Shrubland vegetation topographic facets of Southern California
Hollander, Allan; Underwood, Emma (2021), Shrubland vegetation topographic facets of Southern California, Dryad, Dataset, https://doi.org/10.25338/B8JW59
To approximate the distribution of shrubland species based on their postfire reproductive strategy (resprouter, seeder, and facultative seeder) across Southern California, we created a raster layer subdividing the landscape into a number of different facet classes. This raster dataset is at 30 meters pixel resolution and contains 12 different landscape facet classes based on vegetation and physiography. Specifically, the facets included several different vegetation types based on the California Wildlife Habitat Relations (WHR) classification (three shrubland categories, annual grasslands, valley-foothill riparian woodland, and ‘other’ vegetation types) which were intersected with aspect (two classes: north or south facing) and topography (summit, ridges, slopes, valleys, flats, and depressions). The combination of factors is intended to capture warmer, more exposed vegetation types dominated by seeder species (occurring on south-facing slopes, summits and ridges) versus cooler, less exposed vegetation types associated with resprouter species (occurring on north-facing slopes, valleys, depressions, and flats).
The dataset is a key input into a tool developed for resource managers to aid in the prioritization of restoration activities in shrublands postfire. The tool is available at https://github.com/adhollander/postfire and described in the following technical guide:
Underwood, Emma C., and Allan D. Hollander. 2019. “Post-Fire Restoration Prioritization for Chaparral Shrublands Technical Guide.” https://github.com/adhollander/postfire/blob/master/Postfire_Restoration_Priorization_Tool_Technical_Guide.pdf
The following are the GIS processing workflow steps used to create this dataset. A diagram illustrating this workflow is in the attached file collection (SoCal_Veg_Topo_Facets_Workflow.png).
1) Compile GIS layers. There were two input layers to the GIS workflow, a 30 meter digital elevation model for California (dem30) and a vegetation raster layer of the state from the California Department of Forestry and Fire Protection (fveg15). The 30 meter DEM was downloaded from the USGS National Map (https://www.usgs.gov/core-science-systems/national-geospatial-program/national-map). The vegetation data is the FVEG dataset published in 2015 by the California Department of Forestry and Fire Protection's Fire and Resource Assessment Program (https://frap.fire.ca.gov/media/10894/fveg15_1.zip). This is a 30 meter raster representation of statewide vegetation using the California Wildlife Habitat Relationships vegetation classification system (https://wildlife.ca.gov/Data/CWHR).
2) Import data into GIS. Both data layers were imported into GRASS 7 for further processing, using a mask of the Southern California study region (encompassing the Angeles, Cleveland, Los Padres, and San Bernardino National Forests) to filter processing to the study footprint.
3) Calculate aspect for elevation model. Using the command r.slope.aspect, we generated a raster layer (aspect) giving the topographic aspect (0-360 degrees) of slopes across the study region.
4) Generate north-south aspect layer. Using the command r.mapcalc, we subdivided the aspect layer into north and south-facing slopes through creating a raster layer (nsaspect) with two categories for north and south.
5) Generate geomorphons for study region. The geomorphon raster layer derives from the dem30m surface and classifies the landscape into 10 discrete landform types, examples being ridges, slopes, hollows, and valleys. The algorithm for geomorphon classification uses a pattern recognition approach based on line of sight analysis (Jasiewisc and Stepinski 2013) and was generated using the r.geomorphons extension for GRASS 7.
6) Merge geomorphons with north-south aspect layer. In this step we combined the north-south aspect layer with the geomorphons layer to create a layer entitled nsgeomorphon2a. In so doing we grouped the geomorphon types spurs, slopes, and hollows into a single “slope” category and assigned these to north-facing slopes and south-facing slopes depending upon the value of the north-south aspect layer.
7) Regroup merged layer into three groupings. In this step we took the merged nsgeomorphon2a layer and assigned the classes in it to three different physiographic groups, namely 1) flats 2) valleys, depressions, and north-facing slopes/spurs/hollows/footslopes/shoulders and 3) summits and ridges and south-facing slopes/spurs/hollows/footslopes/shoulders. This grouped layer was named nsgeomorphon2d.
