Views
3 years ago

Simulating the effects of different disturbance regimes on Cortaderia ...

Simulating the effects of different disturbance regimes on Cortaderia ...

BIOLOGICAL CONSERVATION

BIOLOGICAL CONSERVATION 128 (2006) 128– 135 129abundance ong>ofong> alien plant species, or a better performance ong>ofong>particular plant taxa, in human-disturbed habitats than innatural ones (Hobbs and Huenneke, 1992; Pysek, 1994; Robinsonet al., 1995; Lambrinos, 2002). Disturbances provide both‘‘windows ong>ofong> opportunity’’ (sensu Myster, 1993) for ong>theong> entranceong>ofong> alien propagules and increases in ong>theong> quantity ong>ofong> resourcesper capita, including space for establishment andgrowth (Hobbs and Huenneke, 1992; Davis et al., 2000). Furong>theong>rmore,human-disturbed areas are also ong>ofong>ten associatedwith human activities (e.g., agriculture, gardening, road systems,tourism) that promote ong>theong> arrival ong>ofong> alien species (Frenotet al., 2001; Hansen and Clevenger, 2005).Disturbances are complex. Disturbance ong>regimesong> better explainong>theong> set ong>ofong> spatial and temporal ong>disturbanceong> characteristicsoccurring in a territory, including extent, shape,frequency, season, intensity and severity, among oong>theong>rs (Sousa,1984; Moloney and Levin, 1996). Though ong>theong> ong>disturbanceong>regime has been successfully explored in many ecologicalcontexts (e.g., Moloney and Levin, 1996; Pausas, 1999; Lloretet al., 2003), it has rarely been applied to ong>theong> invasion processby alien plants (Vujnovic et al., 2002; With, 2002; Huston,2004), despite that it is a driver ong>ofong> fluctuating resource availability(Davis et al., 2000; Davis and Pelsor, 2001), and thus itmay increase invasibility.To consider both ong>theong> spatial and temporal aspects ong>ofong> a ong>disturbanceong>regime on invasion, a simulation approach is appropriatebecause it allows working at large spatial and temporalscales (Higgins and Richardson, 1996). Most ong>ofong> ong>theong> modelsused for predicting ong>theong> rate ong>ofong> plant invasion (Marco et al.,2002; Cannas et al., 2003) are based on ong>theong> arrival ong>ofong> ong>theong>new species in an empty (bare) land (Higgins et al., 1996).However, this situation is not always realistic; usually, invasionstarts in a vegetated area. Therefore, simulations shouldbe based on ong>theong> life history and competitive attributes ong>ofong> bothong>theong> invading species and ong>theong> dominant species in ong>theong> ‘‘host’’ecosystem (Cannas et al., 2003).Cortaderia selloana (Schultes) Asch. & Graebner) (pampasgrass) is a perennial, tussock grass native to South Americanpampas (Argentina, souong>theong>rn Brazil, and Uruguay) that hasbeen introduced in many regions ong>ofong> ong>theong> world as a gardenplant because ong>ofong> its showy plumes, and also for dryland forage,soil amendment, and windbreaks (Harradine, 1991). Inmany regions, including some Mediterranean ecosystems,C. selloana (Cortaderia hereafter) has become invasive (Bossardet al., 2000; Lambrinos, 2001, 2002). In Spain, this specieshas escaped into wetlands, oldfields and coastalgrasslands. More specifically, in ong>theong> protected coastal wetlandareas ong>ofong> Catalonia (NE Spain), this species has becomea major conservation concern because its invasion ong>ofong> oldfieldsprevents ong>theong>ir restoration back to wetlands wheresome rare or endemic species occur. Cortaderia can reach2–4 m in height, forming a large fountain-like tussock withsharply serrated leaves. The inflorescence consists ong>ofong> showyplume-like heads at ong>theong> end ong>ofong> a stiff stem. Each plant canproduce about a million wind-dispersed viable seeds (Connorand Edgar, 1974; Lambrinos, 2002). Invasion by Cortaderiacan drastically alter ecosystem properties (e.g., flammability,diversity, food webs), as has been demonstrated for ong>theong> morphologicallysimilar Cortaderia jubata (Lem.) Stapf (Bossardet al., 2000; Lambrinos, 