Transcription
Vegetative Dynamic Modeling for the SAMBI DSL ProjectTodd S. EarnhardtBiodiversity and Spatial Information Center, NC State University, Raleigh, NC 27695INTRODUCTIONIn the Southeastern U.S. rapid urbanization, climate change, and the direct and indirect impacts of thosetwo dynamics on ecosystems are major challenges to developing long-term conservation strategies. In thisproject we modeled future landscape conditions under a variety of climate change scenarios in order toprovide a context for land use and management decisions. To simulate vegetation dynamics, we usedspatially explicit state-and-transition models and ESSA Technologies’ Tool for Exploratory LandscapeScenario Analyses (TELSA; ESSA 2008a, ESSA 2008b, Kurz et al. 2000). Our vegetation dynamicsprojections were integrated with projected urban growth and sea level rise projections to develop the finaldataset showing projected landscape conditions. Those projections were then used as input in thelandscape prioritization decision support tools being developed at Auburn University.METHODSTELSA ModelingThere are four major inputs to TELSA: (1) a polygon map of vegetation types, (2) a non-spatial state-andtransition model for each vegetation type, (3) an initial age for each polygon, and (4) an initial structuralstage for each polygon. The vegetation polygons are tracked through time with respect to their condition(state and stage) as a result of three general processes; succession, disturbance, and management.Succession is deterministic and disturbances are stochastic (e.g., fire, insect outbreak) processes aremodeled.The SAMBI region was divided into 20 sub-regions based on Omernik Level IV ecoregion boundaries(U. S. EPA 2011), and vegetation dynamics were simulated across each sub-region separately (Figure 1).This was required due to computational limitations but also served as an additional hierarchical levelwithin the modeling framework where transition parameters could be adjusted to reflect the localdynamics more precisely.1
Figure 1. SAMBI Subregions Based on Omernik Level IV EcoregionsPolygon Land Cover MapFor each sub-region, polygons of land cover patterns were created using eCognition Elements 4.0software (Definiens Imaging 2004) with a leaf on tassel-capped transformation of ca. 2001 Landsat TMimagery (Huang et al. 2002). The resulting polygons were then converted into a raster layer thatmaintained the unique ID of each polygon. Polygons were then attributed with the Southeast GapAnalysis Project (www.basic.ncsu.edu/segap/) ecological systems land cover type using a zonal majorityfunction (Figure 2). A similar process was followed to attribute percent canopy cover and succession classfrom the LANDFIRE project (Rollins et al. 2008), and ownership (USGS National Gap Analysis Program2008) to each polygon as well.2
Figure 2. Reclass of Land Cover Polygons Using The Zonal Majority FunctionState and Transition ModelsThe state-and-transition models we used were modified from the LANDFIRE Vegetation DynamicsDevelopment Tool (VDDT) models (Smith et al. 2009). Those models had been developed in a series ofregional workshops where ecologists familiar with the vegetation described the state and stage for eachvegetation type and assigned probabilities of each disturbance (Figure 3). Each stage box has notation forclass id, age range, sub-category id, and canopy closure class.Figure 3. State-Transition Box Model3
Some VDDT models have an Alternative Succession transition pathway. This pathway allows for analternative to the deterministic transition of the ecological system under certain timeline and disturbanceconditions. This pathway functions similarly to a deterministic pathway in that it has a probability of 1and will occur if the specific time and disturbance criteria are met. In the Atlantic Coastal Plain example,an Alternative Succession pathway exists for a transition from box A (post replacement) to box B (midsuccession, closed canopy) if 10 years pass without any fire disturbance. If there is a fire, then it will takethe polygon in the post replacement condition 14 years before transition to the mid succession, closedcanopy condition. We developed two additional VDDT models to simulate the dynamics in publically andprivately