the 

effort should reveal characteristics of the Massachusetts 

landscape that are particularly vulnerable or resilient to the 

interactive effects of land-use and climate change. 

Cryosphere loss. The Earth’s cryosphere, which includes sea, 

lake, and river ice; glaciers; seasonal snow; and ice-rich permafrost, 

harbors over 80% of the freshwater on the planet. 

The cryosphere cools the planet through its albedo; regulates 

the global sea level; stores substantial stocks of carbon; 

insulates soil from subfreezing air temperatures; and serves 

as a seasonally refreshed water supply for human consumption, 

irrigation, nutrient transport, and waste disposal. The 

prospect of accelerated cryosphere loss under a warming climate 

portends great ecological change and poses enormous 

threats to these ecosystem services, with attendant social and 

economic costs. A 1-meter sea-level rise, now thought to be 

unavoidable with a 600–1000 parts per million atmospheric 

carbon dioxide peak over the coming century (Solomon 

et al. 2009), alone represents an estimated economic impact 

of $1 trillion that will be borne disproportionately by North 

America (Anthoff et al. 2010). 

The extent and rates of cryosphere loss are increasingly 

well monitored, and our ability to project the future rates 

of cryosphere decline is improving. However, the ecological 

consequences—and especially the nature and extent of and 

economic impacts on human society and institutions— 

are still poorly understood. Cryosphere loss represents an 

inadvertent press event caused by human decisions—driven 


y policies and markets—to extract energy from fossil fuel 

and to clear forests and other carbon-storing ecosystems 

for economic development. Changes in wintertime temperatures 

and snowfall will dramatically affect community 

structure and ecosystem processes in high-latitude and 

alpine ecosystems, but the effects will be felt even in arid, 

low-latitude ecosystems that depend on mountain meltwater 

for seasonal water supplies—riverine, floodplain, 

agricultural, and urban ecosystems in particular. Many of 

these effects will be social, since some of the most populous 

cities and productive farmland in North America depend 

on these water supplies. 

Examples of cryosphere loss and its effects abound across 

the LTER Network, from Arctic sites undergoing long-term 

permafrost melt to Antarctic and northern lake sites losing 

ice cover and terrestrial sites experiencing shorter periods of 

snow cover and more-frequent freeze–thaw events. Alpine 

communities, such as that at the Niwot Ridge LTER site, 

illustrate the degree of subregional connectivity involved 

(figure 4): Hydrological connectivity is driven by the duration 

and timing of the seasonal snowpack and snowmelt, 

and under a warming climate, increasing windborne dust 

will accelerate snowpack and glacial melt, which will result 

in the snowline’s moving to a higher elevation, which will in 

turn decrease hydrologic connectivity. With elevated nitrogen 

inputs from windborne dust and Denver air pollution, 

plant species diversity will decrease as alpine areas shrink, 

shrubland will expand, and the landscape will become more 

homogeneous. Exacerbating these trends is the regional 

outbreak of mountain pine beetles that will remove a large 

portion of the subalpine forest. 

Cryosphere change and its consequences are played out 

as long-term trends that in many places will be difficult to 

discern from short-term variability in climate and other 

environmental factors and in social dynamics, such as population 

and economic change. Long-term sites provide the 

perspective necessary to detect trends that would otherwise 

not be visible against this variability. And networked sites 

provide the potential for comparative tests of hypotheses 

that link cryosphere loss, ecosystem services, and human 

decisions, using, for example, the PPD model (figure 3). 

Key research questions (Fountain et al. 2012 [in this 

issue]) include (a) how climate regulation is affected by 

feedbacks from thawing permafrost and sea ice, especially 

because of the release of vast stores of carbon and changes in 

albedo, and what the implications are for regional and global 

economies and policies, including sovereignty; (b) what the 

economic implications of snow and ice loss are, including 

the future of winter recreation and related cultural activities; 

(c) how changing snowpack—the amount and timing 

of water storage and delivery—in the western United States 

will influence the economies of this region, and whether 

the impacts will be disproportionately imposed on disadvantaged 

groups; and (d) what the cultural mechanisms 

are by which cryosphere loss influences public opinion and 

what policies and legal instruments are most effective for 

Articles 

Figure 4. Expected changes in hydrologic connectivity 

related to cryosphere loss at the Niwot Ridge Long Term 

Ecological Research Network site. As windborne dust 

deposition increases in a warming climate, snowpack 

and glacial melt will be accelerated, which will result 

in a higher snowline, a shrunken alpine area, and the 

expansion of shrubland, exacerbated by a climate-induced 

mountain pine beetle outbreak that is now decimating the 

subalpine forest. This figure was created by Eric Parish. 

environmental protection, impact mitigation, and adaptation 

in the face of climate change. The PPD model provides 

an effective means for linking cryosphere change with the 

ecosystem dynamics that lead to altered ecosystem services 

and then with subsequent human activities that may— 

through policies, behaviors, and markets—either slow or 

hasten cryosphere loss. 

Coastal-zone climate change. Because they are at the interface 

of continental and oceanic realms, coastal systems are 

expected to be especially affected by climate change and 

to experience effects from both land and sea. With more 

than 50% of the US population living in coastal counties, 

many changes will play out in human communities and 

economies. Coastal-zone research sites, including nine 


Articles 

LTER Network and many additional partner sites, differ 

in their biophysical vulnerability to the coastal impacts 

of climate change. Some ecosystems along the US eastern 

seaboard will be more affected by sea-level rise and 

storm-surge severities, and others will be more affected 

by ocean acidification (e.g., coral reef communities in the 

south Pacific), the loss of sea ice (Antarctica), or changes in 

water temperature and freshwater inflows. Human vulnerabilities 

will also differ among regions, which arises from 

differences in coastal population density and demographic 

composition and from the location and resilience of the 

regions’ built infrastructure, which ranges from cities to 

drilling platforms. All of these effects and vulnerabilities 

need to be considered in concert in order to provide a comprehensive 

understanding of coastal-zone climate change 

and the potentials for future adaptation. 

The need to understand and anticipate the effects of 

climate change, assess the vulnerabilities of natural and 

human elements of coastal systems, and adapt to or mitigate 

the effects of changes is prompting new efforts for integration 

across academic disciplines and creation of partnerships 

among academic, public, and governmental constituents. 

As for cryosphere loss, long-term studies are needed in 

order to document patterns and consequences of coastalzone 

change. Unlike cryosphere loss, however, coastal-zone 

change is likely to be strongly episodic in response to storm 

events that are projected to be increasingly severe and whose 

inland consequences will, in any case, be magnified because 

of sea-level rise. Posing questions relevant to networked sites 

in the context of a common model allows for a fundamental 

understanding more difficult to gain from shorter-term or 

more geographically discrete research. 

Key questions include (a) how the presses and pulses 

associated with coastal climate change—altered water temperature, 

precipitation, runoff, sea level, solar radiation, wind 

and wave climates, pH, and salinity—affect the structure and 

function of coastal ecosystems and what attributes affect the 

vulnerability of those ecosystems; (b) how climate-induced 

changes in coastal systems affect critical ecosystem services 

such as carbon sequestration, wildlife habitat, food-web 

support, and storm protection; (c) what attributes of human 

systems such as built infrastructure; land use; governance 

structures; and population demographics, including wealth 

and ethnicity, interact to influence human vulnerabilities to 

coastal climate change and how these interact with changes 

in ecosystem services to prompt responses of adaptation and 

mitigation; and (d) how mitigation and adaptation strategies, 

such as coastal engineering and reductions in greenhouse 

gases, would feed back to affect climate drivers and the structure 

and function of coastal systems. Effectively addressing 

these questions requires an approach that acknowledges and 

deeply explores the linked socioecological processes that 

underlie the delivery of almost all of the ecosystem services 

provided by coastal-zone environments. 

Many LTER Network sites are located in coastal zones 

at different latitudes along the eastern and western US 

seaboards, as well as in the South Pacific and Antarctic 

Oceans, and provide a diversity of geomorphologies and 

degrees of human influence, ranging from urban to exurban, 

rural, and natural. They are therefore well positioned 

to address a subset of these key questions, most of which 

will require a combination of long-term baseline data 

and experiments designed to predict the consequences of 

sea-level rise for the ecology of coastal communities. 

Toward an environmentally literate populace 

Society’s ability to understand and act on the coupled 

natural and human systems on which we depend, built 

on a foundation of complex scientific inquiry, is key to a 

sustainable future. And it is the public—decisionmakers at 

all levels, from landowners and local officials to national 

leaders—who must act. The importance of an environmentally 

literate public is hard to overstate: From the grocery 

store and car dealership to the voting booth and corporate 

boardroom, individuals make choices that have far-reaching, 

collective consequences. Education helps to ensure that 

those choices are based at least in part on evidence and reasoning 

underpinned by solid scientific research. 

The LTER approach to research, combined with an ability 

to implement long-term educational initiatives, has allowed 

for unique approaches to the training of future researchers 

and to the conveyance of ecological concepts and insights 

to a broad constituency. At individual network sites, educational 

activities range from K–12 students and teachers 

engaged in schoolyard ecology to undergraduates involved 

in field classes and research internships and to graduate 

students and postdoctoral scholars learning to frame questions 

and to conduct research in long-term and sometimes 

cross-site contexts. Public outreach in many forms reaches 

working professionals, as well as the general public, often 

through the education of those best positioned to communicate 

with the general public. This outreach is increasingly 

placed in a socioecological context, which illustrates 

the natural–human system couplings that are central to 

addressing major environmental issues and that are often 

hidden to many. 

The vision for education in the LTER Network includes 

leveraging both long-term and cross-site perspectives to 

advance fundamental science learning by K–12, undergraduate, 

and graduate students and developing programs 

for key constituent and underrepresented groups. These 

groups include K–12 teachers, university students, education 

policymakers, and the professional public, which includes 

policymakers, natural-resource managers, the working 

media, and others whose success depends on access to and 

imparting of sound ecological knowledge. 

A long-term framework for environmental science literacy. Environmental 

science literacy—the capacity to participate 

in and make decisions through evidence-based discussions 

of socioecological systems—is essential not only for 

many science careers but also for responsible citizenship. 


Environmental science literacy requires citizens to understand, 

evaluate, and respond to multiple sources of information. 

The development of an environmental science literacy 

framework is crucial for providing this capacity among K–12 

students, a key constituency that represents both future 

STEM (science, technology, engineering, and mathematics) 

professionals and the 75% of the US population that will not 

earn a higher degree. It is also important for framing information 

provided to university students, STEM professionals 

and the general public. 

The development of an environmental literacy framework 

requires that we understand stakeholders’ current 

state of knowledge in core areas and how math and science 

concepts are best used to provide a desired level of literacy. 

In this context, stakeholders range from K–12 students to 

the voting public. We know, in general, that for most stakeholders, 

the state of knowledge is low, which is reflected 

both in standardized K–12 test scores (Gonzales et al. 2008) 

and in college-level assessments of ecological concepts (e.g., 

Hartley et al. 2011). We also know that there are troubling 

demographic disparities and, in particular, persistent gaps 

in science and mathematics achievement between white students 

and students of color (Vanneman et al. 2009). Building 

a capacity for principle-based environmental reasoning in 

all stakeholders and broadening the participation of underrepresented 

groups in environmental science careers should 

be important components of all science education efforts 

(George et al. 2001, ESA 2006). 

In K–12 education, the term learning progressions describes 

increasingly sophisticated ways of reasoning about an area 

of study, typically organized around a set of core topics 

that can be used to organize an integrated understanding 

of larger, complex issues (Duschl et al. 2007). A current 

effort within the network to build learning progressions 

into the K–12 science curriculum is being tested in 22 

school districts across the country with an LTER Network– 

associated NSF Math and Science Partnership (MSP) award. 

In districts around four network sites, MSP participants— 

scientists and science educators working with a diverse mix 

of K–12 science teachers and students—are developing 

learning progressions around key science strands. These 

strands include carbon, water, and biodiversity, plus a 

mathematical strand that addresses quantitative reasoning 

and the mathematics of modeling and a citizenship strand 

focused on the roles of culture and place. All of these strands 

are deeply embedded in state science and mathematics standards 

and are connected by the theme of education for citizenship: 

how students take on roles as consumers and voters 

using evidence-based reasoning about personal decisions 

that have environmental consequences. Multidisciplinary 

themes focused on the human impacts of land-use, ecosystem 

structure, and ecosystem services offer rich experiences 

in STEM education that include atmospheric science, soil 

science, geology, agronomy, ecology, hydrology, computer 

science, and systems modeling. Placing these strands in 

a simplified PPD model (figure 3) provides an easily 

Articles 

understood context for showing coupled human–ecosystem 

interactions. 

Two observations have emerged thus far from our LTERbased 

studies of learning progression. First, the PDD model 

embraced by the LTER Network (Collins et al. 2011) emphasizes 

that socioecological systems are organized as dynamic 

hierarchical systems. This basic tenet defines specific ways in 

which subjects or entities of interest are organized and interact 

with one another. Embedded in the concept is the notion 

of scale, wherein the boundaries between levels of interaction 

are defined by differences in the geographies and rates 

at which entities interact. Second, socioecological processes 

may include multiple principles that operate simultaneously 

in the social and ecological realms. 

Our research in student learning and understanding of 

ecological systems indicates that students and teachers fail to 

adopt hierarchical reasoning when questioned about ecological 

systems and principles. The conservation of matter and 

energy is not understood. Processes operating at one scale 

are assumed to operate at the same scope and magnitude at 

other scales. The nature of the interconnectedness of systems 

is overstated (e.g., removal of one species leads to system 

collapse). Human social hierarchies and human agency are 

conflated with ecological hierarchies and processes and are 

applied to natural systems. 

LTER Network research on learning progressions provides 

insights into how to advance student understanding 

of socioecological systems and also provides the scaffolding 

of science and social-science principles that is needed for 

students’ environmental literacy—a key to understanding 

human agency and its application to the scale of action 

that the challenges demand and to engaging the broader 

public. 

Engaging the broader public. Extending this understanding to 

older stakeholders—both the voting public in general and 

working STEM professionals, such as land managers, policy 

analysts, and public- and private-sector decisionmakers— 

presents a different set of challenges (Driscoll et al. 2012). 

Recent calls for a renewed effort by ecologists to engage in 

public outreach (e.g., Groffman et al. 2010) have noted that 

effective communication outside the classroom is influenced 

by learners’ interests, prior knowledge, social networks, and 

values and beliefs. This requires issues to be framed in ways 

that resonate with the public and messages to be delivered 

in ways that acknowledge the importance of emerging forms 

of media and informal learning environments (NRC 2009). 

Effective communication can also involve partnerships with 

boundary organizations that specialize in fostering the use 

of science knowledge in environmental policymaking and 

management (Osmond et al. 2010). These organizations 

range from university-based extension programs at landgrant 

universities to individual site-based efforts. 

One such site-based effort is the Science Links program 

at the Hubbard Brook LTER site. The program is explicitly 

aimed at communicating basic science findings at Hubbard 


Articles 

Brook to interested audiences that range from the general 

public to congressional staffers and is a particularly apt 

example of a way to leverage limited funding to broaden the 

impact of ecological findings. Outreach is initiated early in 

a project and is informed by a group of policy and natural 

resource management advisors who help to craft a message 

that is most relevant to the audience at hand. Such efforts 

can help to shorten an otherwise distressing lag between 

recognizing and addressing important environmental problems, 

such as those associated with acid rain (Likens 2010), 

as well as to help an increasingly dubious public to gain confidence 

in their ability to understand—and ultimately act 

on—complex environmental issues. More broadly, efforts to 

shape the public’s perception of natural areas and to increase 

awareness of the linkages between human and natural systems 

(e.g., Foster 1999) are equally vital, and both targeted 

and more-generalized efforts to build public environmental 

literacy are important priorities for LTER. 

Information for the future 

Cyberinfrastructure describes the network of computing 

environments that support advanced data acquisition, storage, 

management, integration, mining, visualization, and 

other information-processing services (Atkins et al. 2003). 

When used for scientific purposes, cyberinfrastructure is a 

technical solution to the problem of efficiently connecting 

data, computers, and people. 

The development of cyberinfrastructure is integral to the 

success of all environmental networks, including the LTER 

Network: Data for modeling and forecasting are essential 

for identifying the effects of accelerated and abrupt changes, 

and the explosion of real-time data availability calls for 

near-real-time analysis and distribution if those data are to 

be their most useful (AC-ERE 2009). Equally compelling 

is the need for data repositories and archives that allow 

the detection and synthesis of long-term trends and the 

effective integration of data from different networks and 

researchers. 

That less than 1% of ecological data is accessible after 

the publication of derived results (Reichman et al. 2011) 

reveals the social and technological challenges of curating 

that environmental data. Data dispersion, heterogeneity, and 

provenance (Jones MB et al. 2006), coupled with cultural 

norms that provide few rewards, make ecological information 

systems difficult to design, implement, and incentivize. 

Nevertheless, fueled by the emergence of environmental 

observatories charged with collecting and making openly 

available data from a variety of sensors (e.g., NRC 2004) 

and by the success of efforts to assemble and synthesize 

networked data toward broader-scale tests of ecological 

theory (e.g., Mittelbach et al. 2001, Suding et al. 2005), new 

efforts to develop centralized ecological information systems 

are under way. The LTER Network, with its 30 years of 

experience in environmental information management, has 

been a pioneer in these efforts—a microcosm of the hard 

challenges and substantive benefits of networked ecological 

data—and will be among those linked by centralized efforts 

such as DataONE (Michener et al. 2011). 

The distributed data repositories of the 26 network sites 

reflect the vast diversity of ecological data and the breadth 

of approaches to environmental data-management systems 

as developed by field stations, museums, academic institutions, 

state and local governments, and individual scientists. 

Core LTER Network data conform to consistent metadata 

standards (Michener 2006) and are held by individual sites 

within Web-accessible catalogs. These data repositories, 

plus a network-wide policy of open data access, make data 

available to those wishing to assemble cross-site syntheses. 

Although open access has been crucial to the success of 

cross-site studies such as those noted earlier (e.g., Mittelbach 

et al. 2001, Parton et al. 2007), the structure of site data often 

differs from catalog to catalog, which makes the discovery 

and subsequent integration of semantically similar data a 

task that is, at best, inconvenient. What is needed is a central 

repository that maintains the veracity and provenance of site 

data but allows single-portal access. 

Early examples include the climate and hydrology portals 

for LTER data. Centralized access to data from 26 sites provides 

an ability to detect and synthesize patterns and trends 

without the pain of querying 26 separate data catalogs with 

different keyword vocabularies and reporting units. Lowered 

transaction costs makes synthesis practical—and, in some 

cases, possible—when it had not been so previously, providing 

in this case a capacity to integrate multiple climate- and 

hydrologic-system components across disparate ecosystems 

and biomes to provide novel insights (Jones JA et al. 2012 

[in this issue]). What we need next is an ability to perform 

these syntheses for all system components. The LTER 

Network Information System (NIS) is being developed to 

address this need. 

The LTER NIS will provide access to data from the 

26 network sites through a single point of access and at the 

same time ensure the long-term preservation of site data 

through centralized stewardship. Site data will continue to 

be curated at individual sites but will also be exposed to 

harvest by the NIS on a frequent, periodic basis. Further 

data processing will then provide common formats that 

can be easily queried by cross-network portals, such as 

EcoTrends (Peters et al. 2011), that are designed to create 

derived long-term data products and by storage systems 

such as DataONE that will provide a centralized facility 

for storing data from multiple sources, including the NIS. 

Importantly, the system will be scalable: Adding data from 

additional sites, whether they are existing field stations, sensor 

networks, future LTER sites, or individual field projects, 

will be straightforward. 

The nature of LTER data—and by extension, ecological 

data in general—makes a single rigid method for storing 

and accessing them impractical. By design, many long-term 

data sets are collected using common protocols at regular 

intervals from specified locations. As priorities, resources, 

and technologies shift, however, intervals change, protocols 


are improved, and locations sometimes become inappropriate 

or insufficient. Robust sampling programs will have 

precautions and methods in place to protect the veracity 

of long-term observations, including a data-management 

system sufficiently nimble to allow these changes. An additional 

challenge in ecological science, however, is archiving 

and exposing experimental data—data that may be collected 

over a short term, with additional protocols and different 

experimental treatments in a sampling matrix that may not 

correspond much with that of the long-term collection. 

This adds an additional important burden on information 

systems that aspire to address ecological science needs but 

also provides an invaluable opportunity for future users to 

query the full suite of observations available for a particular 

site or region. 

Informal users also need to be accommodated. The best 

information system will make derived products available 

to a variety of potential users—not just scientists but also 

educators, students, decisionmakers, and the public. So 

long as data within the system are fully exposed to all portal 

developers, this accommodation will be straightforward, as 

might be the accommodation of data from nontraditional, 

more-uncertain sources, such as citizen science networks 

(Cohn 2008). 

Conclusions 

The US LTER Network enters its fourth decade with a 

sound record of scientific achievement in the ecological sciences. 

At each of the network’s 26 sites, we have learned an 

extraordinary amount about the organisms and processes 

important at the biome it represents, about the way the 

site’s ecosystems respond to disturbance, and about human 

influences and long-term environmental change. Cross-site 

observations and experiments are increasingly revealing 

how key processes, organisms, and ecological attributes 

are organized and behave across major environmental 

gradients. In total, research in the LTER portfolio is contributing 

substantially to our basic knowledge of ecological 

interactions and to our ability to forecast change and test 

ecological theory. 

Against this backdrop, the LTER Network is undertaking a 

new kind of transdisciplinary science—one that ranges from 

local to global in scope, that blends ecological and social 

science theories, methods, and interpretations in order to 

better understand and forecast environmental change in an 

era when no ecosystem on Earth is free from human influence. 

Furthermore, the network is increasingly focused on 

conveying those results to an engaged audience of decisionmakers 

that can apply it. The LTER Network’s PPD model 

provides a unifying framework for better understanding 

coupled natural–human systems across regions and temporal 

scales and a means for examining feedbacks and testing 

hypotheses about, first, how humans perceive the critical 

services provided by ecosystems; second, how these perceptions 

change behaviors and institutions; and third, how 

these changes in turn feed back to affect ecosystem structure 

Articles 

and function and the ability of these systems to indefinitely 

sustain their delivery of services. 

Environmental literacy is an important ongoing legacy 

of LTER Network science and will remain so. Learners 

at all levels have benefited from LTER Network involvement: 

graduate and undergraduate students, K–12 students 

and educators, working professionals involved in land 

and resource management, policymakers, and the general 

public. Future efforts will be directed toward ensuring 

that these groups understand the linkages and feedbacks 

between social and ecological systems to better inform their 

ability to make evidence-based environmental decisions at 

all levels. 

Advances in cyberinfrastructure are required in order to 

manage and organize the rapidly growing volume of ecological 

information and in order to enable integration and synthesis 

of that information over time. The LTER Network has 

led the ecological community in developing protocols and 

practices for documenting, curating, and sharing data, and 

it is now building the NIS, which will collect and curate data 

from LTER Network and other sites for storage in formats 

that can be queried by applications built to provide users 

with derived long-term data. Data in the system will thus be 

available to scientists, educators, students, decisionmakers, 

and the public for research, decision support, teaching, and 

informal education opportunities. 

The LTER Network’s primary mission is to use long-term 

observations and experiments to generate and test ecological 

theory at local to regional scales. Progress in solving environmental 

problems that today seem intractable depends 

on fundamental, long-term, integrated research that will 

generate a synthetic understanding of highly dynamic socioecological 

systems. Likewise, the early discovery of tomorrow’s 

surprises depends on long-term research that provides a 

capacity to detect new trends. Extending these capacities to 

continental scales will provide the necessary experimental 

context within which to address the causes and consequences 

of change documented both by the LTER Network and by 

the emerging constellation of environmental observatories. 

Acknowledgments 

The LTER Network owes its success to the several thousand 

scientists who have used its sites and data to conduct 

groundbreaking ecological research and to the support and 

leadership provided by the National Science Foundation and 

state and federal agency partners. The network’s principal 

partners include the US Forest Service, the Agricultural 

Research Service of the US Department of Agriculture, the 

US Fish and Wildlife Service’s Bureau of Land Management, 

and the US Geological Survey. We thank three anonymous 

reviewers for insightful comments on an earlier version of 

this article. 

References cited 

[AC-ERE] US National Science Foundation Advisory Committee for 

Environmental Research and Education. 2009. Transitions and Tipping 


Articles 

Points in Complex Environmental Systems. US National Science 

Foundation. 

Adler PB, et al. 2011. Productivity is a poor predictor of plant species 

richness. Science 333: 1750–1753. 

Anthoff D, Nicholls RJ, Tol RSJ. 2010. The economic impact of substantial 

sea-level rise. Mitigation and Adaptation Strategies for Global Change 

15: 321–335. 

Atkins DE, Droegemeier KK, Feldman SI, Garcia-Molina H, Klein 

ML, Messerschmitt DG, Messina P, Ostriker JP, Wright MH. 2003. 

Revolutionizing Science and Engineering through Cyberinfrastructure. 

National Science Foundation. Report no. cise051203. 

Baker JP, Hulse DW, Gregory SV, White D, Van Sickle J, Berger PA, Dole D, 

Schumaker NH. 2004. Alternative futures for the Willamette River 

basin, Oregon. Ecological Applications 14: 313–324. 

Beaulieu JJ, et al. 2011. Nitrous oxide emission from denitrification in 

stream and river networks. Proceedings of the National Academy of 

Sciences 108: 214–219. 

Callahan JT. 1984. Long-term ecological research. BioScience 34: 

363–367. 

Carpenter SR. 2008. Emergence of ecological networks. Frontiers in Ecology 

and the Environment 6: 228. 

Clark CM, Cleland EE, Collins SL, Fargione JE, Gough L, Gross KL, 

Pennings SC, Suding KN, Grace JB. 2007. Environmental and plant 

community determinants of species loss following nitrogen enrichment. 

Ecology Letters 10: 596–607. 

Cleland EE, et al. 2008. Species responses to nitrogen fertilization in 

herbaceous plant communities, and associated species traits. Ecology 

89: 1175. 

Cleland EE, Clark CM, Collings SL, Fargione JE, Gough L, Gross KL, 

Pennings SC, Suding KN. 2011. Natural patterns of invasion in herbaceous 

plant communities are related to soil nitrogen and the functional 

similarity of native species. Journal of Ecology 99: 1327–1338. 

Cohn JP. 2008. Citizen science: Can volunteers do real research? BioScience 

58: 192–197. 

Collins SL, et al. 2011. An integrated conceptual framework for 

social–ecological research. Frontiers in Ecology and the Environment 

9: 351–357. 

Driscoll CT, Lambert KF, Chapin FS III, Nowak DJ, Spies TA, Swanson FJ, 

Kittredge DB Jr, Hart CM. 2012. Science and society: The role of longterm 

studies in environmental stewardship. BioScience 62: 354–366. 

Duschl RA, Schweingruber HA, Shouse AW, eds. 2007. Taking Science 

to School: Learning and Teaching Science in Grades K–8. National 

Academies Press. 

[ESA] Ecological Society of America. 2006. Women and Minorities in 

Ecology II (WAMIE II): Committee Report, March 2006. ESA. 

Evans TP, Kelley H. 2004. Multi-scale analysis of a household level 

agent-based model of landcover change. Journal of Environmental 

Management 72: 57–72. 

Foster DR. 1999. Thoreau’s Country: Journey through a Transformed 

Landscape. Harvard University Press. 

Fountain AG, Campbell JL, Schuur EAG, Stammerjohn SE, Williams MW, 

Ducklow HW. 2012. The disappearing cryosphere: Impacts and ecosystem 

responses to rapid cryosphere loss. BioScience 62: 405–415. 

Franklin JF, Bledsoe CS, Callahan JT. 1990. Contributions of the Long-Term 

Ecological Research Program. BioScience 40: 509–523. 

George YS, Neale DS, Van Horne V, Malcom SM. 2001. In Pursuit 

of a Diverse Science, Technology, Engineering, and Mathematics 

Workforce: Recommended Research Priorities to Enhance Participation 

by Underrepresented Minorities. American Association for the 

Advancement of Science. 

Gholz HL, Wedin DA, Smitherman SM, Harmon M, Parton WJ. 2000. 

Long-term dynamics of pine and hardwood litter in contrasting environments: 

Toward a global model of decomposition. Global Change 

Biology 6: 751–766. 

Gonzales P, Williams T, Jocelyn L, Roey S, Kastberg D, Brenwald S. 2008. 

Highlights from TIMSS 2007: Mathematics and Science Achievement 

of U.S. Fourth- and Eighth-Grade Students in an International Context. 

National Center for Education Statistics, US Department of Education. 

Report no. NCES 2009-001 Revised. 

Groffman PM, Stylinski C, Nisbet MC, Duarte CM, Jordan R, Burgin A, 

Previtali MA, Coloso J. 2010. Restarting the conversation: Challenges at 

the interface between ecology and society. Frontiers in Ecology and the 

Environment 8: 284–291. 

Hartley LM, Wilke BJ, Schramm JW, D’Avanzo C, Anderson CW. 2011. 

College students’ understanding of the carbon cycle: Contrasting 

principle-based and informal reasoning. BioScience 61: 65–75. 

Helton AM, et al. 2011. Thinking outside the channel: Modeling nitrogen 

cycling in networked river ecosystems. Frontiers in Ecology and the 


Hoag DL, Keske-Handley C, Ascough II J, Koontz L. 2005. Decision making 

with environmental indices. Pages 159–182 in Burk AR, ed. New Trends 

in Ecology Research. Nova Science. 

Hobbie JE, Carpenter SR, Grimm NB, Gosz JR, Seastedt TR. 2003. The US 

Long Term Ecological Research Program. BioScience 53: 21–32. 

Isern AR, Clark HL. 2003. The ocean observatories initiative: A continued 

presence for interactive ocean research. Marine Technology Society 

Journal 37: 26–41. 

Johnson JC, Christian RR, Brunt JW, Hickman CR, Waide RB. 2010. 

Evolution of collaboration within the US Long Term Ecological 

Research Network. BioScience 60: 931–940. 

Jones JA, et al. 2012. Ecosystem processes and human influences regulate 

streamflow response to climate change at long-term ecological research 

sites. BioScience 62: 390–404. 

Jones MB, Schildhauer MP, Reichman OJ, Bowers S. 2006. The new bioinformatics: 

Integrating ecological data from the gene to the biosphere. 

Annual Review of Ecology, Evolution, and Systematics 37: 519–544. 

Kane ES, et al. 2008. Precipitation control over inorganic nitrogen import– 

export budgets across watersheds: A synthesis of long-term ecological 

research. Ecohydrology 1: 105–117. 

Keller M, Schimel DS, Hargrove WW, Hoffman FM. 2008. A continental 

strategy for the National Ecological Observatory Network. Frontiers in 

Ecology and the Environment 6: 282–284. 

Knapp AK, et al. 2012. Past, present, and future roles of long-term 

experiments in the LTER network. BioScience 62: 377–389. 

Kratz TK, et al. 2006. Toward a global lake ecological observatory network. 

Publications of the Karelian Institute 145: 51–63. 

Likens GE. 2010. The role of science in decision making: Does evidencebased 

science drive environmental policy? Frontiers in Ecology and the 

Environment 8: e1–e9. 

Liu J, et al. 2008. Complexity of coupled human and natural systems. 

Science 317: 1513–1516. 

Magnuson JJ, et al. 2000. Historical trends in lake and river ice cover in the 

Northern Hemisphere. Science 289: 1743–1746. 

McKnight DM. 2010. Overcoming “ecophobia”: Fostering environmental 

empathy through narrative in children’s science literature. Frontiers in 

Ecology and the Environment 8: e10–e15. 

Michener WK. 2006. Meta-information concepts for ecological data 

management. Ecological Informatics 1: 3–7. 

Michener WK, et al. 2011. DataONE: Data Observation Network for 

Earth—Preserving data and enabling innovation in the biological and 

environmental sciences. D-Lib Magazine 2011, 17: 1–9. (17 January 

2012; www.dlib.org/dlib/january11/michener/01michener.html) 

Mittelbach GG, Steiner CF, Scheiner SM, Gross KL, Reynolds HL, Waide 

RB, Willig MR, Dodson SI, Gough L. 2001. What is the observed 

relationship between species richness and productivity? Ecology 82: 

2381–2396. 

Mulholland PJ, et al. 2008. Stream denitrification across biomes and its 

response to anthropogenic nitrate loading. Nature 452: 202–205. 

[NRC] National Research Council. 2004. NEON: Addressing the Nation’s 

Environmental Challenges. National Academies Press. 

———. 2009. Learning science in informal environments: People, places, 

and pursuits. National Academies Press. 

Olson DM, et al. 2001. Terrestrial ecoregions of the world: A new map of life 

on earth. BioScience 51: 933–938. 


Osmond DL, et al. 2010. The role of interface organizations in science 

communication and understanding. Frontiers in Ecology and the 


Parton W, et al. 2007. Global-scale similarities in nitrogen release patterns 

during long-term decomposition. Science 315: 361–364. 

Pennings SC, Clark CM, Cleland EE, Collins SL, Gough L, Gross KL, 

Milchunas DG, Suding KM. 2005. Do individual plant species show 

predictable responses to nitrogen addition across multiple experiments? 

Oikos 110: 547–555. 

Peters DPC, Groffman PM, Nadelhoffer KJ, Grimm NB, Coffins SL, 

Michener WK, Huston MA. 2008. Living in an increasingly connected 

world: A framework for continental-scale environmental science. 

Frontiers in Ecology and the Environment 6: 229–237. 

Peters DPC, et al. 2011. Long-term Trends in Ecological Systems: A Basis 

for Understanding Responses to Global Change. US Department of 

Agriculture, Agricultural Research Service. 

Peterson BJ, et al. 2001. Control of nitrogen export from watersheds by 

headwater streams. Science 292: 86–90. 

Redman CL, Grove JM, Kuby LH. 2004. Integrating social science into 

the Long Term Ecological Research (LTER) Network: Social dimensions 

of ecological change and ecological dimensions of social change. 

Ecosystems 7: 161–171. 

Reichman OJ, Jones MB, Schildhauer MP. 2011. Challenges and opportunities 

of open data in ecology. Science 331: 703–705. 

Robertson GP. 2008. Long-term ecological research: Re-inventing network 

science. Frontiers in Ecology and the Environment 6: 281. 

Smith MD, Knapp AK, Collins SL. 2009. A framework for assessing ecosystem 

dynamics in response to chronic resource alterations induced by 

global change. Ecology 90: 3279–3289. 

Solomon S, Plattner G-K, Knutti R, Friedlingstein P. 2009. Irreversible 

climate change due to carbon dioxide emissions. Proceedings of the 

National Academy of Sciences 106: 1704–1709. 

Suding KN, Collins SL, Gough L, Clark C, Cleland EE, Gross KL, Milchunas 

DG, Pennings S. 2005. Functional- and abundance-based mechanisms 

explain diversity loss due to N fertilization. Proceedings of the National 

Academy of Sciences 102: 4387–4392. 

Thompson JR, Foster DR, Scheller R, Kittredge D. 2011 The influence of 

land use and climate change on forest biomass and composition in 

Massachusetts, USA. Ecological Applications 21: 2425–2444. 

Articles 

Thompson JR, Wiek A, Swanson FJ, Carpenter SR, Fresco N, Hollingsworth 

T, Spies TA, Foster DR. 2012. Scenario studies as a synthetic and integrative 

research activity for long-term ecological research. BioScience 

62: 367–376. 

Vanneman A, Hamilton L, Anderson JB. 2009. Achievement Gaps: How 

Black and White Students in Public Schools Perform in Mathematics 

and Reading on the National Assessment of Educational Progress. 

National Center for Education Statistics, US Department of Education. 

Report no. NCES 2009-455. 

Waide RB, Willig MR, Steiner CF, Mittelbach G, Gough L, Dodson GI, Juday 

GP, Parmenter R. 1999. The relationship between primary productivity 

and species richness. Annual Review of Ecology and Systematics 30: 

257–300. 

G. Philip Robertson (robertson@kbs.msu.edu) is a professor with the W. K. 

Kellogg Biological Station and with the Department of Crop and Soil Sciences 

at Michigan State University, in Hickory Corners. Scott L. Collins is a professor 

in the Department of Biology at the University of New Mexico, in 

Albuquerque. David R. Foster is a professor in the Department of Organismic 

and Evolutionary Biology at Harvard University and director of the Harvard 

Forest in Petersham, Massachusetts. Nicholas Brokaw is a professor with the 

Institute for Tropical Ecosystem Studies at the University of Puerto Rico, in San 

Juan. Hugh W. Ducklow is a senior scientist and director of The Ecosystems 

Center at the Marine Biological Laboratory in Woods Hole, Massachusetts. 

Ted L. Gragson is a professor and chair of the Department of Anthropology 

at the University of Georgia, in Athens. Corinna Gries is a research scientist 

at the Center for Limnology at the University of Wisconsin, in Madison. 

Stephen K. Hamilton is a professor in the Department of Zoology and the 

W. K. Kellogg Biological Station of Michigan State University, in Hickory 

Corners. A. David McGuire is a professor in the Department of Biology 

and Wildlife at the University of Alaska, in Fairbanks, and is affiliated with 

the US Geological Survey. John C. Moore is a professor and director of the 

Natural Resource Ecology Laboratory at Colorado State University, in Fort 

Collins. Emily H. Stanley is a professor with the Center for Limnology at the 

University of Wisconsin, Madison. Robert B. Waide is professor of biology at 

the University of New Mexico, in Albuquerque, and executive director of the 

Long Term Ecological Research Network. Mark W. Williams is a professor in 

the Department of Geography at the University of Colorado, in Boulder. 


Articles 

Science and Society: The Role of 

Long-Term Studies in Environmental 

Stewardship 

Charles T. DrisColl, KaThleen F. lamberT, F. sTuarT Chapin iii, DaviD J. nowaK, Thomas a. spies, 

FreDeriCK J. swanson, DaviD b. KiTTreDge Jr., anD Clarisse m. harT 

Long-term research should play a crucial role in addressing grand challenges in environmental stewardship. We examine the efforts of five Long 

Term Ecological Research Network sites to enhance policy, management, and conservation decisions for forest ecosystems. In these case studies, 

we explore the approaches used to inform policy on atmospheric deposition, public land management, land conservation, and urban forestry, 

including decisionmaker engagement and integration of local knowledge, application of models to analyze the potential consequences of policy 

and management decisions, and adaptive management to generate new knowledge and incorporate it into decisionmaking. Efforts to enhance 

the role of long-term research in informing major environmental challenges would benefit from the development of metrics to evaluate impact; 

stronger partnerships among research sites, professional societies, decisionmakers, and journalists; and greater investment in efforts to develop, 

test, and expand practice-based experiments at the interface of science and society. 

Keywords: boundary spanning, environmental policy and management, Long Term Ecological Research Network, science communication 

The growing urgency and complexity of challenges to 

global sustainability demands new approaches for engaging 

the intellectual capital of expert communities worldwide. 

To meet this demand, the scientific and social-science communities 

must expand their capacity to work at the interfaces 

between ecological science and environmental policy, 

natural-resource management, and conservation. The need 

for stronger, more-reliable linkages between science and 

society is well documented in both popular media and the 

academic literature (e.g., Lubchenco 1998). 