2000).We used a spatially explicit model to explore Cortaderiainvasion patterns in relation to several ong>disturbanceong> ong>regimesong>that mimic relevant management scenarios occurring in ong>theong>Catalonian region (NE Spain), also typical ong>ofong> oong>theong>r Mediterraneanbasin areas. Our main questions were: (1) To what extentdoes invasion depend on ong>disturbanceong> frequency? (2) Doong>differentong> temporal patterns ong>ofong> ong>disturbanceong> (i.e., regular vs.irregular) modify invasion? (3) Is ong>theong> extent ong>ofong> ong>theong> ong>disturbanceong>more relevant than ong>theong> ong>disturbanceong> frequency? (4) Dochanges in ong>theong> spatial variability ong>ofong> ong>theong> ong>disturbanceong> (i.e., randomvs. fixed and aggregated vs. sparse occurrence ong>ofong> ong>disturbanceong>)affect invasion? To our knowledge this is ong>theong> first timethat ong>differentong> components ong>ofong> a ong>disturbanceong> regime are disentangledto predict ong>theong>ir effect on ong>theong> degree ong>ofong> communityoccupancy by a wind-dispersed alien plant.2. Methods2.1. ModellingWe used a spatial-explicit version ong>ofong> FATE (Moore and Noble,1990) called FATELAND as implemented in ong>theong> LASS modellingenvironment (Pausas and Ramos, 2004, 2006; available atwww.ceam.es/lass). FATE is a non-spatial semi-quantitativevegetation dynamics model specifically designed for ong>disturbanceong>-proneecosystems. FATE was elaborated for Australiancommunities and later tested for Mediterranean ecosystems(Pausas, 1999; Lloret et al., 2003). FATE is based on ong>theong> life historytraits ong>ofong> individual plants at stand level where severalspecies may coexist. Vegetation dynamics are qualitativelypredicted from simple parameters describing species traits.The model runs at annual time steps and simulates cohortsong>ofong> plants that pass through a series ong>ofong> four discrete stages:propagules, seedlings, immature and mature plants. In FATEong>theong>re are three levels ong>ofong> light availability (low, medium, high)determined by ong>theong> vertical structure ong>ofong> ong>theong> vegetation (stratum).Germination and survival are determined on ong>theong> basisong>ofong> ong>theong>se light availability levels, and changes in light availabilityare due to species overtopping one anoong>theong>r (for furong>theong>r detailssee Moore and Noble, 1990; Pausas, 1999; Cousins et al.,2003).FATELAND incorporates ong>theong> spatial component missing inFATE; that is, FATELAND corresponds to a 2-dimensional arrayong>ofong> cells, and ong>theong> FATE model operates in each cell. Differentspecies may co-occur in each cell. FATELAND also includes aspatially explicit dispersal module and a ong>disturbanceong> module.Thus, FATELAND is especially appropriate for modellingspatio-temporal dynamics in disturbed ecosystems. WhileFATE is a full deterministic model, in FATELAND, stochasticity isincluded throughout ong>theong> dispersal module, which is basedon ong>theong> probability ong>ofong> dispersal (p) according to ong>theong> negativeexponential equation: p = A · Exp(k · Distance/MaxDistance),where Distance is a random number between 0 and MaxDistance,MaxDistance is ong>theong> maximum distance to ong>theong> edge ong>ofong>ong>theong> landscape, and k and A are ong>theong> species-specific inputparameters for modelling dispersal (Table 1); k provides ong>theong>shape ong>ofong> ong>theong> curve and A ong>theong> magnitude ong>ofong> ong>theong> dispersal.A fecundity value is used as a number ong>ofong> iterations in thisdispersal module. Disturbance is simulated as ong>theong> removalong>ofong> all vegetation from ong>theong> disturbed cell, following a set ong>ofong>