owned managed pine stands. In the SAMBI region, intensively managed pine stands (almostexclusively Loblolly Pine) are a major part of the forested landscape are readily identifiable on thelandscape due to their conspicuous spectral signature and spatial pattern. The primary disturbance in thesemodels is related to management activities (clearcuts, thinning, planting). We labeled these managed pinestands as public based on an overlay operation with the USGS Protected Areas Database layer. Thedifference between vegetation dynamics in publicly owned versus privately owned stands was a longerrotation time in the public lands.One hundred forty of the 168 ecological systems and land use classes in the baseline land cover map(SEGAP citation) were modeled in TELSA for the SAMBI region. The remaining 20 classes wereconsidered un-projectable because they were not ecological systems (e.g. water, agriculture) or did nothave a LANDFIRE VDDT model associated with them. These classes were held constant throughout themodel run and remained unchanged unless the polygons were converted to urban or inundated by sealevel rise. State transition probabilities are provided for age and distrubance in Appendices A and B.Climate Change Influence on Fire RegimeThe vegetation dynamics models for the SAMBI incorporated the influence of projected climate changevia changes to the fire regime. Time series regression was first used to estimate the relationship betweenclimate and the amount of monthly acreage burned by wildfires. Monthly temperature and precipitationwere used as predictors for both the current month’s model estimate as well as observed temperature andprecipitation for previous months to capture the long-term moisture budget. Once the parameters of thestatistical model were estimated, empirically downscaled climate projections were used to project changesin the amount of acreage burned for a given CO2 emissions scenario. The uncertainty in the climateprojections was quantified and propagated using model weights derived from Bayesian Model Averaging(BMA; Raftery et al. 2005; Bhat et al. 2011). The predictive variance for the probabilistic fire regimechange projection was estimated using the Expectation Maximization algorithm (Dempster 1977) basedon comparisons of the hindcast downscaled GCM output and the observed fire conditions. The predictivevariance estimate was obtained by holding the GCM weights constant based on the BMA analysis. Themodeled relationship was then used to project future wildfire frequencies under anticipated climateconditions projected by three emissions scenarios; A2, B1, A1B. The three climate scenarios represent agradient of projected atmospheric CO2 levels and the associated effects of those levels on temperatureand rainfall patterns. Changes in fire disturbance resulting from shifts in the climate would be reflected inthe state and stage of vegetation. For example, increased fire disturbance would tend to keep systemssubject to fire disturbance in an early succession stage.4
Initial Age and State Class AssignmentEach polygon was assigned an initial ecological system label, age, succession class, and structure class.The ecological system, stage, and structure information were from the zonal majority operationsdescribed previously. An initial age for the polygons was assigned based on the US Forest Service FIAdatabase forest type age distributions. The ecological systems were cross-walked to FIA forest types. Adatabase query generated a proportional area for each age class in the FIA survey cycle whose date wasclosest to 2001. The area proportions were applied to the total area of each ecological system in eachregion thus establishing target areas for each age and ecological system combination. For each forest type,the ecological system polygons were randomly assigned an age value and the area of that polygon wasadded to that group’s cumulative area total. This was repeated until the area target for that age andecological system combination was reached. Provisions were made to ensure that ages were consistentwith the implicit boundaries of the ecological systems succession stage and canopy cover (e.g., an earlysuccession stand would not be assigned an age of 88 years).The State Class ID that TELSA uses