The US National Science Foundation (NSF) recognized 

the importance of translating the benefits of research for 

society by establishing its “broader impacts” review criterion 

in 1997. In 2002, the NSF’s 20-year review of the Long Term 

Ecological Research (LTER) Network recommended that 

the LTER Network program assume a more powerful and 

pervasive role in informing environmental solutions at local, 

national, and international levels. In 2005, the Ecological 

Society of America (ESA) established the foundation for 

its Earth Stewardship initiative (Chapin et al. 2011) when it 

recommended that ecologists must play a greatly expanded 

role in communicating their research and influencing policies 

and decisions that affect the environment. It is critical 

for the LTER Network and for other ecosystem research 

programs to move beyond broad calls for action and to build 

deliberate and effective long-term relationships between 

ecological science and environmental decisionmaking. In 

this article, we illustrate, as do the other authors in this 

special section (e.g., Thompson et al. 2012 [in this issue]), 

the growing role that LTER is playing to enhance science 

engagement with local, regional, and national policy and 

management issues. To develop such programs, the scientific 

community needs to build experience and learn from 

practical examples of effective synthesis and integration of 

LTER to meet the needs of society. In this article, we present 

and discuss five case studies of work at the interface of science, 

policy, and management from forested LTER Network 

sites across the United States. We distill a set of common 

strategies, lessons, and recommendations for improving and 

expanding interface efforts to improve the ability to meet the 

grand challenges in environmental science of our time. 

Effective science interface efforts 

Although the integration of science and society is often 

viewed as a relatively narrow issue of a need for more and 

better science communication, programs that build stronger 

interfaces between science and society require attention to 

the full range of boundary-spanning activities, such as public 

engagement, decision-relevant synthesis, distillation of 

results, and science translation and dissemination, through 





a variety of media to meet the needs of diverse audiences 

(Cash et al. 2003, Driscoll et al. 2011). Boundary spanning 

refers to “practices and processes that facilitate bringing 

science and society closer together in order to produce 

‘useful’ information—that is, information that is salient, 

credible, and legitimate” (McNie et al. 2008, p. 9). Building 

credibility, salience, and legitimacy with stakeholders helps 

to solidify long-term relationships and increases the influence 

of scientific research in the decisionmaking process 

over time (Cash et al. 2003). 

The role of the LTER Network 

After 30 years of coordinated research and education, the 

LTER Network is well positioned to facilitate the integration 

of science and society by using its highly credible, longterm 

science to support engagement with decisionmakers 

to frame relevant questions for research and synthesis that 

can inform environmental policy and conservation. The 

LTER Network consists of 26 research sites throughout the 

United States and a few outside the United States, some of 

which have been operating for three decades or longer. The 

long-term ecosystem measurements and experiments that 

are a hallmark of the LTER Network address important 

environmental issues in coupled human–natural systems, 

including climate change, land use, pollution, and the loss 

of biodiversity (Knapp et al. 2012 [in this issue], Thompson 

et al. 2012 [in this issue]). Another distinguishing feature of 

the LTER Network is its core of researchers at each site, who 

are attentive to the well-being and future of their respective 

bioregions. 

The LTER Network’s new Strategic Communication Plan 

(LTER Network 2010a) and Strategic and Implementation 

Plan (LTER Network 2010b) call for the network to reach 

out to decisionmakers at local, regional, national, and international 

levels. The communication plan specifically recommends 

engaging decisionmakers in the framing of cross-site 

synthesis and equipping these efforts with full-scale communication 

capacity, funding a supplement program for 

LTER Network sites to develop local and regional programs 

for public engagement and outreach, and partnering with 

existing LTER Network sites that have established science 

journalism programs to develop sustained outreach to the 

media. 

The value of long-term monitoring and research 

Environmental policy and management issues play out over 

decades or longer and benefit from the continuous advances 

in understanding that are derived from long-term research. 

Policy development is an iterative process that requires 

ongoing assessment, reevaluation, adaptive management, 

and consideration of future scenarios (Driscoll et al. 2010). 

For example, although the Clean Air Act was first passed 

by Congress in 1972, the development of amendments and 

rules to implement the act are ongoing and rely on quantitative 

information to evaluate the effectiveness of pollution- 

control measures and to guide program management (Lovett 

Articles 

et al. 2007). Long-term measurements that link decreases in 

emissions with changes in soil and water quality and the 

health of aquatic and terrestrial ecosystems are vital to 

assessing the extent to which air pollution regulations meet 

the intent of the act (Driscoll et al. 2001). 

Similarly, effective natural-resource management is adaptive 

and draws on lessons from past decisions and management 

experience distilled from the results of long-term 

measurements and experiments, regional surveys, and 

modeling (Spies et al. 2010). The practice of forestry in landscapes 

that support multiple uses must adapt to new knowledge 

regarding the nature and effects of climate change, 

forest management, land-use trends, intense storms, fire, 

and other disturbances. This understanding must include 

the impacts of these often interacting pulse and press stressors 

on management goals and ecosystem services, such 

as fiber production, biological diversity, carbon storage, 

trace-gas production and consumption, water quantity and 

quality, and recreation. Detailed, long-term measurements 

tied directly to management-relevant forest experiments 

have improved the scientific basis for forest management 

and policy. Important examples include the guiding principles 

for the conservation of old-growth forests (Franklin 

et al. 1981) and regional- and continental-scale carbon 

budgets important to climate-change mitigation (Lovett 

et al. 2007). 

The five case studies presented here represent examples of 

outreach activities at selected forested LTER Network sites. 

We chose this suite of case studies because they have active 

programs for engaging decisionmakers, represent a range of 

policy and management issues, and use different approaches 

to achieve their outreach goals for a common ecosystem type. 

Reviewing efforts across forest sites provides the opportunity 

to consider how audiences, management and policy 

issues, and communication approaches vary across diverse 

regional research sites and programs. Specifically, the case 

studies incorporate the impacts of atmospheric deposition 

on forested ecosystems (Hubbard Brook), land-use change 

and forest conservation in a predominantly private-lands 

landscape (Harvard Forest), endangered species and public 

lands management (Andrews), urban forestry in developed 

landscapes (Baltimore), and forest stewardship in 

the context of changing fire and climate regimes (Bonanza 

Creek). These case studies represent only some of the many 

science–policy integration efforts that exist across the LTER 

Network (for other examples, see the Translating Science 

for Society brochure at http://intranet2.lternet.edu/sites/ 

intranet2.l ternet.edu/files/documents/Network_Publications/ 

Brochures/nsf0533.pdf ). In each of these cases, the ability 

to tap into core strengths of the LTER Network, such as 

long-term research that is relevant to policy and management 

issues, advanced information-management systems, 

and stores of long-term data, has proven essential to the 

synthesis and distillation of science for use in policy and 

management decisions related to coupled human–natural 

systems. 


Articles 

Case studies in linking LTER science with policy, 

conservation, and management 

Below, we describe several case studies that link LTER science 

with policy, conservation, and management. 

Air pollution effects on ecosystems: The Hubbard Brook Research 

Foundation Science Links Program. Air pollution can have 

marked effects on the structure and function of ecosystems 

through elevated atmospheric deposition of sulfur, oxidized 

and reduced nitrogen compounds and mercury, and 

high concentrations of tropospheric ozone. Recent efforts 

to channel this knowledge into decisionmaking through 

organized outreach and communication have increased 

the influence of long-term research on air-quality management 

in the United States (Driscoll et al. 2010). The LTER 

Network, through its long-term measurements and experiments 

(Driscoll et al. 2001), has been particularly effective 

in addressing policy issues concerning air pollution and 

atmospheric deposition effects on ecosystems. 

The effects of air pollution on forest and aquatic ecosystems 

have been a research focus since the inception of the 

Hubbard Brook Ecosystem Study and the Hubbard Brook 

LTER site. The value of long-term measurements of the 

chemistry of precipitation and streamwater at the Hubbard 

Brook Experimental Forest in documenting trends in acidic 

deposition and in assessing the effectiveness of the federal 

Clean Air Act represents an important example of the connections 

between long-term research and air-quality policy 

(figure 1). The Hubbard Brook Research Foundation (HBRF) 

launched Science Links in 1998 to build on this legacy and to 

develop new initiatives linking ecosystem science with public 

policy (http://hubbardbrookfoundation.org/12-2). 

Science Links projects are state-of-the-science synthesis 

efforts of an environmental issue in the context of current 

policy discussions. The first three Science Links projects 

addressed air pollution impacts on ecosystems, including the 

effects of acid, nitrogen, and mercury deposition (Driscoll 

et al. 2001, 2011). Science Links projects involve teams of 

around 12 scientific experts, selected on the basis of their 

experience and disciplinary coverage, and a team of policy 

advisers. The science teams define the scope of the project, 

analyze relevant databases and conduct model calculations. 

The policy advisers are engaged in dialogue from the outset 

to frame policy-relevant questions, discuss the alternatives 

analyzed, and provide input on Science Links products. 

A communication and outreach plan is integral to the 

success of Science Links projects. The written plan provides 

a roadmap to facilitate an exchange between scientists and 

policy stakeholders as well as direct outreach to journalists. 

The centerpiece of any Science Links project is the translation 

report aimed at congressional and government-agency 

staff involved in policy development. These reports are 

structured to facilitate communication of the major findings, 

with the conclusions presented first in clear, straightforward 

terms, followed by supporting information with 

layered details. A proactive media strategy has been critical 

to the impact of Science Links projects. Accurate and widespread 

media coverage has brought attention to Science 

Links results and verified the societal importance of the 

findings for policymakers. The initial public release of a 

Science Links report is followed by additional interviews, 

seminars, and briefings for up to a year. Science Links 

projects have also been coupled with the Hubbard Brook 

LTER site educational activities through the development 

of supplemental teacher guides. 

There are several dimensions to quantifying the impact 

of Science Links projects (Driscoll et al. 2011). The scientific 

impact can be measured by the number of citations 

in the scientific literature; the six Science Links journal 

articles have been cited more than 1300 times in the peerreviewed 

literature. The media impact can be measured by 

the extent and quality of media cover. Science Links initiatives 

have been covered in more than 475 media stories 

and have appeared in major news outlets, including an 

opinion–editorial piece in The New York Times. The impacts 

on policy are more difficult to quantify. Moreover, they are 

356 BioScience • April 2012 / Vol. 62 No. 4 www.biosciencemag.org 

a 

b 

Figure 1. Relationships between annual volume-weighted 

concentrations of sulfate (a) and nitrate (b) in precipitation 

at the Hubbard Brook Experimental Forest and emissions 

of sulfur dioxide and nitrogen oxides, respectively, in the 

emission source area of the northeastern United States. 

The emission source area used is specified in Driscoll and 

colleagues (2001). Abbreviations: NO x , nitrogen oxides; 

R 2 , regression coefficient; SO 2 , sulphur dioxide; Tg/year, 

teragrams per year; µeq/L, microequivalencies per liter.

often beyond the control of the scientists, regardless of the 

process used to link science to policy. Timing is everything 

in this dance between science and management. Forms of 

evidence of policy uptake include reference to Science Links 

findings in proposed legislation (e.g., the Clean Power Act, 

the National Mercury Monitoring Act), legal briefs (e.g., the 

Northeast States New Source Review case against the US 

Environmental Protection Agency), and media accounts 

of major policy and court decisions (e.g., the Interstate 

Transfer Rule for nitrogen oxide emissions and the remand 

of the Clean Air Mercury Rule and its trading provisions). 

Beyond this evidence, policymakers and program managers 

routinely comment on the usefulness of Science Links in 

improving the scientific basis for decisionmaking because 

of its reliance on rigorous long-term research and effective 

translation. 

Uniting conservation science and policy: Examples from the 

Harvard Forest LTER site. The Harvard Forest has oriented its 

Articles 

long-term studies around forest management and conservation 

questions in New England since its inception in 

1907. When it was established by Harvard University, its 

objectives were to serve as a model forest to demonstrate 

the practice of forestry, an experiment station for research 

in forestry, and a field laboratory for students (Fisher 1921). 

Today, Harvard Forest scientists remain dedicated to the 

founding tenet of drawing on insights gained through the 

historical and retrospective study of forests, natural disturbance, 

and land use (Fisher 1933, Foster 2000), and the 

Harvard Forest serves as a central gathering spot for meetings 

and workshops among forest managers, conservationists, 

policymakers, and scientists in the Northeast. 

Long-term research at Harvard Forest has informed many 

important conservation efforts, as well as policy and management 

decisions in the region (figure 2; Foster et al. 

2010). For instance, a review of land-ownership history 

and conservation patterns led to the creation of the North 

Quabbin Partnership and to an increase of conservation 

Figure 2. Harvard Forest research and conservation linkages: primary research and synthesis examples related to the 

Wildlands and Woodlands Initiative. 


Articles 

land in the region from 36.8% in 1993 to 45.1% in 2010 

(Golodetz and Foster 1997). Research on the potential for 

local constraints on forestry to displace harvesting pressures 

to other, more sensitive parts of the world has broadened 

public acceptance of forestry in the region (Berlik et al. 2002). 

Surveys have documented how underrepresented old-growth 

forests are in southern New England, which has aided in the 

preservation of the few remaining sites (Orwig et al. 2001, 

D’Amato et al. 2006). Many of these linkages grew out of 

strong informal ties between scientists and stakeholders built 

by serving on local, state, and regional committees. 

In 2005, the Harvard Forest launched its Wildlands and 

Woodlands (W&W) Initiative, which emphasizes decisionrelevant 

synthesis, communication, and stakeholder partnerships. 

The knowledge gained from dozens of studies at 

the Harvard Forest was synthesized into a series of W&W 

publications that were aimed at nonscientists and that called 

for stemming the loss of forest cover now occurring in all 

six New England states as large areas (e.g., in Maine) experience 

significant shifts in landownership. The publications 

call for balancing the preservation of wildlands with large 

areas of actively managed woodlands and for promoting 

civic engagement through landowner-conceived woodland 

councils (Foster et al. 2005, 2010). 

Since 2005, the W&W Initiative has produced two major 

reports, two update publications, and a Web site (www. 

wildlandsandwoodlands.org), with the purpose of raising 

awareness about the pace and consequences of land-cover 

change. Both W&W reports had extensive stakeholder input, 

and the second garnered comments from several hundred 

agency, nongovernmental-organization (NGO), landowner, 

and industry representatives. Harvard Forest has since 

teamed up with the nonprofit organization Highstead to 

form a partnership with more than 60 participating groups 

to sustain stakeholder engagement and to help implement 

the vision of the W&W Initiative. The reports were accompanied 

by press releases; webinars; stakeholder briefings; 

and, in May 2010, a public event with Harvard University’s 

Kennedy School of Government. 

Assessing the societal impact of Harvard Forest research 

over the past 100 years is beyond the scope of this case 

study. However, we compiled information on the impact 

of W&W communication to shed light on the value of this 

coordinated outreach effort. In the two months following 

its release, the 2010 report generated 137 media and 

newsletter stories and 62 visits per day to the new W&W 

Web site, including visitors from 35 countries from five 

continents. By contrast, Harvard Forest garnered 21 non- 

W&W news stories between 2008 and 2010. W&W authors 

participated in 21 briefings, presentations, and workshops 

in the nine months since publication, which expanded the 

project’s influence and reach. These W&W synthesis and 

communication efforts have contributed to several notable 

policy and management advances, including the decision by 

the state of Massachusetts to establish permanent wildland 

reserves, the introduction of a conservation-finance bill in 

the Massachusetts General Assembly to accelerate the pace 

of conservation, and the launching of an innovative effort 

to aggregate multiple parcels into a single project with the 

goal of conserving approximately 10,000 acres of forest in 

western Massachusetts. The W&W efforts also fueled new 

research, including the establishment of new long-term 

study plots across sites with diverse histories, ownership, and 

management objectives; and a new cross-site LTER proposal 

on the Future Scenarios of Forest Change (see Thompson 

et al. 2012 [in this issue]). 

Sustained research–management partnerships at the Andrews 

Forest LTER site. The H. J. Andrews Experimental Forest and 

LTER site in the Oregon Cascade Range contains many of 

the iconic and hotly debated elements of Pacific Northwest 

forests: old-growth trees; northern spotted owls; and cold, 

clear, fast streams. Societal conflicts over the future of the 

vast tracts of federal forestlands in the region have been 

profoundly affected by science findings from the Andrews 

Forest and, in turn, have strongly influenced the course of 

science in the region and more broadly. 

The research history of the Andrews Forest, stretching 

back to its establishment in 1948, reflects a commitment 

to long-term ecological and watershed research by the US 

Forest Service and with NSF-funded programs under the 

International Biological Program in the 1970s, followed by 

LTER Network since 1980. These integrated science programs 

have produced high-quality studies and long-term 

records that underpin interpretations of ecosystem and 

environmental change and sustain an interdisciplinary cadre 

of scientists, all of whom are essential in investigating ecosystems 

that change abruptly and also gradually over time 

scales of decades and centuries. The context of extensive 

federal forestlands (e.g., US Forest Service, US Bureau of 

Land Management) provides an audience of land managers 

who are required to guide management using current 

science. And, if they fail to do so, litigants and the courts 

remind them. 

A central feature of the Andrews Forest program is a 

research–management partnership that develops, tests, demonstrates, 

and critically evaluates alternative approaches to 

management so that when the policy window opens, new, 

scientifically and operationally credible approaches to management 

are ripe for broad adoption (http:// andrewsforest. 

oregonstate.edu/resmgt.cfm?topnav=35). This partnership 

involves the research community centered on the Andrews 

Forest LTER site and land managers of the Willamette 

National Forest. The partnership has made substantial 

impacts on forest management and policy on topics such 

as the characteristics of and conservation strategies for oldgrowth 

forest ecosystems (Franklin et al. 1981, Spies and 

Duncan 2009); the ecological roles and management implications 

of dead wood on land and in streams (Gregory et al. 

1991); the ecology and population dynamics of the northern 

spotted owl (Forsman et al. 1984); the effects of forest cutting 

and roads on streamflow, including floods (Jones 2000); 


Articles 

Figure 3. Andrews Experimental Forest research and links with management and policy for forestlands and watersheds. 

interactions of road and stream networks (Jones et al. 2000); 

and interactions of climate change with management and 

policy (Spies et al. 2010). 

With roots in the early 1950s and the assignment of the first 

scientist to the Andrews Forest, the research–management 

partnership has become a continuous, place-based learning 

program with balanced, reciprocal communication 

between the management and research communities and 

their respective cultures. To facilitate communication, the 

research–management interface is staffed with a research 

liaison position at the Willamette National Forest, which 

facilitates outreach to land managers and the public. The 

technical findings of research and management experience 

are communicated through diverse media, such as journal 

articles, including ones jointly composed by scientists and 

land managers (Cissel et al. 1999); publications prepared for 

land managers and the general public (e.g., in the Science 

Findings and Science Update series of the US Forest Service; 

www.fs.fed.us/pnw/publications/scifi.shtml, www.fs.fed.us/ 

pnw/publications/sci-update.shtml); workshops; and field 

tours. In some cases, social scientists have examined the 

effectiveness of communications on challenging topics, such 

as the use of historic disturbance regimes to guide future 

land management (Shindler and Mallon 2009). The net 

effect of this communication program is a continuing public 

discussion of the future of forest and watershed management 

and policy in the region. 

The impacts of long-term research from the Andrews 

Forest and the research–management partnership are 

manifest in federal agencies’ management of forest stands 

and landscapes throughout the Pacific Northwest and more 

broadly (figure 3). In particular, the Northwest Forest 

Plan, which drew heavily from research from the Andrews 

Forest, ushered in a new era of ecosystem-based management 

on 10 million hectares of federal lands in northern 

California and western Oregon and Washington (FEMAT 

1993, USDA and USDI 1994). Andrews Forest–based science 

on old-growth forests, forest–stream interactions, aspects 

of biodiversity, and the roles of dead wood in forests and 

streams helped shape new federal land-management policies 

(FEMAT 1993). Several of the key publications have been 

cited in the scientific literature more than 1000 times each, 

which indicates the influence of the concepts in the environmental 

sciences. Since 1994, individual research themes have 

continued to influence management practices in the region 

( figure 3). Publications from the Andrews Forest–based work 

are widely cited in planning documents for timber sales and 

fuel-treatment projects on National Forests and US Bureau 

of Land Management districts across the Pacific Northwest. 

The impact of the research–management partnership has 

drawn social scientists to examine the dynamics, motivations, 

and public perception of these science–management– 

policy connections (Lach et al. 2003). 

Tools for assessing services and values to improve urban naturalresources 

stewardship: Baltimore Ecosystem Study long-term 

data. Information on natural resources in urban areas is 

often lacking and limits the ability of planners and managers 


Articles 

to properly steward or incorporate natural-resources services 

within urban ecosystems. Long-term research is currently 

being conducted in the Baltimore area to foster a better 

understand how urbanization affects natural system processes 

(e.g., Pickett and Cadenasso 2006). Baltimore, through 

its participation in the LTER Network, was one of the first 

cities to have its entire forest and tree structure assessed, 

along with the concomitant ecosystem services and values 

(e.g., pollution removal, carbon storage and sequestration, 

effects on building energy use; see, e.g., Nowak et al. 2008). 

It is also the first city (along with Syracuse, New York) to 

establish (in 1999) permanent vegetation-monitoring plots 

to assess long-term vegetation changes (Nowak et al. 2004). 

These data provide critical information for better understanding 

of urban vegetation systems, their environmental 

effects, and how these ecosystems are changing. These data 

have also helped in the development and testing of publicdomain 

software tools designed to aid managers and the 

general public in assessing urban trees and their associated 

ecosystem services and values. Data collected in Baltimore 

and other cities in the mid to late 1990s led to the development 

of software to assess urban forest structure and functions: 

the UFORE (urban forest effects) model (Nowak and 

Crane 2000). Through time, a diverse collaboration developed 

among numerous partners to expand the development 

of this and other urban forest computer programs into a 

suite of free software tools known as i-Tree (www.itreetools. 

org), which was released in 2006. 

The information provided by i-Tree software has been 

used to inform management and policies throughout 

the world in relation to urban forestry. The influence of 

i-Tree results from the use of the model and local data 

by consultants, managers, and local citizens to guide 

management and policies decisions related to issues such 

as emerald ash borer protection (Siyver 2009), building 

financial support for urban forestry programs (Society 

of Municipal Arborists 2008), linking local tree data 

with the US Conference of Mayors Climate Protection 

Agreement (Hyde 2009), public outreach campaigns (e.g., 

billboards) on the benefits of trees (Siyver 2009), developing 

urban forest strategic management plans (McNeil 

and Vava 2006), and helping secure financing for tree 

planting and management (e.g., Ibrahim 2009). Most of 

the data collected and analyzed through i-Tree are used to 

encourage municipal, county, and state leaders to establish 

or improve urban forestry programs, to recognize 

the role that trees play among urban natural resources, 

and to focus funds to improve stewardship. New tools 

were released in early 2010 (version 4.0) that include new 

approaches to help integrate science into local policy decisions 

related to streamflow, tree pests, local tree cover and 

effects, and related ecosystem services. 

Information and results from i-Tree, its analyses, and 

impacts are generally communicated by the research partners 

and users to others through public presentations, 

reports and articles, webinars, the i-Tree Web site, and word 

of mouth. To assist in communicating project results, i-Tree 

automatically produces a standard report with graphics that 

users can export and customize for their own use (figure 4). 

Users can also report ideas, questions, or problems back to 

the i-Tree team, which are then used to update or develop 

future versions. 

To date, more than 8200 unique users in 99 countries have 

downloaded the software. Use of i-Tree has grown at about 

30% per year since its release in August 2006. i-Tree Web site 

traffic has increased about tenfold since the release of version 

3.0 in June 2009 and continues to increase. Currently, 

about 20,000 unique users access the Web site every three 

months. Focused surveys of users have been conducted to 

help determine the types of impacts. Between 50 and 100 

journal articles and reports have been published in which 

i-Tree was used, and the numbers have increased annually. 

New programs in development are focused on temporal and 

spatial modeling of forest effects, and the Baltimore longterm 

permanent field-plot data are critical to the development 

of these new tools. International urban forest data 

standards are also in development to aid in sharing and in 

the use of the programs among nations. 

Climate-change impacts on wildfire: Bonanza Creek engagement 

with fire managers and indigenous communities. Alaska is 

warming twice as quickly as the global average, with little 

change in precipitation (Chapin et al. 2006a). The resulting 

drying of the boreal forest has increased the annual area 

burned, primarily through increased frequency of dry years 

and larger wildfires, which have important consequences for 

changes in forest cover and the closely coupled human and 

ecological communities (figures 5 and 6; Kofinas et al. 2010). 

Bonanza Creek scientists collaborate with fire managers and 

indigenous communities to share knowledge for predicting 

and adapting to changing fire regimes. 

Working with fire managers, Bonanza Creek LTER ecologists 

have developed predictive models that provide a scientific 

foundation for fire-management decisions. Spatially 

explicit models of climate and wildfire suggest that, by 2050, 

a “typical” fire year in interior Alaska will be similar to the 

most extreme fire years in the historical record (www.snap. 

uaf.edu). These models were developed through extensive 

input from climatologists, ecologists, and fire managers 

(Duffy et al. 2005). 

At the community level, village tribal councils have invited 

Bonanza Creek ecologists to collaborate in developing new 

ecosystem-management strategies to respond to increasing 

wildfire risk. These strategies include the sustainable harvest 

of flammable black spruce stands near communities to heat 

public buildings, create new jobs, and generate secondary 

successional habitat that favors moose—an important food 

source (Chapin et al. 2008, Kofinas et al. 2010). Bonanza 

Creek social scientists and ecologists have also participated 

in federally mandated community wildfire protection planning 

by conducting interviews and focus groups among 

local residents and resource managers. These interviews 


Figure 4. Example of the user interface for i-Tree. The page shows the i-Tree canopy survey page for urban forests. 

demonstrated that local residents trusted managers to plan 

community-level wildfire protection but felt disenfranchised 

in regional wildfire planning for the surrounding 

lands, because their knowledge and concerns about future 

subsistence opportunities and places of cultural value were 

overlooked (Ray 2010). 

Fire-modeling results are communicated to fire managers 

and the public through participation in annual wildfire 

strategic-planning workshops, agency meetings with the 

public, joint agency LTER planning of prescribed burns, and 

production of site-specific 2-kilometer-resolution climate 

projections and fire-risk projections on request (www.snap. 

uaf.edu). 

Community workshops coorganized by tribal councils 

and Bonanza Creek ecologists allow an exchange of local, 

traditional, and scientific knowledge about wildfire ecology. 

This dialogue has enriched understanding by the LTER scientists 

of the ecological and societal consequences of climate 

change. The trust that develops through community partnerships 

enables Bonanza Creek researchers to learn from 

Articles 

and contribute to societal responses to a rapidly changing 

socioecological environment. 

The Bonanza Creek LTER site is forging new ground in 

identifying climate-change impacts that require immediate 

management and community action. The associated metrics 

of impact are therefore recent and qualitative. Judging from 

the ESA’s Sustainability Science Award for Chapin and colleagues’ 

(2006b) article, in which they described the socioecological 

framework for this research, the Bonanza Creek 

LTER site is contributing to fundamental science and to new 

approaches for integrating community knowledge and concerns 

in socioecological research (figure 6). Bonanza Creek 

collaborations contributed to Alaska fire managers’ capacity 

to adapt federal guidelines on the basis of fire issues of the 

lower 48 US states to conditions and issues that are relevant 

to Alaska. Managers use the fire-risk model and routinely 

invite Bonanza Creek ecologists to participate in the training 

of wildfire managers, which indicates that they value the 

practical relevance of LTER. Bonanza Creek ecologists and 

Alaskan indigenous leaders have formed the Working Group 


Articles 

Ecoregions 

Subarctic 

tundra 

Arctic tundra 

Boreal forest 

Coastal rainforest 

on Rural Alaska Self-Reliance, a collaboration to implement 

community visions of adaptation to global change. This collaboration 

suggests that indigenous leaders value and trust 

their interactions with Bonanza Creek scientists. 

Discussion of the case studies. Boundary-spanning efforts can 

facilitate the bridging of science and society by producing 

information that is salient, credible, and legitimate (Cash 

et al. 2003, McNie et al. 2008), which ultimately enriches 

scientific research through stakeholder engagement, the 

expansion of public awareness, and the improvement of 

the scientific basis for decisionmaking. The five LTER case 

studies presented here offer experiences and lessons to help 

answer the question of what characterizes successful collaborative 

outreach efforts. The case studies suggest that 

efforts to build a stronger interface between science and 

society are shaped in part by three overarching attributes 

that pertain to all ecosystems but vary in detail among 

ecosystems: (1) Landscape and social context refers to the 

pattern of land ownership (e.g., private versus public) and 

the role of the different types of knowledge (e.g., local versus 

expert) that influence the framing of environmental issues, 

the management objectives, and the science used in the 

decisionmaking process. (2) Issue definition involves 

Indigenous Languages and Peoples 

Central Yup'ik 

Deg Xinag 

Eyak–Tlingit 

Tlingit 

determining the relevance of particular long-term research to 

policy and management issues at local, regional, or national 

scales (e.g., local fire- or fuel-management issues, regional 

air-quality concerns, federal forestland policy) and the extent 

to which individual actions or government actions are central 

to resolving the issues of concern. (3) Communication 

pathways entail understanding which communication approaches 

are most appropriate for specific decisionmakers, 

and the choice of pathway is determined in part by the 

context and issues addressed (e.g., direct briefings between 

scientists and policymakers; outreach to the media; working 

groups with managers; discussions with local communities, 

including tribes). 

In addition to these three overarching attributes that 

distinguish individual efforts, a set of common elements of 

successful science communication efforts emerges from the 

case studies: 

In all of the case studies, boundary-spanning efforts were 

built on credible, multidecade, interdisciplinary science, and 

peer-reviewed publications. These efforts combine retrospective 

analysis; long-term measurements and experiments; 

quantitative modeling; and, increasingly, scenarios planning. 

For example, the ability of the HBRF Science Links projects 

to assess the impacts of air-quality regulations and the extent 


Ahtna 

Eyak 

Haida 

Tsimshian 

Eskimo–Aleut Athabascan 

Iñupiaq 

Gwich'in Holikachuk 

Unanga{ (Aleut) 

Sugpiaq (Alutiiq) Dena'ina Tanacross 

St. Lawrence Island Yupik 

Hän 

Upper Kuskokwim 

Upper Tanana 

Lower Tanana 

Denaakk'e 

(Koyukon) 

Map by C. West, M. Wilson and J. Kerr - ISER 

Figure 5. Cultural (linguistic) groups and ecoregions are closely coupled in the boreal forest region of Alaska (Chapin 

2009). The Bonanza Creek Long Term Ecological Research Network site uses understanding of socioecological responses 

to climate change as a platform for exploring and implementing adaptation options that rural Athabascan communities 

would find consistent with their history and current commitment to sustainable subsistence lifestyles. Source: Reprinted 

from F. Stuart Chapin III, “Managing ecosystems sustainably: The key role of resilience.” Pages 29–53 in Chapin FS III, 

Kofinas GP, Folke C, eds. Principles of Ecosystem Stewardship: Resilience-Based Natural Resource Management in a 

Changing World (2009), with permission from Springer.

Community concern 

about global change 

Informal 

input 

Informal dialogue 

LTER or community 

request for collaboration 

Tribal council approval 

Focus groups, interviews 

and knowledge sharing 

based on shared concerns 

Revisions Initial results 

Community review 

LTER interest in 

Socioecological change 

Formal 

input 

Community actions Joint publications 

Figure 6. Processes of interaction between the Bonanza 

Creek Long Term Ecological Research (LTER) Network site’s 

scientists and indigenous communities in sharing ecological 

knowledge. Informal discussions between scientists and 

community members lead to a formal request to the village 

tribal council to explore specific questions (e.g., climatechange 

effects on fire regime). One or more rounds of 

interaction involving interviews or focus groups, a review 

of the findings by the community, and a revision based 

on feedback. This leads to a formal report to the village 

council and joint publications by scientists and community 

members, as well as informal input to the village council, 

other community members, and LTER scientists. 

to which ecosystems have recovered was entirely dependent 

on the existence of long-term precipitation, soil and stream 

chemistry, and relevant biological measurements. These data 

enabled scientists to analyze changes in atmospheric deposition 

and associated chemical and biological responses, to 

establish impact thresholds, and to apply dynamic models 

to evaluate the extent to which future emissions reductions 

would achieve policy objectives. 

Among the most important activities is the collaboration 

of scientists and decisionmakers at the outset of and 

throughout a research effort. This interaction helps to define 

issues and questions salient to decisionmakers, to identify 

sources of knowledge beyond traditional scientific data sets, 

and to envision outputs that best meet user needs. This 

Articles 

process also enriches scientific research. For example, the 

Bonanza Creek research framework was expanded through 

interactions with community groups to larger temporal and 

spatial scales and integration of cultural dimensions. This 

led to the recognition of critical thresholds of the resistance 

of the boreal socioecological system to climatic and socioeconomic 

changes. 

Although scientists are accustomed to publishing focused 

research in peer-reviewed publications, these case studies 

point clearly to the need for policy- and managementrelevant 

synthesis and distillation to support the effective 

use of science in the policy and management processes. This 

problem-oriented synthesis is necessary but not sufficient 

for promoting knowledge sharing and should be accompanied 

by work to translate the key findings into compelling 

terms relevant to stakeholders. For example, by pulling 

together disparate findings from across dozens of articles 

produced over a decade or more, the Harvard Forest W&W 

publications have drawn public and stakeholder attention to 

that body of work and catalyzed conservation initiatives far 

beyond what any single study could accomplish. 

Successful outreach should not be an afterthought but 

a major and well-funded initiative with adequate staffing 

and supporting expertise ranging from traditional print 

publications to media, including Web-based, outreach. 

Innovative online tools that promote interaction and social 

networking and that are open source and easily accessible 

are increasingly important communication vehicles. For 

example, the i-Tree project built a program interface that 

is easy to use, open to all, supported, and free: The i-Tree 

partnership has built a platform to which others can contribute, 

and new peer-reviewed tools can be added and 

then supported through the existing i-Tree partnership and 

model structure. 

Partnerships are critical to sustaining reciprocal flow of 

information among scientists, citizen leaders, managers, and 

policymakers; to applying scientific findings to policy and 

management through an adaptive process; and to fueling 

processes in which stakeholder experiences and knowledge 

inform research. For example, the research–management 

partnership developed by the Andrews LTER site and the 

US Forest Service provides a platform for sustained, placebased 

learning with substantial attention to communications 

with many audiences. Similarly, the Baltimore LTER 

i-Tree project also functions as a partnership that meets 

regularly and has open discussions, working toward meeting 

the needs of the urban community. In both cases, the 

involvement of a public entity (the US Forest Service) has 

been instrumental in coordinating and managing the activities 

of the partnerships. 

In addition to these lessons, the case studies presented 

here demonstrate the need for stronger metrics to measure 

the impact of science communications and outreach to 

decisionmakers. In general, metrics for evaluating publicengagement 

outreach efforts can be divided into three 

categories: output, uptake, and impact. The five case studies 


Articles 

present outputs such as the number of publications and 

presentations given to nonscientific audiences. They also 

provide strong evidence of uptake such as media coverage, 

scientific citations, and Web site visitation. Quantifying the 

impact on policy, conservation, and stewardship decisions 

remains elusive. 

Developing and applying meaningful metrics of impact 

is a common challenge. Under the auspices of the White 

House Office of Science and Technology Policy, the National 

Institutes of Health and the NSF are developing metrics of 

impacts for science, called STAR METRICS (Science and 

Technology for America’s Reinvestment: Measuring the 

Effects of Research on Innovation, Competitiveness, and 

Science; Lane and Bertuzzi 2011). Several metrics have been 

proposed to measure the usefulness of scientific knowledge, 

many of which are applied in these case studies (e.g., 

download or hit rates, media coverage, citations in federal 

or state regulations). But in the area of broader impacts or 

social outcomes, such as those in health, safety, and the environment, 

recommendations are under development by an 

interagency working group. Impact metrics for science are 

an important gap in understanding that should be remedied 

by the STAR METRICS program and other science-policy 

research efforts. 

Conclusions 

If science is to aid in the advance toward a more resilient and 

sustainable society, we must experiment with more effective 

means of integrating ecological research and decisionmaking. 

As is evidenced by the five case studies presented 

here for forest ecosystems and by many other examples, 

the LTER Network has an important and unique role to 

play in addressing the grand challenges in environmental 

and sustainability science. The LTER Network and associated 

research, with its long-term interdisciplinary focus, 

its focus on place-based study, its geographic distribution, 

its sophisticated information-management systems, and its 

public-outreach capabilities, are well suited to boundaryspanning 

initiatives that address emerging environmental 

issues related to changes in biogeochemistry, biological 

diversity, climate change, ecohydrology, infectious disease, 

and land use. Policy-relevant synthesis and science communication 

should be a focus of the LTER Network, and these 

activities, in turn, would probably promote cross-site and 

network-wide coordination of matters important to both 

science and society. This work could be enhanced by partnerships 

with established scientific societies that are dedicated 

to similar work. For example, the ESA is advancing a 

partnership among academic societies, agencies, and NGOs 

“to foster Earth Stewardship by (a) clarifying the science 

needs for understanding and shaping trajectories of change 

at local-to-global scales; (b) communicating the basis for 

Earth Stewardship to a broad range of audiences, including 

natural and social scientists, students, the general public, 

policymakers, and other practitioners; and (c) formulating 

pragmatic strategies that foster a more sustainable trajectory 

of planetary change by enhancing ecosystem resilience and 

human well-being” (Chapin et al. 2011, p. 45). 

Harnessing the power of long-term ecological studies 

to address the grand challenges in environmental science 

will require learning from and building on existing efforts 

to better integrate scientific research with societal concerns. 

The NSF can facilitate this process by expanding the 

bounds of informal education to include the engagement of 

decisionmakers and journalists in order to provide the 

requisite research and learning needed to develop, test, and 

expand these critical experiments at the interface of science 

and society. 


The effort and contributions of KFL were generously supported 

by the Bullard Fellowship program of Harvard 

University’s Harvard Forest and by a grant from Highstead. 

This is a contribution of the Long Term Ecological Research 

Network, which is supported by the US National Science 

Foundation and by the US Forest Service. We appreciate 

the efforts of David R. Foster in shepherding this series of 

articles. 


Berlik MM, Kittredge DB, Foster DR. 2002. The illusion of preservation: 

A global environmental argument for the local production of natural 

resources. Journal of Biogeography 29: 1557–1568. 

Cash DW, Clark WC, Alcock F, Dickson NM, Eckley N, Guston DH, 

Jäger J, Mitchell RB. 2003. Knowledge systems for sustainable development. 

Proceedings of the National Academy of Sciences 100: 

8086–8091. 

Chapin FS III. 2009. Managing ecosystems sustainably: The key role of 

resilience. Pages 29–53 in Chapin FS III, Kofinas GP, Folke C, eds. 

Principles of Ecosystem Stewardship: Resilience-Based Natural Resource 

Management in a Changing World. Springer. 

Chapin FS III, Oswood MW, Van Cleve K, Viereck LA, Verbyla DL, eds. 

2006a. Alaska’s Changing Boreal Forest. Oxford University Press. 

Chapin FS III, Lovecraft AL, Zavaleta ES, Nelson J, Robards MD, Kofinas 

GP, Trainor SF, Peterson GD, Huntington HP, Naylor RL. 2006b. Policy 

strategies to address sustainability of Alaskan boreal forests in response 

to a directionally changing climate. Proceedings of the National 


Chapin FS III, et al. 2008. Increasing wildfire in Alaska’s boreal forest: 

Pathways to potential solutions of a wicked problem. BioScience 58: 

531–540. 

