Data from: Developing spatially explicit and stochastic measures of ecological departure
Data files
Apr 15, 2024 version files 180.68 MB
-
README.md
11.06 KB
-
Results.zip
19.19 MB
-
Spatial_NRV_Development-TJR_IJWF.zip
161.47 MB
-
SpatialED_scripts_IJWF.zip
9.37 KB
Abstract
Background: Ecological departure is a metric applied to mapped ecological systems measuring dissimilarity between the distributions of observed and expected proportions of non-stochastic reference vegetation classes within an area.
Aims: We created spatially explicit measures of ecological departure incorporating stochasticity for each ecological system and all ecological systems from a central Nevada USA landscape.
Methods: Spatially explicit ecological departures were estimated from a radius from each pixel governed by a distance-decay function within a moving window. Variability was introduced by simulating replicate climate time series for each spatial reference condition and calculating departure per replicate.
Key results: Single system spatial ecological departure was highly and extensively departed, except for one area of low-elevation groundwater-dependent systems. Variance of spatial ecological departure was extensively low, except in areas of lower ecological departure, despite vegetation differences among replicates. The multiple-system ecological departure exhibited lower ecological departure.
Conclusions: Spatial ecological departure was warranted for efficient land management as results were concordant between non-spatial and spatial metrics; however, rapid coding languages will be required.
README
README: Developing Spatially Explicit and Stochastic Measures of Ecological Departure
https://doi.org/10.5061/dryad.3n5tb2rrm
This study created the new metric of spatially explicit ecological departure where stochasticity is modeled through spatial replicates of the reference condition of vegetation. To create the metric, the study (a) modeled alternative (i.e., stochastic) future spatial reference vegetation of each ecological system using state-and-transition simulation models applied to a central Nevada USA landscape and (b) created R scripts to perform spatially explicit operations (i.e., focused on each pixel of the raster vegetation maps) to calculate the new metrics for single systems and all systems within a moving window. This archive of the dataset contains the fully populated spatial state-and-transition simulation database in the ST-Sim/Syncrosim package and software that produced the reference vegetation (raster results provided) at year 700 and all the R scripts to estimate the spatially explicit and stochastic ecological departure.
Description
- Data Preparation Remote sensing vegetation layers: All work was based on mapped vegetation in raster format. We repurposed vegetation layers mapped in 2014 by contractor Spatial Solutions Inc. for The Nature Conservancy and Barrick Gold Corp. Raster vegetation layers were in geotiff format representing the 76,828-ha basin and range valley in central Nevada called the Northern Newark Valley that extends from the southern Ruby Range on the eastern side, across the northern Newark Valley and Huntington Valley and up the slopes of the Diamond Range on the west side (39o53’04.38” N, 115 o40’39.19” W). The general remote sensing methodology is explained in Provencher et al. (2021; cited in manuscript) applied to a different landscape. The original multispectral and pan-sharpened imagery was obtained from the Spot 6/7 satellite at 1.5m resolution (included) that was resampled to 60m resolution for ease of computing in 2015. This archived dataset includes the resampled 60-m layer for ecological systems and vegetation classes.
Spatial processing: All raster data were processed either with R scripts or ESRI’s ARC Pro. Reclassification of roads to NO-DATA and non-reference vegetation classes to the closest reference vegetation class per ecological system in the original 60m vegetation raster using ARC Pro. This process created the initial conditions of vegetation not containing post-European vegetation classes (see below for file names). These vegetation layers were used in the reference condition simulations (see below)
- Structure of prepared data
├── Results: Results of simulations are in two formats and also found in the zipped file "Results.zip".