to model the polygons is a combination of an ecological systemcode, succession stage code (early [post-disturbance], mid, or late stage), and canopy structure code (earlysuccession, closed, or open). The polygons were tracked with respect to their age and condition (state andstage) as a result of three general processes; succession, disturbance, and management.Integration of Urban Growth and Sea Level Rise Model OutputAfter the map preparation and initial state class assignment, we used TELSA to simulate vegetationdynamics for 100 years for each climate scenario (2001-2100). Mapped output was generated at decadaltime intervals. In a separate modeling effort, urban growth and sea level rise were simulated for theSAMBI region. These data layers were incorporated into the vegetation succession model output usingzonal majority operations. Urban growth predictions took precedence over any transitions projected in theSLAMM output layers.Urban growth projections were modeled independently with SLEUTH-3r model (Jantz et al, 2009) (seeUrban Growth Modeling; basic.ncsu.edu/dsl/) and were used to identify vegetation polygons that wereprojected to transition from a vegetated to an urban cover type. Polygons were classified as urban if 50%or more of the pixels within the polygon had a greater than or equal to 50% probability of being convertedto urban.We incorporated sea level rise projections from Sea-Level Affecting Marshes Model (SLAMM; Clough2008) (see Sealevel Rise Modeling; basic.ncsu.edu/dsl/) into the land cover maps independent of theTELSA modeling. For each TELSA output map, SLAMM data were used to reflect the conversion ofland cover classes to alternate classes due to sea level rise projections. Each initial polygon wascharacterized with respect to the SLAMM data using a majority function. If the majority of the pixelswithin a vegetation polygon were “changed” pixels, the polygon was toggled to the appropriate class.When the transition had a one to one relationship with an ecological system, that system label was appliedto the polygon. In some cases, there was a one-to-many relationship of a SLAMM class to ecologicalsystems. In these cases, we used a spreadsheet script to analyze tables of adjacent polygons to determinethe new ecological system label to apply to the SLAMM transformed polygons. If an adjacent polygonwas labeled with an ecological system that satisfied the SLAMM crosswalk criteria, then the transformedpolygon was assigned that label. Otherwise, the polygon was assigned the most common occurringrelated ecological system.5
Figure 4. Integrated Land Cover Model Representing 2000 and 20606
Once the SLEUTH and SLAMM outputs had been integrated with the vegetation dynamics projections,the TELSA output tables were joined to the spatial data layers to create a grid for each 10 year time step.This was done for each of the climate scenarios by sub-region. The sub-region grids were then mergedinto one contiguous data layer for the SAMBI region (Table 1).Table 1. Final Vegetation Dynamics Datasets (www.basic.ncsu.edu/dsl)File NameDataset NameYearClimate Change ScenarioDSL-SAMBI lc00ab0.zipa1b 20002000a1bDSL-SAMBI lc10ab0.zipa1b 20102010a1bDSL-SAMBI lc20ab0.zipa1b 20202020a1bDSL-SAMBI lc30ab0.zipa1b 20302030a1bDSL-SAMBI lc40ab0.zipa1b 20402040a1bDSL-SAMBI lc50ab0.zipa1b 20502050a1bDSL-SAMBI lc60ab0.zipa1b 20602060a1bDSL-SAMBI lc70ab0.zipa1b 20702070a1bDSL-SAMBI lc80ab0.zipa1b 20802080a1bDSL-SAMBI lc90ab0.zipa1b 20902090a1bDSL-SAMBI lc99ab0.zipa1b 21002100a1bDSL-SAMBI lc00a20.zipa2 20002000a2DSL-SAMBI lc10a20.zipa2 20102010a2DSL-SAMBI lc20a20.zipa2 20202020a2DSL-SAMBI lc30a20.zipa2 20302030a2DSL-SAMBI lc40a20.zipa2 20402040a2DSL-SAMBI lc50a20.zipa2 20502050a2DSL-SAMBI lc60a20.zipa2 20602060a2DSL-SAMBI lc70a20.zipa2 20702070a2DSL-SAMBI lc80a20.zipa2 20802080a2DSL-SAMBI lc90a20.zipa2 20902090a2DSL-SAMBI lc99a20.zipa2 21002100a2DSL-SAMBI lc00b10.zipb1 20002000b1DSL-SAMBI lc10b10.zipb1 20102010b1DSL-SAMBI lc20b10.zipb1 20202020b1DSL-SAMBI lc30b10.zipb1 20302030b1DSL-SAMBI lc40b10.zipb1 20402040b1DSL-SAMBI lc50b10.zipb1 20502050b1DSL-SAMBI lc60b10.zipb1 20602060b1DSL-SAMBI lc70b10.zipb1 20702070b1DSL-SAMBI lc80b10.zipb1 20802080b1DSL-SAMBI lc90b10.zipb1 20902090b1DSL-SAMBI lc99b10.zipb1 21002100b17