Chapin FS III, Pickett STA, Power ME, Jackson RB, Carter DM, Duke C. 

2011. Earth stewardship: A strategy for social-ecological transformation 

to reverse planetary degradation. Journal of Environmental Studies and 

Sciences 1: 44–53. 

Cissel JH, Swanson FJ, Weisberg PJ. 1999. Landscape management using 

historical fire regimes: Blue River, Oregon. Ecological Applications 9: 

1217–1231. 

D’Amato AW, Orwig DA, Foster DR. 2006. New estimates of Massachusetts 

old-growth forests: Useful data for regional conservation and forest 

reserve planning. Northeastern Naturalist 13: 495–506. 

Driscoll CT, Lawrence GB, Bulger AJ, Butler TJ, Cronan CS, Eagar C, 

Lambert KF, Likens GE, Stoddard JL, Weathers KC. 2001. Acidic deposition 

in the northeastern United States: Sources and inputs, ecosystem 

effects, and management strategies. BioScience 51: 180–198. 

Driscoll CT, Cowling EB, Grennfelt P, Galloway JM, Dennis RL. 2010. 

Integrated assessment of ecosystem effects of atmospheric deposition: 

Lessons available to be learned. EM Magazine November 2010: 6–13. 


Driscoll CT, Lambert KF, Weathers KC. 2011. Integrating science and policy: 

A case study of the Hubbard Brook Research Foundation Science Links 

program. BioScience 61: 791–801. 

Duffy PA, Walsh JE, Graham JM, Mann DH, Rupp TS. 2005. Impacts of 

large-scale atmospheric–ocean variability on Alaskan fire season severity. 

Ecological Applications 15: 1317–1330. 

[FEMAT] Forest Ecosystem Management Assessment Team. 1993. 

Forest Ecosystem Management: An Ecological, Economic, and Social 

Assessment. US Department of Agriculture Forest Service. 

Fisher RT. 1921. The management of the Harvard Forest, 1909–1919. 

Harvard Forest Bulletin 1: 1–27. (18 January 2012; http://harvardforest. 

fas.harvard.edu/publications/pdfs/HFpubs/b1.pdf ) 

———. 1933. New England forests: Biological factors. Pages 213–223 in 

Wood W, ed. New England’s Prospect. American Geographical Society. 

Special Publication no. 16. 

Forsman ED, Meslow EC, Wight HM. 1984. Distribution and biology of the 

spotted owl in Oregon. Wildlife Monographs no. 87: 3–64. 

Foster DR. 2000. From bobolinks to bears: Interjecting geographical history 

into ecological studies, environmental interpretation, and conservation 

planning. Journal of Biogeography 27: 27–30. 

Foster DR, Orwig DA. 2006. Preemptive and salvage harvesting of New 

England forests: When doing nothing is a viable alternative. Conservation 

Biology 20: 959–970. 

Foster DR, Aber JD, Melillo JM, Bowden RD, Bazazz FA. 1997. Forest 

response to disturbance and anthropogenic stress: Rethinking the 1938 

hurricane and the impact of physical disturbance vs. chemical and 

climate stress on forest ecosystems. BioScience 47: 437–445. 

Foster D[R], Kittredge D[B], Donahue B, Motzkin G, Orwig D[A], Ellison 

AM, Hall B, Colburn EA, D’Amato A. 2005. Wildlands and Woodlands: 

A Vision for the Forests of Massachusetts. Harvard University, Harvard 

Forest. 

Foster DR, Aber JD, Cogbill CV. 2010. Wildlands and Woodlands: A Vision 

for the New England Landscape. Harvard University, Harvard Forest. 

Franklin JF, Cromack K Jr, Denison W, McKee A, Maser C, Sedell J, Swanson 

F, Juday G. 1981. Ecological characteristics of old-growth Douglas-fir 

forests. US Department of Agriculture, Forest Service, Pacific Northwest 

Forest and Range Experiment Station. General Technical Report 

no. PNW-GTR-118. 

Golodetz AD, Foster DR. 1997. History and importance of land use and 

protection in the North Quabbin Region of Massachusetts (USA). 

Conservation Biology 11: 227–235. 

Gregory SV, Swanson FJ, Mckee WA, Cummins KW. 1991. An ecosystem 

perspective of riparian zones. BioScience 41: 540–551. 

Hall B, Motzkin G, Foster DR, Syfert M, Burk J. 2002. Three hundred 

years of forest and land-use change in Massachusetts, USA. Journal of 

Biogeography 29: 1319–1335. 

Halpern CB, Haugo RD, Antos JA, Kaas SS, Kilanowski AL. 2012. 

Grassland restoration with and without fire: Evidence from a 

tree-removal experiment. Ecological Applications. Forthcoming. 

doi:10.1890/11-1061.1 

Harmon ME, et al. 1986. Ecology of coarse woody debris in temperate 

ecosystems. Pages 133–302 in MacFadyen A, Ford ED, eds. Advances in 

Ecological Research, vol. 15. Academic Press. 

Hyde G. 2009. President’s message. City Trees: Journal of the Society of 

Municipal Arborists 45 (6): 5. 

Ibrahim M. 2009. US benefit analysis snared over $220m for trees. 

Horticulture Week 01 October 2009. (18 January 2012; www.hortweek. 

com/news/942310). 

Jones JA. 2000. Hydrologic processes and peak discharge response to forest 

removal, regrowth, and roads in 10 small experimental basins, western 

Cascades, Oregon. Water Resources Research 36: 2621–2642. 

Jones JA, Swanson FJ, Wemple BC, Snyder KU. 2000. Effects of roads on 

hydrology, geomorphology, and disturbance patches in stream networks. 

Conservation Biology 14: 76–85. 

Kittredge DB, Finley AO, Foster DR. 2003. Timber harvesting as ongoing 

disturbance in a landscape of diverse ownership. Forest Ecology and 


Articles 

Kizlinski ML, Orwig DA, Cobb RC, Foster DR. 2002. Direct and indirect 

ecosystem consequences of an invasive pest on forests dominated by 

eastern hemlock. Journal of Biogeography 29: 1489–1503. 

Knapp AK, et al. 2012. Past, present, and future roles of long-term experiments 

in the LTER network. BioScience 62: 377–389. 

Kofinas GP, Chapin FS [III], BurnSilver S, Schmidt JI, Fresco NL, Kielland 

K, Martin S, Springsteen A, Rupp TS. 2010. Resilience of Athabascan 

subsistence systems to interior Alaska’s changing climate. Canadian 

Journal of Forest Research 40: 1347–1359. 

Lach D, List P, Steel B, Shindler B. 2003. Advocacy and credibility of ecological 

scientists in resource decisionmaking: A regional study. BioScience 

53: 170–178. 

Lane J, Bertuzzi S. 2011. Measuring the results of science investments. 

Science 331: 678–680. 

Levitt JN, Lambert KF. 2006. Report on the Woodlands and Wildlands 

Conservation Finance Roundtable. Harvard University, Harvard 

Forest. 

Lovett GM, Burns DA, Driscoll CT, Jenkins JC, Mitchell MJ, Rustad L, 

Shanley JB, Likens GE, Haeuber R. 2007. Who needs environmental 

monitoring? Frontiers in Ecology and the Environment 5: 

253–260. 

[LTER Network] Long Term Ecological Research Network. 2010a. LTER 

Strategic Communication Plan: Bridging to Broader Audiences. LTER 

Network. (8 February 2012; http://intranet2.lternet.edu/sites/intranet2. 

lternet.edu/files/documents/LTER%20History/Planning%20Documents/ 

Final%20LTER%20Strategic%20Communication%20Plan%20-%20 

Nov%2011%202010.pdf ) 

———. 2010b. Strategic and Implementation Plan. LTER Network. 

(18 January 2012; http://intranet2.lternet.edu/documents/lter-strategicand-implementation-plan) 

Lubchenco J. 1998. Entering the century of the environment: A new social 

contract for science. Science 279: 491–497. 

Mayer AL, Kauppi PE, Angelstam PK, Zhang Y, Tikka PM. 2005. Importing 

timber, exporting ecological impact. Science 308: 359–360. 

McDonald RI, Motzkin G, Foster DR. 2008. The effect of logging on 

vegetation composition in western Massachusetts. Forest Ecology and 

Management 255: 4021–4031. 

McNeil J, Vava C. 2006. Oakville’s urban forest: Our solution to our 

pollution. Town of Oakville, Parks and Open Space Department, 

Forestry Section. (8 February 2012; www.itreetools.org/resources/reports/ 

Oakville%27s%20Urban%20Forest.pdf ) 

McNie EC, van Noordwijk M, Clark WC, Dickson NM, Sakuntaladewi 

N, Suyanto JL, Joshi L, Leimona B, Hairiah K, Khususiyah N. 2008. 

Boundary Organizations, Objects and Agents: Linking Knowledge 

with Action in Agroforestry Watersheds. Center for International 

Development at Harvard University and World Agroforestry Centre. 

Working Paper no. 80. 

Nowak DJ, Crane DE. 2000. The Urban Forest Effects (UFORE) Model: 

Quantifying urban forest structure and functions. Pages 714–720 

in Hansen M, Burk T, eds. Integrated Tools for Natural Resources 

Inventories in the 21st Century. US Department of Agriculture, Forest 

Service, North Central Research Station. General Technical Report 

no. NC-212. 

Nowak DJ, Kurodo M, Crane DE. 2004. Urban tree mortality rates and tree 

population projections in Baltimore, Maryland, USA. Urban Forestry 

and Urban Greening 2: 139–147. 

Nowak DJ, Crane DE, Stevens JC, Hoehn RE, Walton JT, Bond J. 2008. 

A ground-based method of assessing urban forest structure and ecosystem 

services. Arboriculture and Urban Forestry 34: 347–358. 

Nudel M. 2003. Better conservation through partnerships. Exchange: 

Journal of the Land Trust Alliance 22: 17–21. 

Orwig DA, Cogbill CV, Foster DR, O’Keefe JF. 2001. Variations in oldgrowth 

structure and definitions: Forest dynamics on Wachusett 

Mountain, Massachusetts. Ecological Applications 11: 437–452. 

Pickett STA, Cadenasso ML. 2006. Advancing urban ecological studies: 

Frameworks, concepts, and results from the Baltimore Ecosystem Study. 

Austral Ecology 31: 114–125. 


Articles 

Ray L. 2010. Can Community-Based Natural Resource Management 

Improve Wildfire Policy Planning in Interior Alaska? Addressing Value 

Differences, Ineffective Participatory Processes, and Conflicts over 

Traditional Ecological Knowledge. Doctoral Dissertation. Department 

of Geography. Clark University, Worcester, Massachusetts. 

Shindler B, Mallon AL. 2009. Public Acceptance of Disturbance-Based 

Forest Management: A Study of the Blue River Landscape Strategy 

in the Central Cascades Adaptive Management Area. US Department 

of Agriculture, Forest Service, Pacific Northwest Research Station. 

Research Paper no. PNW-RP-581. 

Siyver D. 2009. Milwaukee UFORE: Taking it to the streets. City Trees: 

Journal of the Society of Municipal Arborists 45 (5): 10–12. 

Society of Municipal Arborists. 2008. More than cows and cowboys: Urban 

forestry in Casper, Wyoming. City Trees: Journal of the Society of 

Municipal Arborists Sept/Oct: 44(5): 6–10. 

Spies TA, Duncan SL, eds. 2009. Old Growth in a New World: A Pacific 

Northwest Icon Reexamined. Island Press. 

Spies TA, Giesen TW, Swanson FJ, Franklin JF, Lach D, Johnson KN. 2010. 

Climate change adaptation strategies for federal forests of the Pacific 

Northwest, USA: Ecological, policy, and socio-economic perspectives. 

Landscape Ecology 25: 1185–1199. 

BRINGING 

BIOLOGY 

JOIN 

AIBS 

www.aibs.org 

AIBS INDIVIDUAL MEMBERS 

RECEIVE PRINT AND/OR 

ELECTRONIC ACCESS 

TO BIOSCIENCE. AIBS 

ORGANIZATIONAL MEMBERS 

PARTICIPATE IN THE NATIONAL 

DIALOGUE ABOUT HOW TO 

ENSURE A BRIGHT FUTURE 

FOR BIOLOGY. 




62: 367–376. 

[USDA and USDI] US Department of Agriculture, US Department of the 

Interior. 1994. Record of Decision for Amendments to Forest Service 

and Bureau of Land Management Planning Documents within the 

Range of the Northern Spotted Owl. USDA Forest Service. 

Charles T. Driscoll (ctdrisco@syr.edu) is affiliated with the Department of 

Civil and Environmental Engineering at Syracuse University, in New York. 

Kathleen F. Lambert and Clarisse M. Hart are affiliated with Harvard University’s 

Harvard Forest, in Petersham, Massachusetts. F. Stuart Chapin III is 

affiliated with the Institute of Arctic Biology at the University of Alaska, Fairbanks. 

David J. Nowak is affiliated with the US Department of Agriculture, 

Forest Service, Northern Research Station, in Syracuse, New York. Thomas A. 

Spies and Frederick J. Swanson are affiliated with the US Forest Service, Pacific 

Northwest Research Station, in Corvallis, Oregon. David B. Kittredge, Jr., is 

affiliated with the Department of Environmental Conservation at the University 

of Massachusetts–Amherst. 

TO INFORMED 

DECISION 

MAKING 

JOIN THE 

COMMUNITY... 

As an umbrella organization 

AIBS works with biologists and 

their professional organizations, 

to ensure that reliable 

information is used when 

decisions are made–in public 

policy, research funding, and 

the public forum. 

To learn more about our 

impact and to join AIBS, visit 

www.aibs.org. 

1900 CAMPUS COMMONS DR. 

STE 200 

RESTON, VA 20191 

WWW.AIBS.ORG 


Articles 

Scenario Studies as a Synthetic and 

Integrative Research Activity for 

Long-Term Ecological Research 

Jonathan R. thompson, aRnim Wiek, FRedeRick J. sWanson, stephen R. caRpenteR, nancy FResco, 

teResa hollingsWoRth, thomas a. spies, and david R. FosteR 

Scenario studies have emerged as a powerful approach for synthesizing diverse forms of research and for articulating and evaluating alternative 

socioecological futures. Unlike predictive modeling, scenarios do not attempt to forecast the precise or probable state of any variable at a given 

point in the future. Instead, comparisons among a set of contrasting scenarios are used to understand the systemic relationships and dynamics 

of complex socioecological systems and to define a range of possibilities and uncertainties in quantitative and qualitative terms. We describe five 

examples of scenario studies affiliated with the US Long Term Ecological Research (LTER) Network and evaluate them in terms of their ability 

to advance the LTER Network’s capacity for conducting science, promoting social and ecological science synthesis, and increasing the saliency of 

research through sustained outreach activities. We conclude with an argument that scenario studies should be advanced programmatically within 

large socioecological research programs to encourage prescient thinking in an era of unprecedented global change. 

Keywords: socioecological systems, science synthesis, participatory engagement, futures 

Public investment in science increasingly comes with the 

expectation that research will address the complex 

challenges that society faces and will contribute to strategies 

that sustain the vitality and integrity of socioecological 

systems (Reid et al. 2010). During the past 30 years, as the 

US National Science Foundation (NSF)–sponsored Long 

Term Ecological Research (LTER) Network has grown from 

6 to 26 research sites, ranging from Alaska to Antarctica and 

from urban ecosystems to forests to coral reefs. The LTER 

Network has increasingly used innovative approaches to 

connect science to society: from direct engagement with 

policymakers to educational programs in public schools 

and software applications that inform the public about the 

structure, function, and state of socioecological systems 

(Bestelmeyer et al. 2005, Driscoll et al. 2012 [in this issue], 

Robertson et al. 2012 [in this issue]). Scenario studies are 

one such approach and are emerging as a powerful tool for 

synthesizing the results of LTER science and for engaging 

with stakeholders to consider socioecological futures. There 

is a rich history involving the use of scenarios to encourage 

prescient thinking, largely born out of military and corporate 

planning (Kahn 1962, Wack 1985). Scenarios offer a 

framework for practitioners to integrate diverse modes of 

knowledge and to explicitly recognize those components 

of complex systems that are understood and those that are 

uncertain. In this way, scenario studies can describe multiple 

ways in which our shared socioecological future may unfold, 

sometimes including a visioning process that is focused on 

the specific attributes of one preferred future condition. 

Although the first type of scenario studies yield impartial 

descriptions of a range of possible future states, visioning 

describes desirable future states (or visions)—for instance, 

according to sustainability principles or stakeholder agreement 

(Swart et al. 2004, Carpenter and Folke 2006, Weaver 

and Rotmans 2006). Both types of scenario are contrasted 

with probability-based approaches to forward-looking 

research below. 

We refer here to scenario studies in general as any strategy 

for describing plausible future conditions while explicitly 

incorporating relevant science, societal expectations, and 

internally consistent assumptions about major drivers, relationships, 

and constraints (Xiang and Clarke 2003, Iverson 

Nassauer and Corry 2004, Raskin et al. 2005, Bolte et al. 2007, 

Mahmoud et al. 2009). Unlike predictive modeling, scenario 

studies acknowledge the uncertainty inherent in socioecological 

systems and therefore do not attempt to forecast the 

precise or probable state of any variable at a given point in 

the future. Instead, comparisons among a set of contrasting 

scenarios are used to understand the systemic interrelation 

and dynamics of complex socioecological systems 





Articles 

and to define a range of possibilities and uncertainties in 

quantitative and qualitative terms. The value of scenario 

studies lies in the process of embedding alternative states of 

the future into a transparent problem-solving framework 

(Swart et al. 2004). In this way, scenario studies may help 

to anticipate change in systems characterized by high levels 

of irreducible uncertainty and low levels of controllability 

(Bennett et al. 2003, Peterson et al. 2003a) and evoke new 

integrative perspectives and novel concepts of ecological 

change (Carpenter and Folke 2006, Carpenter et al. 2006). 

Although environmental scenarios are developed for a wide 

range of specific purposes, scenario studies can serve three 

widely accepted functions: education and public information, 

scientific exploration, and decision support and strategic 

planning (Alcamo and Henricks 2008, Henrichs et al. 

2010). 

In many large science- and environmental-assessment 

programs, scenarios have been used to describe and underpin 

analyses of alternative futures. The Intergovernmental 

Panel on Climate Change’s (IPCC) emission scenarios 

(Naki enovi and Swart 2000) and the Millennium Ecosystem 

Assessment’s scenarios (MA 2005) are perhaps the most well 

known, but there are many others (e.g., Sala et al. 2000, 

Raskin et al. 2005). Several LTER sites, for example, are 

deeply involved in scenario studies and even more plan to 

be. A recent LTER Network–sponsored workshop brought 

together 32 social and ecological researchers representing 

16 LTER sites from around the United States that were 

actively engaged in some aspect of scenario studies for the 

region surrounding their sites (Thompson and Foster 2009). 

The participants reaffirmed what may seem self-evident: 

LTER-based science has several characteristics amenable 

to developing regional socioecological scenarios. Credible 

socioecological scenarios require a deep understanding of 

long-term environmental dynamics—a signature strength 

of LTER science—and a tight coupling of ecological and 

social research—an emerging strength of and direction for 

the LTER Network (Collins et al. 2011, Robertson et al. 2012 

[this issue]). 

Looking forward to the next 30 years of LTER—and, 

more generally, to other research programs concerned 

with understanding socioecological systems—we argue 

that the application of scenario studies can advance prescient 

thinking in an era of unprecedented rates of global 

change. For example, an overarching goal set forth in the 

LTER Network’s Strategic and Implementation Plan is to 

use its deep understanding of complex socioecological systems 

to help anticipate ecological, evolutionary, and social 

responses to future environmental change and to inform 

societal strategies to adapt to this change (LTER Network 

2011). This aligns seamlessly with the primary functions of 

scenario studies (i.e., underpinning decision processes, collaborative 

learning, and scientific exploration). More specifically, 

scenarios can enable site-based research programs to 

explore possible as well as desirable and sustainable future 

states of their respective regions. And through sustained 

partnerships with land managers, policymakers, and other 

stakeholders, participatory scenarios can increase the societal 

relevance of their research. Moreover, in scenario studies 

in which the socioecological future of their regions is 

formally considered, new research needs can be identified 

while research in temporal and spatial dimensions is scaled 

up. Finally, developing and analyzing future scenarios 

would provide a platform for working with many scientific 

networks, including the National Ecological Observatory 

Network (NEON) and Urban Long Term Research Areas 

(ULTRA). 

Driscoll and colleagues (2012) explore the novel ways in 

which scientists have delivered their research findings to 

policymakers and managers where they can inform decisions. 

They describe several important communications 

approaches that allow new science to span boundaries and 

to address new challenges. In the present article, we examine 

scenario studies as one such boundary-spanning approach, 

using several examples of scenario studies from LTER sites 

in order to identify approaches that may be more broadly 

applicable. More specifically, we evaluate one emerging and 

four mature scenario studies in terms of three questions that 

address the value of scenarios for advancing socioecological 

science, programs, and outreach: 

A science question: How do the scenario studies relate the past to 

the future? Related to this question are those of what attributes 

of the socioecological system will change a lot, what 

attributes will change a little, and why. To articulate a range 

of alternative futures in a plausible and credible manner, it is 

necessary to understand the relevant history and trajectory 

of environmental and social change of the system of interest 

and its component parts. Evaluating future scenarios in 

light of recent changes, then, leads to an understanding of 

system attributes that are more or less resilient or vulnerable 

to future change. This feature of scenario studies is especially 

valuable as socioecological change approaches “tipping 

points” and the need to anticipate and mitigate future 

change becomes acute (Scheffer et al. 2001, Rockström et al. 

2009). 

A programmatic question: How do the scenario studies relate to the 

more traditional long-term science occurring within LTER? Related 

to this question is how they can advance science synthesis. 

Scenarios take many forms and, as is shown below, they may 

have narrow or expansive thematic scope. But in all the case 

studies, scenarios are either informed by or are used as a 

platform for applying the core long-term research coming 

from the individual LTER sites. Consequently, scenarios can 

be a compelling approach to science synthesis and for crosssite 

comparative analyses. 

An outreach question: How do the scenario studies affect the 

region and regional society through real-world changes or capacity 

building? Future scenarios draw together stakeholders 

affected by the hypothesized future changes. By engaging 


in scenario studies, scientists can place themselves in a new 

relationship with society. It can be a means for linking 

science and decisionmaking in order to support future 

and hopefully sustainable trajectories for socioecological 

systems, whether they are wild landscapes, natural resource 

lands, or urban areas. 

Case studies 

The case studies showcase diverse ways in which scenario 

studies are being employed as a vehicle to understand 

the drivers of socioecological change, to engage with 

regional stakeholders, and to consider shared socioecological 

futures. Each is distinct in its motivation, methodology, 

and outcomes. To provide a consistent organization for 

evaluation, we describe each using a seven-part conceptual 

framework (figure 1, table 1) that includes (1) a trigger 

or context (i.e., what precipitated the study and in what 

environment [e.g., academic, professional] it occured), (2) a 

goal (i.e., what the leaders and participants in the scenario 

study hoped to achieve), (3) construction or methodology 

(i.e., what scenario construction method was used), (4) collaboration 

(i.e., who was involved in scenario development 

[e.g., regional or national stakeholders, experts]), (5) scenarios 

(i.e., what scenarios [i.e., topics] were generated and 

what the mode or representation [e.g., narratives, visuals, 

maps] was), (6) use (i.e., in what ways the scenarios 

were used after they were completed [e.g., research, planning 

and decisionmaking, education and training]), and 

(7) impacts (i.e., what social impacts the application of the 

scenarios yielded [e.g., it improved the network, influenced 

policy decisions, changed professional practice, or increased 

capacity]). 

Figure 1. Analytical framework for evaluating diverse forms of scenario studies 

conducted throughout the US Long Term Ecological Research Network. 

Articles 

Northern Highlands Lake District scenarios: The North Temperate 

Lakes LTER site. The Northern Highlands Lake District 

(NHLD) is located in northern Wisconsin. During the past 

several decades, the population of the region has expanded 

significantly, which has resulted in the construction of new 

housing and infrastructure and a marked increase in tourism 

and recreation, which are important components of 

the region’s economy (Peterson et al. 2003b). These developments 

have placed additional pressure on the lakes and 

surrounding forests, which has caused concern about the 

long-term sustainability of these resources and the economic 

activity they support. Partially on the basis of the social- 

and natural-science infrastructure developed by scientists 

at the North Temperate Lakes (NTL) LTER site, the NHLD 

was selected as one of a few pilot study regions for the 

Millennium Ecosystem Assessment Scenarios Working 

Group (Carpenter 2008). Participants in the working group 

were drawn from local businesses, nongovernmental organizations 

(NGOs), tribal organizations, and government, as 

well as technical experts from the NTL, to engage in wideranging 

conversations about the vulnerabilities and potential 

strengths of the region, as well as the fears and hopes for 

the landscape’s future. Through an iterative process, facilitators 

distilled diverse storylines into four composite scenarios 

that portrayed alternative drivers of change over the 

coming 25 years. They commissioned an artist to develop 

illustrations for the storylines and allowed the stakeholders 

to refine the scenarios. The four scenarios represented very 

different pathways that the NHLD might follow with regard 

to future population trends, zoning requirements, and recreation 

policies. They were designed to be internally coherent, 

logically consistent, and plausible. In this way, it was possible 

to explore each scenario individually, 

as well as the similarities 

and differences among them. 

The scenarios were presented as 

illustrated stories in a short book 

and on a Web site. After the scenarios 

were settled, simulation 

modeling was begun in order 

to couple the qualitative stories 

to the anticipated quantitative 

changes in ecological systems. 

Together, the scenarios and outputs 

from simulations were used 

to spark conversations about the 

future in meetings with individual 

decisionmakers, in large 

public meetings, and in online 

forums. The facilitators also ran 

an online survey to gather public 

reactions to the scenarios. In 

subsequent years, the scenarios 

have been used in several university 

classes through role-playing 

adaptive-management games. 


Articles 

Table 1. Features of the US Long Term Ecological Research (LTER) Network sites and their associated projects discussed in this article. 

Scenarios: topic (1) 

and format (2) Uses Impacts 

Method (1) and 

steps (2) Collaboration 

Triggers and 

context Goals 

Project Title, LTER site, space, 

and record length 

1. Increased network 

2. Influenced policy 

choices 

3. Increased 

outreach capacity 

of NTL 

1. Role-playing 

adaptive-management 

games 

2. Simulation models 

1. Landscape change 

2. Narratives (stories); 

artistic renditions; 

book; Web page 

Input from 

regional 

stakeholders 

1. Workshops 

2. Iteratively 

review scenarios 

and combine 

scenario elements 

1. Create contrasting 

scenarios 

2. Stakeholder 

outreach and 

collaborative learning 

3. Provide input for 

LTER 

1. Academic 

2. Pilot scenario 

study for Millennium 

Ecosystem 

Assessment 

Northern Highlands Lake District 

LTER site: North Temperate 

Lakes (NTL) 

Space: 5300 square kilometers 

(km2 ) 

Time: 25 years forward 

1. Influenced forest 

practices in North 

America 

1. Development of 

forest management 

plan 

2. Begin plan 

implementation 

3. Public 

demonstration project 

1. Forest landscape 

structure and function 

under alternative 

disturbance regimes 

2. Narratives 

(technical); maps; 

report 

Research- 

Forest 

Service 

managers 

collaboration 

1. Assess fire 

history 

2. Write 

alternative 

scenarios 

3. Formally 

analyze 

4. Elicit feedback 


scenarios 

2. Test history as 

management guide 


management and policy 

1. Professional 

(forest planning) 

2. Alternative forest 

management 

Blue River Landscape Plan and 

Study 

LTER site: Andrews Forest (AND) 

Space: 23 km2 Time: 500 years back 

200 years forward 

1. Incorporated into 

general plan 

2. Influenced 

policies 

3. Built planners and 

citizens’ capacity 

1. Revision of 

general plan 


training, fundraising 

3. Primary and 

university classes 

1. Urban development 

(economy, 

environment, society, 

infrastructure) 

2. Narratives 

(technical, stories); 

illustrations; graphs; 

report 

Input from 

regional 

stakeholders 

1. Review 

scenario literature 

2. Elicit vision 

statements 

3. Formally 

analyze 

4. Elicit feedback 

1. Create sustainability 

vision and contrasting 

scenarios 


outreach and 



General Plan 


(urban planning) 

2. Revision of the 

city’s general plan 

2. Future Vision and Scenarios 

for Phoenix 

LTER site: Central Arizona 

Phoenix (CAP) 

Space: 1344 km2 Time: 40 years forward 

1. Incorporated into 

report to governor’s 

panel on climate 

change 

2. Informed 

decisions 

(communities) 

1. Modeling of 

ecological impacts; 

planning 

2. Energy projects 

3. University classes 

Expert-driven 1. Climate and 

biophysical 

parameters 

2. Reports; interactive 

Web mapping 

1. Downscale 

climate scenarios 

2. Model 

biophysical 

parameters 


scenarios 


outreach and 


3. Create research and 

communication tool on 

climate change 

1. Academic 

2. Societal need for 

regional information 

regarding climate 

change 

Scenarios Network for Alaska/ 

Arctic Planning (SNAP) 

LTER site: Bonanza Creek (BNZ) 

Space: AK / 1.7 million km2 Time: 20 years back 

100 years forward 

1. Forest conversion, 

energy production, 

climate mitigation, 

conservation, etc. 

2. Narratives 

(technical), maps, 

graphs, visualizations 

Input from 

national 

and regional 

stakeholders 

1. Elicit narratives 

2. Parameterize 

interactive 

simulation models 


scenarios 

2. Provide input 

for cross-site LTER 

comparative research 

1. Academic 

2. Identify regional 

characteristics that 

alter continentalscale 

drivers of 

global change 

North American Forest Scenarios 

LTER sites: Harvard Forest, 

Coweeta, NTL, AND, BNZ 

Space: 15 landscapes / 

5600–27,000 km2 Time: 50 years forward 


This is an emerging project whose use and 

impacts have yet to be realized.

Overall, the process of scenario planning increased the 

connectivity of the NTL group to other groups actively 

engaged in the region. The process also had the immediate 

effect of increasing the extent of networking among diverse 

stakeholders who had long histories in the NHLD but little 

prior interaction. This connectivity improved the outreach 

capacity of the NTL and introduced the science to more 

potential users. 

Blue River Landscape Plan and Study scenarios: The Andrews 

Forest LTER site. In the early 1990s, forest policy in the 

Pacific Northwest abruptly shifted from several decades of 

management in which commodity (timber) production was 

emphasized to one that was focused on species conservation 

(Duncan and Thompson 2006). The Northwest Forest Plan 

(NWFP) was the policy instrument that codified this change. 

The NWFP was itself the product of a scenario approach 

in which scientists, including several from the Andrews 

Forest (AND) LTER site, prepared 10 management scenarios 

(alternatives) for the Clinton administration (FEMAT 1993), 

with one selected as the basis of the NWFP. Within the 

plan, 10 large parcels (40,000–200,000 hectares) of federal 

forestlands were delineated as adaptive management areas 

(AMAs), in which scientists and land managers were charged 

with examining alternative strategies to meeting conservation 

and timber-production goals. At the Central Cascades 

AMA, which contains the AND site, LTER scientists teamed 

with land managers of the Willamette National Forest 

to construct a landscape-scale forest-management plan. 

They used science-based visioning to describe an ecologically 

desirable but unconventional approach to long-term 

management that was informed by historic fire regimes 

and landscape dynamics rather than by standard reservedesign 

criteria. The resulting scenario, called the Blue River 

Landscape Plan and Study (BRLP; Ecoshare 2011), patterned 

timber harvests on the historical wildfire regime (Cissel 

et al. 1999). The BRLP’s resulting “disturbance-based” or 

“historical range of variability” approach began with a 

dendrochronology-based fire-history study that spanned 

the previous 500 years (Weisberg 1998), which was used to 

semiquantitatively generate a map of three fire-regime types 

distinguished by fire frequency and severity and constrained 

in part by topography. A forest planner used this fire-regime 

map and a map of the then-present distribution of forest 

age classes to project harvests 200 years into the future. The 

plan included novel approaches to individual-tree and patch 

retention and to cutting-rotation length to emulate the 

historical wildfire severity and frequency. This disturbancebased 

approach to landscape management was contrasted 

with expected future management under the NWFP, which 

is based on the management of unharvested reserves and 

matrix land (i.e., actively harvested areas between reserves), 

for its conservation value for selected species (Cissel et al. 

1999). Public reaction to the disturbance-based plan was 

assessed through surveys and field-tour discussions. The 

public had a significant level of acceptance, although the 

Articles 

concept and vocabulary were unfamiliar to many (Shindler 

and Mallon 2009). The BRLP has since served as an actual 

management plan intended for implementation, as well as 

a demonstration project used for critical discourse within 

science, land-management, and public circles. The BRLP 

helped reveal to the various stakeholders that an understanding 

of natural-disturbance regimes can lead to a viable 

approach to harvesting and the conservation of biodiversity 

and ecosystem functions. Lessons being learned from the 

BRLP have been widely communicated and may be used for 

new applications as society charts the management of public 

lands through a dynamic future socioecological system 

(Spies and Duncan 2009). 

Future vision and scenarios for Phoenix: Central Arizona–Phoenix 

LTER site. In autumn 2009, the city of Phoenix, the Central 

Arizona–Phoenix (CAP) LTER site, and the School of 

Sustainability at Arizona State University initiated a research 

project entitled “The Future of Phoenix: Crafting Sustainable 

Development Strategies.” (Wiek et al. 2012). The purpose of 

the combined visioning and scenario study was to create— 

through a collaboration of expert facilitators, scientists, 

and stakeholders—a vision and contrasting scenarios that 

captured a spectrum of possible future developments of 

Phoenix. The vision described Phoenix as a desirable and 

sustainable socioecological system, as was determined by 

stakeholder agreement and sustainability principles (Gibson 

2006). The contrasting scenarios described alternative—less 

desirable and less sustainable—future states and represented 

what might happen if the vision is not achieved (Withycombe 

and Wiek 2011). In the vision that resulted from this participatory 

research process, Phoenix was described as being 

comprised of vibrant communities, where healthy food, 

clean water, fresh air, excellent educational opportunities, 

satisfying jobs, and public transit options are available to 

all citizens; where strong local businesses take advantage of 

local assets to build a diverse and community-oriented urban 

economy; where governance is open and transparent and 

reflects the values of all people, regardless of their power or 

influence; and where the urban ecological system is preserved 

and cared for, so that it can be sustained for generations to 

come. In the contrasting scenarios, two alternative future 

states of Phoenix are described: Phoenix behind the times, in 

which the city acknowledges critical challenges from climate 

change to environmental degradation and social tensions 

but cannot seem to keep pace with other regions in creating 

a healthy socioecological system, and Phoenix overwhelmed, 

in which the city ignores long-term challenges, upholds the 

growth paradigm, and continues to overextend its capacities 

and to jeopardize sustainable future development pathways. 

The vision was presented in three formats: stories with photographs, 

a narrative, and a map of priorities. The scenarios 

were presented as newspaper cover pages with illustrations 

and as narratives. Visioning and scenario construction combined 

several methods, such as consistency analysis, diversity 

analysis, sustainability appraisal, and trade-off analysis 


Articles 

(Wiek et al. 2006, Withycombe and Wiek 2011). In the study, 

the knowledge, preferences, and values of a broad range of 

experts and regional stakeholders were elicited, deliberated, 

and integrated. The core stakeholder-engagement activities 

were 15 meetings in all parts of the city, designed to elicit 

vision statements, and a visioning workshop, in which the 

elicited vision statements were revisited, systemically linked, 

and reprioritized on the basis of potential conflicts. More 

than 100 citizens, city planners, business representatives, 

nonprofit organization members, and researchers participated 

in this workshop. The vision and the scenarios are now 

being used by the city administration in planning, training, 

and fund raising, as well as for teaching purposes in public 

schools and at Arizona State University. Overall, the project 

spurred a rich public discussion about the future directions 

in which Phoenix may be headed and about the degree to 

which these directions align with a desirable and sustainable 

future vision. The substance of the vision and scenarios has 

been incorporated into the current public-hearing draft of 

the city’s general plan (City of Phoenix 2010), which is the 

city’s most important guide for long-term planning and 

development. It provides direction for planning and allows 

researchers, policy analysts, and stakeholders to evaluate the 

effectiveness of policies and actions. 

The Scenarios Network for Alaska Planning: The Bonanza Creek 

LTER site. Arctic and boreal forests are warming about twice 

as fast as the global average, creating widespread concern and 

interest in the patterns and consequences of climate change, 

especially among northern residents. At the Bonanza Creek 

(BNZ) LTER site, scenarios have been used as both a research 

and a communications tool to explore the consequences of 

recent and projected climate change in Alaska. Rapid change 

experienced in Alaska has focused public attention on these 

scenarios, which in turn has led to the establishment of 

a research partnership: the Scenarios Network for Alaska 

Planning (SNAP; www.snap.uaf.edu), which comprises the 

BNZ, the University of Alaska, and several state and federal 

agencies, local communities, and nonprofits. SNAP has 

developed scenarios of future climate as high-resolution 

maps of mean monthly historical and projected temperature 

and precipitation (between 1900 and 2100) that account for 

the known effects of topography and the movement of air 

masses. By downscaling the five global circulation models 

that were shown to perform best in the far north and by providing 

outputs for three emission scenarios defined by the 

IPCC, SNAP offers stakeholders a range of possible climate 

futures, rather than a single prediction. Moreover, discussion 

and interpretation of uncertainty play a large role in SNAP 

projects. The actual technical work of downscaling the 

climate scenarios is an expert-driven process. However, all 

SNAP projects are collaborations between SNAP researchers 

and land managers or other stakeholders. These stakeholders 

help determine how the data will be linked to landscape 

models, existing data sets, and local knowledge and 

how the resulting scenarios will be used and interpreted. 

For example, in collaboration with US Fish and Wildlife 

Service, The Nature Conservancy, and other partners, SNAP 

researchers have used climate-change scenarios to model 

potential biome shifts and changes in the ranges of endemic 

and invasive species. In the resulting report (Murphy et al. 

2010), barren-ground caribou, Alaska marmots, trumpeter 

swans, and reed canary grass were used as indicator species 

to assess multiple possible futures, given the possible range 

of climate-change impacts. In addition, the climate scenarios 

have been used directly by Alaska communities in order to 

inform decisionmaking about the future sustainability of 

hydroelectric generation and other energy project plans 

(Cherry et al. 2010). 

The American Forest Futures Projects: The Harvard Forest, 

Andrews Forest, Bonanza Creek, Coweeta, and North Temperate 

Lakes LTER sites. Scenario planning is the focus of American 

Forest Futures Projects, an emerging group of cross-site 

LTER projects that are being advanced across the major 

forested regions of the United States in collaboration with 

major NSF-sponsored programs (including, e.g., five LTER 

Network sites, NEON, ULTRA), the US Forest Service, the 

US Geological Survey, the Smithsonian Institution, and 

several NGOs (e.g., The Nature Conservancy, the Ecological 

Society of America, the Heinz Center). The projects were 

designed to address pressing scientific questions related to 

how and why forested regions vary in their socioecological 

responses to global change (www.wildlandsandwoodlands. 

org/node/133). In a parallel thrust, the researchers designed 

the projects as a means to advance several burgeoning 

network-level LTER Network priorities, such as national- 

and regional-scale stakeholder engagement, development 

of models and visualization tools, and communication with 

policymakers. The projects rely on tiered scenarios developed 

at national and regional scales to address at least four 

dominant themes: economic development, energy exploitation 

(e.g., bioenergy), climate mitigation and adaptation, 

and landscape-scale conservation. 