│ ├── 8-DigitRasterYr700_2022: The folder contains 21 geotiff rasters of vegetation (system and vegetation class are combined in one vegetation code, which is found in the papers supplemental file S1 - see paper and in the "State Attribute Values" menu under "Advanced" tab of the ST-Sim database). These rasters at year 700 were used to calculate spatially explicit and stochastic ecological departure with the R scripts. ST-Sim automatically labels the files with a generic code: for example, "scn3855.sa_SYSxCLASS-8-Digit-Code.it1.ts700," which translates to "scenario 3855" for scn3855 (same code in ST-Sim database), "sa_" is state attribute, SYSxCLASS-8-Digit-Code means the state attribute code is from the attribute type "SYSxCLASS-8-Digit-Code" defined in ST-Sim's "State Attribute Values" menu, .it2. is iteration 2, and ts700 is timestep 700.
│ └── State Classes_NRV-700yrs.xlsx: The non-spatial tabular (Excel format) area per vegetation class per ecological system and per iteration (= Monte-Carlo replicates) at year 700. These non-spatial data were not used for our analyses but can be used to calculate non-spatial ecological departure as describe in the paper.
├── Spatial NRV Development-TJR_IJWF: Contains the entire simulation database. Syncrosim/ST-Sim is freeware downloaded from www.apexrms.com after registration. ApexRMS provides online documentation and online or in-person training. The ST-Sim state-and-transition simulation package added to the Syncrosim simulation platform (version 2.3.12, updated to 2.4.18 reported for the archive) was used to create the reference condition by simulating the initial condition rasters for 700 years.
│ ├── GIS
│ │ ├── TJRDEM_60m.tif: USGS Digital Elevation Modelsexpressed as percent slope (60m resolution to affect fire spread)
│ │ ├── TJR_LO_NRVinput.tif: Land ownership at 60m resolution (for initial conditions but not used)
│ │ ├── TJR_ReferenceOnly_CLA_2022.tif: Current vegetation classes for initial conditions at 60m resolution. This raster was obtained from remote sensing.
│ │ └── TJR_ReferenceOnly_SYS_2022.tif: Ecological systems for initial conditions at 60m resolution. This raster was obtained from remote sensing.
│ ├── Results
│ │ ├── 8-DigitRasterYr700_2022: Same as above under "Results"
│ │ └── State Classes_NRV-700yrs.xlsx: Same as above under "Results"
│ ├── Results.zip: Zipped folder of results described above
│ ├── Spatial_NRV_TJR.ssim: Syncrosim library (version 2.3.12, updated to 2.4.18) that runs the ST-Sim simulation package. After installation of the appropriate version of Syncrosim, opening this library will give the user full access to the model and its menus. This simulation library must be accompanied by the unaltered input folder "Spatial_NRV_TJR.ssim.input" (see belwo) and the unaltered output folder "Spatial_NRV_TJR.ssim.output" (see below). "Spatial_NRV_TJR.ssim" was fully loaded with all rasters and menus and ready to run in parallel processing mode. A powerful server will be required (we use AWS Cloud servers) with at least 24 cores and 356 GB of memory. The limiting factor is memory; therefore, combined serial and parallel may be needed with restricted memory availability. Variability among replicates was provided by the "External Variables" menu and translated to effects on ecological processes in the "Distribution" and "Transition Multipliers" menu of ST-Sim (these require advanced training to understand).
│ ├── Spatial_NRV_TJR.ssim.input: Folder is managed by Syncrosim library "Spatial_NRV_TJR.ssim" and must not be changed by users; otherwise, the simulation will not work at all or not properly.
│ │ ├── Project-256
│ │ │ ├── colormap_map1_stsim_avgtp-572_AUTOGEN.bin: Color identity of average transition probablities in mapped results
│ │ │ ├── colormap_stsim_sc.txt: Color identity of state classes for results in menu "Definitions"
│ │ │ ├── colormap_stsim_str.txt: Color identity of transition groups for results in menu "Definitions"
│ │ │ ├── colormap_stsim_tg-1012.txt, etc. Color identity code of transition groups for results in menu "Definitions"
│ │ ├── Scenario-1477: Folder containing initial conditions. The number of the scenario was automatically assigned by Syncrosim.
