The use of chronosequences in studies of ecological succession ...

More documents

Recommendations

Info

728 L. R. Walker et al.build-up of species-specific pathogens in the root zone canaccelerate species replacement and hence vegetation change(Van der Putten, Van Dijk & Peters 1993). More recent studieson abandoned fields of known age that differ in time sinceabandonment have led to significant insights about howbelow-ground communities and plant–soil feedbacks serve asdrivers of species replacement and vegetation successionaldevelopment (e.g. De Deyn et al. 2003; Kardol, Bezemer &van der Putten 2006). Our understanding of the successionaldevelopment of soil biotic communities has also advancedthrough studies of recently exposed glacial substrates. Thesesubstrates are initially composed of simple, heterotrophic,microbial communities (Bardgett & Walker 2004; Bardgettet al. 2007) and photosynthetic and nitrogen-fixing bacteria(Schmidt et al. 2008) that over time develop more complex,fungal-based food webs (Ohtonen et al. 1999; Bardgett et al.2007). Also, advances have emerged from applying the chronosequenceapproach to substrates of differing decay stage andtherefore age, such as fungal communities on decaying leaves(Frankland 1998) and microarthropod communities on decayingtree stumps (Seta¨la¨ & Marshall 1994).CONVERGENCE OF SERES AND VEGETATIVESTRUCTUREThere are multiple potential trajectories for succession, includingsingle or multiple pathways that can be parallel, convergentor divergent, but that also can be cyclic or form complex networks(Fig. 1; Walker & del Moral 2003). Single and cyclicpathways are the most easily adapted to space-for-time inferencesbecause they typically have few dominant species andfew stages (Watt 1947). Chronosequences can also be usefulfor the study of convergent seres, particularly when convergenceoccurs early in succession. Whenever multiple pathwaysare present along a chronosequence, sufficient within-stagesampling is required to detect the pathways and avoid erroneousinferences about non-existent pathways (Fig. 1). In thecase of an incomplete chronosequence (missing stages), additionalhistorical, retrospective, observational or experimentaldata are critical before robust inferences can be made aboutthe missing links.Convergence occurs as a reduction in heterogeneity of speciescomposition among sites over time or as a growing resemblanceamong different trajectories (Christensen & Peet 1984;del Moral 2007). Convergence is most likely where there issome biological legacy from the initial disturbance, where adeterministic sequence of species or life-forms is driven by biologicalprocesses or where environmental conditions are predictable(Nilsson & Wilson 1991; Inouye & Tilman 1995;Wilson, Allen & Lee 1995). Decreasing beta-diversity is oneway to measure convergence along a sere (del Moral & Jones2002). Convergence to a dominant growth form such as tussockgrasses, dense shrub lands, or trees can potentially reducethe typically stochastic processes of dispersal and establishment,and distinctly alter ecosystem properties and environmentalconditions (Walker & del Moral 2003). For example,where succession proceeds from relatively open vegetation toclosed forest canopy, one might expect a convergence (reductionof variation) among stands of plant traits such as specificleaf area and root : shoot ratios, soil microbiological traitssuch as the relative biomass of bacteria and fungi or environmentalchanges such as amount of understorey light and soiland air temperatures. Despite some evidence of predictabledirectional shifts in these variables (Tilman 1988; Wood &Morris 1990; Chapin et al. 1994; Llambí et al. 2003; Bardgett& Walker 2004), more needs to be done to investigate convergenceamong stands along the lines of the study by Fukamiet al. (2005) on the convergence of plant functional traits duringsecondary succession. Trait convergence is also complicatedby spatial heterogeneity in most plant (Armesto, Pickett& McDonnell 1991) and soil (Boerner, DeMars & Leicht 1996)communities, and a lack of uniformity in the effects of similarstructures such as trees on the environment (Binkley & Giardini1998). Such spatial variability compounds the difficulty ofinterpreting temporal variability within sites and suggests theneed for caution in interpreting chronosequences, even thosebasedonconvergenceofvegetative structure. Although convergence(especially of life- and growth forms) is a commonphenomenon in some long-term seres (Rydin & Borgegård1991; Poli Marchese & Grillo 2000), other seres show increasedheterogeneity of life-forms as we discuss later.Glacier Bay, Alaska, USA, is a well-studied sere that illustratesmany of the points we make about convergence, includingthe need for multiple sources of information, intensesampling and an understanding of the role of the dominantplant species. The retreating glaciers at Glacier Bay haveexposed moraines that have been dated by geological records,direct observation and repeat photography (Vancouver 1798;Field 1947; Goldthwait 1966). A chronosequence of early successionalplants has been validated through permanent plotsinitiated by Cooper (1923) and several additional observationaland experimental studies (summarized in Chapin et al.1994). However, links to the next stage are less well established.Detailed sampling determined that the early successionalplants (notably the nitrogen-fixing Alnus) do not always precedestands of Picea (Fastie 1995), the dominant tree specieson moraines > 200 years old, as previously assumed. Piceaforests contribute greatly to soil acidification (Alban 1982) andpromote a retrogressive stage (Wardle, Walker & Bardgett2004) when Picea stands degenerate after about 10 000 years(Ugolini & Mann 1979; Noble, Lawrence & Streveler 1984)and understorey diversity increases. Assuming that early successionalstages converge to Picea forests (a likely, althoughnot directly observed linkage), concerns about Alnus–Piceasequences during the first 200 years become less critical whenaddressing longer time-scales where Picea and its accompanyingecosystem-level effects predominate. Therefore, for measuresof soil biota, soil fertility and plant physiognomyencompassing several millennia, the exact replacementsequence for plant species at hundred-year scales is of marginalimportance, if the processes of interest have converged. Moreimportantatthelongertime-scales are the frequency, intensityand spatial distributions of fire, insect outbreaks, logging andother disturbances that destroy forests and initiate secondaryÓ 2010 The Authors. Journal compilation Ó 2010 British Ecological Society, Journal of Ecology, 98, 725–736
Chronosequences, succession and soil development 729succession, because of the presence of residual forest soil followingsuch disturbances (Walker & del Moral 2003).LONG-TERM AND RETROGRESSIVE SERESOver time frames encompassing thousands to millions of years,dramatic shifts can occur in soil properties and accompanyingplant, animal and microbial communities. These changesnegate the previously held assumption that plant communitiesreach a stable and self-replacing climax (Whittaker 1953). Atsuch temporal scales, chronosequences are usually the onlytool available to interpret changes in ecosystem processes, suchas net primary productivity and rates of decomposition, nutrientmineralization and nutrient immobilization (Vitousek2004; Wardle, Walker & Bardgett 2004; Wardle et al. 2008).Long-term chronosequences have also long been recognized asvaluable for understanding processes of soil formation anddevelopment over time (Walker & Syers 1976), often independentlyof their application to plant and soil biological communities.However, the linkages between long-term soildevelopment, shorter-term changes in microbial and faunalcommunities and vegetation development are relativelypredictable (Wardle 2002; Bardgett et al. 2005), making thechronosequence approach a reasonable template for interpretationof change at many temporal scales.Predictable shifts during stages of progressive successioninclude increasing plant and soil microbial biomass, nutrientavailability and rates of nutrient cycling (Chapin, Matson &Mooney 2003). While such increases can continue for thousandsof years (Vitousek 2004; Walker & Reddell 2007), in theabsence of catastrophic disturbances that reset the system, ecosystemretrogression can occur, which involves a markeddecrease in nutrient availability, often accompanied by reductionsin plant biomass (Walker et al. 2001; Wardle, Walker &Bardgett 2004). This pattern has been widely documented inmany climates and vegetation types, with the possible exceptionsof arid systems (Lajtha & Schlesinger 1988; but see Selmants& Hart 2008) and tropical lowland rain forests (Ashton1985; Kitayama 2005). Retrogression is typically driven byconversion of soil nutrients and especially phosphorus to lessavailable forms, and in some cases leaching of nutrients belowthe rooting zone or the development of impermeable soil pansleading