The researchers have convened a national advisory board 

of experts, consisting of federal agency and national NGO 

representatives, to develop national-scale scenarios describing 

the anticipated drivers and the changes associated with 

each of the themes. After they are finalized, the suite of 

national-scale scenarios will be reinterpreted at regional 

scales around the country by engaging regional stakeholders 

(e.g., natural-resource managers, scientists, agency officials, 

conservation professionals) in the development of narratives 

in which the local manifestation of each of the national 

scenarios is described. This process of downscaling global 

scenarios to local scales can be an effective approach for 

participatory capacity building and for motivating behavioral 

changes in local stakeholders and decisionmakers (Shaw 

et al. 2009). The regional scenarios will, in turn, be used to 

define land-use assumptions and to parameterize spatially 

interactive landscape-simulation models (e.g., Thompson 

et al. 2011) within study a series of study landscapes 


(5600–27,000 sqaure kilometers) dispersed throughout five 

forest regions: the Northeast, the Lake States, the Southeast, 

the Pacific Northwest, and Alaska. Using the scenarios and 

simulation outputs, the group plans to conduct a series of 

within- and across-region comparisons to evaluate how ecosystem 

attributes and services change in response to similar 

national scenarios of global change along multiple social 

and ecological gradients, including productivity and topography, 

land tenure and demographics, land-use history and 

policies, disturbance regimes, and climate. This approach for 

coupling qualitative and quantitative scenarios has informed 

prescient planning and policy and has generated a rich set of 

fundamental research questions (see, e.g., Spies et al. 2007, 

Schmitt Olabisi et al. 2010). Although it is still in its early 

stages, the American Forest Futures Projects have already 

spurred dialogue among dozens of agencies and stakeholder 

groups in an effort to describe an envelope of plausible 

futures for forested landscapes across the country. 

Discussion of the case studies 

From these case studies, we offer responses to our initial 

three questions: 

How do the scenario studies relate the past to the future? Unlike 

predictive modeling, in scenario studies, no level of confidence 

is asserted that any particular changes will occur. 

Instead, several possible changes are integrated into a set 

of potential future pathways in which the major drivers are 

logical and consistent across scenarios. In doing so, scenario 

studies provide useful contexts for addressing questions 

about how past change may or may not help understand 

future change, given a range of possible societal dynamics 

(e.g., population growth, shifting demographics, land-use 

change) and possible environmental dynamics (e.g., climate, 

invasive species, major disturbance events). The BRLP, for 

example, was based on the concept that important elements 

of the future will represent an extension of historic trajectories 

of change and also incorporated a detailed understanding 

of historic landscape change over many centuries in 

response to disturbance by wildfire, flood, and landslides. 

The future landscape is expected to include continued 

responses of forest and stream systems to those past disturbance 

regimes, as well as future regimes resulting from direct 

and indirect human influences and natural processes. The 

vision of management based on historical ecosystem dynamics 

acted as a medium for discussing these complexities 

among scientists, land managers, and the public. By examining 

a range of plausible futures, scientists and stakeholders 

could consider the range of social and ecological uncertainty 

and learn about the attributes that drive change, even if 

they do not know exactly what level of change will occur. 

The value of considering how historical trends may diverge 

along multiple alternative pathways was also evident in the 

NHLD, where big shifts in the status quo were foreseen from 

both internal and external forces. Internally, concern was 

focused on political gridlock and a lack of capacity to make 

Articles 

collective decisions. Externally, concern was focused on the 

demographic and economic drivers of massive development 

in the Northern Highland. These internal and external social 

drivers were thought to affect the resilience of the region to 

climate change and other biophysical drivers. In the NHLD 

scenarios that were seen as optimistic by most respondents, 

the social and political system of the region changed in ways 

that facilitated resilient responses to large-scale biophysical 

changes. 

Evaluating multiple scenarios that are informed but not 

constrained by history can have great advantages over the 

predictive modeling of future conditions, which can pose 

serious challenges even in seemingly straightforward analyses 

and when long-term data exist. For example, we live in 

a period of high concern over climate change, yet in many 

cases, climate-change signals are difficult to interpret amid 

the temporal variability of climate at multiple temporal 

scales, even with 50-year historical records (Jones et al. 2012 

[in this issue]). In Alaska, where the impacts of climate 

change have been felt by communities for many years, gathering 

information about how land managers and residents 

are experiencing and reacting to change offers crucial information 

about how to address future change. Those who are 

living with changes, such as increased fire risk, treeline shift, 

or severe coastal erosion, can offer information about how 

these changes are already being experienced and managed 

and can make specific requests about what kind of climate 

information and what models would be most useful. The 

SNAP scenarios span a range of potential climate futures, 

informed by the best available science, but without the 

false precision of predictive models. The climate scenarios 

can, in turn, be integrated into socioecological models and 

narrative stories that define potential futures in a way that 

managers and residents can relate to and plan for. The use of 

scenario studies is therefore a productive means for relating 

historical landscape change to forward-looking analyses of 

unpredictable socioecological systems. 

How do the scenario studies relate to the more traditional 

long-term science occurring within the LTER Network? Scenario 

studies are often a form of synthesis of both biophysical and 

social science. Framing plausible alternative depictions of 

the future and evaluating them promotes highly integrative 

thinking. The social, political, and land-management contexts 

in which LTER sites are embedded increasingly demand 

a level of integrated analysis that is uncommon in traditional 

scientific research but that is central to scenario studies. In 

each of the case studies, the strengths of traditional LTER 

science—notably, monitoring and evaluating the impacts 

of long-term environmental change—were integrated into 

the scenario studies. For example, in the American Forest 

Futures Projects, the future response of forest ecosystems to 

climate and land-use change are evaluated over time using 

ecosystem-process models that are developed and constrained 

by LTER. National and regional stakeholders define 

the narratives that guide their assumptions about land use 


Articles 

and other societal responses, thus linking the scenarios to 

the legacy of “hard science” traditionally upheld by LTER. 

Likewise, in Alaska, where climate change is a topic that has 

reached a relatively high level of public awareness, SNAP 

and the BNZ are already engaged in the type of synthesis 

that forges connections between LTER and exogenous social 

and economic pressures. In working with groups such as the 

National Park Service, the Fairbanks North Star Borough 

Climate Change Task Force (www. investfairbanks.com/ 

Taskforces/climate.php), and Alaska’s governor’s Subcabinet 

on Climate Change (www.snap.uaf.edu/projects/governorssubcabinet-climate-change-0), 

SNAP has received input from 

partners who are concerned about the interplay among 

multiple ecological factors in the context of budgets, short 

planning cycles, and public criticism. Scenario planning has 

proven to be a useful tool in this context, because it allows 

scientists to offer the best available information in a way that 

incorporates uncertainty and that allows planners to assess 

risk on their own terms. 

Importantly, in all of the case studies discussed here, 

knowledge transfer and syntheses flowed in both directions, 

whereby the science informed the scenarios and, in turn, 

the scenarios informed the science. For example, within 

the NTL program, the scenarios contributed to a shift 

toward explicit long-term thinking about socioecological 

change that has influenced research and the planning of site 

activities. Recent projects have been focused on thresholds 

for abrupt change in fisheries, food webs, and invasivespecies 

dynamics, for example. Climate change has been a 

long-standing interest at the NTL that is also serving as an 

organizing focus for hypothesis-driven long-term research. 

Similarly, the Future Vision and Scenarios for Phoenix 

study serves as a pilot project and will be continued with 

even more emphasis on the integration of previous CAP 

work. To this end, a formal synthesis workshop will be conducted 

in order to make the links to the other CAP working 

groups explicit and to plan future contributions. The study 

has also been proven to stimulate collaboration between 

CAP researchers and other research groups at Arizona State 

University that are already engaged or interested in scenario 

studies in the CAP region (across different spatial levels and 

topic areas). 

How do the scenario studies impact the region and the 

regional communities through real-world changes or capacity 

building? Scenario studies are a distinctive form of science 

engagement with society and one that has only recently 

been employed in LTER. They are much more participatory 

than traditional outreach that consists of a delivery of scientific 

findings to policymakers, managers, and the public. 

For example, the Future Vision and Scenarios for Phoenix 

study has served as a powerful medium to enhance the 

engagement between CAP researchers and regional stakeholders. 

Although CAP research has included stakeholder 

engagement since its inception in 1997, the dominant mode 

of operation has been one-way elicitation. The scenario 

study, in contrast, has demonstrated how more interactive 

engagement can enhance the interest and ownership for 

the challenges as well as potential solutions across different 

stakeholder groups. The ongoing scenario work further 

expands the participatory-research methodology through 

advanced collaborative visualization, walking audits, and 

exploration courses. Similarly, the BRLP scenario has proven 

to be very useful process for generating discussion among 

varied stakeholders—pro- and antilogging groups alike—of 

the dynamics of forest ecosystems and the relevance of 

natural variability for future management. Having a management 

scenario based on historic disturbance regimes 

has fostered public understanding of landscape dynamism 

at a time when some elements of the public seek to freeze 

components of the landscape as though it were a museum 

diorama. 

LTER scenario studies have also led to novel uses of 

technology to engage the public. Taking just a few examples 

from the case studies, the American Forest Futures Projects 

are developing a Web-based course modeled after the highly 

successful LTER Network–sponsored “From Yardstick to 

Gyroscope” class (http://news.lternet.edu/article170.html), 

in which the science of scenario studies will be taught to 

university students and landscape planners. As part of the 

BNZ SNAP project, a Web-based community charts tool 

has been developed (www.snap.uaf.edu/community-charts) 

that allows Alaska residents to compare a range of possible 

future climate scenarios for their own village (from among 

more than 350 in the state) and a Google Earth interface 

(www.snap.uaf.edu/google-earth-maps) that lets users zoom 

in on their own region and define which models they would 

like to explore. At CAP, as part of the Future Vision and 

Scenarios for Phoenix study, the Decision Theater at Arizona 

State University is to be used to evaluate the effectiveness 

of those participatory processes for capacity building and 

decisionmaking. 

Overall, the scenario studies presented here demonstrate 

how considering our shared socioecological future 

can motivate sustained engagement between science and 

society. We believe that the long-term dimension of the US 

LTER Network makes it a highly suitable venue for scenario 

studies. Indeed, the 30-year history of the LTER program 

provides time for development of social networks with key 

parts of society (e.g., decisionmakers, NGOs, government 

agencies with lands responsibilities, interested members of 

the public). Finally, members of the LTER Network’s science 

community live and work in their studied landscapes and 

are themselves stakeholders with a deep personal stake in 

its future. 

Conclusions 

As the case studies show, scenarios can be an effective 

approach for synthesizing science in a major research 

program and can lead to an improved understanding 

of socioecological change. Scenario studies offer a flexible 

framework for integrating the best available ecological 


science with the myriad of uncertainties that are inherent 

in global change. Using scenarios, large ecological research 

programs, such as the LTER Network, NEON, and ULTRA, 

can lead societal discussion regarding the future of their 

landscapes. The core strengths of the LTER Network in 

particular—its history of long-term, place-based studies; its 

community of scholars committed to integrative research 

across disciplines and service to society; and its diversity of 

landscapes, stakeholders, and disturbance regimes—make 

it ideally suited to leading scenario studies in each of the 

landscapes in which LTER sites are present. As such, we suggest 

that scenario studies be advanced in collaboration with 

many other research groups and agencies as a network-wide 

activity to promote research in socioecological systems and 

cross-site comparative analyses across the network. 


Support for this work was provided by US National Science 

Foundation Long Term Ecological Research Network awards 

to Harvard University, Arizona State University, Oregon 

State University, the University of Alaska, and the University 

of Wisconsin and by the US Forest Service. 


Alcamo J, Henrichs T. 2008. Towards guidelines for environmental scenario 

analysis. Pages 13–35 in Alcamo J, ed. Environmental Futures: The 

Practice of Environmental Scenario Analysis. Elsevier. 

Bennett EM, Carpenter SR, Peterson GD, Cumming GS, Zurek M, Pingali P. 

2003. Why global scenarios need ecology. Frontiers in Ecology and the 


Bestelmeyer S, Dailey S, Elser M, Hembree P, Landis C, O’Connell K, 

Simmons B, Sommer S, Steiner S. 2005. Handbook for LTER Education. 

US Long Term Ecological Research Network. 

Bolte JP, Hulse DW, Gregory SV, Smith C. 2007. Modeling biocomplexity: 

Actors, landscapes and alternative futures. Environmental Modeling 

and Software 22: 570–579. 

Carpenter S[R]. 2008. Seeking adaptive change in Wisconsin’s ecosystems. 

Pages 407–421 in Waller DM, Rooney TP, eds. The Vanishing 

Present: Wisconsin’s Changing Lands, Waters, and Wildlife. University 

of Chicago Press. 

Carpenter SR, Folke C. 2006. Ecology for transformation. Trends in Ecology 

and Evolution 21: 309–315. 

Carpenter SR, Bennett EM, Peterson GD. 2006. Scenarios for ecosystem 

services: An overview. Ecology and Society 11: 29. 

Cherry JE, Walker S, Fresco N, Trainor S, Tidwell A. 2010. Impacts 

of Climate Change and Variability on Hydropower in Southeast 

Alaska: Planning for a Robust Energy Future. University of Alaska, 

Fairbanks. (31 January 2012; http://ine.uaf.edu/accap/documents/seak_ 

report_final.pdf ) 

Cissel JH, Swanson FJ, Weisberg PJ. 1999. Landscape management using 

historical fire regimes: Blue River, Oregon. Ecological Applications 9: 

1217–1231. 

City of Phoenix. 2010. Phoenix General Plan Update: Transitioning to 

a Sustainable Future. Public Hearing Draft. December 2010. City 

of Phoenix, Phoenix, Arizona. (8 September 2011; http://phoenix.gov/ 

citygovernment/planres/cityplan/planphx/index.html) 

Collins SL, et al. 2011. An integrated conceptual framework for long-term 

social-ecological research. Frontiers in Ecology and the Environment 

9: 351–357. 

Driscoll CT, Lambert KF, Chapin FS III, Nowak DJ, Spies TA, Swanson FJ, 

Kittredge DB Jr, Hart CM. 2012. Science and society: The role of longterm 

studies in environmental stewardship. BioScience 62: 354–366. 

Articles 

Duncan SL, Thompson JR. 2006. Forest plans and ad hoc scientist groups in 

the 1990s: Coping with the Forest Service viability clause. Forest Policy 

and Economics 9: 32–41. 

[FEMAT] Forest Ecosystem Management Assessment Team. 1993. 

Forest Ecosystem Management: An Ecological, Economic, and Social 

Assessment. US Forest Service. 

Gibson RB. 2006. Sustainability assessment: Basic components of a 

practical approach. Impact Assessment and Project Appraisal 24: 

170–182. 

Henrichs T, Zurek M, Eickout B, Kok K, Raudsepp-Hearne C, Ribeiro T, 

van Vuuren D, Volkery A. 2010. Scenario development and analysis for 

forward-looking ecosystem assessments. Pages 151–220 in Ash N, et al., 

eds. Ecosystems and Human Well-Being: A Manual for Assessment 

Practitioners. Island Press. 

[Ecoshare] Ecoshare: Interagency Clearinghouse of Ecological Information. 

2011. Blue River Landscape Strategy. Ecoshare. (31 January 2012; 

http://ecoshare.info/projects/central-cascade-adaptive-managementpartnership/landscape-studies/blue-river-landscape-strategy) 

Iverson Nassauer J, Corry RC. 2004. Using normative scenarios in landscape 

ecology. Landscape Ecology 19: 343–356. 

Jones JA, et al. 2012. Ecosystem processes and human influences regulate 

streamflow response to climate change at long-term ecological research 

sites. BioScience 62: 390–404. 

Kahn H. 1962. Thinking about the unthinkable. Horizon Press. 

[LTER Network] US Long Term Ecological Research Network. 2011. Strategic 

and Implementation Plan: Long-term Ecological Research Network 

(LTER): Research and Education. LTER Network. (31 January 2012; 

http://intranet2.lternet.edu/sites/intranet2.lternet.edu/files/documents/ 

LTER%20History/Planning%20Documents/LTER_SIP_Dec_05_2010. 

pdf ) 

Mahmoud M, et al. 2009. A formal framework for scenario development in 

support of environmental decision-making. Environmental Modeling 

and Software 24: 798–808. 

[MA] Millennium Ecosystem Assessment. 2005. Ecosystems and Human 

Well-being: Scenarios. Island Press. 

Murphy K, Huettmann F, Fresco N, Morton J. 2010. Connecting Alaska 

Landscapes into the Future. US Fish and Wildlife Service. (8 September 

2011; www.snap.uaf.edu/files/SNAP-connectivity-2010-complete.pdf ) 

Naki enovi N, Swart R. 2000. Emissions Scenarios. Intergovernmental 

Panel on Climate Change. 

Peterson GD, Cumming GS, Carpenter S. 2003a. Scenario planning: A 

tool for conservation in an uncertain world. Conservation Biology 17: 

358–366. 

Peterson GD, Beard TD Jr, Beisner BE, Bennett EM, Carpenter SR, 

Cumming GS, Dent CL, Havlicek TD. 2003b. Assessing future ecosystem 

services: A case study of the Northern Highlands Lake District, 

Wisconsin. Conservation Ecology 7 (3, Art. 1). 

Raskin P, Monks F, Riberio T, van Vuuren D, Zurek M. 2005. Global scenarios 

in historical perspective. Pages 35–45 in Carpenter S, Pingali P, 

Bennett E, Zurek M, eds. Ecosystems and Human Well-being, Vol. 2: 

Scenarios. Island Press. 

Reid WV, Chen D, Goldfarb L, Hackmann H, Lee YT, Mokhele K, Ostrom 

E, Raivio K, Rockström J, Schellnhuber HJ, Whyte A. 2010. Earth system 

science for global sustainability: Grand challenges. Science 330: 

916–917. 

Robertson GP, et al. 2012. Long-term ecological research in a humandominated 

world. Bioscience 62: 342–353. 

Rockström J, et al. 2009. A safe operating space for humanity. Nature 461: 

472–475. 

Sala OE, et al. 2000. Global biodiversity scenarios for the year 2100. Science 

287: 1770–1774. 

Scheffer M, Carpenter S, Foley JA, Folke C, Walker B. 2001. Catastrophic 

shifts in ecosystems. Nature 413: 591–596. 

Schmitt Olabisi LK, Kapuscinski AR, Johnson KA, Reich PB, Stenquist B, 

Draeger KJ. 2010. Scenario visioning and participatory system dynamics 

modeling to investigate the future: Lessons from Minnesota 2050. 

Sustainability 2: 2686–2706. 


Articles 

Shaw A, Sheppard S, Burch S, Flanders D, Wiek A, Carmichael J, Robinson J, 

Cohen S. 2009. Making local futures tangible—Synthesizing, downscaling, 

and visualizing climate change scenarios for participatory capacity 

building. Global Environmental Change 19: 447–463. 

Shindler B, Mallon A. 2009. Public acceptance of disturbance-based 

forest management: A study of the Blue River Landscape Strategy in 

the Central Cascades adaptive management area. US Department of 

Agriculture, US Forest Service, Pacific Northwest Research Station. 

Research Report no. PNW-RP-581. 

Spies TA, Duncan SL. 2009. Old Growth in a New World: A Pacific 

Northwest Icon Reexamined. Island Press. 

Spies TA, Johnson KN, Burnett KM, Ohmann JL, McComb BC, Reeves GC, 

Bettinger P, Kline JD, Garber-Yonts B. 2007. Cumulative ecological and 

socioeconomic effects of forest policies in coastal Oregon. Ecological 

Applications 17: 5–17. 

Swart RJ, Raskin P, Robinson J. 2004. The problem of the future: 

Sustainability science and scenario analysis. Global Environmental 

Change 14: 137–146. 

Thompson J[R], Foster D[R]. 2009. Report to the LTER Network Office 

on the Scenarios of Future Landscape Change Working Group 

Meeting April 1–2, 2009. (8 September 2011; http://lno.lternet.edu/files/ 

lno/Workship%20Report%20-%20Landscape%20Change_Thompson_ 

Foster.pdf ) 

Thompson JR, Foster DR, Scheller R, Kittredge D[B Jr]. 2011. The influence 

of land use and climate change on forest biomass and composition in 

Massachusetts, USA. Ecological Applications 21: 2425–2444. 

Wack P. 1985. Scenarios: Uncharted waters ahead. Harvard Business Review 

63: 73–89. 

Weaver PM, Rotmans J. 2006. Integrated sustainability assessment: What 

is it, why do it and how? International Journal of Innovation and 

Sustainable Development 1: 284–303. 

Weisberg PJ. 1998. Fire history, fire regimes, and development of forest 

structure in the central western Oregon Cascades. PhD dissertation. 

Oregon State University, Corvallis. 

Wiek A, Binder CR, Scholz RW. 2006. Functions of scenarios in transition 

processes. Futures 38: 740–766. 

Wiek A, Ness B, Brand FS, Schweizer-Ries P, Farioli F. 2012. From complex 

systems analysis to transformational change: A comparative appraisal 

of sustainability science projects. Sustainability Science. doi:10.1007/ 

s11625-011-0148-y 

Withycombe L, Wiek A. 2011. What if the future does not play out as imagined 

in the vision? Advances in sustainability planning and research 

through contrasting, non-intervention scenarios. Project Report. School 

of Sustainability, Arizona State University. (http://start.lab.asu.edu/) 

Xiang W-N, Clarke KC. 2003. The use of scenarios in land-use planning. 

Environment and Planning 30: 885–909. 

Jonathan R. Thompson (thompsonjr@si.edu) is affiliated with the Smithsonian 

Institution’s Conservation Biology Institute, in Front Royal, Virginia. Arnim 

Wiek is affiliated with the School of Sustainability at Arizona State University, 

in Tempe. Frederick J. Swanson and Thomas A. Spies are affiliated with 

the US Forest Service’s Pacific Northwest Research Station, in Corvallis, 

Oregon. Stephen R. Carpenter is affiliated with the Center for Limnology 

at the University of Wisconsin, Madison. Nancy Fresco is affiliated with the 

School of Natural Resources and Agricultural Science at the University of 

Alaska, Fairbanks. Teresa Hollingsworth is affiliated with the Boreal Ecology 

Cooperative Research Unit at the University of Alaska, Fairbanks, and with 

the US Forest Service’s Pacific Northwest Research Station, also in Fairbanks. 

David R. Foster is affiliated with Harvard University’s Harvard Forest, in 

Petersham, Massachusetts. 


Past, Present, and Future Roles 

of Long-Term Experiments in the 

LTER Network 

AlAn K. KnApp, MelindA d. SMith, SArAh e. hobbie, Scott l. collinS, tiMothy J. FAhey, 

Gretchen J. A. hAnSen, douGlAS A. lAndiS, KiMberly J. lA pierre, Jerry M. Melillo, 

tiMothy r. SeAStedt, GAiuS r. ShAver, And JAcKSon r. WebSter 

Articles 

The US National Science Foundation–funded Long Term Ecological Research (LTER) Network supports a large (around 240) and diverse portfolio 

of long-term ecological experiments. Collectively, these long-term experiments have (a) provided unique insights into ecological patterns and 

processes, although such insight often became apparent only after many years of study; (b) influenced management and policy decisions; and 

(c) evolved into research platforms supporting studies and involving investigators who were not part of the original design. Furthermore, this 

suite of long-term experiments addresses, at the site level, all of the US National Research Council’s Grand Challenges in Environmental Sciences. 

Despite these contributions, we argue that the scale and scope of global environmental change requires a more-coordinated multisite approach to 

long-term experiments. Ideally, such an approach would include a network of spatially extensive multifactor experiments, designed in collaboration 

with ecological modelers that would build on and extend the unique context provided by the LTER Network. 

Keywords: climate change, global change, long-term research, LTER Network, multifactor experiments 

More than 30 years ago, Odum (1977) described ecology 

as a uniquely integrative and synthetic scientific 

endeavor, and the diversity of research approaches employed 

today reflects this perspective (e.g., Rees et al. 2001). 

Collectively, ecologists conduct research that includes virtually 

every ecosystem type on Earth, with studies that span 

a broad range of spatial and temporal scales. Moreover, 

the most successful research programs generally employ a 

mixture of complementary approaches, including retrospective 

studies, observations, experiments (natural and 

manipulative), gradient studies, synthetic analyses, and 

modeling. The research portfolios of the 26 sites within the 

US National Science Foundation (NSF)–funded Long Term 

Ecological Research (LTER) Network (http://lternet.edu) 

exemplify this spatially and temporally extensive and multifaceted 

approach to ecology (Hobbie et al. 2003), although 

the degree to which the sites allocate time and resources to 

particular research approaches varies on the basis of the 

questions being addressed, site-specific constraints, and the 

culture of the discipline (e.g., oceanographers study systems 

differently than do forest ecologists). 

Site-based manipulative experiments have long been a 

cornerstone of ecological research. Although long-term 

experiments (box 1) have a foundational history in ecology 

(e.g., the Rothamstad fertilization study begun in 1856), 

short-term experiments are much more common for a 

variety of reasons (Tilman 1989). Short-term experiments 

provide information on how a system is regulated at the 

time and place of the manipulation (i.e., the initial limiting 

factors and their interactions). However, in complex systems 

with multiple components operating on different time scales 

of response, the initial trajectories of response of either the 

whole system or of individual components (e.g., single species) 

will not necessarily indicate either the direction or the 

magnitude of long-term change. Long-term experiments, 

when coupled with measurements of key processes, will be 

more likely to enable interpretation of the entire trajectory 

of system response, as well as the trajectories of change 

in system components (i.e., species or pools of organic 

matter and elements). A key difference between long-term 

and short-term experiments is that long-term experiments 

provide insight into the causes of the changes in the slope 

of responses, the causes of the inflection points, and the 

magnitude of the long-term change, whereas short-term 

experiments are focused only on the initial trajectories. 

Thus, long-term experiments can address mechanisms and 

temporal dynamics in a complementary fashion, particularly 

when the experiments are combined with other research 

approaches (e.g., observational and gradient studies). In 

doing so, they can help elucidate how historical influences 





Articles 

Box 1. Definition of long-term ecological experiments. 

We define experiments simply as planned or active manipulations 

of a single factor or of several factors (drivers or components 

of ecological systems) in order to test hypotheses. 

Such manipulations may occur at spatial scales that vary from 

small plots to whole systems (watersheds or entire lakes). This 

definition excludes consideration of natural field experiments 

such as those involving floods, droughts, hurricanes, wildfires, 

and other uncontrolled natural events, unless they can serendipitously 

be incorporated into a planned experimental design 

(see the text). 

More challenging is the definition of long term. Ideally, 

ecological or life-history criteria (e.g., process and turnover 

rates, life span of organisms) or the time scales of ecological 

phenomena would be used to define an experiment as long 

term. Operationally, however, the duration of experiments, 

which typically require financial resources to be initiated and 

continued, is to a large extent constrained by funding cycles. 

Given the Long Term Ecological Research (LTER) Network 

focus of this review, we view a long-term experiment as one 

that is planned to exceed or already exceeds six years in length, 

which would span two traditional LTER Network funding 

cycles and is also a sufficient length to allow both mechanisms 

and temporal dynamics to be identified. This duration is also 

consistent with the requisite period of continuous measurements 

for support by the US National Science Foundation’s 

Long-Term Research in Environmental Biology program. 

that shape the present can be merged with a mechanistic 

understanding of contemporary processes to enable forecasting 

the future of ecological systems (Carpenter 2002). In 

contributing to this understanding, long-term experiments 

are critical for informing and validating mechanistic models 

and for linking short-term process studies with long-term 

observations to improve models for forecasting future 

scenarios. 

Long-term experiments have been an integral part of 

the NSF-funded LTER program since its inception in 1980 

(Callahan 1984). Indeed, the LTER program’s early mission 

statement specifically called for the creation of a legacy of 

“well-designed and well-documented long-term observations, 

experiments, and archives of samples and specimens” 

(Hobbie et al. 2003, p. 22). Although there are certainly 

many valuable long-term experiments worldwide (Rees et al. 

2001), a large and diverse collection has been developed 

within the US LTER Network. Now, after more than three 

decades of research, an assessment of the types of long-term 

experiments within the LTER Network and their contributions 

is warranted. Below, we review the motivation for and 

the role of long-term experiments in the LTER Network, 

we summarize the collective lessons learned from these 

long-term experiments, and we envision the future role 

for long-term experiments in advancing ecological science 

and addressing critical questions facing society as the LTER 

Network continues into its fourth decade. 

Early long-term experiments of the LTER Network 

Long-term experiments served as the basis for the establishment 

of the research programs at several of the initial LTER 

sites funded in the 1980s. For example, the AND, CWT, and 

HBR LTER sites (see table 1 for site abbreviations) began as 

US Department of Agriculture (USDA) Forest Service experimental 

forests. At these sites, the responses of hydrology, 

energy flow, and nutrient cycling to changes in forest structure 

caused by disturbance had been quantified experimentally 

using paired watersheds (e.g., Hewlett and Helvey 1970) and 

the small-watershed approach (Likens 1985) for a number 

of years prior to their being included in the LTER program 

(figure 1). Similarly, the USDA Agricultural Research Service 

established long-term livestock grazing experiments at several 

of their experimental ranges in the early 1900s, two of which 

became LTER program sites (JRN and SGS; see figure 1). The 

response of arctic tundra vegetation to long-term manipulations 

of nutrient availability and temperature (Chapin et al. 

1995) also provided the foundation for continuing studies of 

global-change effects at the ARC site (figure 1). 

Research programs at several other founding LTER sites 

were built around newly established long-term experiments 

designed to test a diverse set of hypotheses, with the expectation 

of a long-term funding commitment. For example, the 

role and mechanisms whereby soil resources regulate community 

structure have been examined in plot-level manipulations 

at the CDR site since its inception (figure 1). The 

long-term consequences of nitrogen (N) deposition, climate 

warming, and hurricanes for nutrient cycling, carbon (C) 

dynamics, and forest productivity have been assessed at HFR 

(Foster et al. 1997). Moreover, experiments evaluating the 

interactions of fire, ungulate grazing, and climate variability 

in driving patterns and processes in tallgrass prairie formed 

the basis for long-term experiments at KNZ (figure 1). 

Finally, the core experiment in row-crop agriculture at KBS 

involved the long-term imposition of a gradient of varying 

management inputs into midwestern US cropping systems. 

In all cases, the expectation that key ecological responses to 

these manipulations might not become evident for many 

years justified the long-term nature of the experimental 

design for these LTER programs. 

Long-term experiments today in the LTER Network 

In order to provide a more comprehensive overview of 

long-term experiments in the LTER Network, we surveyed 

each of the 26 LTER sites (table 1) to quantify the number, 

types, and durations of long-term experiments supported 

by the LTER Network. In addition, we convened a two-day 

working-group meeting with participants from of a subset 

of LTER sites in February 2011. On the basis of responses 

(from 100% of the sites) to the survey questions, a total of 

239 long-term experiments (completed and ongoing) were 

identified within the LTER Network (table 1). Of these, less 

than 10% exceed 30 years in duration, which indicates that 

most long-term experiments were initiated during the time 

frame of the LTER Network (figure 2). Diversity among the 


Articles 

Table 1. Selected results from a questionnaire survey of the 26 sites in the Long-Term Ecological Research (LTER) Network, 

site acronyms, and site age. 

Site name Site acronym 

experiments was manifest in the wide array of resources 

and environmental factors manipulated (e.g., soil nutrients, 

water, and temperature were the most commonly manipulated, 

carbon dioxide [CO 2 ] more rarely), as well as alterations 

in biotic attributes (e.g., species removal and additions, 

herbivory) and disturbance regimes (e.g., fire, logging, windstorms, 

pathogen outbreaks; figure 2). However, in only 28% 

(66 of 239) of the long-term experiments was more than 

one of these factors manipulated concurrently (table 1). 

As befits the chronic nature of many global-change drivers 

(Smith et al. 2009), the number of sites that employed press 

(i.e., continuous alteration) treatments in long-term experiments 

was 80%, with 75% also supporting long-term pulse 

experiments. Collectively, all of the Grand Challenges for the 

environment identified by the National Research Council 

(NRC 2001) are being addressed across the LTER Network, 

Year established 

in LTER Network 

Number of 

long-term 

experiments 

Number of 

multifactor 

experiments 

Multisite 

experiments 

(yes [y] or no [n]) 

Andrews Experimental Forest AND 1980 23 1 y 

Arctic ARC 1987 10 7 y 

Baltimore Ecosystem Study BES 1997 1 0 n 

Bonanza Creek Experimental Forest BNZ 1987 18 3 y 

California Current Ecosystem CCE 2004 2 0 n 

Cedar Creek Ecosystem Science Reserve CDR 1981 36 9 y 

Central Arizona-Phoenix CAP 1997 2 0 n 

Coweeta Hydrological Laboratory CWT 1980 5 1 y 

Florida Coastal Everglades FCE 2000 5 1 n 

Georgia Coastal GCE 2000 1 0 n 

Harvard Forest HFR 1988 18 2 y 

Hubbard Brook Experimental Forest HBR 1987 5 1 n 

Jornada Basin JRN 1981 15 2 y 

Kellogg Biological Station KBS 1987 12 8 y 

Konza Prairie Biological Station KNZ 1980 18 8 y 

Luquillo Experimental Forest LUQ 1988 4 1 y 

McMurdo Dry Valleys MCM 1993 12 3 n 

Moorea Coral Reef MCR 2004 1 1 n 

Niwot Ridge NWT 1980 8 6 y 

North Temperate Lakes NTL 1980 6 2 n 

Palmer Station PAL 1991 0 0 n 

Plum Island Ecosystems PIE 1998 3 2 y 

Santa Barbara Coastal SBC 2000 1 0 y 

Sevilleta SEV 1988 12 3 y 

Shortgrass Steppe SGS 1981 17 4 y 

Virginia Coast Reserve VCR 1987 4 1 y 

Note: The principal investigators (PIs) were asked to provide information about long-term ecological experiments (as defined in box 1) at their site. 

Among other questions, the PIs were asked to report the number of long-term experiments, the number of experiments that included more than one 

manipulated factor, and if any of these long-term experiments were conducted at multiple sites. Additional results from this survey are provided in 

figure 2 and in the text. 

although long-term experiments related to infectious disease 

are notably underrepresented (figure 2). Finally, a majority 

of the sites (16 of 26) participated in multisite or networklevel 

experiments (i.e., experiments that spanned multiple 

sites), which may or may not have included other LTER 

sites. Notable examples include the Nutrient Network, the 

Long-Term Intersite Decomposition Team (LIDET), and the 

International Tundra Experiment. 

Our assessment of long-term experiments during the 

working-group meeting revealed that many of the earliest 

long-term experiments were designed as single-factor 

manipulations with the goal of gaining an improved understanding 

of the long-term nature of change (primarily at 

the process level) in response to either short-term (pulse) or 

long-term (press) manipulations. In the early years of LTER, 

assessing recovery from disturbance using pulse experiments 


Articles 

Figure 1. Examples of the diversity of long-term experiments established or already ongoing in the earliest cohorts of sites 

(more than 20 years old; see table 1) in the Long Term Ecological Research (LTER) Network. (a) Small watersheds with 

different forest harvest treatments at the HBR site (photograph: Hubbard Brook Research Foundation); (b) snowfence 

experiment designed to alter soil moisture and growing-season length at the NWT site (photograph: Mark Williams, Niwot 

Ridge LTER Program, University of Colorado); (c) a long-term ungrazed exclosure (left) adjacent to a grazed grassland at the 

SGS site (photograph: Alan K. Knapp); (d) warming, nutrient addition, and shade experiments at the ARC site (photograph: 

Sarah E. Hobbie); (e) biodiversity manipulations at the CDR site (photograph: David Tilman); (f) watershed-scale fire 

experiments at the KNZ site (photograph: Alan K. Knapp). Site abbreviations are defined in table 1. 

(i.e., a single perturbation or manipulation event) was the 

most common type of experiment (Callahan 1984), such as 

deforestation studies at HBR and CWT or wind-disturbance 

studies at HRF. These studies were motivated by the need 

to develop or test theory in order to enable a deeper understanding 

of the ways in which key drivers altered ecosystem 

structure and function. The goal was to complement historical 

observations with knowledge and insight that were 

unattainable from traditional short-term experiments and 

space-for-time studies. Other experiments were initiated 

to understand the impacts of anthropogenic disturbances 

such as logging, grazing, and acid deposition. Although the 

initial focus was on disturbances, the more-recent trend is 

for long-term experiments to have stronger relevance to 

global-change drivers or to be linked explicitly to policy and 

major issues in natural-resource management (figure 3). 

The working group concluded that, given the number and 

diversity of long-term experiments in the LTER Network, 

Callahan’s (1984) vision of the LTER Network as one that 

could address the “ serious contradiction between the time 

scales of many ecological phenomena and the support to 

finance their study” (p. 363) has been fulfilled. 

Synergies between the LTER Network and long-term 

experiments 

Although long-term experiments have been supported by 

many other programs and have been conducted successfully 

at a wide range of sites around the globe, long-term 

experiments conducted at LTER sites have many important 

advantages. First, long-term measurements at LTER sites 

provide essential context for interpreting experimental 

responses. Long-term observations can also provide pretreatment 

and reference-system data, as well as the basis 

for identifying ecological surprises, such as extreme climatic 

events (Lindenmayer et al. 2010, Smith 2011). Such 

data are critical for interpreting experimental results over 


Figure 2. (a) The number of long-term experiments across 

the 26 Long-Term Ecological Research (LTER) Network 

sites according to their duration. Experiments more than 

30 years in duration were initiated prior to the existence 

of the LTER Network, whereas those less than 6 years 

in duration have been recently initiated with plans for 

continuation for more than 6 years. (b) Proportion of sites 

in the LTER Network with long-term experiments in which 

different types of factors were manipulated. Resource 

manipulations are in the left side of the panel, and other 

types of manipulations are shown in the right side (Biotic = 

species removals/additions, herbivory, etc.; Disturbance = 

fire, hurricane simulation, etc.; Land use = agriculture, 

forest clear-cutting, etc.; Other = salinity, soil pH, etc.). 

(c) Proportion of LTER Network sites with long-term 

experiments in which the National Research Council’s 

Grand Challenges in Environmental Sciences (NRC 2001) 

were addressed. These data are from a survey of all 26 

LTER Network–site principal investigators (see table 1). 

Articles 

a background of natural variability—for separating signal 

from noise; however, these data sets are usually impossible 

to maintain through typical funding programs. At the 

HBR site, for example, researchers were able to quantify 

the magnitude of soil calcium (Ca) depletion caused by 

twentieth-century acid deposition only through 40 years 

of continuous biogeochemical measurements (Likens et al. 

1996). These observations, in turn, inspired a long-term 

experiment in which soil Ca availability was returned to 

preindustrial levels (figure 3), with the long-term trajectory 

of forest biomass and demography providing the context 

for interpreting responses to the experimental treatment. 