130 BIOLOGICAL CONSERVATION 128 (2006) 128– 135designed spatial and temporal ong>disturbanceong> scenarios (seebelow).We built an artificial landscape ong>ofong> 70 · 70 (4900) squarecells, each one equivalent to 3 · 3 m, with two speciestypes: ‘‘Grass’’ and ‘‘Cortaderia’’ (Table 1). Grass type wouldinclude ong>differentong> herbaceous species commonly co-occurringwith Cortaderia in Spain. In ong>theong>se grasslands, graminoidspecies are dominant (e.g., Festuca arundinacea Schreber, Elymuspungens (Pers.) Melderis, Cynodon dactylon (L.) Pers.),although forbs also occur (e.g., Trifolium pratense L.). TheGrass type species is assumed to usually establish ashort-term persistent seed bank (based on species closelyrelatedto those found in ong>theong> study area; Grime et al.,1988). Since ong>theong>y represent an early successional stage, weassume a low-to-moderate ability to recruit under closedcanopy (i.e., at low resource levels, Table 1). We also assumethat ong>theong> Grass type has less dispersal ability than Cortaderia;ong>theong> dispersal parameters for ong>theong> latter species were derivedfrom Saura and Lloret (2005). The Cortaderia type representsong>theong> attributes ong>ofong> ong>theong> invasive species Cortaderia selloana.This species produces a great amount ong>ofong> long-distance dispersedseeds (Connor and Edgar, 1974; Lambrinos, 2002;Saura and Lloret, 2005). We assume that ‘‘Grass’’ plantshave a large self-replacing ability, resulting in a longerlife-span than Cortaderia. Although Cortaderia sprouts afteraboveground removal, we did not include this life historytrait in ong>theong> modelling because our ong>disturbanceong>s simulateddrastic events in which all living plants would be totally removed.Empty cells may occur after ong>disturbanceong>s or plantdeath. All ong>theong> landscape is potentially available for invasionby Cortaderia, if ong>theong> establishment conditions are appropriate(Table 1).2.2. ScenariosThe initial situation corresponds to a landscape in which allong>theong> cells are occupied by ‘‘Grass’’ plants, and Cortaderia occursin a narrow front (1 cell wide) in one border ong>ofong> ong>theong> landscape.This mimics ong>theong> case found in many natural invaded zonesTable 1 – Main species attributes used for modelling ong>theong>invasion ong>ofong> Cortaderia selloana in a landscapedominated by ong>theong> species type ‘‘Grass’’Traits Grass CortaderiaLifespan (years) 100 25 ± 3Age at maturation (years) 2 4Innate dormancy? (years) Yes (2) NoGermination rate Low, Medium, High None, Low, LowSeedling survival No, Yes, Yes Yes, Yes, YesImmature survival No, Yes, Yes Yes, Yes, YesMature survival No, Yes, Yes Yes, Yes, YesDispersal capacity (A) 0.3 0.9Dispersal shape (k) 2.4 6.9Fecundity 1 500 15 000The three values for germination and survival refer to low, mediumand high resource levels. See ong>theong> method section for furong>theong>rdetails.near urban areas where Cortaderia is used as an ornamentalplant. From this initial landscape we applied three sets ong>ofong> ong>disturbanceong>scenarios (A, B, C) within a 55-year timeframe bycombining ong>differentong> spatial and temporal ong>disturbanceong> patterns(Table 2). In all cases, ong>theong> ong>disturbanceong> started at year 5.(A) Scenarios ong>ofong> increasing ong>disturbanceong> frequency. Weselected regular and sparse ong>disturbanceong> scenarios fromno ong>disturbanceong> (A0) to 5 (A5), 10 (A10), 20 (A20) and 40(A40) ong>disturbanceong>s ong>ofong> 30 randomly chosen cells at eachong>disturbanceong> event. Therefore, ong>theong> ong>disturbanceong> frequencieswere (from A5 to A40): every 10, 5, 2 and 1 year,and ong>theong> inter-ong>disturbanceong> periods were 9, 4, 1 and 0 years,respectively.(B) Scenarios ong>ofong> increasing ong>disturbanceong> frequency anddecreasing extent ong>ofong> each ong>disturbanceong> (i.e., number ong>ofong>disturbed cells). Five (B5), 10 (B10), 20 (B20) and 40 (B40)ong>disturbanceong>s ong>ofong> 120, 60, 30 and 15 cells respectively. Notethat ong>theong> total number ong>ofong> disturbed cells for ong>theong> simulatedtimeframe is ong>theong> same (i.e. 600 cells) in each scenario.Disturbance frequency and inter-ong>disturbanceong> periodswere as in (A) above.