LITERATURE CITEDBhat, S.K., M. Haran, A. Terando, R. Tonkonojenkov, and K. Keller, 2011: Climate projections usingBayesian model averaging and space-time dependence. Journal of Agricultural and EnvironmentalStatistics, 16, 606-628.Comer, P., D. Faber-Langendoen, R. Evans, S. Gawler, C. Josse, G. Kittel, S. Menard, M. Pyne, M. Reid,K. Schulz, K. Snow, and J. Teague. 2003. Ecological Systems of the United States: A WorkingClassification of U.S. Terrestrial Systems. NatureServe, Arlington, Virginia.Clough, J. S. 2008. SLAMM 5.0.1. Technical documentation and executable program downloadable .html.Definiens Imaging 2004. eCognition Elements Users Guide 4. Definiens Imaging GmbH, München,Germany. 26 pp.Dempster, A.P., N.M. Laird, and D.B. Rubin, 1977: Maximum likelihood from incomplete data via theEM algorithm. Journal of the Royal Statistical Society, 39B, 1-39.ESSA Technologies Ltd. 2008a. TELSA: Tool for Exploratory Landscape Scenario Analyses, ModelDescription, Version 3.6. Vancouver, BC. 64 pp.ESSA Technologies Ltd. 2008b. TELSA – Tool for Exploratory Landscape Scenario Analyses: User’sGuide Version 3.6. Prepared by ESSA Technologies Ltd., Vancouver, BC. 235 pp.Huang, C., B. Wylie, L. Yang, C. Homer, and G. Zylstra. 2002. Derivation of a tasseled captransformation based on Landsat 7 at-satellite reflectance. International Journal of Remote Sensing,23(8)1741-1748.Jantz, C. A., S.J. Goetz, D. Donato and P. Claggett. Designing and Implementing a Regional UrbanModeling System Using the SLEUTH Cellular Urban Model. Computers, Environment and UrbanSystems (2009), doi:10/1016/j.compenvurbsys.2009.08.03Kurz,W.A., S.J. Beukema, W. Klenner, J.A. Greenough, D.C.E. Robinson, A.D. Sharpe, and T.M.Webb2001. TELSA: The Tool for Exploratory Landscape Scenario Analyses.Computers and Electronics inAgriculture, Computers and Electronics in Agriculture 27(1-2) 227-242.Raftery, A.E., T. Gneiting, F. Balabdaoui, and M. Polakowski, 2005: Using Bayesian model averaging tocalibrate forecast ensembles. Monthly Weather Review, 133, 1155-1174.Rollins, Matthew G.; Frame, Christine K., tech. eds. 2006. The LANDFIRE Prototype Project: nationallyconsistent and locally relevant geospatial data for wildland fire management. Gen. Tech. Rep. RMRSGTR-175. Fort Collins: U.S. Department of Agriculture, Forest Service, Rocky Mountain ResearchStation. 416 p.Smith, J., K. Blankenship, D. Johnson, S. Simon,C. Ryan, R. Swaty, M. Bucher, M. Brod and J. Patton.Adapting LANDFIRE Vegetation Dynamics Models. August 2009. Last accessed 24 April 2012 andfire-vegetation-dynamics-models].U.S. Environmental Protection Agency. 2011. Level III and IV Ecoregions of the Conterminous UnitedStates. U.S. EPA Office of Research and Development. [Last accessed 24 April 2012 http://www.epa.gov/wed/pages/ecoregions/level iii iv.htm#Level IV]U.S.G.S. National Gap Analysis Program. 2008. Southeast GAP Regional Land Cover 2001.Biodiversity and Spatial Information Center, North Carolina Cooperative Fish and Wildlife ResearchUnit, NC State University (http://www.basic.ncsu.edu/segap/ last accessed 18 October 2010).8
Appendix A. State Age Transition D305721106611Post replacement Early-development Southern Coastal Plain Dry Upland Hardwood Forest0Post replacement Early-development Southern Piedmont Dry Oak-(Pine) Forest - Hardwood Modifier015306621Open canopy Mid-development Southern Piedmont Dry Oak-(Pine) Forest - Hardwood Modifier107611Post replacement Early-development Florida Longleaf Pine Sandhill- Open Understory107711Post replacement Early-development Florida Longleaf Pine Sandhill - Scrub/Shrub Understory015307621Open canopy Mid-development Florida Longleaf Pine Sandhill- Open Understory015307721107911Open canopy Mid-development Florida Longleaf Pine Sandhill - Scrub/Shrub UnderstoryPost replacement Early-development EGCP Maritime Forest015307921Open canopy Mid-development EGCP Maritime Forest109011014309021014309121012309521109611Post replacement Early-development Atlantic Coastal Plain Fall-Line Sandhills Longleaf Pine Woodland - LoblollyModifierPost replacement Early-development Atlantic Coastal Plain Fall-line Sandhills Longleaf Pine Woodland - OpenUnderstory ModifierPost replacement Early-development East Gulf Coastal Plain Interior Upland Longleaf Pine Woodland - OpenUnderstory