The dramatic recovery of growth, health, and regeneration 

of sugar maple (Acer saccharum) in response to a moderate 

increase in soil Ca provided further evidence that humaninduced 

soil Ca depletion has caused widespread decline 

in this highly valued forest tree and in forest productivity 

overall (Juice et al. 2006). In this case, the deployment of a 

highly sensitive tracer in the Ca treatment provided additional 

insights into the pathways of Ca flux through forested 

watersheds (Dasch et al. 2006). 

In addition to long-term observations revealing patterns 

and inspiring experimentation to elucidate mechanisms, 

post-experiment monitoring at LTER sites can improve 

understanding and prompt the design of new studies 

(Janzen 2009). For example, a decade following the cessation 

of 10 years of experimental N additions at the 

CDR site, species richness had recovered, whereas other 

aspects of community structure had not. This led to a 

short-term experiment in which the manipulation of soil 

N availability, plant litter, and seed dispersal revealed the 

mechanisms related to recovery from chronic N fertilization 

(Clark and Tilman 2010). Similarly, ecosystem recovery was 

monitored for 10 years following 6 years of experimental 

acidification of Little Rock Lake (part of the NTL site) in 

northern Wisconsin (figure 3). This monitoring revealed 

that total zooplankton biomass recovered in one year, but 

community composition exhibited sustained differences. 

Approximately 40% of zooplankton taxa exhibited a lag in 

recovery after pH returned to the levels at which a biological 

response was first observed (Frost et al. 2006). Continued 

monitoring of the recovery of this system revealed complex 

trophic interactions and hysteresis effects—responses that 

would not have appeared with only a year or two of posttreatment 

measurements. Such opportunities arise because 

LTER Network sites provide consistent access to instrumentation 

and infrastructure, stable funding levels, and 

well-trained personnel who are available to continue the 

measurements. Furthermore, the information-management 

capabilities of LTER Network sites ensure that data will be 

collected using consistent protocols and will be archived 

and readily available for use by other researchers (Ingersoll 

1997). Easily accessible data with well-documented metadata 

facilitate cross-site comparisons that are integral for 

expanding the results of experiments to broader spatial 

scales. 


Articles 

Figure 3. Examples of long-term experiments related to global change in the Long-Term Ecological Research (LTER) 

Network. (a) Application of calcium silicate to an experimental watershed at the HRB site to mechanistically examine 

acid-rain impacts (photograph: US Forest Service); (b) the multifactor BioCON experiment at the CDR site with 

manipulations of carbon dioxide, nitrogen, and biodiversity (photograph: David Tilman); (c) Lake acidification curtain 

in a lake at the NTL site to assess whole-system impacts of changes in pH (photograph: Carl J. Watras); (d) soil-warming 

experiment at the HFR site (plot delineated by the lack of snow in winter; photograph: Jerry M. Melillo); (e) automated 

trace-gas chambers for assessing the impacts of different management practices in field crops at the KBS site (photograph: 

Julie E. Doll, Kellogg Biological Station LTER Program); (f) litter-exclusion experiment at the CWT site designed to allow 

the examination of the importance of terrestrial energy inputs into streams, which also has relevance for evaluating the 

impacts of mountaintop mining practices (photograph: Sue L. Eggert). Site abbreviations are defined in table 1. 

The value of long-term ecological experiments 

On the basis of the working-group meeting and the broader 

site survey, we highlight below the diverse array of insights 

and contributions provided by long-term ecological experiments 

within the LTER Network. Our goal is to extend our 

review beyond the core scientific benefits of this suite of 

studies to include unique opportunities and broader impacts 

afforded by these long-term experiments. 

Insights into long-term responses. Obviously, but not trivially, 

one of the greatest scientific benefits of long-term site-based 

experiments is that they elucidate how ecological systems 

respond to experimental treatments over the long term. 

Indeed, such experiments have repeatedly demonstrated 

that long-term responses to treatments can differ markedly 

from short-term responses (Tilman 1989, Debinski and Holt 

2000). For example, in the BioCON experiment at the CDR 

site (biodiversity, CO 2 , and N are manipulated; see figure 3), 

Reich and colleagues (2006) tested the hypothesis that the 

availability of N would constrain the response of productivity 

in an N-poor grassland community to elevated CO 2 . 

Although this hypothesized interaction between N and CO 2 

eventually became apparent, it was not until the fourth year 

of treatment. If the experiment had been discontinued before 

the fourth year (i.e., within the time frame of typical funding 

cycles), researchers might have concluded that N availability 

had no effect on the response of these communities to elevated 

CO 2 . Similarly, in a manipulation of soil temperature 

at the HFR site (figure 3), the conclusion regarding the influence 

of warming on soil respiration would have been very 

different had the experiment ended after its first few years 

(figure 4). In that experiment, warming strongly increased 


Figure 4. Long-term dynamics of soil carbon-dioxide 

release from the soil-warming experiment at the Harvard 

Forest Long-Term Ecological Research site. The data shown 

are the proportional change in carbon lost between soils 

in the heated plots (H, heated by soil-warming cables) 

and those not warmed (but disturbed similarly to the 

heated plots, DC). Note that from 1991 to 1997, there was 

a strong effect of soil warming, but this effect diminished 

substantially as the experiment continued. See Melillo and 

colleagues (2002) for further details. 

soil respiration (by about 25%) in the first five years of the 

study (Melillo et al. 2002). However, by the tenth year of 

treatment, the warming stimulation had declined to less than 

5% above ambient controls. Conclusions based on the initial 

results of this experiment would have led to an overestimate 

of the positive feedback to climate warming resulting from 

enhanced soil organic matter decomposition. Elucidation 

of the mechanisms underlying these responses was made 

possible by coupling these single-factor studies with those in 

which soil and N were simultaneously manipulated, whereas 

additional microbial studies enhanced ecosystem-scale 

understanding (Contosta et al. 2011). 

Occasionally, experimentally induced shifts in ecosystems 

can take years to appear, which can lead to ecological 

surprises (Lindenmayer et al. 2010), as exemplified by the 

long-term phosphorus (P)–addition experiment conducted 

at the Upper Kuparuk River, Alaska (part of the ARC site). 

In this case, although P fertilization stimulated epilithic 

algal production immediately, a major increase in bryophyte 

production (with subsequent effects on nutrient cycling and 

higher trophic levels) became apparent only after a decade 

of treatment—a response that was completely unexpected 

(figure 5; Slavik et al. 2004). In other instances, the lack of 

response even after a decade or more of treatments may 

lead to an important new understanding of patterns and 

processes. For example, inspired by the dramatic response 

of desert grasslands to an exclusion of small mammals in 

Portal, Arizona (e.g., Brown and Heske 1990), replicate 

Articles 

Figure 5. Long-term changes in epilitihic chlorophyll 

(top panel, in micrograms [µg] per square centimeter 

[cm 2 ]) and bryophyte cover (bottom panel, percentage) 

in reference and phosphorus-fertilized reaches of the 

Kuparuk River at the Arctic Long Term Ecological Research 

site. In this long-term experiment, bryophyte cover 

was negligible in both reference and fertilized reaches 

prior to 1992 but increased dramatically in response to 

chronic phosphorus additions afterward, with concurrent 

reductions in epilithic chlorophyll. The chlorophyll 

data are not shown in 1988 because the samples were 

contaminated by green algal filaments. Means are shown, 

and the error bars represent the positive standard error. 

The data are updated from Slavik and colleagues (2004). 

small mammal exclosures were established in a number 

of grassland and shrubland sites, including the SEV site. 

Although no significant differences in vegetation composition 

and dynamics have been observed to date at the SEV 

site (cf. Báez et al. 2006), major changes have occurred in 

other arid-land ecosystems (Meserve et al. 2003). The differences 

in responses among sites reflect the degree to which 

top-down control of community structure can vary among 

arid-land ecosystems and can prompt new experiments to 

better understand context dependence in the function of 

communities and ecosystems. 

Besides elucidating how ecological responses to experimental 

treatments can change over time, long-term experiments 

can provide unique opportunities to uncover the 

mechanisms underlying such dynamics. In some instances, 

initial system responses may be dominated by the disturbance 

associated with initiating an experiment or because of 


Articles 

legacies of pretreatment conditions. Therefore, it may take 

a while before system responses are indicative of ecological 

processes. For example, patterns of nitrate loss from agricultural 

systems that vary in management intensity are notably 

variable, in part because many studies are initiated before 

treatments have fully equilibrated or are conducted over relatively 

short periods of time. This has led to widely conflicting 

reports and little agreement about the best management 

practices to decrease nitrate loss. In contrast, the long-term 

cropping systems experiment at the KBS site (figure 3) has 

enabled comparisons of nitrate losses from conventional, notill, 

low-input, and organic cropping systems. Each of these 

treatments had six years to equilibrate before measurement, 

and they have been assessed for 11 years, which accounts for 

interannual variability (Syswerda et al. 2008). This study has 

revealed consistent and marked differences in nitrate losses, 

with the low-input and organic systems having about half the 

nitrate losses of the conventionally managed systems. 

Long-term experiments can reveal the importance of indirect 

effects in ecosystem processes not apparent in the short 

term, as well as responses that may change over the long term 

(Tilman 1989). At the NTL site, the experimental acidification 

of a small seepage lake (figure 3) produced numerous 

changes driven by indirect interactions related to changes in 

the food web rather than by direct consequences of lower 

pH. For example, the rotifer Keratella taurocephela has displayed 

low acid tolerance in laboratory studies but increased 

in density throughout the acidification experiment as a result 

of decreased invertebrate-predator abundance (Gonzalez and 

Frost 1994). Keratella taurocephela also underwent morphological 

changes in the acidified basin as a result of reduced 

predation pressure. Responses driven by food-web interactions 

were often the opposite of expectations based on laboratory 

studies of pH tolerance and generally exhibited a time lag; 

these unexpected results would not have been observed on time 

scales shorter than the response time of all trophic levels. 

Other opportunities from long-term experiments. Unavoidably, 

long-term experiments occur against a backdrop of longterm 

trends; stochastic ambient conditions, including 

climate variability and extremes and infrequent disturbances; 

and changes in the abundances of predators and 

pathogens at scales greater than the experimental units. 

Although the inability to control ambient conditions can be 

challenging, variability in background conditions can sometimes 

also prove fortuitous (Tilman 1989). For example, 

long-term experiments can be particularly valuable if they 

coincide with climate or weather extremes that provide new 

ecological understanding. Of course, the longer an experiment 

is conducted, the greater the chances that ambient 

conditions will vary in ways that produce insights and even 

inspire new experiments. During the Ca-addition studies 

at the HBR site (figure 3), an intense ice storm damaged 

sugar maple trees, which allowed researchers to document 

improved wound repair as one of the major responses to 

the alleviation of Ca deficiency and, presumably, a principal 

mechanism underlying increased growth rates (Huggett 

et al. 2007). At the CWT site, Yeakley and colleagues 

(2003) designed an experiment to investigate the importance 

of riparian Rhododendron species to nutrient export 

to streams. They instrumented treatment and reference 

hillslopes, made two years of pretreatment measurements, 

and then removed all Rhododendron stems from a 10-meter 

(m) strip along 30 m of stream. Less than two months 

later, Hurricane Opal downed most of the large trees on 

the reference hillslope. Over the course of the study, they 

found that the hurricane impact on canopy trees had far 

greater effects on nutrient export than did the experimental 

Rhododendron removal. Finally, a wildfire at the KNZ site 

in 1996 reset most of the long-term fire treatments in the 

60 watersheds at the site (figure 1) by burning them all on 

the same date. This afforded the opportunity to sample 

a large number of sites affected by the same fire but with 

a wide array of longer-term fire histories. This sampling 

revealed the importance of the amount of time since a prior 

fire in determining aboveground net primary productivity 

responses to fire (figure 6) and helped researchers interpret 

results from other experiments regarding the role of soil N 

in postfire responses (Blair 1997, Knapp et al. 1998). 

Scientists conducting long-term experiments can also 

take advantage of dramatic changes in community structure, 

such as those resulting from extirpation or biological invasion. 

For example, at the KBS site, long-term monitoring 

of predaceous lady beetles (Coccinellidae) in experimental 

agricultural treatments has revealed three exotic species 

additions since 1988, including the multicolored Asian lady 

beetle (Harmonia axyridis). The arrival of the soybean aphid 

(Aphis glycines) in 2000 reunited these two Asian species in 

a new context and resulted in surprising dynamics. Prior to 

2000, H. axyridis was a common species; however, after 2000, 

it became dominant (figure 7). Moreover, it demonstrated 

classic predator–prey cycling with high abundances following 

years of aphid outbreak (Heimpel et al. 2010); process-level 

studies demonstrated strong top-down control of A. glycines 

by coccinellids (Costamagna and Landis 2006). A further 

surprise was that biological control of the soybean aphid 

was regulated by the structure of the surrounding landscape: 

Suppression was positively correlated with landscape 

diversity at the 1.5-kilometer scale (Gardiner et al. 2009). 

Finally, long-term experiments can provide opportunities 

to address novel questions that were not part of 

the original motivation for the experiment. Long-term 

N-enrichment studies at the ARC site were initially established 

as part of a broader suite of treatments designed 

to assess resource limitation and the response of tundra 

ecosystems to global-change factors (figure 1; Chapin et al. 

1995). After 20 years of adding N, the researchers shifted 

their focus toward asking how much of the cumulative N 

added still remained in the plots and how this affected C 

pools (Mack et al. 2004). Somewhat surprisingly, these plots 

had lost significant C, despite greater C inputs (net primary 

production) and aboveground C stocks with N addition. Net 


Figure 6. Response in aboveground net primary production 

(ANPP, in grams [g] per square meter [m 2 ]) to a wildfire at 

the Konza Prairie (KNZ) Long-Term Ecological Research 

site. The KNZ site includes a number of watersheds that are 

experimentally burned at different frequencies (from annual 

fire to no fire). A wildfire burned most of the site in 1991 

(the no-fire data are from areas that escaped the wildfire), 

which allowed the researchers to gain insight into how 

long-term fire history (including the time since the last fire) 

influenced ANPP responses. The mean for each time interval 

is displayed, and the error bars represent the standard error. 

Source: Adapted from Knapp and colleagues (1988). 

ecosystem C losses resulted from soil C losses in deeper 

horizons, presumably from enhanced decomposition caused 

by N addition. 

Broader contributions of long-term experiments. Beyond the 

basic ecological understanding that results from individual 

experiments, long-term experiments can be valuable by contributing 

to synthetic activities, testing and inspiring ecological 

theory and models across systems, and informing policy 

and management. With respect to synthesis, combining the 

results of multiple studies in meta-analyses or data syntheses 

can elucidate how factors such as climate, species composition, 

and edaphic conditions interact with treatments to 

influence ecological responses, which adds value to the original, 

individual experiments. For example, although numerous 

studies have demonstrated declines in plant species richness 

in response to N addition, synthesis of such results across a 

number of LTER sites revealed that the relative magnitude 

of species loss varied greatly across sites; warmer sites and 

those with high cation-exchange-capacity (CEC) soils exhibited 

lower declines in richness than sites that were colder 

or had low CEC (Clark et al. 2007). Sites with high abundance 

of C 4 grasses had a greater productivity response to N 

addition than did other sites, which was in turn associated 

with greater proportional declines in species richness. 

Long-term experiments conducted at multiple sites, 

although less common, are particularly powerful ways to 

Articles 

achieve general, synthetic understanding of ecological processes. 

Such studies can identify important influences on 

ecological processes at large spatial scales. They are more 

powerful than ad hoc meta-analytical syntheses, because they 

can eliminate variation in experimental methodologies across 

sites. For example, the LIDET studied long-term decomposition 

(10 years) by deploying common substrates across 27 

sites, including 16 LTER Network sites (Harmon et al. 2009). 

The LIDET experiment yielded new insights into decomposition 

processes including a new understanding of the rates 

and controls of decomposition of more slowly decomposing 

litter fractions. Such insights would have been difficult to 

derive from the meta-analysis of published decomposition 

studies because most of those studies last one year at most 

and because substrate variability confounds site-to-site variability 

in factors such as climate (Adair et al. 2010). 

Long-term experiments also take on added value when 

they inform and are informed by ecological theory and 

models. Tilman’s (1982) study of resource competition 

that developed the resource-ratio hypothesis of competitive 

interactions was enriched by the close interplay between 

long-term experiments and the development of theory at the 

CDR site. Similarly, empirical research on resource limitation 

at the ARC site has occurred in close connection with 

the development of theory on multiple-resource limitation 


Mean abundance of Harmonia axyridis 

(number per trap per week) 

Figure 7. Abundance of adult multicolored Asian lady 

beetles (Harmonia axyridis) at the Kellogg Biological 

Station Long Term Ecological Research site from 1989 to 

2010. This invasive species was first detected in 1994 and, 

until 1999, its mean abundance was approximately 0.2 

adults per trap per week (a). Following the arrival of the 

soybean aphid (Aphis glycines), major aphid outbreaks 

(arrows) occurred every other year from 2001 to 2005, 

prompting strong numerical responses by Harmonia in the 

subsequent years and overall greater mean abundance (b). 

Since 2007, no aphid outbreaks have occurred, and a 

new pattern of intermediate Harmonia abundance may 

be forming (c). The data are updated from Heimpel and 

colleagues (2010).

Articles 

(Herbert et al. 1999). Multiple-resource-limitation theory, in 

turn, has influenced the development of Earth-system models 

that explore the implications of N constraints on C cycling at 

a global scale (e.g., Gerber et al. 2010). Given the number and 

diversity of long-term experiments in the LTER Network, the 

opportunity certainly exists for additional theory and model 

development capable of linking a broad array of patterns and 

processes across temporal and spatial scales. 

There are broader practical outcomes of many long-term 

experiments when they influence the design of policy and 

management strategies. For example, the litter-exclusion 

experiment at the CWT site (figure 3) clearly demonstrated 

the importance of tree-leaf litter to the organisms living in 

headwater streams (Wallace et al. 1997). The importance 

of maintaining the integrity of headwater streams has been 

used as an argument against mountaintop mining practices 

in the central Appalachian region (Meyer and Wallace 2001) 

and cited in court cases (e.g., P. C. Chambers, US District 

Judge, Memorandum Opinion and Order on Civil Action 

No. 3:05-0784 in the Huntington Division of the US District 

Court of the Southern District of West Virginia, 23 March 

2007). Experimental demonstration of the dramatic impacts 

of deforestation on soil and surface-water chemistry at the 

HBR site (Bormann et al. 1974) raised awareness of the 

consequences of intensive forest harvest for environmental 

quality and influenced the development of best management 

practices for these forests. Simulations of hurricanes and 

mortality resulting from pests and pathogens demonstrated 

that the ecosystem consequences of salvage and restoration 

management of forests—both post-windstorm and in 

advance of insect infestation or disease—are often much 

greater than the impacts of the disturbances themselves and 

led to the argument that leaving forests alone is often a viable 

management alternative from an ecological perspective 

(Foster and Orwig 2006). Finally, the results from long-term 

experiments, coupled with gradient studies, field observations, 

and data from monitoring networks, have been used 

to estimate the critical loads of N for freshwater and terrestrial 

ecosystems of the United States (Pardo et al. 2011). 

Long-term experiments can provide a unique opportunity 

to directly evaluate the consequences of policy changes 

using an adaptive-management approach. In adaptivemanagement 

experiments, the researcher attempts to learn 

about managed systems by experimentally changing policy 

and assessing the outcomes on the appropriate time scale. 

At the NTL site, several policies were enacted as part of a 

long-term experiment designed to reduce P inputs and to 

control algal blooms in Lake Mendota, Wisconsin. Initial 

changes to land-use practices revealed little effect of management 

on the P inputs to the lake, possibly because of resuspension 

of P stored in sediments. Subsequent manipulation 

of the Lake Mendota food web to create a trophic cascade 

was more effective in reducing algal blooms, especially when 

they were accompanied by fishing-regulation changes and 

additional land-use practices implemented in the 1990s and 

2000s (Carpenter et al. 2006). 

Future policy can also be informed by results from 

long-term experiments. Work at the KBS site suggests 

that agricultural nitrous oxide (N 2 O) emissions (figure 3) 

increase exponentially with increasing rates of N fertilizer 

use, after the rates exceed the N-uptake capacity of the crop. 

In intensive agricultural systems, application of N in excess 

of plant needs is rather common, because fertilizer is inexpensive 

relative to commodity prices, and producers tend to 

hedge against the risk of insufficient N in order to achieve 

maximum yields. Using the concept of maximum return 

to N, Millar and colleagues (2010) developed a transparent 

N 2 O-reduction protocol suitable for incentivizing N 2 O 

reductions without affecting crop yields. They estimated 

that if the protocol were widely adopted as a part of future C 

cap-and-trade markets, the protocol could reduce N 2 O from 

fertilized row-crop agriculture by more than 50%. 

Long-term experiments as platforms for new and unplanned 

research. Because of their multiyear nature, long-term 

experiments can serve as platforms for research that goes 

well beyond the goals originally envisioned. As a result, they 

provide opportunities for novel studies that take advantage 

of imposed treatments, as well as for more detailed 

process-level studies to uncover mechanisms behind longterm 

patterns. For example, the soil-warming studies at the 

HFR site have become a platform for new studies of soil 

N; the consequences of garlic mustard invasion, an exotic 

species that inhibits mycorrhizae and the growth of some 

tree seedlings; and for microbial studies of the mechanisms 

responsible for the pattern of increased heterotrophic soil 

respiration followed by a diminishing response to warming 

(Bradford et al. 2008). This latter research showed that the 

apparent acclimation of soil respiration at the ecosystem 

scale results from the combined effects of reductions in soil 

C pools and microbial biomass and changes in the thermal 

response of microbial respiration. Mass-specific respiration 

rates were lower when seasonal temperatures were higher, 

which suggests that rate reductions under experimental 

warming probably occurred through temperature-induced 

changes in the microbial community. 

The long-term nutrient-addition studies that were conducted 

at multiple LTER sites to understand the role of N 

limitation, as well as to test ecosystem response to enhanced N 

deposition, have similarly provided a rich template for more 

short-term mechanistic studies. At the NWT site, for example, 

these studies have identified the importance of plant species 

traits in affecting community change (Suding et al. 2006) 

and demonstrated how plant–microbial feedbacks influence 

species coexistence (Ashton et al. 2008). Experiments initially 

designed to study vegetation change have been used to provide 

important insights as to how increased N availability can 

affect soil C storage (Neff et al. 2002). 

Long-term experiments are increasingly serving as platforms 

for ecological metagenomics studies. At the SEV site, 

for example, microbial ecologists are examining the metagenomic 

responses of rhizosphere microbes in a fully crossed 


multifactor experiment that includes increased winter rainfall, 

N amendment, and nighttime warming (Collins et al. 

2010). At the CDR site, metagenomic studies have shown 

a marked divergence of microbial communities in grassland 

communities developed under ambient as opposed to 

elevated CO 2 in terms of both composition and function, 

with communities under elevated CO 2 exhibiting increased 

abundance of the genes involved in labile C decomposition 

and C and N fixation (He et al. 2010). 

A call for a new generation of long-term experiments 

We have argued that well-designed long-term experiments 

focused on key questions and important processes can have 

tremendous value for individual sites. These long-term 

experiments can provide the understanding necessary for 

forecasting, coping with, and mitigating the consequences of 

human-driven global changes to the environment (Collins 

et al. 2011). However, the scope and pace of change occurring 

in ecological systems today—and forecast for the future— 

are, by all accounts, unprecedented in human history (Palmer 

et al. 2004, Solomon et al. 2007, Smith et al. 2009). Because 

of the global scale of altered biogeochemical cycles from 

increased atmospheric CO 2 concentrations, nutrient enrichment 

and depletion, climatic change (means and extremes; 

Smith 2011), exotic species introductions, and land-use 

change, all ecosystems are, and will continue to be, affected 

by these alterations (Solomon et al. 2007). Independently 

conducted site-based studies in which global-change factors 

(i.e., N or temperature) were manipulated and modeling 

studies both show that responses to these global-change factors 

may vary dramatically among ecosystems—from little or 

no response to a substantial change in function (e.g., Weltzin 

et al. 2003). But experimental ecologists need to think beyond 

the site level and provide a more complete understanding of 

how and why ecosystems differ in their susceptibility and 

sensitivity to these changes. Such understanding is critical for 

forecasting the ecological consequences of global change at 

regional to continental spatial scales and yet it is a challenge 

that ecologists have not fully met (Smith et al. 2009). 

This lack of comprehensive understanding occurs, in part, 

because ecologists have historically conducted disparate, 

largely independent experiments that tend to differ markedly 

with regard to (a) what is manipulated, (b) how much 

and for how long manipulations occur, and (c) what (and 

how) response variables are measured. This is certainly true 

of long-term experiments conducted in the LTER Network. 

As a result, syntheses across these studies can be difficult, 

since there is no way of knowing how much these different 

approaches contribute to the range of responses observed 

among ecosystems (Knapp et al. 2004). As Callahan (1984) 

pointed out more than 25 years ago, “even similar projects 

are not often comparable unless effort and resources have 

been devoted to making them so. This inherent tendency 

away from comparability becomes more prominent among 

projects conducted at locations that are geographically and 

biologically disjunct” (p. 363). Recently, renewed calls have 

Articles 

been made for establishing unified sampling protocols 

and conducting simultaneous multisite, multifactor experiments 

to address the most pressing global-change issues 

(Janzen 2009, Luo et al. 2011; e.g., The Nutrient Network, 

http://nutnet.science.oregonstate.edu). We echo and extend 

these calls by proposing that these multisite experiments 

should be planned as long-term experiments capable of 

elucidating both ecological dynamics and ecological mechanisms. 

As is appropriate, these should take advantage of the 

long-term observations and contextual understanding extant 

within the LTER Network as well as from nonnetwork sites 

and other existing and emerging observatories (Carpenter 

2008, Robertson et al. 2012 [in this issue]). 

Designing multisite experiments will not be without challenges. 

For example, how one scales treatment levels across 

disparate ecosystems can influence how these different ecosystems 

respond to “common treatments.” For experiments 

that alter CO 2 , treatments can be designed to increase CO 2 

by either a constant amount or a constant proportion with 

an expectation of comparable results because essentially all 

terrestrial ecosystems have very similar atmospheric CO 2 

concentrations. But for resources such as water and N, initial 

stocks and turnover rates can vary by orders of magnitude 

among mesic or xeric and fertile or infertile ecosystems. 

Therefore, experiments designed to increase or decrease 

resources by a constant proportion of availability and those 

designed to change them by a constant amount will likely 

lead to very different responses among sites. As was noted 

above, long-term data from observations and monitoring 

at LTER sites and the rich array of shorter-term processlevel 

studies that each site has conducted can be invaluable 

for providing the appropriate context for planning such 

experiments, as well as for interpreting their results. Longterm 

data can also provide the breadth of understanding 

necessary for devising more-detailed studies that are often 

necessary for elucidating key mechanisms. Other issues that 

must be addressed, and that LTER scientists grapple with 

today, include determining the end point for long-term 

experiments (e.g., when do the costs outweigh the new 

knowledge gained by continuing an experiment?) and how 

to optimally balance the trade-off between the spatial scale 

of manipulations and the number of replicates. Small-plotbased 

long-term experiments, such as those at the CDR site, 

can be designed to be statistically robust with many replicates. 

This is not feasible with large-scale whole-watershed 

manipulations, such as those at the HBR site (figure 1). The 

scale of manipulations also determines the degree to which 

long-term experiments can serve as platforms for additional 

research. These are important tradeoffs in design that need 

to be considered for future long-term experiments. 

We further advocate that such multisite long-term experiments 

should be designed in close collaboration with ecological 

modelers in order to alleviate the more vexing 

uncertainties in today’s models (Luo et al. 2011). Scenario 

planning (Peterson et al. 2003) can also be a valuable tool 

for designing experiments capable of providing information 


Articles 

relevant to stakeholders in addition to ecologists. These longterm 

experiments should be as large in scale as is feasible in 

order to permit the evaluation of ecological processes not possible 

in small-plot experiments (Smith et al. 2009). Designing 

large-scale experiments will present many unique challenges, 

since they are not simply big versions of the small-scale 

manipulations that ecologists have conducted in the past. 

Innovation in experimental design, statistical analysis, and 

engineering will be necessary. But as was noted above, with 

space set aside to accommodate unanticipated use, these networks 

of large-scale experiments would also serve as valuable 

research platforms for the broader ecological community. A 

highly coordinated and spatially extensive network of multifactor 

long-term experiments that adopts such a comparative 

approach would provide understanding and quantitative 

response data on ecological change at a temporal and spatial 

scale heretofore unavailable to ecologists. Such a network 

of experiments would complement the monitoring-based 

approach that the emerging National Ecological Observatory 

Network (NEON) has adopted by providing experimentally 

defined units of accelerated or amplified ecological change. 

Data from multisite, long-term experiments can be fused 

with models (Luo et al. 2011) to make more-robust forecasts 

for a broad range of ecosystems. Such forecasts can then be 

tested against the real-time tracking of ecological change 

provided by NEON and other observatory networks—and, 

of course, by the LTER Network. 

For long-term experimental networks as they are envisioned 

above to be realized, it is clear that clever designs, 

collaborations between ecologists and sensor and infrastructure 

engineers (Collins et al. 2006), and efficient 

deployment will be necessary from both scientific and economic 

perspectives. But as society demands the knowledge 

needed to cope with and mitigate global changes, ecologists 

would be remiss in forgoing the opportunity to build on the 

legacy of long-term experiments reviewed above to meet 

these challenges. 


We thank the principal investigators of the 26 Long Term 

Ecological Research Network sites for providing information 

regarding long-term experiments at their sites, Steve 

Carpenter and Tim Kratz for insight regarding North 

Temperate Lake–site experiments, and David Foster for providing 

editorial input and oversight. Robert Holt, Charles 

Driscoll, and an anonymous reviewer provided helpful comments 

on an earlier version of the manuscript. 


Adair EC, Hobbie SE, Hobbie RK. 2010. Single-pool exponential decomposition 

models: Potential pitfalls in their use in ecological studies. 

Ecology 91: 1225–1236. 

Ashton IW, Miller AE, Bowman WD, Suding KN. 2006. Nitrogen preferences 

and plant-soil feedbacks as influenced by neighbors in the alpine 

tundra. Oecologia 156: 625–636. 

Báez S, Collins SL, Lightfoot D, Koontz TL. 2006. Bottom-up regulation 

of plant community structure in an aridland ecosystem. Ecology 87: 

2746–2754. 

Blair JM. 1997. Fire, N availability and plant response in grasslands: A test of 

the transient maxima hypothesis. Ecology 78: 2359–2368. 

Bormann FH, Likens GE, Siccama TG, Pierce RS, Eaton JS. 1974. The export 

of nutrients and recovery of stable conditions following deforestation 

at Hubbard Brook. Ecological Monographs 44: 255–277. 

Bradford MA, Davies CA, Frey SD, Maddox TR, Melillo JM, Mohan JE, 

Reynolds JF, Treseder KK, Wallenstein MD. 2008. Thermal adaptation 

of soil microbial respiration to elevated temperature. Ecology Letters 

11: 1316–1327. 

Brown JH, Heske EJ. 1990. Control of a desert–grassland transition by a 

keystone rodent guild. Science 250: 1705–1707. 

Callahan JT. 1984. Long-term ecological research. BioScience 34: 

363–367. 

Carpenter SR. 2002. Ecological futures: Building an ecology of the long now. 

Ecology 83: 2069–2083. 

———. 2008. Emergence of ecological networks. Frontiers in Ecology 6: 228. 

Carpenter SR, Lathrop RC, Nowak P, Bennett EM, Reed T, Soranno PA. 2006. 

The ongoing experiment: Restoration of Lake Mendota and its watershed. 

Pages 236–256 in Magnuson JJ, Kratz TK, Benson BJ, eds. Long- 

Term Dynamics of Lakes in the Landscape. Oxford University Press. 

Chapin FS III, Shaver GR, Giblin AE, Nadelhoffer KJ, Laundre JA. 1995. 

Responses of arctic tundra to experimental and observed changes in 

climate. Ecology 76: 694–711. 

Clark CM, Cleland EE, Collins SL, Fargione JE, Gough L, Gross KL, 

Pennings SC, Suding KN, Grace JB. 2007. Environmental and plant 

community determinants of species loss following nitrogen enrichment. 

Ecology Letters 10: 596–607. 

Collins SL, et al. 2006. New opportunities in ecological sensing using wireless 

sensor networks. Frontiers in Ecology and the Environment 4: 

402–407. 

Collins SL, Fargione JE, Crenshaw CL, Nonaka E, Elliott JR, Xia Y, Pockman 

WT. 2010. Rapid plant community responses during the summer 

monsoon to nighttime warming in a northern Chihauhuan Desert 

grassland. Journal of Arid Environments 74: 611–617. 

Collins SL, et al. 2011. An integrated conceptual framework for long-term 

social–ecological research. Frontiers in Ecology and the Environment 

9: 351–357. 

Contosta AR, Frey SD, Cooper AB. 2011. Seasonal dynamics of soil respiration 

and N mineralization in chronically warmed and fertilized soils. 

Ecosphere 2 (3, Art. 36). 

Costamagna AC, Landis DA. 2006. Predators exert top-down control of 

soybean aphid across a gradient of agricultural management systems. 


Dasch AA, Blum JD, Eagar C, Fahey TJ, Driscoll CT, Siccama TG. 2006. The 

relative uptake of Ca and Sr into tree foliage using a whole-watershed 

calcium addition. Biogeochemistry 80: 21–41. 

Debinski DM, Holt RD. 2000. A survey and overview of habitat fragmentation 

experiments. Conservation Biology 14: 342–355. 

Franklin JF, Bledsoe CS, Callahan JT. 1990. Contributions of the long-term 

ecological research program. BioScience 40: 509–523. 

Frost TM, Fischer JM, Klug JL, Arnott SE, Montz PK. 2006. Trajectories of 

zooplankton recovery in the Little Rock Lake whole-lake acidification 

experiment. Ecological Applications 16: 353–367. 

Foster DR, Orwig DA. 2006. Preemptive and salvage harvesting of New 

England forests: When doing nothing is a viable alternative. Conservation 

Biology 20: 959–970. 

Foster DR, Aber JD, Melillo JM, Bowden RD, Bazzaz FA. 1997. Forest 

response to disturbance and anthropogenic stress: Rethinking the 1938 

Hurricane and the impact of physical disturbance vs. chemical and 

climate stress on forest ecosystems. BioScience 47: 437–445. 

Gardiner MM, Landis DA, Gratton C, DiFonzo CD, O’Neal M, Chacon 

JM, Wayo MT, Schmidt NP, Mueller EE, Heimpel GE. 2009. Landscape 

diversity enhances the biological control of an introduced crop pest in 

the north-central U.S. Ecological Applications 19: 143–154. 

Gerber S, Hedin LO, Oppenheimer M, Pacala SW, Shevliakova E. 2010. 

Nitrogen cycling and feedbacks in a global dynamic land model. Global 

Biogeochemical Cycles 24 (Art. GB1001). doi:10.1029/2008GB003336 


Gonzalez MJ, Frost TM. 1994. Comparisons of laboratory bioassays and a 

whole-lake experiment: Rotifer responses to experimental acidification. 


Harmon ME, Silver WL, Fasth B, Chen H, Burke IC, Parton WJ, Hart SC, 

Currie WS, LIDET. 2009. Long-term patterns of mass loss during the 

decomposition of leaf and fine root litter: An intersite comparison. 

Global Change Biology 15: 1320–1338. 

He Z, Xu M, Deng Y, Kang S, Kellogg L, Wu L, Van Nostrand JD, Hobbie 

SE, Reich PB, Zhou J. 2010. Metagenomic analysis reveals a marked 

divergence in the structure of belowground microbial communities at 

elevated CO 2 . Ecology Letters 13: 564–575. 

Heimpel GE, Frelich LE, Landis DA, Hopper KR, Hoelmer K, Sezen Z, 

Asplen MK, Wu K. 2010. European buckthorn and Asian soybean 

aphid as part of an extensive invasional meltdown in North America. 

Biological Invasions 12: 2913–2931. 

Herbert DA, Rastetter EB, Shaver GR, Ågren GI. 1999. Effects of plant 

growth characteristics on biogeochemistry and community composition 

in a changing climate. Ecosystems 2: 367–382. 

Hobbie JE, Carpenter SR, Grimm NB, Gosz JR, Seastedt TR. 2003. The US 

long-term ecological research program. BioScience 53: 21–32. 

Huggett BA, Schaberg PG, Hawley GJ, Eagar C. 2007. Long-term calcium 

addition increases growth release, wound closure, and health of sugar 

maple (Acer saccharum) trees at the Hubbard Brook Experimental 

Forest. Canadian Journal of Forest Research 37: 1692–1700. 

Janzen HH. 2009. Long-term ecological sites: Musings on the future, as seen 

(dimly) from the past. Global Change Biology 15: 2770–2778. 

Juice SM, Fahey TJ, Siccama TG, Driscoll CT, Denny EG, Eagar C, Cleavitt 

NL, Minocha R, Richardson AD. 2006. Response of sugar maple to calcium 

addition to northern hardwood forest. Ecology 87: 1267–1280. 

Knapp AK, Briggs JM, Hartnett DC, Collins SL. 1998. Grassland Dynamics: 

Long-Term Ecological Research in Tallgrass Prairie. Oxford University 

Press. 

Knapp AK, et al. 2004. Searching for generality in ecology: Post-hoc 

synthesis of long-term data from North American grasslands and 

South African savannas. Frontiers in Ecology and the Environment 2: 

483–491. 

Likens GE, Driscoll CT, Buso DC. 1996. Long-term effects of acid rain: 

Response and recovery of a forest ecosystem. Science 272: 244–246. 

Lindenmayer DB, Likens GE, Krebs CJ, Hobbs RJ. 2010. Improved probability 

of detection of ecological “surprises.” Proceedings of the National 


Luo Y, et al. 2011. Coordinated approaches to quantify long-term ecosystem 

dynamics in response to global change. Global Change Biology 17: 

843–854. 

Mack MC, Schuur EAG, Bret-Harte MS, Shaver GR, Chapin FS III. 2004. 

Ecosystem carbon storage in arctic tundra reduced by long-term 

nu trient fertilization. Nature 431: 440–443. 

Melillo JM, Steudler PA, Aber JD, Newkirk K, Lux H, Bowles FP, Cartricala 

C, Magill AH, Ahrens T, Morrisseau S. 2002. Soil warming and carboncycle 

feedbacks to the climate system. Science 298: 2173–2176. 

Meserve PL, Kelt DA, Milstead WB, Gutiérrez JR. 2003. Thirteen years of 

shifting top-down and bottom-up control. BioScience 53: 633–646. 

Meyer JL, Wallace JB. 2001. Lost linkages and lotic ecology: Rediscovering 

small streams. Pages 295–317 in Press MC, Huntly NJ, Levin S, eds. 

Ecology: Achievement and Challenge. Blackwell Science. 

Millar N, Robertson GP, Grace PR, Gehl RJ, Hoben JP. 2010. Nitrogen fertilizer 

management for nitrous oxide (N 2 O) mitigation in intensive corn 

(Maize) production: An emissions reduction protocol for US Midwest 

agriculture. Mitigation and Adaptation Strategies for Global Change 

15: 185–204. 

Neff JC, Townsend AR, Gleixner G, Lehman SJ, Turnbull J, Bowman WD. 

2002. Variable effects of nitrogen additions on the stability and turnover 

of soil carbon. Nature 419: 915–917. 