(C) Scenarios with changing ong>disturbanceong> patterns combiningong>theong> following temporal and spatial patterns:Temporal ong>disturbanceong> patterns:• Regular temporal pattern (CR) – Five ong>disturbanceong>s at regularintervals ong>ofong> time (every 10 years). The first ong>disturbanceong>is at year 5 and ong>theong> last one at year 45. Thus ong>theong> inter-ong>disturbanceong>periods consist ong>ofong> 9 years.• Irregular temporal pattern (CI) – The first ong>disturbanceong> is atyear 5 and ong>theong>n at years 23, 25, 43 and 45. Therefore, ong>theong>sequence ong>ofong> inter-ong>disturbanceong> periods is 2, 18 and 2.Spatial ong>disturbanceong> patterns:• Aggregated (Q10f) – a fixed block ong>ofong> 10 · 10 cells (100 cells,ca. 2% ong>ofong> ong>theong> landscape) located in ong>theong> centre ong>ofong> ong>theong> landscapeis disturbed every ong>disturbanceong> year; ong>theong> remaininglandscape is not affected by ong>disturbanceong>.• Aggregated random (Q10r) – a block ong>ofong> 10 · 10 cells (100cells, ca. 2% ong>ofong> ong>theong> landscape) is randomly located withinong>theong> landscape every ong>disturbanceong> year.• Sparse fixed (Q1f) – 100 cells are randomly chosen at ong>theong>first year ong>ofong> ong>disturbanceong> and ong>theong> same cells are disturbedevery ong>disturbanceong> year.• Sparse random (Q1r) – 100 randomly chosen cells are disturbedevery ong>disturbanceong> year.These simulations mimic several ong>disturbanceong> ong>regimesong>associated with specific human activities, including conservationmanagement (Table 3). Dumping resulting from periurbandevelopment is common in coastal wetland areas ong>ofong>Catalonia (NE Spain), producing a pattern ong>ofong> small, scattereddisturbed areas (Table 2, CI Q1 scenarios). Agricultural activitieseliminate Cortaderia from ong>theong> invaded communities but afew individuals remaining in ong>theong> field margins can invadewhen ong>theong> field is abandoned again. The temporal pattern ong>ofong>ong>theong>se abandoned fields is irregular, and ong>theong> size ong>ofong> ong>theong> areas

Historical and Modern Disturbance Regimes of Piñon-Juniper
ECOSYSTEM SERVICES AND DISTUR- BANCE REGIME AS ...
TransienT immuniTy simulaTors For ConDuCTeD DisTurBanCes
effects of land abandonment on animal species in europe - Potsdam ...
effects of land abandonment on animal species in europe - Potsdam ...
effects of land abandonment on animal species in europe - Potsdam ...
effects of land abandonment on animal species in europe - Potsdam ...
effects of land abandonment on animal species in europe - Potsdam ...
Simulating the effects of different fire regimes on plant functional ...
Natural disturbance regimes in northeastern North America ...
Effect of Photoperiodic Regime and Egg Laying Disturbance on the ...
Simulation of Disturbance Activities
Simulating the Effects of Disturbances and on Fuel ... - COMET
HIGH-SEVERITY FIRE REGIMES Objectives - Abiotic Disturbances
MIXED-SEVERITY FIRE REGIMES Objectives - Abiotic Disturbances
The effect of different disturbance treatments on tree ... - Epsilon - SLU
A general model for simulating the effects of landscape ...
Disturbance regimes of stream and riparian systems - a disturbance ...
Effect of Different Watering Regimes on Growth Performance of Five ...
Simulating the effects of climate change on fire regimes in Arctic ...
Effect of disturbance, habitat fragmentation and alien invasive plants ...
Effect of different mowing regimes on butterflies and diurnal ... - Raco
EFFECT OF DIFFERENT TEMPERATURE REGIMES ON SEED ...
EFFECT OF DIFFERENT TEMPERATURE REGIMES ON SEED ...
Different management regimes in a boreal forest ... - Epsilon - SLU
Simulating the effects of using different types of species distribution ...
SIMULATION OF LIGHT REGIMES IN TYPICAL SUBALPINE ...
Vegetation responses to different spatial patterns of soil disturbance ...
Analysis of Natural Wind Disturbance Regimes Resulting from ...
Effect of different cooking regimes on rhubarb polyphenols