ModifierPost replacement Early-development East Gulf Coastal Plain Interior Upland Longleaf Pine Woodland - Scrub/Shrub012309621Open canopy Mid-development Atlantic Coastal Plain Fall-Line Sandhills Longleaf Pine Woodland - LoblollyModifierOpen canopy Mid-development Atlantic Coastal Plain Fall-line Sandhills Longleaf Pine Woodland - OpenUnderstory ModifierOpen canopy Mid-development East Gulf Coastal Plain Interior Upland Longleaf Pine Woodland - OpenUnderstory ModifierOpen canopy Mid-development East Gulf Coastal Plain Interior Upland Longleaf Pine Woodland - Scrub/Shrub109911Post replacement Early-development Southern Coastal Plain Oak Dome and Hammock014309921Open canopy Mid-development Southern Coastal Plain Oak Dome and Hammock110811Post replacement Early-development Southern Piedmont Dry Oak-(Pine) Forest - Mixed Modifier015310821Open canopy Mid-development Southern Piedmont Dry Oak-(Pine) Forest - Mixed Modifier112411Post replacement Early-development Florida Peninsula Inland Scrub05212421Closed canopy Mid-development Florida Peninsula Inland Scrub114011Post replacement Early-development Mississippi River Riparian Forest019214021Closed canopy Mid-development Mississippi River Riparian Forest114121Post replacement Mid-development East Gulf Coastal Plain Dune and Coastal Grassland2999114121Post replacement Mid-development East Gulf Coastal Plain Dune and Coastal Grassland114131Post replacement Late-development East Gulf Coastal Plain Dune and Coastal Grassland30999114131Post replacement Late-development East Gulf Coastal Plain Dune and Coastal Grassland114211Post replacement Early-development East Gulf Coastal Plain Near-Coast Pine Flatwoods - Scrub/Shrub014314221Open canopy Mid-development East Gulf Coastal Plain Near-Coast Pine Flatwoods - Scrub/Shrub114311Post replacement Early-development Central Florida Pine Flatwoods014314321Open canopy Mid-development Central Florida Pine Flatwoods114411Post replacement Early-development Atlantic Coastal Plain Brownwater Stream Floodplain Forest019214421Closed canopy Mid-development Atlantic Coastal Plain Brownwater Stream Floodplain Forest114611Post replacement Early-development East Gulf Coastal Plain Small Stream and River Floodplain Forest019214621Closed canopy Mid-development East Gulf Coastal Plain Small Stream and River Floodplain Forest114711Post replacement Early-development Southern Coastal Plain Blackwater River Floodplain Forest019214721Closed canopy Mid-development Southern Coastal Plain Blackwater River Floodplain Forest114811Post replacement Early-development East Gulf Coastal Plain Large River Floodplain Forest - Forest Modifier019214821Closed canopy Mid-development East Gulf Coastal Plain Large River Floodplain Forest - Forest Modifier115011Post replacement Early-development Southern Coastal Plain Nonriverine Basin Swamp010215021Closed canopy Mid-development Southern Coastal Plain Nonriverine Basin Swamp115111Post replacement Early-development Atlantic Coastal Plain Northern Basin Swamp and Wet Hardwood Forest010215121Closed canopy Mid-development Atlantic Coastal Plain Northern Basin Swamp and Wet Hardwood Forest11521107215221010215221115311Post replacement Early-development Atlantic Coastal Plain Nonriverine Swamp and Wet Hardwood Forest - OakDominated ModifierPost replacement Early-development Atlantic Coastal Plain Nonriverine Swamp and Wet Hardwood Forest - OakDominated ModifierPost replacement Early-development Atlantic Coastal Plain Southern Tidal Wooded Swamp010215321Closed canopy Mid-development Atlantic Coastal Plain Nonriverine Swamp and Wet Hardwood Forest - OakDominated ModifierClosed canopy Mid-development Atlantic Coastal Plain Nonriverine Swamp and Wet Hardwood Forest - OakDominated ModifierClosed canopy Mid-development Atlantic Coastal Plain Southern Tidal Wooded Swamp115411Post replacement Early-development Atlantic Coastal Plain Small Brownwater River Floodplain Forest019215421Closed canopy Mid-development Atlantic Coastal Plain Small Brownwater River Floodplain Forest115511Post replacement Early-development East Gulf Coastal Plain Near-Coast Pine Flatwoods - Open014315521Open canopy Mid-development East Gulf Coastal Plain Near-Coast Pine Flatwoods - Open115611Post replacement Early-development East Gulf Coastal Plain