[NRC] National Research Council. 2001. Grand Challenges in Environmental 

Sciences. National Academies Press. 

Odum EP. 1977. The emergence of ecology as a new integrative discipline. 

Science 195: 1289–1293. 

Articles 

Pardo LH, et al. 2011. Effects of nitrogen deposition and empirical nitrogen 

critical loads for ecoregions of the United States. Ecological Applications 

21: 3049–3082. 

Peterson GD, Cumming GS, Carpenter SR. 2003. Scenario planning: A 

tool for conservation in an uncertain future. Conservation Biology 17: 

358–366. 

Rees M, Condit R, Crawley M, Pacala S, Tilman D. 2001. Long-term studies 

of vegetation dynamics. Science 293: 650–655. 

Reich PB, Hobbie SE, Lee T, Ellsworth DS, West JB, Tilman D, Knops JMH, 

Naeem S, Trost J. 2006. Nitrogen limitation constrains sustainability of 

ecosystem response to CO 2 . Nature 440: 922–925. 


world. BioScience 62: 342–353. 

Slavik K, Peterson BJ, Deegan LA, Bowden WB, Hershey AE, Hobbie JE. 

2004. Long-term responses of the Kuparuk River ecosystem to phosphorus 

fertilization. Ecology 85: 939–954. 

Smith MD. 2011. An ecological perspective on extreme climatic events: A 

synthetic definition and framework to guide future research. Journal of 

Ecology 99: 656–663. 

Smith MD, Knapp AK, Collins SL. 2009. A framework for assessing ecosystem 

dynamics in response to chronic resource alterations induced by 

global change. Ecology 30: 3279–3289. 

Solomon SD, Qin M, Manning Z, Chen M, Marquis KB, Avery T, Tignor M, 

Miller HL, eds. 2007. Climate Change 2007: The Physical Science Basis. 

Cambridge University Press. 

Suding KN, Miller A, Bechtold H, Bowman W. 2006. The consequences 

of species loss on ecosystem nitrogen cycling depends on community 

composition dynamics. Oecologia 149: 141–149. 

Syswerda SP, Basso B, Hamilton SK, Tausig JB, Robertson GP. 2012. 

Long-term nitrate loss along an agricultural intensity gradient in the 

upper Midwest USA. Agriculture, Ecosystems and Environment 149: 

10–19. 

Tilman D. 1982. Resource Competition and Community Structure. 

Princeton University Press. 

———. 1989. Ecological experimentation: Strengths and conceptual problems. 

Pages 136–157 in Likens GE, ed. Long-Term Studies in Ecology. Springer. 

———. 1996. Biodiversity: Population versus ecosystem stability. Ecology 

77: 350–363. 

Wallace JB, Eggert SL, Meyer JL, Webster JR. 1997. Multiple trophic 

levels of a forest stream linked to terrestrial litter inputs. Science 277: 

102–104. 

Weltzin JF, et al. 2003. Assessing the response of terrestrial ecosystems to 

potential changes in precipitation. BioScience 53: 941–952. 

Yeakley JA, Coleman DC, Haines BL, Kloeppel BD, Meyer JL, Swank WT, 

Argo BW, Deal JM, Taylor SF. 2003. Hillslope nutrient dynamics following 

upland riparian vegetation disturbance. Ecosystems 6: 154–167. 

Alan K. Knapp (aknapp@colostate.edu, alan.knapp@colostate.edu) is affiliated 

with the Graduate Degree Program in Ecology and with the Department 

of Biology at Colorado State University, Fort Collins. Melinda D. Smith 

and Kimberly J. La Pierre are affiliated with the Department of Ecology 

and Evolutionary Biology at Yale University, in New Haven, Connecticut. 

Sarah E. Hobbie is affiliated with the Department of Ecology, Evolution 

and Behavior at the University of Minnesota, St. Paul. Scott L. Collins is 

affiliated with the Department of Biology at the University of New Mexico, 

in Albuquerque. Timothy J. Fahey is affiliated with the Department of 

Natural Resources at Cornell University, in Ithaca, New York. Gretchen 

J. A. Hansen is affiliated with the Center for Limnology at the University of 

Wisconsin–Madison. Douglas A. Landis is affiliated with the Department of 

Entomology and with the Great Lakes Bioenergy Research Center at Michigan 

State University, in East Lansing. Jerry M. Melillo and Gaius R. Shaver are 

affiliated with The Ecosystems Center, Marine Biological Laboratory, in Woods 

Hole, Massachusetts. Timothy R. Seastedt is affiliated with the Department of 

Ecology and Evolutionary Biology and with the Institute of Arctic and Alpine 

Research, at the University of Colorado, Boulder. Jackson R. Webster is affiliated 

with the Department of Biological Sciences at Virginia Tech, Blacksburg. 


Articles 

Ecosystem Processes and Human 

Influences Regulate Streamflow 

Response to Climate Change at 

Long-Term Ecological Research Sites 

Julia a. Jones, irena F. Creed, Kendra l. HatCHer, robert J. Warren, Mary betH adaMs, Melinda H. 

benson, eMery boose, Warren a. broWn, JoHn l. CaMpbell, alan CoviCH, david W. CloW, CliFFord n. 

daHM, Kelly elder, CHelCy r. Ford, nanCy b. GriMM, donald l. HensHaW, Kelli l. larson, evan s. 

Miles, KatHleen M. Miles, stepHen d. sebestyen, adaM t. sparGo, asa b. stone, JaMes M. vose, 

and MarK W. WilliaMs 

Analyses of long-term records at 35 headwater basins in the United States and Canada indicate that climate change effects on streamflow are not 

as clear as might be expected, perhaps because of ecosystem processes and human influences. Evapotranspiration was higher than was predicted by 

temperature in water-surplus ecosystems and lower than was predicted in water-deficit ecosystems. Streamflow was correlated with climate variability 

indices (e.g., the El Niño–Southern Oscillation, the Pacific Decadal Oscillation, the North Atlantic Oscillation), especially in seasons when 

vegetation influences are limited. Air temperature increased significantly at 17 of the 19 sites with 20- to 60-year records, but streamflow trends were 

directly related to climate trends (through changes in ice and snow) at only 7 sites. Past and present human and natural disturbance, vegetation 

succession, and human water use can mimic, exacerbate, counteract, or mask the effects of climate change on streamflow, even in reference basins. 

Long-term ecological research sites are ideal places to disentangle these processes. 

Keywords: precipitation/runoff ratio, trend, succession, socioecological systems, Budyko curve 

Although many factors affect streamflow, recent concerns 

have been focused on the effects of climate change on 

streamflow. Increasing temperature, more-severe storms, 

advanced snowmelt, and declining snow cover are associated 

with increased drought and flooding (Groisman 

et al. 2004, Stewart et al. 2005, Huntington 2006, Barnett 

et al. 2008, Karl et al. 2009, McDonald et al. 2011, USDOI 

2011). Nevertheless, many human actions, natural disturbance 

effects, and ecosystem processes complicate, mitigate, 

and potentially counteract the climate effects on streamflow 

(Meybeck 2003, Jones 2011). Relevant human actions 

include ongoing disturbance and legacies of past disturbance, 

as well as global climate change. Understanding how 

climate change, social systems, and ecosystem processes 

affect streamflow is critical for mitigating conflicts between 

economic development and environmental conservation. 

Long-term studies of headwater basins, the source areas 

for water supplies, provide an informative starting point 

for understanding the effects of climate, social factors, and 

ecosystem processes on streamflow (figure 1). The US Forest 

Service (USFS) Experimental Forests and Ranges (EFRs) and 

the US Department of Agriculture Agricultural Research 

Service (ARS) established long-term studies in small basins 

throughout the United States beginning in the early 1900s 

(see supplemental table S1, available online at http://dx.doi. 

org/10.1525/bio.2012.62.4.10). Four EFRs (AND, CWT, HBR, 

LUQ; for the site-name abbreviations, see table S1) and one 

ARS site (JRN) became member sites of the US Long Term 

Ecological Research (LTER) Network as early as 1980. Some 

LTER Network sites also utilize streamflow records from the 

US Geological Survey (USGS) National Water Information 

Service database. Some headwater basin studies participate 

in the USGS Hydrologic Benchmark Network; the USGS 

Water, Energy, and Biogeochemical Budgets (WEBB) program; 

or the Canadian HydroEcological Landscapes and 

Processes (HELP) program. Climate and hydrologic data 

from many sites have been publicly available since the 1990s 

(www.fsl.orst.edu/climhy). 





Figure 1. Selected study basins from US Long Term Ecological Research (LTER) sites, US Forest Service Experimental 

Forests and Ranges, and US Geological Survey (USGS) Water, Energy, and Biogeochemical Budgets (WEBB) sites. 

Although they are less numerous than, for example, the 

USGS reference sites used in studies of climate change (e.g., 

Poff et al. 2007), LTER study sites provide unique insights 

Articles 

into the interacting effects of social systems, ecosystems, 

and climate change on hydrology. In common with all ecosystems 

on Earth, the study basins have a long history of 


Articles 

natural disturbances and human impacts. Most of the study 

basins experienced no management during the period of 

record, but all of them are experiencing succession from past 

human disturbances; many continue to experience natural 

disturbances; and a few have agriculture, forestry, or residential 

development. Moreover, these study sites have matching 

records of climate drivers and hydrologic responses 

from (in most cases) relatively small areas. Therefore, these 

sites and analyses provide a unique opportunity to compare 

hydrologic responses to climate drivers over multiple 

decades and to interpret the responses on the basis of concurrent 

studies of social and ecosystem processes. 

A conceptual model of social systems, climate, 

ecosystems, and water 

The fundamental premise of the present article is that 

streamflow responds to ecosystem processes, which, in 

turn, respond both to climate drivers and to social drivers 

(figure 2). Social systems are a primary driver of streamflow 

through human water use and regulation and may indirectly 

affect streamflow through human-induced climate change. 

However, even in headwater ecosystems lacking human 

residents, social factors, including population dynamics, 

economic development, political conflicts, and resource 

policies, produce ecosystem disturbances, including forest 

harvest or clearance, grazing, agriculture, mining, and fire. 

In turn, these disturbances influence ecological succession, 

evapotranspiration, and streamflow. 

Climate drivers—especially precipitation, temperature, 

snow and ice, and extreme events—also create ecosystem 

disturbances (e.g., wildfires, floods, wind and ice storms). 

Ecosystems continuously respond to disturbances (both 

human and natural) through ecological succession, and disturbances 

and responses differ among biomes. Ecosystems, 

social systems, and climate also respond to streamflow. 

Headwaters provide water for downstream ecosystems and 

communities, and ecosystem processes drive climate through 

evapotranspiration and energy exchange (figure 2). 

Social systems 

Economic development 

Political stability 

Demographic trends 

Resource policies 

Institutions 

Ecosystems 

Disturbance 

Succession 

Biomes 

Water use 

Streamflow 

Climate systems 

Precipitation 

Temperature 

Snow/ice dynamics 

Extreme events 

Figure 2. Conceptual model of social, ecological, and climate 

influences on streamflow. 

Many of the study sites experienced social effects on 

ecosystem processes prior to becoming LTER sites. Most of 

the temperate forest in the eastern half of the United States 

and Canada and tropical forests in Puerto Rico experienced 

forest harvest, land clearance for agriculture, and grazing, 

followed by land abandonment (Swank and Crossley 1988, 

Foster and Aber 2004). Sites in the southwestern United 

States experienced intensive grazing of domestic animals 

during the past four centuries (Peters et al. 2006). Boreal 

forest, temperate wet forest, and tundra sites experienced 

varying fire regimes, mostly driven by climate but, in some 

cases, affected by prehistoric peoples (e.g., Weisberg and 

Swanson 2003). Some sites contain agriculture, forestry, and 

urban development. 

In this study, we primarily examine the relationship between 

climate and streamflow on the basis of energy exchange. In 

cases in which streamflow behavior cannot be explained 

purely by climate, other hydrologic processes, as well as past 

and present human and natural disturbance effects on ecosystems, 

are considered as possible explanations. 

Study sites and questions 

In this study, we examined long-term records of air temperature 

(T), precipitation (P), and streamflow (Q) from 

35 basins at US LTER Network sites, USFS EFRs, USGS 

WEBB sites, and Canadian sites (table S1, figure 3a, 3b). 

The study sites were headwater basins that have mostly 

experienced no management since their records began. 

Nevertheless, the study basins have undergone succession in 

response to earlier natural or human disturbance, and some 

basins experienced natural disturbance during the periods 

of record. The records of T, P, and Q were obtained from 

the US LTER Network’s Climate and Hydrology Database 

Projects (ClimDB/HydroDB; http://climhy.lternet.edu), supplemented 

by USGS streamflow data (http://co.water.usgs. 

gov/lochvale, http://ny.cf.er.usgs.gov/hbn) and data from the 

Canadian HELP program (table S1). 

The study basins ranged from 0.1 to 10,000 square kilometers; 

2 were less than 10 hectares (ha), 5 were 10–100 ha; 

10 were 100–1000 ha; 5 were 1000–10,000 ha; 8 were 

10,000–100,000 ha; 2 were 100,000–1 million ha; and 2 were 

undefined (FCE, MCM; see table S1). The two largest basins 

were near ARC and GCE; eight additional large basins are 

located in the vicinity of CAP, HFR, JRN, KBS, NTL, OLY, 

PIE, and SEV. The smallest basins (smaller than 100 ha) are 

mostly associated with USFS or Canadian sites, at AND, BES, 

CWT, DOR, FER, HBR, MAR, and TLW. 

The study sites represent many biomes (potential vegetation) 

(figure 4). More than half of the sites were temperate 

forest (CAS, CWT, DOR, HBR, HFR, FER, KEJ, MAR, 

MRM, NTL, PIE, TLW), temperate wet forest (AND, CAR, 

MAY, OLY), and boreal forest (ELA, FRA, LVW, TEN, 

UPC). The remainder includes tundra or cold desert (ARC, 

MCM, NWT), warm desert (CAP, JRN, SEV), cool desert or 

woodland (BNZ), woodland or grassland (BES, FCE, GCE, 

KBS, KNZ, SBC), and tropical rainforest (LUQ). Actual 


a 

b 

ARC 

BNZ 

Legend 

Data analysis 

Budyko 

Trend 

Oscillations 

MAY 

CAR 

CAR 

CAR 

CAS 

ARC 

BNZ 

MAY 

CAR 

CAR 

CAR 

CAS 

UPC 

OLY 

AND 

SBC CAP 

OLY 

UPC 

AND 

SBC CAP 

Average annual P−PET (mm) 

< 49 

600–799 

50–99 800–999 

100–199 1000–1999 

200–399 2000–2999 

400–599 > 3.000 

MRM 

TEN 

LVW NWT 

FRA 

MRM 

SEV 

JRN 

TEN 

LVW NWT 

FRA 

SEV 

JRN 

ELA 

MAR 

KNZ 

KNZ 

NTL 

ELA 

MAR 

NTL 

KEJ 

TLW DOR 

HBR 

HFR 

KBS 

PIE 

BES 

FER 

BES 

FER 

Articles 


GCE 

TLW 

GCE 

CWT 

FCE 

DOR 

HFR 

KBS 

PIE 

FER 

BES 

CWT 

FCE 

KEJ 

HBR 

Figure 3. (a) Map of sites used in this analysis. The study-site characteristics and abbreviations are in supplemental table S1, 

available online at http://dx.doi.org/10.1525/bio.2012.62.4.10. The symbols indicate which of the three analyses (Budyko 

curve, trends, and correlation with climate indices [oscillations]) were conducted with data from that site. (b) Map of average 

annual precipitation (P) minus potential evapotranspiration (PET) in millimeters (mm) with study-site locations. 

LUQ 

LUQ

Articles 

Figure 4. Mean annual temperature, mean annual 

precipitation, and biomes of the study sites. The studysite 

abbreviations are in supplemental table S1, available 

online at http://dx.doi.org/10.1525/bio.2012.62.4.10. Data 

for the MCM site are not shown. Abbreviations: °C, degrees 

Celsius; mm, millimeters. 

and potential vegetation may differ because of disturbance 

(table S1). 

In our analyses, we examined three questions, using 

successively more-stringent requirements of data sets and 

interpreted these results in the light of ecological and social 

factors: (1) How is potential evapotranspiration (PET) 

related to actual evapotranspiration (AET) at each site, and 

how do these relationships compare with the theoretical 

Budyko curve (n = 30 sites)? (2) How is streamflow correlated 

with climate indices (e.g., the El Niño–Southern 

Oscillation [ENSO], the Pacific Decadal Oscillation [PDO], 

the North Atlantic Oscillation [NAO]; n = 21 sites)? (3) How 

have temperature, precipitation, and streamflow changed 

over time (n = 19 sites)? 

Energy- and water-balance relationships to observed 

water use 

The values of T, P, and Q from 30 sites with matched T, P, and 

Q (table S1, figure 3) were used to calculate PET and AET 

(i.e., P – Q). These values were plotted on the Budyko curve 

(Budyko 1974), which displays the relationship between 

PET and AET, each indexed by P (figure 5a). Thirty of the 

35 sites had data on T, P, Q, and basin area for a common 

10-year period (1993–2002), although a slightly adjusted 

period was used for 10 of the sites (figure 5b). PET was 

calculated from T (after Hamon 1963) on the basis of the 

number of daylight hours, mean monthly temperature, and 

the saturated vapor pressure. Annual PET was calculated as a 

sum of monthly values. The Budyko curve assumes that the 

water balance is Q = P – ET (evapotranspiration), with no 

significant losses to or gains from groundwater, and that the 

basins are at steady state, unaffected by vegetation dynamics 

(Donohue et al. 2007). 

The distribution of study basins on the Budyko curve 

reveals that observed water use in ecosystems in small basins 

deviated systematically from its expected dependence on 

energy and water balances. As was expected, observed ecosystem 

water use (AET ÷ P) was positively correlated to energy 

and water inputs to evapotranspiration (PET ÷ P) in sites 

with a water surplus (P ÷ PET) and insensitive to increases 

in energy at sites with a water deficit (P ÷ PET), following 

the theoretical Budyko curve (figure 5b). However, only 7 of 

30 sites (ARC, DOR, FRA, HBR, KEJ, KNZ, and OLY) fell 

on the Budyko curve, where observed water use (AET ÷ P) 

was equal to predicted water use (PET ÷ P) (fi gure 5b). Of 

the 19 sites with a moisture surplus (P ÷ PET) that did not 

fall on the Budkyo curve, 14 were above it, with higher than 

expected evapotranspiration (AET ÷ P > PET ÷ P). Of the 

five sites with moisture deficits (P < PET), four fell below the 

Budyko curve, with lower than expected evapotranspiration 

(AET ÷ P < PET ÷ P) (figure 5b). 

This result may indicate that ecosystems evaporate, transpire, 

and store more water than would be expected on the 

basis of temperature and day length at wet sites and less than 

would be expected at dry sites. Ecosystem structure (e.g., 

rooting depth, leaf area) and processes (e.g., adaptations to 

water deficits) may produce lower streamflow in wet sites and 

higher streamflow in dry sites than would be predicted from 

energy and water balances alone. However, other factors may 

also explain the departures of the sites from the theoretical 

Budyko curve. For instance, the PET value estimated from 

climate-station T records may not represent PET over entire 

basins, especially in mountain sites (e.g., AND, NWT, LVW). 

AET ÷ P is also considerably overestimated from P – Q 

in basins in which the groundwater recharge bypasses the 

stream gauge (Graham et al. 2010, Verry et al. 2011). 

When the annual values of T, P, and Q are plotted on the 

Budkyo curve, the interannual variation of AET relative 

to PET varies among biomes (figure 5c, 5d). Variation in 

AET ÷ P was less than in PET ÷ P at the desert sites (CAP, 

SEV) and at forested sites (AND, CAS, CWT, FER, HBR, 

MAR, NTL) (figure 5d). In contrast, at alpine sites (LVW, 

NWT), the interannual variation in AET ÷ P was large relative 

to the variation in PET ÷ P. This behavior of sites relative 

to the Budyko curve implies that ecosystems are capable of 

adjusting AET to compensate for climate variability at desert, 

grassland, and forest sites, but less so at alpine sites. 

Ecosystems have more-similar rates of net primary productivity 

per unit precipitation in dry than in wet years 

(Huxman et al. 2004). Comparisons of long-term AET and 

PET from study basins to the theoretical Budyko curve 

(figure 5) suggest that AET varies in a narrower range than 

would be expected from energy and water balances alone, 

which underscores the importance of ecosystem process 

effects on streamflow. 


a 

Evaporative index (actual 

evapotranspiration ÷ precipitation (AET÷P)) 

c 

Actual evapotranspiration ÷ precipitation 

1.6 

1.4 

1.2 

1.0 

0.8 

0.6 

0.4 

Water limit 

Energy limit 

Budyko curve 

0.2 

PET÷P1 

0.0 

energy limited water limited 

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 

Dryness index (potential evapotranspiration ÷ precipitation (PET÷P)) 

1.2 

1.0 

0.8 

0.6 

0.4 

0.2 

−0.2 

BA 

SBC 

BNZ 

A 

SEV 

CAP 

0.0 

0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 11.0 12.0 

Potential evapotranspiration ÷ precipitation 

Streamflow and regional climate oscillations 

We examined the relationship of three climate indices 

(ENSO, PDO, and NAO) to Q at 21 sites (tables S1 and S2). 

These indices measure multiyear or multidecadal oscillations 

of sea-surface temperatures and atmospheric-pressure 

differentials in the east–central tropical Pacific (ENSO), the 

northern Pacific (PDO), and the northern Atlantic (NAO). 

They are correlated with local and regional temperature, 

precipitation, and streamflow in the United States (Cayan 

et al. 1999, Barlow et al. 2001, Enfield et al. 2001). The sites 

included in this analysis had fewer than10 years of continuous 

(monthly) Q at one or more gauging stations, separated 

A 



ELA MARNTLBES 

GCE 

CAS 

TEN 

FER 

TLW 

CWT 

ARC 

AND 

UPC 

CAR 

KBS 

PIE 

MRM 

NWT 

DOR 

FRA 

OLY 

LVW 

KEJ 

HBR 

LUQ 

KNZ SBC BNZ 

Articles 

into a cool season (November–April) and a warm season 

(May–October). Correlated streamflow records (Pearson’s 

r > .80) were pooled at study sites with multiple stream 

gauges. Climate indices were obtained from online databases 

(NAO, www.cgd.ucar.edu/cas/jhurrell/indices.html; ENSO, 

www.cdc.noaa.gov/ClimateIndices/List; PDO, www.esrl.noaa. 

gov/psd/data/correlation/pdo.data). The streamflow–climate 

oscillation relationships were tested using generalized least 

squares models with autoregressive moving average functions; 

the models were evaluated with the Durbin–Watson test 

statistic and Akaike’s information criterion. The results are 

shown as the sign (+ or –) of the relationship of streamflow 


b 

d 

1.0 

SEV 

P−PET 

CAP 

Articles 

to climate oscillation variables that were significant at 

a < .10 in the models (table S2). 

The cool-season (November–April) and warm-season 

(May–October) streamflow values were significantly correlated 

with at least one climate index or interaction term 

at all sites except FRA, KNZ, NWT, and PIE (table S2). 

Significant correlations of streamflow were more frequent 

with ENSO (11, plus six interactions) and PDO (10, plus 

one interaction) than with NAO (4, plus six interactions) 

(table S2). Significant correlations were also slightly more 

frequent in winter (18) than in summer (14). These findings 

extend Greenland and colleagues’ (2003) analysis of climate 

indices, temperature, and precipitation at US LTER Network 

sites and corroborate the results of other studies. Molles 

and Dahm (1990) noted that streamflow in two rivers in 

New Mexico was significantly higher during El Niño (warm 

sea-surface temperatures in the eastern Pacific) than during 

La Niña conditions. Cayan and colleagues (1999) showed 

that days with high daily precipitation and streamflow 

were more frequent than average in the US Southwest and 

less frequent in the Northwest during El Niño by examining 

effects on snowpack accumulation and the subsequent 

melt. Enfield and colleagues (2001) found that sea-surface 

temperatures in the northern Atlantic are correlated with 

those in the northern Pacific and are associated with variations 

in streamflow in the Mississippi River and in Florida. 

Sea-surface temperature and pressure anomalies originating 

in the North Pacific affect cyclonic circulation over the East 

Coast and summer precipitation, streamflow, and drought 

(Barlow et al. 2001). 

These findings underscore the importance of separating 

the effects on streamflow of climate variability from longterm 

trends. Because the ENSO oscillation has a wavelength 

of 2–7 years, trends in climate and streamflow data sets over 

fewer than 20 years may simply reflect ENSO. Similarly, the 

PDO has a wavelength of 4–16 years (MacDonald et al. 2005), 

with mostly negative PDO in the 1950s to the mid-1970s 

and mostly positive PDO from 1976 to 1998. The NAO was 

predominantly negative between the 1950s and the early 

1970s and was mostly positive between the 1980s and the 

early 1990s. As a result, climate and streamflow trends from 

the 1950s to 2000 may be strongly affected by these climate 

oscillations. For example, at ARC, streamflow increased 

between 1988 and 2003, but declined between 1988 and 2008, 

so lengthening the record shifted the direction of apparent 

change. The lack of statistically significant increases in 

minimum temperature at the tundra ARC and boreal forest 

LVW sites (see the next section) may also be attributed to 

confounding effects of climate oscillations on these relatively 

short-term records. 

Climate oscillations influence ecosystem processes 

through streamflow and moisture. Streamflow was slightly 

more weakly correlated to climate indices in summer than 

in winter, perhaps because precipitation and streamflow are 

more closely related when ecosystems are dormant. ENSO 

is linked to aquatic-community structure in the Southwest 

(Sponseller et al. 2010), PDO is related to salmon returns 

in the Northwest (Mantua et al. 1997), and NAO is linked 

to stream salamander abundance in the Southeast (Warren 

and Bradford 2010). Headwater streamflow records are 

just beginning to be long enough to relate climate variability 

and trends to ecosystem processes and population 

dynamics. 

Climate and streamflow trends at long-term 

headwater basin study sites 

Nineteen sites had long-term records suitable for testing 

trends in T, P, and Q (table S1). Sites were included in the 

analysis if they had overlapping records of T, P, and Q 

that exceeded 20 years. The climate and streamflow record 

lengths used for trend estimation ranged from 20 to just 

over 60 years; five were 20–30 years; one was 30–40 years; 

four were 40–50 years; seven were 50–60 years; and two 

were more than 60 years (table S1). The records exceeding 

40 years are from USGS gauges and nearby climate stations 

(CAP, GCE, JRN, OLY, SBC, SEV), USFS EFRs that became 

US LTER Network sites (AND, CWT, HBR), other LTER 

Network sites (HFR), and USFS EFRs (FER, MAR) that 

did not become LTER Network sites. The records less than 

40 years in length were USGS gauges and nearby climate 

stations at LTER Network sites, WEBB sites, and EFRs (ARC, 

FRA, LUQ, LVW, NTL, and NWT). 

Interannual trends in minimum and maximum daily T, 

P, Q, and runoff ratios (Q:P) were estimated using linear 

regression and the Mann–Kendall nonparametric trend test 

(Helsel and Hirsch 2002). In these analyses, we used the 

period of record or from 1950 onward. Linear regressions 

and Mann–Kendall tests produced almost identical results 

(Hatcher 2011). The water year was defined as 1 October to 

30 September. Tests were conducted using daily data. The 

daily P and Q values were log transformed before analysis. 

Data were tested for autocorrelation before analysis, and 

residuals from linear regression analyses were also tested 

for autocorrelation. Significant trends in annual T, P, and Q 

were defined as 10 or more days (out of 365) with significant 

trends (at a � .025) and no autocorrelation before regression 

or in the residuals, and an average slope of the trend in 

daily values exceeding its standard error. 

Annual minimum or maximum daily temperature 

increased significantly at 17 of the 19 sites (minimum temperatures 

increased at 13 sites and maximum temperatures 

increased at 7 sites), but only two sites experienced significant 

changes in precipitation over the period of available 

record (figure 6). Minimum daily temperatures increased by 

several degrees Celsius since 1980 at NWT and FRA, highelevation, 

snow-dominated sites in the Rocky Mountains, 

but not at the other high-elevation Rocky Mountain site 

(LVW). Minimum daily temperature also increased by several 

degrees Celsius since the 1950s at climate stations near 

JRN and SEV in New Mexico, since the 1950s at a southeastern 

temperate forest site (CWT), and since the 1960s 

at a northern hardwood site (MAR) but not at its neighbor 


Annual precipitation (mm) 

4000 

3500 

3000 

2500 

2000 

1500 

1000 

500 

a 

0 

−15.0 −10.0 −5.0 0.0 5.0 10.0 15.0 20.0 25.0 

Minimum daily temperature (°C) 

1950 est 

1960 est 

1970 est 

1980 est 

1990 est 

2010 est 

(NTL, for which the record began in 1990). Mean annual 

precipitation increased significantly at LUQ and NWT. The 

first day of spring (defined as the last day of freezing temperature) 

moved earlier by between 0.31 and 1.98 days per 

year—that is, by more than 15 days in 50 years—at eight 

sites (AND, ARC, CWT, FER, FRA, HBR, LUQ, MAR, NWT) 

(Hatcher 2011). 

Runoff ratios (Q:P) changed at 8 of 19 sites (figure 6b). 

Tundra and boreal forest sites with ice and permafrost (LVW, 

NWT) experienced increases in runoff ratios, and so did 

temperate deciduous forest sites in the northeastern United 

States (HBR, HFR, PIE), which have a seasonal snowpack. An 

increase in runoff ratio means either that AET has decreased, 

or that there is a net addition of water to the system, such as 

from melting ice or interbasin water transfers. The observed 

increases in runoff ratios at LVW and NWT may be associated 

with the melt of ice, snow, and permafrost in response 

to warming temperatures during seasons in which these 

ecosystems are dormant (not taking up water). However, 

warming did not result in increased runoff ratios at other 

sites with permafrost (ARC, which has a short record) or 

seasonal snowpacks (e.g., AND, FRA, MAR, NTL). Runoff 

Articles 

1950 est 

1960 est 

1970 est 

1980 est 

1990 est 

2010 est 

.1 

0 

A CAP 

−10.0 −5.0 0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 

ratios did not change at most other sites, which mostly lack 

significant snow and ice (Hatcher 2011). 

Streamflow changes vary according to the season and differ 

among various biomes (figure 7). At undisturbed desert 

sites in Arizona (CAP) and New Mexico (SEV), streamflow 

did not change at any time of year (figure 7a, 7b). However, 

in a desert mountain basin northeast of JRN (New Mexico) 

and in a semiarid mountain basin near SBC (southern 

California) containing residential and urban development, 

streamflow increased during low-flow periods (figure 7c, 

7d). In a large basin in coastal Georgia containing agriculture 

and forest plantations (GCE), streamflow declined in 

early and late summer (figure 7e). 

At tundra sites on the North Slope of Alaska and in the 

Rocky Mountains (ARC, NWT; figure 7f, 7g), streamflow 

increased in early spring and late fall, during time periods 

adjacent to freezing periods. Streamflow increased in spring 

at boreal forest sites in the Rocky Mountains (FRA, LVW; 

figure 7h, 7i). At a temperate forest site in western North 

Carolina (CWT), where seasonal snowpacks do not form, 

streamflow did not change at any time of year (figure 7j), but 

at a temperate forest site in West Virginia (FER), streamflow 


Runoff ratio (discharge:precipitation) 

1 

.9 

.8 

.7 

.6 

.5 

.4 

.3 

.2 

b 

Maximum daily temperature (°C) 

Figure 6. (a) Multidecade trends in minimum daily temperature (in degrees Celsius [°C]) and precipitation (in millimeters 

[mm]) at long-term watershed study sites. Each site is designated by the minimum daily temperature and precipitation 

at the beginning and end of the decades spanning the period of record, based on the statistically significant trend 

in that variable over the period of record. The initial observation is connected to or contained within the 2010 

estimated (est) observation for each site. The radius of the 2010 estimated symbol is 0.5°C and 125 mm. Statistically 

significant increases in minimum daily temperature occurred at all sites except ARC, CAP, JRN, PIE, and SEV 

(the study-site characteristics and abbreviations are in supplemental table S1, available online at http://dx.doi. 

org/10.1525/bio.2012.62.4.10). Statistically significant changes in annual precipitation occurred only at LUQ and 

NWT. (b) Multidecade trends in maximum daily temperature and runoff ratios at long-term watershed study sites. 

Each site is designated by the maximum daily temperature and runoff ratio at the beginning and end of the decades 

spanning the period of record, based on the trend in that variable over the period of record. The initial observation is 

connected to or contained within the 2010 estimated observation for each site. The radius of the 2010 estimated symbol 

is 0.5°C and .0125. Statistically significant increases in maximum daily temperature occurred at CAP, FRA, JRN, LUQ, 

NWT, and SEV. Statistically significant changes in annual streamflow occurred at FER, GCE, HFR, HBR, JRN, LVW, 

MAR, NWT, NTL, PIE, and SEV.

Articles 

Proportion change per year 

relative to the mean daily flow 









a b 

CAP – Sycamore Creek – 1961–2010 

SEV – Jemez River – 1954–2010 

.10 

.015 

.05 

.00 

.01 

.005 

.00 

−.05 

−.005 

−.01 

O N D J F M A M J J A S O 

−.01 


Month 

Month 

c JRN – Rio Ruidoso – 1950–2010 

d SBC – San Jose Creek – 1950–2010 

.015 

.08 

.01 

.06 

.005 

.04 

.00 

.02 

−.005 

.00 

−.01 

−.02 

−.015 


−.04 


Month 

Month 

e GCE – Ohoopee River – 1950–2009 

f ARC – Kuparuk River – 1971–2010 

.02 

.20 

.01 

.15 

.10 

.00 

.05 

−.01 

.00 

−.05 

−.02 

−.10 

−.03 


−.15 


Month 

Month 

g 

NWT – Green Lake 4 – 1981–2008 

FRA – East Saint Louis – 1976–2005 

.10 

.04 

.05 

.03 

.02 

.00 

.01 

−.05 

.00 

−.01 

−.10 

−.02 

−.15 


−.03 


Month 

Month 

i LVW – Loch Vale out let – 1984–2010 

j 

CWT – WS18 – 1950–2009 

.10 

.015 

.001 

.05 

.005 

.00 

.000 

−.005 

−.05 

−.010 

−.10 


−.015 


Month 

Month 

Figure 7. Daily changes in streamflow at 19 US Long Term Ecological Research sites, US Forest Service Experimental 

Forests and Ranges, and US Geological Survey Water, Energy, and Biogeochemical Budgets sites arranged by biome (from 

figure 4) as a function of the day of the water year (1 October to 30 September). (a–c) Desert sites (CAP, SEV, JRN); 

(d), (e) savanna sites (SBC, GCE); (f), (g) tundra sites (ARC, NWT); (h), (i) boreal forest sites (FRA, LVW); (j–p) temperate 

forest sites (CWT, FER, HFR, HBR, MAR, NTL, PIE); (q), (r) wet temperate forest sites (AND, OLY); (s) wet tropical 

forest site (LUQ). The vertical axis and the green line are the slope of regression of log-transformed streamflow for each 

day of the water year over the period of record (see supplemental table S1, available online at http://dx.doi.org/10.1525/ 

bio.2012.62.4.10). The vertical axis units are the proportion change per year relative to the mean daily flow. The 

percentage change can be calculated as (1 + p) n , where p is the proportion change and n is the number of years. Note the 

different vertical axis scales. The horizontal black line represents no change (a proportion change of 0); the wiggly black 

lines are the upper and lower bounds on the 97.5% confidence interval. The red dots represent significant increases, and 

the blue dots represent significant decreases in daily streamflow, where a � .025. 












h









k 

increased in the summer (figure 7k). Winter streamflow 

increased at three temperate forest sites in New England (HFR, 

HBR, PIE) and declined at one (NTL) (figure 7l, 7m, 7o, 7p). 

In addition, streamflow increased in March and decreased 

in April at HBR (figure 7m), and it increased in March and 

declined in summer at MAR (figure 7n). At wet temperate forest 

sites in Oregon (AND, OLY), streamflow declined in spring 

(figure 7q, 7r). Streamflow did not change at any time of year 

at a wet tropical forest site in Puerto Rico (LUQ) (figure 7s). 

PIE – Ipswich River – 1938–2010 

Articles 

.08 

FER – WS4 – 1952–2007 

.06 

HFR – Swift River – 1964–2010 

.06 

.04 

.04 

.02 

.02 

.00 

.00 

−.02 

−.02 

−.04 

−.04 


−.06 


Month 

Month 

m 

−.02 

O N D J F M A 

Month 

M J J A S O 

Social, ecological, and climate factors influencing 

streamflow trends 

Multiple social and ecological factors may explain the 

streamflow trends at long-term headwater basin sites, even 

though humans do not directly affect most of these sites 

( figure 2). Economic development, population growth, and 

the use of fossil-fuel resources have increased atmospheric 

carbon dioxide, warmed the Earth, contributed to more- 

intense precipitation events, and increased evapotranspiration 






.02 

AND – WS2 – 1958–2009 

.03 

OLY – Hoko River near Sekiu – 1962–2009 

.01 

.02 

.00 

.01 

.00 

−.01 

−.01 

−.02 

−.02 

−.03 


−.03 


Month 

Month 



.04 

.02 

.00 

−.02 

−.04 





.04 

HBR – WS3 – 1958–2007 

.10 

MAR – S2 – 1961–2009 

.03 

.05 

.02 

.01 

.00 

.00 

−.05 

−.01 

−.10 

−.02 


−.15 


Month 

Month 

.01 

.00 

−.01 

−.02 

−.03 

−.04 

−.05 

−.06 

o 


Month 

Figure 7. (Continued) 

NTL – Trout River near Trout Lake – 1990–2010 

.03 

.02 

.01 

.00 

−.01 

q r 

s 

LUQ – Espiritu Santo – 1975–2009 

−.06 

O N D J F M A 

Month 

M J J A S O 

l 

n 

p

Articles 

(Min et al. 2011, Pall et al. 2011), which in turn have been 

linked to increased flooding and drought (Barnett et al. 

2008, Karl et al. 2009). Yet direct climate-trend effects on 

streamflow in headwater basins may be mitigated by ecological 

processes, including disturbance, succession, and 

vegetation adaptations to water scarcity. In many cases, the 

vegetation—and, therefore, evapotranspiration in headwater 

basins—is affected by human activities, such as past 

logging, grazing, agriculture, and fire suppression. Moreover, 

in some headwater basins, land-use changes, including agriculture 

and exurban expansion, may mitigate or overwhelm 

climate-trend effects on streamflow. 

Some observed trends in streamflow appear to be direct 

effects of climate trends. For example, increased streamflow 

in fall and spring at ARC, FRA, and NWT is probably the 

result of the expanding period of thaw at these tundra and 

boreal forest sites (figure 7f, 7g, 7i). Increases in streamflow 

at these sites may be driven by permafrost melt; changes 

in the chemical composition of streamflow support this 

hypothesis (Bowden et al. 2008, Caine 2010). In addition, 

increased spring streamflow at temperate forest sites with 

seasonal snowpacks in New England (HBR) and in the 

upper Midwest (MAR) is probably the result of earlier 

snowmelt, whereas increased winter streamflow at temperate 

forest sites in New England (HFR) may be the result of 

a shift from snow to rain. These responses are consistent 

with the results of published studies (Hodgkins et al. 2003, 

Stewart et al. 2005, Clow 2010, Campbell et al. 2011). 