to water-logging (Walker & Syers 1976; Vitousek 2004;Coomes et al. 2005; Peltzer et al. in press). Long-term (millennialscale) changes in soil processes track, and are impacted by,mid-term (100–1000 years) to short-term (1–100 years)decreases in litter quality, decomposition rates, nutrient useefficiency and nutrient accumulation in plants (Cordell et al.2001; Richardson et al. 2005; Wardle et al. 2009) and veryshort-term (days to months) alterations in soil microbial andanimal populations (Wardle, Walker & Bardgett 2004; Bardgettet al. 2005; Doblas-Miranda et al. 2008). Therefore, retrogressiondoes not simply involve shifts in community- andecosystem-level properties at longer time-scales, but an integrationof short- to long-term processes that are distinct from progressivesuccession. To the extent that plant and soilcharacteristics of interest are predictable across stages of retrogression,chronosequences remain a valid tool. We use twoexamples to illustrate the benefits of applying the chronosequenceapproach to long-term seres. Each has a relatively shortprogressive phase followed by a much longer retrogressivephase.The current Hawaiian Islands represent an excellent,> 7-Myr chronosequence, because the ecological consequencesof their sequential development over an oceanic hotspotare well-documented (Vitousek 2004), making them idealfor between-island comparisons (Mueller-Dombois & Fosberg1997). Both progressive (Mueller-Dombois 1987) and retrogressive(Wardle, Walker & Bardgett 2004) succession havebeen documented in this system, with progressive successiondominant on the younger Island of Hawaii (0–0.43 Myr) andretrogressive succession more widespread on older islands suchas Maui (0.8–1.3 Myr) and Kauai (5.1 Myr). Within-islandchronosequences have also been characterized on the reliablydated and mapped series of volcanic surfaces on the Island ofHawaii that range from 1 year to > 4000 years old (Drake &Mueller-Dombois 1993; Aplet & Vitousek 1994; Kitayama,Mueller-Dombois & Vitousek 1995). For example, one cancompare succession and soil development on several surfaces(a‘a lava, pahoehoe lava) across a wide range of elevations(900–> 3000 m a.s.l.), spatial scales (local to > 500 km 2 )andclimates. Under such conditions, studies of chronosequencescan thus be designed to meet various assumptions, variationcan be quantified through replication within categories, andmultivariate approaches can correct for incomplete designswhere chronosequence assumptions are not met. Dominationof the Hawaiian forests by a single tree species (Metrosiderospolymorpha), albeit with several ecotypes, further facilitatescomparisons between stages of plant morphology or soil developmentduring both the progressive and retrogressive phasesof succession. However, given the numerous climatic changesand variable allochthonous inputs, such as phosphorus inputsfrom Asian dust, that have occurred during the long history ofthe current Hawaiian Islands (Chadwick et al. 1999), age-specificprocesses necessarily become less precise (Vitousek 2004).The Cooloola Dune sequence in eastern Australia is anotherexample of a long-term sere with a retrogressive phase where achronosequence approach has been useful. The progressivephase lasted for c. 250 000 years as soil carbon, nutrients andforest biomass accumulated, and was followed byc. 350 000 years of retrogression as podzolic soils developed,leaching occurred to 20-m depth and forest productivitydeclined (Thompson 1981; Walker et al. 1981, 2000; Wardle,Walker & Bardgett 2004). The oldest soils support a diverseunderstorey plant community (Wardle et al. 2008) adapted toextreme infertility. As in Hawaii, other disruptions inevitablyoccur over such long time spans (fire is a recurring phenomenoninAustralia),butthechronosequenceas a soil-age gradientremains robust. In both Hawaii and Australia, researchquestions that are best answered in studies of the older stagesshift to the effects of soil age on community and ecosystemprocesses, rather than the generation of hypotheses aboutmechanisms of succession and species replacements bestaddressed in younger seres.Ó 2010 The Authors. Journal compilation Ó 2010 British Ecological Society, Journal of Ecology, 98, 725–736
Page 1 and 2: Journal of Ecology 2010, 98, 725-73
Page 3: Chronosequences, succession and soi
Page 7 and 8: Chronosequences, succession and soi

The use of chronosequences in studies of ecological succession ...

Create successful ePaper yourself

Delete template?

Save as template?