Near-Coast Pine Flatwoods - Offsite Hardwood014315621Open canopy Mid-development East Gulf Coastal Plain Near-Coast Pine Flatwoods - Offsite Hardwood115811Post replacement Early-development Atlantic Coastal Plain Southern Maritime Forest015315821Open canopy Mid-development Atlantic Coastal Plain Southern Maritime Forest115811Post replacement Early-development Atlantic Coastal Plain Southern Maritime Forest020315821Open canopy Mid-development Atlantic Coastal Plain Southern Maritime Forest116111Post replacement Early-development Atlantic Coastal Plain Northern Maritime Forest011116122Post replacement Mid-development Atlantic Coastal Plain Northern Maritime Forest109111109511115211Open canopy Mid-development Southern Coastal Plain Dry Upland Hardwood Forest9
EndAgeToStateClassName59ToStateClassID316131Post replacement Mid-development Atlantic Coastal Plain Northern Maritime Forest12116211Post replacement Early-development Atlantic Coastal Plain Central Maritime Forest116411Post replacement Early-development Southern Coastal Plain Nonriverine Cypress Dome020316221Open canopy Mid-development Atlantic Coastal Plain Central Maritime Forest015316421116511Open canopy Mid-development Southern Coastal Plain Nonriverine Cypress DomePost replacement Early-development Southern Piedmont Large Floodplain Forest - Forest Modifier019216521Closed canopy Mid-development Southern Piedmont Large Floodplain Forest - Forest Modifier116711Post replacement Early-development East Gulf Coastal Plain Interior Upland LL Pine Woodland - Loblolly Modifier012316721Open canopy Mid-development East Gulf Coastal Plain Interior Upland LL Pine Woodland - Loblolly Modifier116811Post replacement Early-development Southern Piedmont Small Floodplain and Riparian Forest019216821Closed canopy Mid-development Southern Piedmont Small Floodplain and Riparian Forest116911Post replacement Early-development Southern Piedmont Mesic Forest015216921Closed canopy Mid-development Southern Piedmont Mesic Forest117011Post replacement Early-development Southern Piedmont Dry Oak-(Pine) Forest - Loblolly Pine Modifier015317021Open canopy Mid-development Southern Piedmont Dry Oak-(Pine) Forest - Loblolly Pine Modifier117111Post replacement Early-development Atlantic Coastal Plain Xeric River Dune05217121Closed canopy Mid-development Atlantic Coastal Plain Xeric River Dune117211Post replacement Early-development East Gulf Coastal Plain Interior Upland Longleaf Pine Woodland - OffsiteHardwood ModifierPost replacement Early-development Atlantic Coastal Plain Fall-line Sandhills Longleaf Pine Woodland - OffsiteHardwood ModifierPost replacement Early-development Atlantic Coastal Plain Northern Tidal Wooded Swamp012317221014317321010217721Open canopy Mid-development East Gulf Coastal Plain Interior Upland Longleaf Pine Woodland - OffsiteHardwood ModifierOpen canopy Mid-development Atlantic Coastal Plain Fall-line Sandhills Longleaf Pine Woodland - OffsiteHardwood ModifierClosed canopy Mid-development Atlantic Coastal Plain Northern Tidal Wooded Swamp07217821117911Post replacement Early-development Atlantic Coastal Plain Nonriverine Swamp and Wet Hardwood Forest Taxodium/Nyssa ModifierPost replacement Early-development Atlantic Coastal Plain Peatland Pocosin04317921Closed canopy Mid-development Atlantic Coastal Plain Nonriverine Swamp and Wet Hardwood Forest Taxodium/Nyssa ModifierOpen canopy Mid-development Atlantic Coastal Plain Peatland Pocosin118011Post replacement Early-development Southern Coastal Plain Seepage Swamp and Baygall015218021Closed canopy Mid-development Southern Coastal Plain Seepage Swamp and Baygall118111Post replacement Early-development Central Atlantic Coastal Plain Wet Longleaf Pine Savanna and Flatwoods014318121Open canopy Mid-development Central Atlantic Coastal Plain Wet Longleaf Pine Savanna and Flatwoods118311Post replacement Early-development Atlantic Coastal Plain Blackwater Stream Floodplain Forest - Forest Modifier019218321Closed canopy Mid-development Atlantic Coastal Plain Blackwater Stream Floodplain Forest - Forest Modifier118411Post replacement Early-development Atlantic Coastal Plain Streamhead Seepage Swamp, Pocosin, and Baygall08218421Closed canopy Mid-development Atlantic Coastal Plain Streamhead Seepage Swamp, Pocosin, and