Some observed trends in streamflow may be the result 

of biological responses to climate change. For example, 

declining streamflow at a woodland site (summer at GCE), a 

boreal forest site (LVW), temperate forest sites (MAR, NTL), 

and wet temperate forest sites (spring at AND, OLY) may 

be the result of increased evapotranspiration in response to 

warmer temperatures. Conifer forests, which occur at MAR, 

NTL, LVW, AND, and OLY, are adapted to photosynthesize 

and respire when conditions are favorable; warmer temperatures 

may lead to an earlier onset of transpiration and 

to declining streamflow (e.g., Moore KM 2010). Streamflow 

trends at wet temperate and boreal forest sites (LVW, AND, 

OLY; figure 7) are restricted to the immediate period of 

snowmelt, and declines may reflect increased evapotranspiration 

(Moore KM 2010, Oishi et al. 2010, Campbell et al. 

2011). In contrast, streamflow trends are largest during the 

nonsnowmelt periods at temperate forest sites in the upper 

Midwest (NTL, MAR), where wetlands (bogs and lakes) 

occupy a large proportion of basin area (Verry et al. 2011). 

Increased evapotranspiration associated with declining ice 

cover (Magnuson et al. 2000) or increased radiation associated 

with decreased precipitation may account for declining 

flows at these sites. 

Some sites experienced no trends in streamflow, despite 

increases in temperature. For example, desert sites (CAP and 

SEV) and the wet tropical site (LUQ) experienced significant 

increases in minimum and maximum daily temperatures, 

no change in precipitation (figure 6), and almost no changes 

in streamflow (figure 7a, 7b, 7s). Vegetation adaptations to 

drought might explain the lack of a streamflow response to 

warming at the desert sites. At the wet tropical forest site, 

the effects of the 1996 Hurricane Hugo on leaf area, evapotranspiration, 

and streamflow (Scatena et al. 1996) may have 

overwhelmed climate-trend effects. 

Some sites experienced trends in streamflow that appear 

to be biological responses to past disturbances. For example, 

forest succession and declining evapotranspiration may 

explain increased summer streamflow at two temperate 

forest sites (FER, HBR), which were logged in the early 

nineteenth century (figure 7k, 7m). At a third temperate 

forest site (CWT), streamflow and runoff ratios have not 

changed, despite increases in air temperature (figure 6, 

figure 7j). Forest succession following disturbances in the 

early 1900s at all three sites (table S1; Swank et al. 2001, 

Adams et al. 2006) and associated changes in species composition 

or leaf area may have influenced streamflow trends. 

Analyses of long-term paired-basin experiment data (e.g., 

Jones and Post 2004) indicate that streamflow continues to 

change over decades or centuries of forest succession after 

disturbance. 

Responses to land use and disturbance, such as advanced 

snowmelt or declining summer streamflow, may be misconstrued 

as responses to climate change. In paired-basin 

experiments (see the discussion of long-term experiments 

in Knapp and colleagues 2012 [in this issue]), forest harvest 

advanced the timing of peak snowmelt and associated 

streamflow by up to three weeks in temperate forest sites 

with a seasonal snowpack (AND, HBR); the effect lasted for 

more than 10 years (Jones and Post 2004). By 25 to 35 years 

after forest harvest in temperate forest basins (AND, CWT, 

HBR), summer streamflow declined by up to 30%–50% 

relative to the reference basins (Hornbeck et al. 1997, Swank 

et al. 2001, Jones and Post 2004). Regenerating species in 

early forest succession may transpire more water per unit of 

leaf area and, in some cases, have greater total leaf area than 

the species that were removed, which would reduce summer 

streamflow (Swank et al. 2001, Moore GM et al. 2004). 

Historic legacies from past disturbance in these long-term 

studies demonstrate that streamflow and timing responses 

to forest disturbance are at least as large as responses associated 

with climate trends over the past 20–60 years at the 

study sites. Daily streamflow during the late summer and 

early fall increased by up to 300% in the 1–5-year period 

after experimental forest harvest (AND, CWT, HBR), but 

most daily changes were on the order of 50% or less (Jones 

and Post 2004). By comparison, trends in daily streamflow 

associated with climate trends at the 19 study sites were on 

the order of 0.005–0.05 of log(Q) per year. The lower value 

is equivalent to changes of 10%–25%, and the higher value 

is equivalent to more than a 100% change over 20–60 years, 

but changes of this magnitude are restricted to a few days 

per year (figure 7). 

Finally, some observed trends in streamflow may be 

direct human effects on the hydrologic cycle. For example, 


increased irrigation using groundwater or water imported 

from other basins may explain increasing streamflow during 

dry seasons at a desert site in New Mexico (JRN) and a 

savanna site in southern California (SBC), which have some 

agriculture and residential development (figure 7c, 7d). 

Increasing winter streamflow trends at a temperate forest 

site (PIE) may reflect urban expansion (Claessens et al. 

2006). Therefore, human effects on streamflow may mimic, 

exacerbate, counteract, or mask climate effects on streamflow, 

making it challenging to determine the vulnerability 

of human communities (sensu Polsky et al. 2007) to variations 

in water supply. 

Headwater basins in this study drain into major river 

systems that supply water to major agricultural areas and 

medium and large cities. Climate change is expected to 

increase the variability of future streamflow and to stress 

municipal water supplies (Milly et al. 2008, Covich 2010, 

McDonald et al. 2011, USDOI 2011). Long-term studies 

of headwater basins can help distinguish biophysical from 

social causes of variability in water supply and, hence, 

the relationships between ecological and social resilience 

(Adger 2000). Water scarcity may be perceived even in 

areas with abundant rainfall, where politics rather than 

true scarcity may govern water restrictions (Hill and Polsky 

2006). Meanwhile, residents, professional policymakers, 

and academics in Phoenix (CAP) implicated population 

growth, climate change, and drought as the most important 

causes of water scarcity, rather than their own wateruse 

habits (Larson et al. 2009). Adding to this research, 

in this study, we suggest that rather complex interactions 

among historical social factors, ecosystem processes, and climate 

influence the long-term water supply from headwater 

basins. 

The role of information management 

Long-term ecological data are critical to answering societal 

questions of national concern and significance. Long-term 

data are the only way to distinguish trends from short-term 

variability in key environmental indicators, such as climate 

and streamflow. However, many long-term data remain 

inaccessible or difficult to access. Many valuable data sets are 

stored in inconvenient file formats with limited metadata. 

Variations in methods, variables, units, measurement scales, 

and quality-control annotation complicate data integration 

and prevent automated approaches to data synthesis. 

Until the 1990s, the difficulty of identifying, accessing, 

and integrating climate and hydrologic data from 

the LTER Network, EFRs, and related networks precluded 

cross-site studies. ClimDB/HydroDB (http://climhy.lternet. 

edu), a collaborative effort between LTER Network and 

USFS information managers, was initiated in 1997 to 

overcome these limitations. ClimDB/HydroDB is a Web 

harvester and data warehouse that provides uniform access 

and visualization of daily streamflow and meteorological 

data through a single portal. Participating sites manage 

original data within their local information systems but 

Articles 

periodically contribute data to the warehouse. Although 

the ClimDB/HydroDB approach is not a complete solution 

to data-access and -integration issues, it has served as an 

effective bridge technology between older, more rigid data- 

distribution models and modern service-oriented architectures 

(Henshaw et al. 2006). 

The LTER Network has made great strides in collecting, 

archiving, and integrating long-term data sets online, 

enabling synthesis activities such as this one, and providing 

an example for other environmental observatories. 

Information managers at LTER Network sites have led the 

development of metadata standards, data dictionaries, and 

software for data integration. The LTER Network datamanagement 

system serves as a model for emerging national 

observatories and existing programs. 

Conclusions 

This study provides an example of the special kinds of science 

that are possible from networks of long-term study 

sites. Climate and streamflow records are sufficiently long 

that averages, variability, and trends can be meaningfully 

analyzed. The climate and streamflow properties are 

simple and comparable because they are consistently measured 

across sites. Climate and streamflow data are broadly 

relevant to ecosystem processes and ecosystem services. 

Above all, multisite synthesis is fostered by and contributes 

to an open, inclusive culture of science collaboration. 

This study showed that actual evapotranspiration was 

predicted by PET at only 7 of 30 sites with 10-year-long 

records (the Budyko curve). Taken individually, these departures 

might simply reflect the inability of a climate station 

to represent the conditions of a whole basin or the fact 

that streamflow depends on groundwater and other forms 

of storage, as well as precipitation and temperature. But 

taken collectively, the departures of these sites from the 

Budyko prediction suggest the intriguing hypothesis that 

water-scarce ecosystems evapotranspire less and waterabundant 

ecosystems evapotranspire more than would 

be predicted from their climates. Moreover, streamflow 

at many of these sites was significantly related to one or 

more climate index (ENSO, NAO, or PDO), which is not 

surprising, but the slightly more frequent significant correlations 

of streamflow with climate indices in winter than 

in summer imply that ecosystem processes mediate climate– 

streamflow coupling. Finally, 17 of 19 sites had significant 

increases in their minimum or maximum daily temperatures 

or both, but streamflow trends were directly related 

to climate trends at only 7 of the sites, all of which have 

permanent or seasonal ice and snow. In contrast, at other 

sites, and during certain seasons at these seven sites, streamflow 

trends were contrary to those expected from climate 

drivers. 

A key finding from this study is that the past and present 

human uses of ecosystems and human water-use practices 

can mimic, exacerbate, counteract, or mask the effects of 

climate change on streamflow. Social factors, including 


Articles 

past land management and disturbance and the resulting 

ecological succession, all influence trends in streamflow, 

even at sites that are considered to be “reference” basins. In 

other words, exogenous (climate) factors are not the only 

drivers of nonstationarity (e.g., Milly et al. 2008) in streamflow 

from headwaters: Ecosystem processes and their social 

drivers are also important controls. In order to understand 

reference conditions for natural flow regimes (e.g., Poff 

et al. 2007), we need to better understand how ecological 

processes and social drivers mediate the expression of 

climate on streamflow. Long-term study sites, where all 

these processes are being studied, are ideal places for this 

ongoing work. 


Funding for this work was provided by the US Long Term 

Ecological Research (LTER) Network and National Science 

Foundation grants to participating sites and by a Natural 

Sciences and Engineering Research Council of Canada 

Discovery Grant to IFC. We thank LTER Network information 

managers for the creation and maintenance of 

ClimDB/HydroDB; the US Forest Service for the initial 

establishment and continued support of climate and basin 

measurements at many of the study sites; the US Geological 

Survey (USGS) Water, Energy, and Biogeochemical Budgets 

program, the USGS Hydrologic Benchmark Network, and 

the USGS National Water Information Service for provision 

of data; and the Networks of Centres of Excellence– 

Sustainable Forest Management Network–funded project on 

HydroEcological Landscapes and Processes (HELP) and the 

participating Canadian experimental basins from which data 

were contributed to the HELP project (Peter Tschaplinski for 

CAR, Tom Clair for KEJ, Fred Beall for TLW and MRM, 

Ray Hesslein for ELA, Peter Dillon for DOR, and Rita 

Winkler for UPC). We thank Merryl Albers, David R. Foster, 

Eveleyn Gaiser, Ann Giblin, Stephen P. Loheide II, Randy K. 

Kolka, Richard V. Pouyat, Sylvia Schaefer, Emily H. Stanley, 

Frederick J. Swanson, Will Wollheim, and three anonymous 

reviewers for comments on the manuscript. 


Adams MB, DeWalle DR, Hom JL, eds. 2006. The Fernow Watershed 

Acidification Study. Springer. 

Adger WN. 2000. Social and ecological resilience: Are they related? Progress 

in Human Geography 24: 347–364. 

Barnett TP, et al. 2008. Human-induced changes in the hydrology of the 

western United States. Science 319: 1080–1083. 

Barlow M, Nigam S, Berbury EH. 2001. ENSO, Pacific decadal variability, 

and U.S. summertime precipitation, drought, and stream flow. Journal 

of Climate 14: 2105–2128. 

Bowden WB, Gooseff MN, Balser A, Green A, Peterson BJ, Bradford J. 

2008. Sediment and nutrient delivery from thermokarst features in the 

foothills of the North Slope, Alaska: Potential impacts on headwater 

stream ecosystems. Journal of Geophysical Research 113 (Art. G02026). 

doi:10.1029/2007JG000470 

Budyko MI. 1974. Climate and Life. Academic Press. 

Caine N. 2010. Recent hydrologic change in a Colorado alpine basin: An 

indicator of permafrost thaw? Annals of Glaciology 51: 130–134. 

Campbell JL, Driscoll CT, Pourmokhtarian A, Hayhoe K. 2011. 

Streamflow responses to past and projected future changes in climate 

at the Hubbard Brook Experimental Forest, New Hampshire, 

United States. Water Resources Research 47 (Art. W02514). 

doi:10.1029/2010WR009438 

Cayan DR, Redmond KT, Riddle LG. 1999. ENSO and hydrologic extremes 

in the western United States. Journal of Climate 12: 2881–2893. 

Claessens L, Hopkinson C, Rastetter E, Vallino J. 2006. Effect of historical 

changes in land use and climate on the water budget of an urbanizing 

watershed. Water Resources Research 42 (Art. W03426). 

doi:10.1029/2005WR004131 

Clow DW. 2010. Changes in the timing of snowmelt and streamflow 

in Colorado: A response to recent warming. Journal of Climate 23: 

2293–2306. doi: 10.1175/2009JCLI2951.1 

Covich AP. 2010. Adaptation to sustain high-quality freshwater supplies 

in response to climatic change. Resources For the Future. Issue Brief 

no. 10-05:1-22. 

Donohue RJ, Roderick ML, McVicar TR. 2007. On the importance 

of including vegetation dynamics in Budyko’s hydrological model. 

Hydrology and Earth System Sciences 11: 983–995. 

Enfield DB, Mestas-Nuñez AM, Trimble PJ. 2001. The Atlantic multidecadal 

oscillation and its relation to rainfall and river flows in the continental 

U.S. Geophysical Research Letters 28: 2077–2080. 

Foster DR, Aber JD, eds. 2004. Forests in Time: The Environmental 

Consequences of 1,000 Years of Change in New England. Yale University 

Press. 

Graham CB, van Verseveld W, Barnard HR, McDonnell JJ. 2010. Estimating 

the deep seepage component of the hillslope and catchment water 

balance within a measurement uncertainty framework. Hydrological 

Processes 24: 3631–3647. 

Greenland D, Goodin DG, Smith RC. 2003. Climate Variability and 

Ecosystem Response at Long-Term Ecological Research Sites. Oxford 

University Press. 

Groisman PY, Knight RW, Karl TR, Easterling DR, Sun B, Lawrimore JH. 

2004. Contemporary changes of the hydrological cycle over the contiguous 

United States: Trends derived from in situ observations. Journal of 

Hydrometeorology 5: 64–85. 

Hamon WR. 1963. Computation of direct runoff amounts from storm rainfall. 

International Association of Scientific Hydrological Publications 

63: 52–62. 

Hatcher KL. 2011. Interacting effects of climate, forest dynamics, landforms, 

and river regulation on streamflow trends since 1950: Examples from 

the Willamette basin and forested headwater sites in the US. Master’s 

thesis. Oregon State University, Corvallis. 

Helsel DR, Hirsch RM. 2002. Statistical methods in water resources. 

Techniques of Water Resources Investigations, Book 4, Chapter A3. 

US Geological Survey. (30 January 2012; http://pubs.usgs.gov/twri/ 

twri4a3) 

Henshaw DL, Sheldon WM, Remillard SM, Kotwica K. 2006. ClimDB/ 

HydroDB: A Web harvester and data warehouse approach to building 

a cross-site climate and hydrology database. In: Proceedings of 

the 7th International Conference on Hydroscience and Engineering, 

Philadelphia, PA, September 2006. (30 January 2012; http://hdl.handle. 

net/1860/1434). 

Hill TD, Polsky C. 2006. Adaptation to drought in the context of suburban 

sprawl and abundant rainfall. Geographical Bulletin 47: 85–100. 

Hodgkins GA, Dudley RW, Huntington TG. 2003. Changes in the timing of 

high flows in New England over the 20th century. Journal of Hydrology 

278: 244–252. 

Hornbeck JW, Martin CW, Eagar C. 1997. Summary of water yield 

experiments at Hubbard Brook Experimental Forest, New Hampshire. 

Canadian Journal of Forest Research 27: 2043–2052. 

Huntington TG. 2006. Evidence for intensification of the global water cycle: 

Review and synthesis. Journal of Hydrology 319: 83–95. doi:10.1016/ 

j.jhydrol.2005.07.003 

Huxman TE, et al. 2004. Convergence across biomes to a common rain-use 

efficiency. Nature 429: 651–654. 


Jones JA. 2011. Hydrologic responses to climate change: Considering geographic 

context and alternative hypotheses. Hydrological Processes 25: 

1996–2000. doi:10.1002/hyp.8004 

Jones JA, Post DA. 2004. Seasonal and successional streamflow response 

to forest cutting and regrowth in the northwest and eastern United 

States. Water Resources Research 40 (Art. W05203). doi:10.1029/ 

2003WR002952 

Karl TR, Melillo JM, Peterson TC, eds. 2009. Global Climate Change 

Impacts in the United States: A State of Knowledge Report from the U.S. 

Global Change Research Program. Cambridge University Press. 



Larson KL, White D, Gober P, Harlan S, Wutich A. 2009. Divergent perspectives 

on water resource sustainability in a public-policy–science context. 

Environmental Science and Policy 12: 1012–1023. 

MacDonald GM, Case RA. 2005. Variations in the Pacific Decadal 

Oscillation over the past millennium. Geophysical Research Letters 32 

(Art. L08703). doi:10.1029/2005GL022478 

Magnuson JJ, et al. 2000. Historical trends in lake and river ice cover in the 

northern hemisphere. Science 289: 1743–1746. 

Mantua NJ, Hare SR, Zhang Y, Wallace JM, Francis RC. 1997. A Pacific interdecadal 

climate oscillation with impacts on salmon production. Bulletin 

of the American Meteorological Society 78: 1069–1079. 

McDonald RI, Green P, Balk D, Fekete BM, Revenga C, Todd M, 

Montgomery M. 2011. Urban growth, climate change, and freshwater 

availability. Proceedings of the National Academy of Sciences 108: 

6312–6317. 

Meybeck M. 2003. Global analysis of river systems: From Earth system 

controls to Anthropocene syndromes. Philosophical Transactions of the 

Royal Society B 358: 1935–1955. doi:10.1098/rstb.2003.1379 

Milly PCD, Betancourt J, Falkenmark M, Hirsch RM, Kundzewicz ZW, 

Lettenmaier DP, Stouffer RJ. 2008. Stationarity is dead: Whither water 

management? Science 319: 573–574. 

Min SK, Zhang X, Zwiers FW, Hegerl GC. 2011. Human contribution to 

more-intense precipitation extremes. Nature 470: 378–381. 

Molles MC, Dahm CN. 1990. A perspective on El Niño and La Niña: 

Global implications for stream ecology. Journal of the North American 

Benthological Society 9: 68–76. 

Moore GM, Bond BJ, Jones JA, Phillips N, Meinzer FC. 2004. Structural 

and compositional controls on transpiration in 40- and 450-year- 

old riparian forests in western Oregon, USA. Tree Physiology 24: 

481–491. 

Moore KM. 2010. Trends in streamflow from old growth forested watersheds 

in the western Cascades. Student research paper. Oregon State 

University, Corvallis. (30 January 2012; http://ir.library.oregonstate.edu/ 

xmlui/handle/1957/15741) 

Oishi AC, Oren R, Novick KA, Palmroth S, Ka GG. 2010. Interannual invariability 

of forest evapotranspiration and its consequence to water flow 

downstream. Ecosystems 13: 421–436. 

Pall P, Aina T, Stone DA, Stott PA, Nozawa T, Hilberts AGJ, Lohmann D, 

Allen MR. 2011. Anthropogenic greenhouse gas contribution to flood 

risk in England and Wales in autumn 2000. Nature 470: 382–385. 

Peters DPC, Bestelmeyer BT, Herrick JE, Fredrickson EL, Monger HC, 

Havstad KM. 2006. Disentangling complex landscapes: New insights 

into arid and semiarid system dynamics. BioScience 56: 491–501. 

Poff NL, Olden LD, Merritt DM, Pepin DM. 2007. Homogenization of 

regional river dynamics by dams and global biodiversity implications. 

Proceedings of the National Academy of Sciences 104: 5732–5737. 

Polsky C, Neff R, Yarnal B. 2007. Building comparable global change 

vulnerability assessments: The vulnerability scoping diagram. Global 

Environmental Change 17: 472–485. 

Scatena FN, Moya S, Estrada C, Chinea JD. 1996. The first five years in 

the reorganiation of aboveground biomass and nutrient use following 

Hurricane Hugo in the Bisley experimental watersheds, Luquillo experimental 

forest, Puerto Rico. Biotropica 28: 424-440. 

Sponseller RA, Grimm NB, Boulton AJ, Sabo JL. 2010. Responses of macroinvertebrate 

communites to long-term flow variability Sonoran Desert 

Articles 

stream. Global Change Biology 16: 2891–2900. doi:10.1111/j.1365- 

2486.2010.02200.x 

Stewart IT, Cayan DR, Dettinger MD. 2005. Changes toward earlier streamflow 

timing across western North America. Journal of Climate 18: 

1136–1155. 

Swank WT, Crossley DA Jr, eds. 1988. Forest Hydrology and Ecology at 

Coweeta. Vol. 66, Ecological Studies. Springer. 

Swank WT, Vose JM, Elliott KJ. 2001. Long-term hydrologic and water 

quality responses following commercial clearcutting of mixed hardwoods 

on a southern Appalachian catchment. Forest Ecology and 


[USDOI] US Department of the Interior. 2011. Reclamation: Managing 

Water in the West. USDOI. 

Verry ES, Brooks KN, Nichols DS, Ferris DR, Sebestyen SD. 2011. Watershed 

hydrology. Pages 193–212 in Kolka RK, Sebestyen SD, Verry ES, Brooks 

KN, eds. Peatland Biogeochemistry and Watershed Hydrology at the 

Marcell Experimental Forest. CRC Press. 

Warren RJ, Bradford MA. 2010. Seasonal climate trends, the North Atlantic 

Oscillation, and salamander abundance in the Southern Appalachian 

Mountain Region. Journal of Applied Meteorology and Climatology 

49: 1597–1603. 

Weisberg PJ, Swanson FJ. 2003. Regional synchroneity in fire regimes of 

western Oregon and Washington, USA. Forest Ecology and Management 

172: 17–28. 

Julia A. Jones (jonesj@geo.oregonstate.edu) is a professor of geosciences at 

Oregon State University, Corvallis, and co-principal investigator (PI) of the 

Andrews LTERL Network site; she studies hydrology and landscape ecology. 

Irena F. Creed is a professor of biology and geography at the University of 

Western Ontario, London, Ontario, Canada; she conducts biogeochemical 

modeling and analysis. Kendra L. Hatcher completed an MS in geography at 

Oregon State University, Corvallis; she studies physical geography and geographic 

information science. Robert J. Warren is a postdoctoral associate at the 

Yale School of Forestry and Environmental Studies, in New Haven, Connecticut, 

and studies organisms and climate variability at the Coweeta LTER Network 

site. Mary Beth Adams is a soil scientist at the US Forest Service (USFS) 

Northern Research Station, in Parsons, West Virginia, studying biogeochemistry 

at the Fernow Experimental Forest. Melinda H. Benson is an assistant 

professor of geography at the University of New Mexico, in Albuquerque, studying 

ecosystem services and resilience. Emery Boose is an information manager 

for Harvard University’s Harvard Forest, in Petersham, Massachusetts; he 

studies climate and hydrology at the Harvard Forest LTER Network site. 

Warren A. Brown is a senior research associate at the Cornell Institute for 

Social and Economic Research at Cornell University, in Ithaca, New York; he 

works on demographic change. John L. Campbell is a research ecologist at the 

USFS Northeastern Research Station, in Durham, New Hampshire; he studies 

climate, hydrology, and biogeochemistry at the Hubbard Brook Experimental 

Forest, an LTER Network site. Alan Covich is a professor at the Odum School 

of Ecology, at the University of Georgia, in Athens, and co-PI of the Luquillo 

Experimental Forest, an LTER Network site; he studies stream ecology. David 

W. Clow is a research hydrologist at the US Geological Survey’s Colorado Water 

Science Center, in Lakewood; he studies alpine hydrology and climate at the 

Loch Vale LTER Network site. Clifford N. Dahm is a professor of biology at the 

University of New Mexico, in Albuquerque; lead scientist of the CALFED Bay- 

Delta Program; and co-PI of the Sevilleta LTER Network site; he studies stream 

ecology, hydrology, and biogeochemistry. Kelly Elder is a research hydrologist 

at the USFS Rocky Mountain Research Station, in Fort Collins, Colorado; he 

studies snow hydrology and climate at the Fraser Experimental Forest, part 

of the USFS Experimental Forests and Ranges program, in Fraser, Colorado. 

Chelcy R. Ford is a research scientist with the USFS Southern Research Station, 

Coweeta Hydrological Laboratory, in Otto, North Carolina; she studies forest 

hydrology at the Coweeta LTER Network site. Nancy B. Grimm is a professor 

in the School of Life Sciences at Arizona State University, in Tempe, and co-PI 

of the Central Arizona–Phoenix LTER Network site; she studies stream ecology 

and urban systems. Donald L. Henshaw is an information technology specialist 

with the USFS Pacific Northwest Research Station in Portland, Oregon; he 


Articles 

designed and manages the LTER Network’s Climate and Hydrology Database 

Projects. Kelli L. Larson is an assistant professor in the School of Geographic 

Science and Urban Planning at Arizona State University, in Tempe; she studies 

water resource geography and governance. Evan S. Miles completed an MS 

in water resources at Oregon State University, in Corvallis; he studies hydrogeology 

and glaciers. Kathleen M. Miles is a PhD student in geography at 

Oregon State University, in Corvallis; she studies hydrology and climate change 

at the Andrews Experimental Forest, part of the LTER Network. Stephen D. 

Sebestyen is a research hydrologist with the USFS Northern Research Station, 

in Grand Rapids, Minnesota; he studies hydrology and biogeochemistry at 

Marcell Experimental Forest, in Grand Rapids. Adam T. Spargo is a research 

technician in the Department of Biology at the University of Western Ontario, 

in London, Ontario, Canada; his role includes working on developing a longterm 

experimental catchment database for Canada. Asa B. Stone is an assistant 

professor at Central New Mexico Community College, in Albuquerque; 

she studies ecological psychology. James M. Vose is the project leader at the 

USFS Southern Research Station, Coweeta Hydrological Laboratory, in Otto, 

North Carolina; he studies watershed hydrology, biogeochemistry, and ecology 

at the Coweeta LTER Network site. Mark W. Williams is a professor of geography 

at the University of Colorado, in Denver; a fellow of the Institute of Arctic 

and Alpine Research; and PI of the Niwot Ridge LTER Network site; he studies 

mountain ecology and alpine hydrology. 


Articles 

The Disappearing Cryosphere: 

Impacts and Ecosystem Responses 

to Rapid Cryosphere Loss 

Andrew G. FountAin, John L. CAmpbeLL, edwArd A. G. SChuur, ShAron e. StAmmerJohn, 

mArk w. wiLLiAmS, And huGh w. duCkLow 

The cryosphere—the portion of the Earth’s surface where water is in solid form for at least one month of the year—has been shrinking in response 

to climate warming. The extents of sea ice, snow, and glaciers, for example, have been decreasing. In response, the ecosystems within the cryosphere 

and those that depend on the cryosphere have been changing. We identify two principal aspects of ecosystem-level responses to cryosphere loss: 

(1) trophodynamic alterations resulting from the loss of habitat and species loss or replacement and (2) changes in the rates and mechanisms of 

biogeochemical storage and cycling of carbon and nutrients, caused by changes in physical forcings or ecological community functioning. These 

changes affect biota in positive or negative ways, depending on how they interact with the cryosphere. The important outcome, however, is the 

change and the response the human social system (infrastructure, food, water, recreation) will have to that change. 

Keywords: cryosphere, ecosystem response, environmental observatories 

Global average air temperature has warmed by 1 degree 

Celsius (°C) over the past century, and in response, the 

cryosphere—the part of the Earth’s surface most influenced 

by ice and snow—is changing. Specifically, alpine glaciers 

are retreating, the expanse of Arctic sea ice has been shrinking, 

the thickness and duration of winter snowpacks are 

diminishing, permafrost has been melting, and the ice cover 

on lakes and rivers has been appearing later in the year and 

melting out earlier. Although these changes are relatively 

well documented, the ecological responses and long-term 

consequences that they initiate are not. Detailed studies 

have identified specific responses to individual components 

cryospheric changes (e.g., polar bear habitat and sea ice 

loss), but a more integrated view across many landscapes 

and types of changes has been lacking. In the present article, 

we draw largely—but not exclusively—from sites of the US 

Long Term Ecological Research (LTER) Network (the special 

section in this issue; see especially Robertson et al. 2012) to 

synthesize our current knowledge of ecosystem responses 

to the changing cryosphere in an attempt to infer broad 

responses and to anticipate the further range of changes that 

we might expect. We contend that place-based, long-term, 

interdisciplinary efforts, such as LTER-type projects, are 

the best suited for tracking such changes and for detecting 

and understanding their cascading effects throughout the 

ecosystem. 

The cryosphere 

For the purposes of this synthesis, the spatial extent of the 

cryosphere for the Northern Hemisphere includes the mean 

February extent of snow cover (measured between 1987 

and 2003) and the mean March extent of sea ice (measured 

between 1979 and 2003). For the Southern Hemisphere, we 

include the mean August and September extents of snow 

and sea ice, respectively. Broad statistics for the cryosphere 

and its changes are provided in table 1 and are depicted in 

figure 1. 

Permafrost (figure 2a) is widespread in the Arctic and 

boreal regions of the Northern Hemisphere, with the permafrost 

zone occupying about 24% of the exposed land area. 

Most of this (78%) occurs in lowlands below 1000 meters 

(m) of elevation, whereas deposits of alpine permafrost 

are widely distributed. Changes in permafrost are typically 

documented by two metrics: temperature and the depth to 

the permafrost, which is defined as the active layer, which 

in turn is the surface layer that thaws seasonally. Since the 

1980s, permafrost temperatures have generally increased 

between 0.5° and 2°C when measured at about 10 m, a 

depth at which seasonal variations cancel each other out and 

thus yield a seasonally constant value (Romanovsky et al. 

2007). At some Russian sites, where many data are available, 

the active layer increased by 1.7–5.5 centimeters (cm) 

per year over the 10-year period between 1997 and 2007 





Articles 

(Mazhitova et al. 2008), whereas other sites have shown little 

change (Zamolodchikov et al. 2008). However, recent data 

have shown that active-layer depth measurements alone may 

obscure the degradation of permafrost, because the ground 

Table 1. Components of the global cryosphere and their areal extent. 

Component Definition and remarks 

Snow Perennial or seasonal cover of 

the land surface: 98% in Northern 

hemisphere 

Glaciers Perennial snow or ice that moves 

Alpine glaciers and ice caps 

surface subsides as permafrost thaws and internal ice melts. 

This subsidence process (called thermokarst) can radically 

restructure surface hydrology by altering the dynamics 

of water bodies, initiating or expanding surface channel 

incision, and drying surface soil 

Extent 

(in 10 6 km 2 ) LTER site(s) a 

1.9 (summer) 

45 (winter) 

ARC, AND, BNZ, CDR, 

KBS, KNZ, HBR, HFR, 

MCM, NTL, NWT, 

SGS, PIE, 

0.53 MCM, NWT, PAL 

Ice sheets 14 MCM, PAL 

Permafrost Subsurface Earth material remaining 

below 0 degrees Celsius for at least 

2 years 

23 ARC, BNZ, MCM, 

NWT 

Lake and river ice Seasonal cover of lakes and rivers ? ARC, AND, BNZ, CDR, 

KBS, KNZ, HBR, HFR, 

MCM, NTL, NWT, 

SGS, PIE 

Sea ice Perennial or seasonal cover of the 

ocean 

19–27 PAL 

Note: All extents are from table 4.1 of Solomon and colleagues (2007). The value for permafrost is the 

region in which permafrost occurs and includes both frozen and unfrozen soils. 

km 2 , square kilometers; LTER, long-term ecological research. 

a See table 2 for the site abbreviations. 

a b 

layers. Observations near Toolik 

Lake, Alaska, have shown rapid 

mass wasting of surface soils 

undergoing thaw, which resulted 

in an increased loading of suspended 

sediment in streams, 

with direct and indirect effects 

on stream biota (Bowden et al. 

2008). In the McMurdo Dry 

Valleys of Antarctica, enhanced 

incision of stream water into 

massive subsurface ice has caused 

one river to flow underground 

for some distance. At Niwot Ridge 

in Colorado, increasing water 

flow and solute concentrations in 

early autumn have been attributed 

to the melting of alpine 

permafrost (Caine 2010). 

One iconic and highly conspicuous 

feature of global warming 

is glacier recession (figure 2b). 

Figure 1. Approximate geographic limits of the cryosphere. (a) January climatology of Northern Hemisphere sea ice 

(measured between 1979 and 2005) and snow extent (measured between 1967 and 2005) with the North Pole referenced 

(the red dot). (b) September climatology of Southern Hemisphere sea ice (measured between 1979 and 2003) and snow 

extent (measured between 1987 and 2002) with the South Pole referenced (the red dot). Source: Reprinted from John 

Maurer, Atlas of the Cryosphere. National Snow and Ice Data Center (2007; http://nsidc.org/data/atlas). 


Articles 

Figure 2. Examples of the cryosphere. (a) Winter eastern forest, Mount Washington, New Hampshire (Photograph: Jerry 

and Marcy Monkman, www.ecophotography.com); (b) Melting sea ice and an iceberg, Charcot Island, Antarctic Peninsula 

(Photograph: Grace K. Saba, Rutgers University); (c) Massive ice exposed by degrading permafrost, Noatak National Preserve, 

Arkansas (Photograph: Edward A. G. Schuur); (d) Dana Glacier, Sierra Nevada, California. The top panel is the glacier in 

1883 (Photograph: I. C. Russell, US Geological Survey); the bottom panel is from 2004 (Photograph: Hassan Basagic). 

Glaciers have been receding worldwide since the end of 

the Little Ice Age in the late 1800s, although regional and 

temporal variations in recession have occurred (Dyurgerov 

and Meier 2000). In recent decades, glacier-mass loss has 

accelerated, with the increased rate ascribed to increased 

temperatures. Glacier change in the United States reflects 

these global trends through area losses over the past century 

of 34%–56% in the Sierra Nevada and the Cascades of 

Oregon and Washington and about 40% at Niwot Ridge, in 

the Colorado Front Range. In contrast to these observations 

and to those elsewhere in the alpine Southern Hemisphere, 

glaciers in the McMurdo Dry Valleys of Antarctica appear 

to be in equilibrium, since their positions have not changed 

since their observation began in 1993 (Fountain et al. 2006). 

The removal of water from long-term storage in glacial ice 

increases summer streamflow and global sea level. However, 

as the mass of glaciers decline, their ability to support summer 

streamflow declines, and they decrease in their ability to 

buffer the watersheds against drought. 

Sea ice (figure 2c) occurs in the Arctic; in the Southern 

Ocean surrounding Antarctica and the Baltic Sea; and in 

part of the northwest Pacific, from the Siberian coast down 

to the Japanese island of Hokkaido. Most sea ice forms 

and melts annually, with perennial multiyear ice restricted 

to the high latitudes of the Arctic and Antarctica. Since 

the advent of continuous satellite monitoring in the late 


Articles 

1970s, widespread decreases in sea ice have been recorded 

throughout most of the Arctic at an average rate of 3% loss 

per decade (Comiso and Nishio 2008). In contrast, decreases 

in Antarctic sea ice have been regionally confined and juxtaposed 

against regions of increasing sea ice, such that the 

average rate of change overall is a slight increase of 1% per 

decade. Changes in sea ice alter the extent and distribution 

of foraging platforms for larger mammals and refuge habitat 

for smaller species. At the Palmer Peninsula, seasonal seaice 

cover has been shrinking at astonishing rates because of 

increases in onshore winds driven by hemispheric changes 

in atmospheric circulation. The duration of sea-ice cover has 

declined by 85 days since 1978 (Stammerjohn et al. 2008). 

Seasonal lake and river ice occur in all temperate regions, 

with durations of days to months, whereas perennial 

ice cover is only found at extremely high latitudes and 

elevations. The date of lake-ice formation and breakup is 

commonly recorded for commercial purposes related to 

shipping, trapping, fishing, ice harvesting, and transportation 

and yields an extensive long-term record (Magnuson 

et al. 2000). Since 1846, lake-ice duration in the Northern 

Hemisphere has decreased by 12 days per century, which 

is equivalent to a warming of 1.2°C per century. A 20-year 

record of ice thickness in late March on an alpine lake in 

the Niwot Ridge LTER site shows a consistent thinning 

of the ice cover at 2.0 cm per year (Caine 2002). Ice cover 

exhibits strong control over exchanges in gases and material, 

solar radiation, and heat between aquatic habitats and 

the atmosphere. The duration of ice exerts a profound 

influence on the patterns of water circulation and thermal 

stratification, which are closely linked to the life cycles of 

aquatic organisms and to the biogeochemical cycling of the 

ecosystem. 

Snow (figure 2d) is the largest component of the cryosphere 

in areal extent. About 98% of the snow-covered land 

on Earth is in the Northern Hemisphere, which contains 

nearly half of the planet’s land surface. In the Southern 

Hemisphere, over 99% of the snow cover is confined to 

Antarctica and largely consists of perennial snow. In the 

Northern Hemisphere, strong negative trends in the extent 

of snow cover have been observed over recent decades (Déry 

and Brown 2007). Increased snowfall and snow depth have 

been reported at the highest-elevation sites of the western 

United States (Williams et al. 1996); however, most locations 

in the Mountain West have experienced snowpack 

declines, and concern has risen about streamflow, water 

yields, and water supply. In the Pacific Northwest, extensive 

snow- covered regions are now deemed at risk in terms of 

their capacity to provide reliable water yields because of 

atmospheric warming, altitudinal shifts in the distribution 

of snow and rain, and declining winter snowpacks (Nolin 

and Daly 2006). Winter snow depths have also been decreasing 

throughout the northeastern United States. For example, 

at the Hubbard Brook LTER site in New Hampshire, 

the maximum snow depth has declined by 25 cm (7 cm 

water equivalent), and snow cover duration has decreased 

by 21 days over the past 53 years (Campbell et al. 2010), 

which has led to major changes in terrestrial and aquatic 

ecosystems. 

One simple metric in the attempt to capture potential 

ecosystem vulnerability to changes in the cryosphere across 

ecosystems is the duration of frost and freezing temperatures 

(table 2, figure 3). Frost days are those with long-term mean 

daily minimum temperatures below 0°C; freeze days are 

those with long-term mean daily maximum temperatures 

below 0°C. Vulnerability can be thought of as susceptible 

to increased or decreased frost or freezing periods. For 

example, ecosystems that do not experience frost, such 

as those in the tropics, are highly vulnerable to cold temperatures. 

Significant ecosystem changes could be expected 

if the climate were to cool, making frost commonplace. 