│ │ │ └── stsim_InitialConditionsSpatial: Contains spatial initial condition rasters of ecologicla systems, state classes, and land ownership in geotiff format
│ │ │ ├── TJR_LO_NRVinput.tif: Same as described above
│ │ │ ├── TJR_ReferenceOnly_CLA_2022.tif: Same as described above
│ │ │ └── TJR_ReferenceOnly_SYS_2022.tif: Same as described above
│ │ ├── Scenario-3817: Folder containing the USGS 60-m Digitial Elevation Model. The number of the scenario was automatically assigned by Syncrosim.
│ │ │ └── stsim_DigitalElevationModel
│ │ │ └── TJRDEM_60m.tif: Same as above
│ │ └── Scenario-3855: Similar to Senario-3817 but generated by Syncrosim/ST-Sim to supply a full suite of rasters not supplied by user.
│ │ ├── stsim_DigitalElevationModel: Same as above
│ │ │ └── TJRDEM_60m.tif: Same as above
│ │ └── stsim_InitialConditionsSpatial: Folder generated by Syncrosim/ST-Sim containing spatial initial condition rasters, including one self-generated age raster not supplied by user.
│ │ ├── It0000-Ts0000-age.tif: Geotiff raster genenrated by the simulation software to assign a random age to each state class per ecological system. Users did not supply this raster.
│ │ ├── TJR_LO_NRVinput.tif: Same as above
│ │ ├── TJR_ReferenceOnly_CLA_2022.tif: Same as above
│ │ └── TJR_ReferenceOnly_SYS_2022.tif: Same as above
│ └── Spatial_NRV_TJR.ssim.output: Folder containing results
│ └── Scenario-3855: Folder containing spatial results. Scenario number was generated by Syncrosim.
│ ├── stsim_OutputSpatialAverageTransitionProbability: Transition (e.g., fire) frequency expressed as an average probability of occurrences per transition for each raster pixel across the number of years and replicates. These results were not used to estimate ecological departure and were used to evaluate model performance.
│ └── stsim_OutputSpatialStateAttribute: Simulated replicated reference condition vegetation rasters where each pixel was codes with an 8-digit unique code corresponding to an ecological system (first 5 digits) and the state class (last 3 digits). These reference rasters were the only data needed to calculate ecological departure in addition to the observed current map of ecological systems and state classes per system.
│
├── SpatialED_scripts_IJWF: Folder containing all R scripts to estimate non-spatial and spatially explicit and stochastic ecological departure. Scripts written with R Version 4.2.2 (also available at https://github.com/sbyer-tnc/IJWF-Spatial-ED-Publication). There are three scripts.
│ ├── nonspatial_ed.R: Estimates the traditional non-spatial ecological departure.
│ ├── spatial_ed_multisystem.R: Estimates the mean and variance of spatially explicit and stochastic ecological of all ecological systems within a moving window in the landscape. The R scripts estimate ecological departure per pixel with a moving window.
│ └── spatial_ed_singlesystem.R: Estimates the mean and variance of spatially explicit and stochastic ecological of each ecological systems within a moving window in the landscape. The R scripts estimate ecological departure per pixel with a moving window.
Methods
The goals of the study were to develop a new spatially explicit and stochastic metric of ecological departure from replicated reference conditions applied to single and all ecological systems within a search radius of each map pixel. Remote-sensed vegetation rasters of ecological systems and their vegetation classes originally created by The Nature Conservancy in Nevada for private industry were repurposed from a 2014 project.
The methodology involved four steps: (a) estimating the distance-adjusted proportions of observed vegetation classes per ecological system within a radius of each raster pixel, (b) simulating replicates of the vegetation reference condition for 700 years and estimating the distance-adjusted proportion of expected vegetation classes per ecological system within the same radius of each raster pixel and per replicate on year 700, (c) estimate the mean and variance of ED by applying Eq. 1 to each map pixel of the observed and expected rasters per replicate; and (d) recalculating the observed and expected distance-adjusted proportions among all ecological systems per map pixel and per replicate and application of Eq. 1.