Baygall118511Post replacement Early-development Atlantic Coastal Plain Small Blackwater River Floodplain Forest019218521Closed canopy Mid-development Atlantic Coastal Plain Small Blackwater River Floodplain Forest118611Post replacement Early-development Atlantic Coastal Plain Dry and Dry-Mesic Oak Forest014318621Open canopy Mid-development Atlantic Coastal Plain Dry and Dry-Mesic Oak Forest118811Post replacement Early-development Atlantic Coastal Plain Mesic Hardwood and Mixed Forest014218821Closed canopy Mid-development Atlantic Coastal Plain Mesic Hardwood and Mixed Forest119011Post replacement Early-development Atlantic Coastal Plain Clay-Based Carolina Bay Forested Wetland010319021Open canopy Mid-development Atlantic Coastal Plain Clay-Based Carolina Bay Forested Wetland119111Post replacement Early-development Atlantic Coastal Plain Southern Wet Pine Savanna and Flatwoods014319121Open canopy Mid-development Atlantic Coastal Plain Southern Wet Pine Savanna and Flatwoods119211014319221119311Post replacement Early-development Atlantic Coastal Plain Fall-line Sandhills Longleaf Pine Woodland - Scrub/ShrubUnderstory ModifierPost replacement Early-development Atlantic Coastal Plain Upland Longleaf Pine Woodland014319321Open canopy Mid-development Atlantic Coastal Plain Fall-line Sandhills Longleaf Pine Woodland - Scrub/ShrubUnderstory ModifierOpen canopy Mid-development Atlantic Coastal Plain Upland Longleaf Pine Woodland119411Post replacement Early-development Southern Coastal Plain Hydric Hammock010219421Closed canopy Mid-development Southern Coastal Plain Hydric Hammock120111Post replacement Early-development Evergreen Plantations or Managed Pine (can include dense successional regrowth)021220121120411Post replacement Early-development Public Evergreen Plantations or Managed Pine (can include dense successionalregrowth)Post replacement Early-development Atlantic Coastal Plain Large Natural Lakeshore021220421121911010321921Closed canopy Mid-development Evergreen Plantations or Managed Pine (can include dense successionalregrowth)Closed canopy Mid-development Public Evergreen Plantations or Managed Pine (can include dense successionalregrowth)Open canopy Mid-development Atlantic Coastal Plain Large Natural Lakeshore122811Post replacement Early-development Floridian Highlands Freshwater Marsh01222821Closed canopy Mid-development Floridian Highlands Freshwater Marsh123111Post replacement Early-development Southern Coastal Plain Herbaceous Seepage Bog010223121Closed canopy Mid-development Southern Coastal Plain Herbaceous Seepage Bog123311Post replacement Early-development East Gulf Coastal Plain Treeless Savanna and Wet Prairie014323321Open canopy Mid-development East Gulf Coastal Plain Treeless Savanna and Wet Prairie125011Post replacement Early-development East Gulf Coastal Plain Tidal Wooded Swamp010225021Closed canopy Mid-development East Gulf Coastal Plain Tidal Wooded Swamp136911Post replacement Early-development Southern Coastal Plain Nonriverine Swamp and Wet Hardwood Forest Taxodium ModifierPost replacement Early-development Southern Coastal Plain Nonriverine Basin Swamp - Okefenokee Bay/GumModifier07236921010237021Closed canopy Mid-development Southern Coastal Plain Nonriverine Swamp and Wet Hardwood Forest Taxodium ModifierClosed canopy Mid-development Southern Coastal Plain Nonriverine Basin Swamp - Okefenokee Bay/GumModifier117311117711117811137011Open canopy Late-development Atlantic Coastal Plain Northern Maritime Forest10
EndAgeToStateClassName10ToStateClassID237121Post replacement Early-development Southern Coastal Plain Nonriverine Basin Swamp - Okefenokee Pine Modifier0137711Post replacement Early-development Southern Coastal Plain Nonriverine Basin Swamp - Okefenokee Nuphar Modifier137811Post replacement Early-development Southern Coastal Plain Nonriverine Basin Swamp - Okefenokee Clethra Modifier0102377210102378211559205731Closed canopy Mid-development Southern Coastal Pla
The predictive variance estimate was obtained by holding the GCM weights constant based on the BMA analysis. The . When the transition had a one to one relationship with an ecological system, that system label was applied to the polygon. In some cases, there was a one-to-many relationship of a SLAMM class to ecological