Alternatively, ecosystems accustomed to long frozen periods, 

such as polar and high alpine ecosystems, are highly vulnerable 

to warm temperatures. Those ecosystems exposed to 

moderate periods of frost or freezing would be expected to 

be less vulnerable to changes in temperature. We focus on 

the warming climate, and as such, the tropical ecosystems 

will not be directly exposed to cryospheric losses, whereas 

polar and high alpine ecosystems may be the most vulnerable 

to such change. In table 2 and figure 3, we can see the 

vulnerability of the major ecosystem research sites under 

study by US scientists. 

Ecosystem responses to the loss of snow and ice 

The various components of the cryosphere provide physical 

habitat for diverse organisms. Iconic examples include 

polar bears and penguins in sea ice and pikas in rock glaciers 

(rock debris with ice filling the void spaces between 

the rocks; the mass flows downhill), but many other species 

ranging in size from microbes to whales inhabit permafrost, 

glaciers, sea-ice, and snow-covered landscapes. As these 

habitats shrink and disappear, resident species are forced to 

migrate, often tracking the distribution of receding frozen 

habitats across the landscape. Since different organisms 

respond and move at different rates (e.g., trees versus penguins), 

cryosphere recession can have many consequences: 

the fragmentation of animal and plant communities and 

the development of new assemblages, disruption of seasonally 

synchronized phenological connections among species, 

and losses in biodiversity and the associated changes 

in ecosystem function (Parmesan 2006). Although these 

processes are occurring at unprecedented rates in response 

to rapid climate warming, it has required decades of coordinated 

observations to document significant change and 

to uncover the mechanisms linking climate forcing to ecosystem 

responses. 

Prolonged, systematic studies of this type are a key contribution 

of LTER. The LTER Network of sites facilitates longterm 

observations, experiments, and comparative studies 

that enable us to identify common processes and mechanisms 

across diverse ecosystems. The highly interdisciplinary 

nature of LTER helps to quickly reveal interpretations of 


Table 2. Cryosphere processes at US Long Term Ecological Research (LTER) Network sites and related LTER sites. 

Site Abbreviation Record 

Continental 

glacier 

causes and consequences and needed adjustments of monitoring 

approaches to catch signals of previously unmonitored 

or unanticipated system behaviors. Here, we present 

some notable examples of ecological and biogeochemical 

changes in response to cryosphere loss. 

Changes in populations and trophodynamic implications. Decadalscale 

declines or distributional shifts in snow- and icedependent 

species are now extensive and well documented 

(Chapin et al. 2005). When ice-dependent species suffer 

habitat loss, the changes in frozen habitats (glaciers, sea 

ice, snowpacks, and permafrost) impose both bottom-up 

and top-down forcings on terrestrial and aquatic ecosystems. 

Ice loss affects ecosystems directly through the loss 

of physical habitat and through alterations in thermal 

conditions and indirectly by altering light and nutrient 

supply to primary producers. Both Arctic and Antarctic 

sea ice harbor a resident microbial community of diatoms, 

other phytoplankton, bacteria, and protozoan grazers that 

contributes to the total primary production of polar seas. 

Like sea ice, ice and rock glaciers are habitats for specially 

adapted species that may disappear as glaciers retreat and 

their cold, glacier-fed streams disappear. The American pika 

Alpine 

glacier Sea ice Lake ice 

Continuous 

permafrost 

Discontinuous 

permafrost 

Seasonal 

snow 

McMurdo MCM 1 1988–2009 X X X X 

Arctic Tundra ARC 1 1988–2008 X X X 

Niwot Ridge NWT 1 1952–2006 X X X X 

Bonanza Creek BNZ 1 1988–2009 X X 

Palmer PAL 1 1989–2010 X X X X 

Loch Vale LVW 2 1993–2008 X X X X 

Marcell MAR 1961–2010 X 

North Temperate 

Lakes 

NTL 1 1978–2010 X X 

Hubbard Brook HBR 1,3 1964–2007 X 

Harvard Forest HFR 1 1964–2002 X 

Kellogg KBS 1 1988–2010 X 

Shortgrass Steppe SGS 1 1969–2010 X 

Sevilleta SEV 1,4 1991–2010 X 

Articles 

Fernow FER 3 1951–2007 X 

Konza KNZ 1 1982–2011 X 

Plum Island PIE 1 1950–2004 X X 

Coweeta CWT 1,3 1950–2009 X 

Baltimore BES 1 2000–2009 X 

Andrews Forest AND 1,3 1957–2007 X X 

Jornada JRN 1,3 1983–2009 X 

Olympic OLY 5 1962–2009 X 

Transient 

snow 

Note: The superscript number next to the abbreviation refers to the sponsoring agency for that site: 1, US LTER Network site; 2, US Geological Survey 

Water, Energy, and Biochemical Budgets site; 3, US Department of Agriculture Experimental Forest and Range site; 4, National Wildlife Refuge; 5, State 

Experimental Forest. Record refers to the length of the air-temperature record used to estimate frost and freezing duration at each location. 

(Ochotona princeps)—although it is not as well known or 

charismatic as penguins or polar bears—is attaining new 

status as a poster child for glacier loss and climate change. 

Pikas do not hibernate and use subsurface microclimates 

in rocky debris to persist where surface conditions would 

preclude their survival. Despite this adaptation, some local 

pika extinctions in the Northwest have been linked to cold 

exposure caused by a loss of insulating snow cover (Ray et al. 

2012). Permafrost thaw also results in habitat disappearance 

for its resident species. Because permafrost occurs in so 

many different habitats in different stages of development, 

its loss may trigger primary or secondary successions. 

Widespread past and projected future reductions in 

snow-cover extent, duration, depth, and water equivalent 

can also have extensive ecological repercussions. Many plant 

and animal species are adapted to snow-cover conditions 

and will perish if they are unable to migrate or tolerate 

less snow cover. Even so, not all animals that live in cold 

environments respond negatively to reductions in snow 

cover. For example, ungulates such as white-tailed deer, 

mule deer, elk, and caribou expend less energy and are less 

susceptible to predation when snowpacks are shallower. 

Some of the species most susceptible to snow-cover loss 


Articles 

Figure 3. Duration of frost and freezing periods at LTER and related long-term research 

sites, from climate records. See table 2 for the site abbreviations. Frost days are those 

with long-term mean daily minimum temperatures below 0 degrees Celsius (°C); freeze 

days are those with long-term mean daily maximum temperatures below 0°C. 

are those that overwinter below ground, since snow insulates 

the subsurface and moderates its temperature. The 

shortgrass steppe in the western United States receives little 

snowfall and therefore presents an endpoint in the spectrum 

of snow cover. At this semiarid, high-elevation site, 

snowfall from November to late February has little effect 

on ecological processes; however, large snowfalls in March 

and April (after the ground has thawed) strongly influence 

the subsequent productivity by controlling the availability 

of water and nutrients (Cayan et al. 2001). Some plants 

are photosynthetically active in shallow spring snowpacks, 

giving them a competitive advantage in regions with short 

growing seasons (Starr and Oberbauer 2003). As the climate 

warms, the disappearance of snow cover and the increased 

length of the growing season may benefit some plants, 

provided that other requirements for growth are not limiting. 

Many plants in alpine and tundra regions are reliant 

on snow for water and nutrients and therefore are found 

in the greatest abundance where the range of snow depths 

is optimal (Walker et al. 1993). Although the snow-free 

period will lengthen in a warmer climate, the lack of snow 

cover during colder months will increase soil temperature 

variation, making roots susceptible to winter injury. Soil 

freezing can directly and adversely affect roots by causing 

cellular damage and can also sever fine roots through frost 

heaving. Reduced nutrient uptake as a result of root injury 

has been shown to lower nutrient retention and to increase 

hydrologic fluxes from soils during the growing season 

(Fitzhugh et al. 2001). 

Changes in habitat and productivity 

regimes can ripple up 

the trophic ladder, as is demonstrated 

extensively in marine 

food webs. Changing snow 

and ice conditions alter habitat 

suitability for many bird species 

(e.g., petrels, Adelie penguins 

[Pygoscelis adeliae]) and 

limit the physical space available 

for habitation (Micol and 

Jouventin 2001, Weimerskirch 

et al. 2003). The huge populations 

of krill in Antarctic marginal 

sea-ice zones serve in turn 

as a major food resource for a 

suite of large predators, including 

seabirds, seals, and whales. 

Sea-ice microbial communities 

serve as a principal food source 

for juvenile krill, which also 

hide from predators in underice 

cryptic spaces. Therefore, the 

regional decline in the duration 

and extent of sea-ice cover in the 

Bellingshausen and Amundsen 

Seas has resulted in declining 

abundance and ranges of Antarctic krill (Euphausia superba), 

possibly the most numerous metazoan species on Earth. 

Atkinson and colleagues (2004) documented large-scale, 

order-of- magnitude declines in krill populations over 50 years 

in the South Atlantic sector of the Antarctic seas. Meanwhile, 

the number of salps—pelagic, gelatinous, ice-avoiding tunicates 

with few predators—has increased; they have, in effect, 

replaced krill as an intermediate species in Antarctic marine 

food chains. One of the best-studied examples of the response 

of predator populations to sea-ice loss is the Adelie penguin, 

a true Antarctic penguin with strong fidelity to sea ice as a 

platform for foraging activity (Ducklow et al. 2007). Since 

1975, Adelie penguins nesting near Palmer Station have 

declined by about 80% in response to a host of environmental 

changes, including habitat loss and altered food availability 

(figure 4). Fraser and Hofmann (2003) demonstrated that 

penguin chicks weighing less than 300 grams at fledging 

had a reduced probability of surviving past the first year. 

They suggested that changes in the sea-ice season shifted the 

period of maximum krill stocks away from the penguins’ 

peak foraging season. Over the same period, two subpolar 

species—chinstrap (Pygoscelis antarcticus) and gentoo penguins 

(Pygoscelis papua)—have successfully immigrated to the 

region and now constitute half of the total penguin population 

in the region. The mechanisms behind these shifts and 

their long-term outcome are unclear (Trivelpiece et al. 2011). 

The recent loss of sea ice could boost primary productivity in 

the Arctic Ocean by a factor of two or three. In the northern 

Bering Sea, primary-productivity changes caused by warming 


Year 

Figure 4. Populations of ice-dependent (Adelie) and icetolerant 

(chinstrap and gentoo) penguins near Palmer 

Station, Antarctica, measured between 1976 and 2009. 

Source: Adapted from Ducklow and colleagues (2007). 

and sea-ice loss have resulted in a dramatic reorganization of 

the ecosystem (Grebmeier et al. 2006). This shallow marine 

ecosystem was formerly characterized by high primary productivity 

and efficient export to the bottom, which supports a 

high stock of benthic prey for diving ducks and walruses. With 

the loss of sea ice and the warming of the water column, the 

export of surface productivity into the benthos has declined, 

which has caused a switch from a system with top predators 

sustained by benthic prey to one dominated by pelagic fish. 

Changes in biogeochemical cycles. Changes in the extent, seasonality, 

and duration of cryosphere components affect the 

cycling of nutrients in land and ocean ecosystems. Glacier 

retreat and rock glacier shrinkage expose new landscapes 

that are typically carbon poor yet nutrient rich because 

of rock weathering. Microbial life—particularly nitrogen 

fixers—occupy these new landscapes (Nemergut et al. 2007), 

which increases the nitrate levels of streams and lakes down 

valley. These conditions are transient and slowly change 

as higher plants occupy the landscape over time scale of 

decades to centuries. High alpine waters are typically oligotrophic 

and are therefore susceptible to ecological changes 

that result from increases in nitrogen export from the land 

(Baron et al. 2009). Williams and colleagues (2007) characterized 

the nutrient content in the outflow of the Green 

Lake 5 rock glacier, located in the Green Lakes Valley of the 

Colorado Front Range. The nitrate concentrations from 

the rock glacier are among the highest reported for highelevation 

surface waters. These extreme nitrate concentrations 

appear to be characteristic of rock glacier outflows in 

the Rocky Mountains (Williams et al. 2007). Fluorescence 

index values and dissolved organic matter (DOM) measurements 

are consistent with a switch from terrestrial DOM in 

the summer to an increasingly aquatic-like microbial source 

during the autumn months. Glacier melting has also been 

implicated in the regulation of phytoplankton species composition 

in Antarctic coastal regions where diatoms—the 

Articles 

preferred food for Antarctic krill (see above)—are being 

replaced by less-palatable cryptophytes. Glacial inputs 

change light availability by stabilizing the surface-water 

column and possibly stimulate growth selectively by adding 

limiting micronutrients (Dierssen et al. 2002). Melting 

glaciers and sea ice also transfer airborne pollutants stored 

in the snow and ice to the marine environment. 

Perhaps the most important result from the reduction in 

duration of lake ice in a warming climate is less-frequent 

oxygen-depletion events and the associated adverse biological 

consequences (Prowse et al. 2006). For river ice, large 

fluxes of allocthonous detrital material and nutrients are 

flushed into the river water column because of channel scour 

during ice breakup and flooding. Geomorphically, at the 

Pine Island LTER site in coastal Massachusetts, the formation 

and transport of river ice are important factors in determining 

salt marsh platform elevation and have implications 

for responses to rising sea level. The delivery of sediment to 

the marsh through ice rafting (Wood et al. 1989), the compression 

of the marsh surface as a function of ice thickness, 

and the scour of vegetation are winter processes that will 

change as less river ice forms and its transport into fringing 

salt marshes declines in the coming decades. 

In cold regions, nutrient cycling is closely coupled with 

snowpack dynamics, with much of the annual export of 

stream nutrients occurring in winter, when biological uptake 

is low. Changes in the snowpack alter hydrology, which 

affects the amount, timing, and magnitude of spring snowmelt. 

The resulting changes in streamflow not only affect 

nutrient transport but, when they are combined with 

changes in temperature, also affect aquatic habitats, causing 

potential shifts in species assemblages. Nutrients accumulate 

in the snowpack over winter and are released in an ionic 

pulse during the first portion of snowmelt (Johannessen and 

Henriksen 1978). Although snowmelt can be an important 

source of nutrients and water early in the growing season, 

it can also cause episodic acidification in areas with high 

atmospheric deposition and poorly buffered soils (e.g., 

Schaefer et al. 1990). The soils beneath the snowpack are also 

an important source of nutrients during winter and early 

spring. The snowpack regulates soil temperatures, keeping 

them warm enough for many biologically mediated reactions. 

Snow fence experiments, which enhance winter snow 

accumulation, have shown that higher rates of soil microbial 

respiration and nitrogen mineralization occur under deeper 

snowpacks in subalpine forest and Arctic tundra environments 

because of warmer soil temperatures that result from 

the thermally insulating effects of the snow (Schimel et al. 

2004). In contrast, in boreal spruce and temperate hardwood 

forests, thin winter snowpacks increase the frequency and 

depth of soil freezing, which results in elevated summer 

nitrogen emissions that are probably a result of reduced 

nitrate uptake by damaged roots and by root decomposition 

(Fitzhugh et al. 2001, Maljanen et al. 2010). Fluxes of carbon 

dioxide (CO 2 ) mirror those of nitrogen, and the timing 

and magnitude of the nitrogen and CO 2 fluxes in all cases 


Articles 

depends on plant species. Elevated nitrogen mineralization 

contributes significantly to fluxes of greenhouse gases and 

to the cycling of nitrogen and carbon. In the mixed-grass 

prairie of North America, greater snowpack increased soil 

moisture by midsummer, which resulted in increased soil 

respiration (Chimner and Welker 2005) and increased plant 

invasions (Blumenthal et al. 2008). 

But most of the attention to the biogeochemical consequences 

of cryosphere loss has been focused on the potential 

for changes in carbon storage and release. Arctic permafrost 

contains twice the CO 2 found in the atmosphere, which 

dramatically demonstrates the potential for altering the 

climate as further warming occurs. Site-specific information 

can provide some indication as to the future release 

rate of carbon from thawing permafrost. A recent group 

of studies was focused on an upland thermokarst site near 

Denali National Park in Alaska, where changes in plant and 

soil processes were studied as a function of time since the 

thermokarst disturbance was initiated. The studies showed 

that increased permafrost thaw and ground-surface subsidence 

increased net and gross primary productivity as 

plant growth was stimulated by a thaw (Vogel et al. 2010). 

Plant species composition changed along with changes in 

plant growth rates as graminoid-dominated moist acidic 

tundra shifted to shrub-dominated tundra with increased 

rates of thawing. The increased carbon uptake by plants 

initially offset the greater ecosystem respiration, such that 

this thermokarst was a net sink of carbon 15 years after the 

initiation of the thaw, even though decomposition of older 

carbon deep in the soil was already taking place (Schuur 

et al. 2009). Over more decades of thaw, plant growth rates 

remained high, but increased old soil carbon losses eventually 

offset the greater carbon uptake, and this thermokarst 

became a net source of carbon to the atmosphere. 

In a contrasting study of lowland thermokarst in three 

Canadian peatlands, the carbon accumulation in surface 

soil organic matter was higher in unfrozen bogs and in 

areas where permafrost had degraded than in areas where 

permafrost was intact (Turetsky et al. 2007). This growth 

in surface soil carbon accumulation was consistent with 

the Alaskan upland study, but the equivalent net ecosystem 

carbon exchange measurements were not available to 

determine whether the thawed permafrost peat ecosystems 

were overall net sources or sinks of carbon. Permafrost 

thaw in this lowland system promoted the release of methane 

(CH 4 ) because waterlogged conditions predominated 

in Sphagnum moss lawns that replaced the feather moss 

(Hylocomium splendens) and black spruce (Picea mariana) 

forest in locations where permafrost degraded. This CH 4 

release was hypothesized to potentially offset the observed 

surface soil carbon accumulation for at least for 70 years, 

until plant and ecosystem succession in the moss lawn created 

conditions more like those in the unfrozen bogs, which 

stored surface soil carbon but released only small amounts 

of CH 4 . The release of CH 4 is a common pathway of carbon 

loss in lowland thermokarst, where drainage is restricted 

( Myers-Smith et al. 2007), and CH 4 has 25 times greater heattrapping 

capacity than CO 2 on a century timescale. However, 

decreased total carbon emissions in anaerobic systems can 

partially offset the increased radiative forcing of CH 4 release, 

which possibly makes the net radiative forcing of increased 

carbon losses in lowland and upland thermokarst more 

similar than what the difference in heat-trapping capacity 

between CO 2 and CH 4 would initially suggest. 

The oceanic sink for anthropogenic CO 2 is large and 

lessens the potential greenhouse effect by limiting CO 2 

accumulation in the atmosphere. The current (2009) net 

annual carbon uptake by the ocean is 2.3 ± 0.4 petagrams 

(Pg) of carbon per year, compared to 2.4 Pg of carbon 

per year for land, but the land uptake is partially offset by 

the 1.1 ± 0.7 Pg of carbon per year in releases caused by 

deforestation (Le Quéré et al. 2009). As the ocean warms and 

its inventory of CO 2 increases, the oceanic sink is expected 

to weaken. Oceanic CO 2 uptake is governed by gas exchange 

across the air–sea interface, so the regional allocation of 

CO 2 uptake is primarily a function of the area of sea surface 

involved. This fraction has been predicted to increase as ice 

melts and productivity increases, which will expose new 

ocean areas to solar irradiance (Peck et al. 2010). The Arctic 

Ocean constitutes just 3% of the total ocean area and is 

mostly covered by sea ice, which blocks gas exchange, but it 

accounts for 5%–14% of the total ocean CO 2 uptake. New 

observations suggest, however, that the recent dramatic loss 

of sea ice has been accompanied by decreased rather than 

increased CO 2 uptake (Cai et al. 2010), which is counter to 

current understanding and predictions. The rapid, diverse, 

and complex changes wrought by cryosphere loss constitute 

a major scientific challenge that demands new large-scale 

observing systems on land and in the ocean to provide new 

observational infrastructure as a resource for coordinated 

experimental studies performed by the LTER Network and 

other scientists. 

Effects on humans from the loss of snow and ice 

Cryosphere loss will result in far-reaching social, economic, 

and geopolitical impacts, but a detailed treatment 

is beyond the scope of this synthesis. Most attention has 

been devoted to the impacts associated with a loss of snow 

cover, glacier melting, and sea-level rise, which are treated 

elsewhere (Kundzewicz et al. 2007). Thawing permafrost 

will also have important social consequences, because it can 

destabilize engineered structures and can cause destructive 

slides, flows, and slumps. Changes in snow cover can 

have important consequences for humans and may affect 

many diverse activities, including agriculture, recreation, 

tourism, engineering, commerce, and energy production. 

For example, the New Hampshire ski industry has abandoned 

low-elevation ski areas in the southern part of the 

state since the 1970s, in part because of climate warming, 

in favor of ski areas at higher elevations in the north 

(Hamilton et al. 2003). Skiing contributes about $1 billion 

annually to the economy of Utah, but recent climate change 


evaluations of the ski industry there suggest that it is at risk 

in the next several decades (Lazar and Williams 2008). A 

similar impact is anticipated for the Pacific Northwest. The 

most important effect is the influence on streamflow. In 

many semiarid regions of the world, such as the southwestern 

United States, snowmelt from mountain snowpacks is 

the dominant source of water for human consumption and 

irrigation. Therefore, changes in the amount and timing of 

snowmelt in mountainous areas could affect stream ecosystem 

services, such as drinking- water supply, wastewater 

assimilation, and hydropower. Lesser amounts of snow could 

also have an impact on agriculture and the ability to produce 

food, both through an increased occurrence of drought 

and through an inadequate supply of water for irrigation. 

Some evidence suggests that changes in snowmelt may also 

increase the risk of forest fires. In the western United States, 

earlier snowmelt dates correspond to increased wildfire frequency, 

because soils and vegetation are becoming drier and 

the period of potential ignition is lengthening (Westerling 

et al. 2006). Estimates of sea-level rise to 2100 have been 

continually revised upward since the 2007 report of the 

Intergovernmental Panel on Climate Change (Solomon 

et al. 2007) as new data and modeling have been developed. 

At the time of this writing, the rise in sea level by the end of 

the century is projected to be about 1 m (Pfeffer et al. 2008). 

The economic cost of a 1-m rise in sea level is estimated to 

exceed $1 trillion (Anthoff et al. 2010), with enormous social 

and political dislocations as residents of low-lying regions 

are forced to move to higher ground. 

The Arctic has emerged as a key laboratory for the 

study of climate change impacts on human communities, 

partly because it is host to the world’s largest indigenous 

population that maintains a subsistence lifestyle (Kofinas 

et al. 2010) and partly because of the rapidly manifesting 

impacts on infrastructure, transportation, and international 

relations. The complex interplay among climate, biogeochemical, 

ecological, and sociopolitical factors responding to 

cryosphere loss in the Arctic and around the world demands 

new levels of interdisciplinary collaboration and new models 

for scientific study (Driscoll et al. 2012 [in this issue]). 

A system-level understanding of the global cryosphere is 

fundamental to predicting the future course of the Earth’s 

socioecological system and to laying out a course for human 

social, political, and economic adaptation to climate change. 

As was demonstrated in this article and others in this issue, 

socioecological ecosystem science as pioneered by the US 

LTER Network is a key component of our current and future 

understanding of cryospheric change. 

Conclusions 

Earth is distinguished in the solar system by the coexistence 

of water in its three phases: solid (frozen), liquid (melted), 

and gas (evaporated). The solid phase—the global cryosphere 

in all its components: glaciers; snow; permafrost; sea, 

lake, and river ice—is arguably the most rapidly changing 

element of the Earth system. Cryosphere loss can be viewed 

Articles 

as a planetary-scale redistribution of solid water into its 

liquid and gas phases. This large-scale reorganization will 

trigger changes in the balance of positive and negative 

feedbacks in the climate system (e.g., changes in planetary 

albedo), with far-reaching consequences for ecosystems 

and society, including changes in sea level, precipitation, 

and water availability. The current geophysical rates of 

cryosphere loss are now well documented but our lack in 

understanding of the relevant mechanisms limits our ability 

to predict the future course of change, with potentially grave 

consequences for society. In particular, we lack long-term 

observations and system-level experiments in which the 

linkages between changes in physical habitat and climate 

on one hand and ecosystem structure and biogeochemical 

functions on the other are addressed. LTER Network sites 

have pioneered coordinated observations and experimental 

manipulations of ecosystems and elemental cycles (Knapp 

et al. 2012 [in this issue]). An expansion of our fundamental 

knowledge of the phenologies and processes governing ecosystem 

responses to climate change is a necessary first step 

in creating future scenarios of change and human responses 

to it (Thompson et al. 2012 [in this issue]). This new understanding 

will continue to come from LTER sites situated in 

all the major cryosphere systems (table 1). 


The authors wish to thank the funding provided by the 

National Science Foundation’s (NSF) Long Term Ecological 

Research (LTER) Network for supporting our long-term 

studies, in which we track the ecosystem response to the disappearing 

cryosphere. The LTER Network office supported 

the workshop from which this article is derived. Table 2 and 

figure 3 were kindly contributed by Julia A. Jones, Kendra 

L. Hatcher, and Evan S. Miles. Valuable contributions were 

made by Mike Antolin, Anne Giblin, John Hobbie, John 

Magnusson, Anne Nolin, Bruce Peterson, Bill Sobczak, and 

Will Wolheim. We greatly appreciate all who participated. 

Fred Swanson offered many helpful suggestions, and David 

Foster carefully edited the manuscript (and it is much 

improved). NSF LTER Site Grants OPP 0823101, OPP 

1115245, DEB 1114804, DEB-1026415, DEB-0620579, and 

DEB-1027341 supported the authors during the preparation 

of this article. 


Anthoff D, Nicholls R, Tol RSJ. 2010. The economic impact of substantial 

sea-level rise. Mitigation and Adaptation Strategies for Global Change 

15: 321–335. 

Atkinson A, Siegel V, Pakhomov E, Rothery P. 2004. Long-term decline in 

krill stock and increase in salps within the Southern Ocean. Nature 432: 

100–103. 

Baron JS, Schmidt TM, Hartman MD. 2009. Climate-induced changes in 

high elevation stream nitrate dynamics. Global Change Biology 15: 

1777–1789. 

Blumenthal D, Chimner RA, Welker KM, Morgan JA. 2008. Increased snow 

facilitates plant invasion in mixedgrass prairie. New Phytologist 179: 

440–448. 


Articles 

Bowden WB, Gooseff MN, Balser A, Green A, Peterson BJ, Bradford J. 

2008. Sediment and nutrient delivery from thermokarst features in the 

foothills of the North Slope, Alaska: Potential impacts on headwater 

stream ecosystems. Journal of Geophysical Research 113 (Art. G02026). 

doi:10.1029/2007JG000470. 

Cai W-J, et al. 2010. Decrease in the CO 2 uptake capacity in an ice-free 

Arctic Ocean basin. Science 329: 556–559. 

Caine N. 2002. Declining ice thickness on an alpine lake is generated by 

increased winter precipitation. Climatic Change 54: 463–470. 

———. 2010. Recent hydrologic change in a Colorado alpine basin: An 

indicator of permafrost thaw? Annals of Glaciology 51: 130–134. 

Campbell JL, Ollinger SV, Flerchinger GN, Wicklein H, Hayhoe K, Bailey 

AS. 2010. Past and projected future changes in snowpack and soil frost 

at the Hubbard Brook Experimental Forest, New Hampshire, USA. 

Hydrological Processes 24: 2465–2480. 

Cayan DR, Kammerdiener SA, Dettinger MD, Caprio JM, Peterson DH. 

2001. Changes in the onset of spring in the western United States. 

Bulletin of the American Meteorological Society 82: 399–415. 

Chapin FS III, et al. 2005. Polar systems. Pages 717–743 in Hassan R, Scholes 

R, Ash N, eds. Ecosystems and Human Well-being: Current State and 

Trends. Island Press. 

Chimner RA, Welker JM. 2005. Ecosystem respiration responses to experimental 

manipulations of winter and summer precipitation in a 

Mixedgrass Prairie, WY, USA. Biogeochemistry 73: 257–270. 

Comiso JC, Nishio F. 2008. Trends in the sea ice cover using enhanced and 

compatible AMSR-E, SSM/I, and SMMR data. Journal of Geophysical 

Research 113 (Art. C02S07). doi:10.1029/2007JC004257. 

Déry SJ, Brown RD. 2007. Recent Northern Hemisphere snow cover extent 

trends and implications for the snow-albedo feedback. Geophysical 

Research Letters 34 (Art. L22504). doi:22510.21029/22007GL031474. 

Dierssen HM, Smith RC, Vernet M. 2002. Glacial meltwater dynamics 

in coastal waters west of the Antarctic peninsula. Proceedings of the 


Driscoll CT, Lambert KF, Chapin FS III, Nowak DJ, Spies TA, Swanson 

FJ, Kitteridge DB Jr, Hart CM. 2012. Science and society: The role 

of long-term studies in environmental stewardship. BioScience 62: 

354–366. 

Ducklow HW, Baker K, Martinson DG, Quetin LB, Ross RM, Smith RC, 

Stammerjohn SE, Vernet M, Fraser W. 2007. Marine pelagic ecosystems: 

The west Antarctic peninsula. Philosophical Transactions of the Royal 

Society of London B 362: 67–94. 

Dyurgerov MB, Meier MF. 2000. Twentieth century climate change: 

Evidence from small glaciers. Proceedings of the National Academy of 

Sciences 97: 1406–1411. 

Fitzhugh RD, Driscoll CT [Jr], Groffman PM, Tierney GL, Fahey TJ, Hardy 

JP. 2001. Effects of soil freezing disturbance on soil solution nitrogen, 

phosphorus, and carbon chemistry in a northern hardwood ecosystem. 

Biogeochemistry 56: 215–238. 

Fountain AG, Nylen TH, MacClune KL, Dana GL. 2006. Glacier mass balances 

(1993–2001), Taylor Valley, McMurdo Dry Valleys, Antarctica. 

Journal of Glaciology 52: 451–462. 

Fraser WR, Hofmann EE. 2003. A predator’s perspective on causal links 

between climate change, physical forcing and ecosystem response. 

Marine Ecology Progress Series 265: 1–15. 

Grebmeier JM, Overland JE, Moore SE, Farley EV, Carmack EC, Cooper LW, 

Frey KE, Helle JH, McLaughlin FA, McNutt SL. 2006. A major ecosystem 

shift in the northern Bering Sea. Science 311: 1461–1464. 

Hamilton LC, Rohall DE, Brown BC, Hayward GF, Keim BD. 2003. Warming 

winters and New Hampshire’s lost ski areas: An integrated case study. 

International Journal of Sociology and Social Policy 23: 52–73. 

Johannessen M, Henriksen A. 1978. Chemistry of snow meltwater: Changes 

in concentration during melting. Water Resources Research 14: 

615–619. 



Kofinas GP, Chapin FS III, BurnSilver S, Schmidt JI, Fresco NL, Kielland 

K, Martin S, Springsteen A, Rupp TS. 2010. Resilience of Athabascan 

subsistence systems to interior Alaska’s changing climate. Canadian 

Journal of Forest Research 40: 1347–1359. 

Kundzewicz ZW, Mata LJ, Arnell NW, Döll P, Kabat P, Jiménez B, Miller K, 

Oki T, Sen Z, Shiklomanov IA. 2007. Freshwater resources and their 

management. Pages 173–210 in Parry ML, Canziani OF, Palutikof JP, 

van der Linden PJ, Hanson CE, eds. Climate Change 2007: Impacts, 

Adaptation and Vulnerability. Contribution of Working Group II to the 

Fourth Assessment Report of the Intergovernmental Panel on Climate 

Change. Cambridge University Press. 

Lazar B, Williams M. 2008. Climate change in western ski areas: Potential 

changes in the timing of wet avalanches and snow quality for the 

Aspen ski area in the years 2030 and 2100. Cold Regions Science and 

Technology 51: 219–228. 

Le Quéré C, et al. 2009. Trends in the sources and sinks of carbon dioxide. 

Nature Geoscience 2: 831–836. 

Magnuson JJ, et al. 2000. Historical trends in lake and river cover in the 

Northern Hemisphere. Science 289: 1743–1746. 

Maljanen M, Alm J, Martikainen PJ, Repo T. 2010. Prolongation of soil frost 

resulting from reduced snow cover increases nitrous oxide emissions 

from boreal forest soil. Boreal Environment Research 15: 34–42. 

Maurer J. 2007. Atlas of the Cryosphere. National Snow and Ice Data Center. 

(26 January 2012; http://nsidc.org/data/atlas). 

Mazhitova G, Malkova G, Chestnykh O, Zamolodchikov D. 2008. Recent 

decade thaw depth dynamics in the European Russian Arctic based on 

the Circumpolar Active Layer Monitoring (CALM) data. Pages 433–438 

in Kane DL, Hinkel KM, eds. Proceedings of the Ninth International 

Conference on Permafrost. Institute of Northern Engineering, University 

of Alaska, Fairbanks. 

Micol T, Jouventin P. 2001. Long-term population trends in seven Antarctic 

seabirds at Pointe Géologie (Terre Adélie): Human impact compared 

with environmental change. Polar Biology 24: 175–185. 

Myers-Smith IH, McGuire AD, Harden JW, Chapin FS III. 2007. Influence 

of disturbance on carbon exchange in a permafrost collapse and adjacent 

burned forest. Journal of Geophysical Research 112: 11. 

Nemergut DR, Anderson SP, Cleveland CC, Martin AP, Miller AE, Seimon 

A, Schmidt SK. 2007. Microbial community succession in unvegetated, 

recently-deglaciated soils. Microbial Ecology 53: 110–122. 

Nolin AW, Daly C. 2006. Mapping “at risk” snow in the Pacific Northwest. 

Journal of Hydrometeorology 7: 1164–1171. 

Parmesan C. 2006. Ecological and evolutionary responses to recent climate 

change. Annual Review of Ecology, Evolution, and Systematics 37: 

637–669. 

Peck LS, Barnes DKA, Cook AJ, Fleming AH, Clarke A. 2010. Negative feedback 

in the cold: Ice retreat produces new carbon sinks in Antarctica. 

Global Change Biology 16: 2614–2623. 

Pfeffer WT, Harper JT, O’Neel S. 2008. Kinematic constraints on glacier 

contributions to 21st-century sea-level rise. Science 321: 1340–1343. 

Prowse TD, Wrona FJ, Reist JD, Gibson JJ, Hobbie JE, Lévesque LMJ, 

Vincent WF. 2006. Climate change effects on hydroecology of Arctic 

freshwater ecosystems. Ambio: A Journal of the Human Environment 

35: 347–358. 

Ray C, Beever E, Loarie S. 2012. Retreat of the American pika: Up the mountain 

or into the void? In Brodie JF, Post E, Doak D, eds. Conserving 

Wildlife Populations in a Changing Climate. University of Chicago 

Press. Forthcoming. 


world. BioScience 62: 342–353. 

Romanovsky VE, Gruber S, Instanes A, Jin H, Marchenko SS, Smith SL, 

Trombotto D, Walter KM. 2007. Frozen ground. Pages 181–200 in 

United Nations Environment Programme (UNEP). Global Outlook for 

Ice and Snow. UNEP. 

Schaefer DA, Driscoll CT Jr, van Dreason R, Yatsko CP. 1990. The episodic 

acidification of Adirondack lakes during snowmelt. Water Resources 

Research 26: 1639–1647. 

Schimel JP, Bilbrough C, Welker JM. 2004. Increased snow depth affects 

microbial activity and nitrogen mineralization in two Arctic tundra 

communities. Soil Biology and Biogeochemistry 36: 217–227. 


Schuur EAG, Vogel JG, Crummer KG, Lee H, Sickman JO, Osterkamp TE. 

2009. The effect of permafrost thaw on old carbon release and net 

carbon exchange from tundra. Nature 459: 556–559. 

Solomon SD, Qin M, Manning Z, Chen M, Marquis KB, Avery T, Tignor M, 

Miller HL, eds. 2007. Climate Change 2007: The Physical Science Basis. 

Cambridge University Press. 

Stammerjohn SE, Martinson DG, Smith RC, Iannuzzi RA. 2008. Sea ice 

in the western Antarctic Peninsula region: Spatio-temporal variability 

from ecological and climate change perspectives. Deep Sea Research 

55: 2041–2058. 

Starr G, Oberbauer SF. 2003. Photosynthesis of Arctic evergreens under 

snow: Implications for tundra ecosystem carbon balance. Ecology 84: 

1415–1420. 




62: 367–376. 

Trivelpiece WZ, Hinke JT, Miller AK, Reiss CS, Trivelpiece SG, Watters GM. 

2011. Variability in krill biomass links harvesting and climate warming 

to penguin population changes in Antarctica. Proceedings of the 


Turetsky MR, Wieder RK, Vitt DH, Evans RJ, Scott KD. 2007. The disappearance 

of relict permafrost in boreal north America: Effects on peatland 

carbon storage and fluxes. Global Change Biology 13: 1922–1934. 

Vogel JG, Schuur EAG, Trucco C, Lee H. 2010. The carbon cycling response of 

arctic tundra to permafrost thaw and thermokarst development. Journal 

of Geophysical Research 114 (Art. G4). doi:10.1029/2008JG000901 

Walker DA, Halfpenny JC, Walker MD, Wessman CA. 1993. Long-term 

studies of snow–vegetation interactions. BioScience 43: 287–301. 

Weimerskirch H, Inchausti P, Guinet C, Barbraud C. 2003. Trends in bird 

and seal populations as indicators of a system shift in the Southern 

Ocean. Antarctic Science 15: 249–256. 

Articles 

Westerling AL, Hidalgo HG, Cayan DR, Swetnam TW. 2006. Warming and 

earlier spring increase western U.S. forest wildfire activity. Science 313: 

940–943. 

Williams MW, Losleben M, Caine N, Greenland D. 1996. Changes in 

climate and hydrochemical responses in a high-elevation catchment 

in the Rocky Mountains, USA. Limnology and Oceanography 41: 

939–946. 

Williams MW, Knauf M, Cory R, Caine N, Liu F. 2007. Nitrate content and 

potential microbial signature of rock glacier outflow. Earth Surface 

Processes and Landforms 32: 1032–1047. 

Wood ME, Kelley JT, Belknap DF. 1989. Patterns of sediment accumulation 

in the tidal marshes of Maine. Estuaries 12: 237–246. 

Zamolodchikov D, Kotov A, Karelin D, Razzhivin V. 2008. Recent climate 

and active layer changes in northeast Russia: Regional output 

of Circumpolar Active Layer Monitoring (CALM). Pages 2021–2026 

in Kane DL, Hinkel KM, eds. Proceedings of the Ninth International 

Conference on Permafrost. Institute of Northern Engineering, University 

of Alaska, Fairbanks. 

Andrew G. Fountain (andrew@pdx.edu) is a professor in the Department of 

Geology at Portland State University in Portland, Oregon. John L. Campbell is 

an ecologist at the US Department of Agriculture Forest Service in Durham, New 

Hampshire. Edward A. G. Schuur is an associate professor in the Department 

of Biology at the University of Florida, Gainesville. Sharon E. Stammerjohn is 

an assistant professor in the Department of Ocean Sciences at the University 

of California, Santa Cruz. Mark W. Williams is a professor in the Department 

of Geography at the University of Colorado, Boulder. Hugh W. Ducklow is a 

senior scientist and director of The Ecosystems Center at the Marine Biological 

Laboratory in Woods Hole, Massachusetts.

biology join - Coweeta LTER - University of Georgia

Create successful ePaper yourself

Delete template?

Save as template?