8) Reclass vegetation layer to main habitat types. The vegetation layer fveg15 contains information about many details of the vegetation, including canopy size, canopy cover, and main habitat type. This reclass step extracts the main habitat type into a separate raster named fveg15whr.
9) Combine vegetation layer with physiography layer. Using the command r.cross, we combined the layers fveg15whr and nsgeomorphon2d into a new layer nsgeoxfvegwhr with a separate category for each combination of the raster values from the two input layers.
10) Reclass combined layer into small set of groupings. Taking the nsgeoxfvegwhr layer, we recategorized the 196 combinations of raster values into a set of 12 different combinations using the command r.reclass. This layer is named nsgeoxfvegnbclasses. The 12 different classes generated as an output are the following, with their raster values paired with their classes:
0 Annual grassland: south-facing slopes; summits; ridges
1 Annual grassland: north-facing slopes; valleys; depressions; flats
2 Chamise-redshanks chaparral: south-facing slopes; summits; ridges
3 Chamise-redshanks chaparral: north-facing slopes; valleys; depressions; flats
4 Mixed or montane chaparral: south-facing slopes; summits; ridges
5 Mixed or montane chaparral: north-facing slopes; valleys; depressions; flats
6 Valley-foothill riparian: south-facing slopes; summits; ridges
7 Valley-foothill riparian: north-facing slopes; valleys; depressions; flats
8 Coastal scrub: south-facing slopes; summits; ridges
9 Coastal scrub: north-facing slopes; valleys; depressions; flats
10 Other: south-facing slopes; summits; ridges
11 Other: north-facing slopes; valleys; depressions; flats
11) Export dataset. Using the command r.out.gdal, we exported the nsgeoxfvegnbclasses layer as the raster geotiff file SoCal_Veg_Topo_Facets.tif.
The GRASS commands used for these 11 steps are below:
r.in.gdal input="/home/adh/CARangelands/Vegetation/fveg15_11.tif" output="fveg15" memory=300 offset=0
r.proj input="dem1sec_calif" location="CAllnad83" mapset="statewide" output="dem30m" method="bilinear" memory=300 resolution=30
r.slope.aspect elevation=dem30m@statewide slope=slope aspect=aspect
r.mapcalc 'nsaspect = if(aspect <= 180, 1, 2)'
r.geomorphon --overwrite dem=dem30m@statewide forms=SoCalgeomorphons search=11 skip=4 flat=1 dist=0
r.mapcalc --overwrite 'nsgeomorphon = if((SoCalgeomorphons@socalNF == 5 ||| SoCalgeomorphons@socalNF == 6 ||| SoCalgeomorphons@socalNF == 7) &&& nsaspect == 1, 11, if(((SoCalgeomorphons@socalNF == 5 ||| SoCalgeomorphons@socalNF == 6 ||| SoCalgeomorphons@socalNF == 7) &&& nsaspect == 2), 12, SoCalgeomorphons@socalNF))'
r.reclass input=nsgeomorphon2a@socalNF output=nsgeomorphon2d rules=/home/adh/SantaClaraRiver/PostfireRestoration/jupyter/datasets/nsgeomorphon-reclass2d.lut
r.reclass input="fveg15@statewide" output="fveg15whr" rules="/home/adh/CARangelands/Vegetation/fveg15whr.lut"
r.cross --overwrite input=fveg15whr@statewide,nsgeomorphon2d@socalNF output=nsgeoxfvegwhr
r.reclass --overwrite input=nsgeoxfvegwhr@socalNF output=nsgeoxfvegnbclasses rules=/home/adh/SantaClaraRiver/PostfireRestoration/datasets/fvegwhrtonbclasses.lut
r.out.gdal --overwrite input=nsgeoxfvegnbclasses@socalNF output=SoCal_Veg_Topo_Facets.tif format=GTiff type=Byte createopt=COMPRESS=DEFLATE
The primary file in this dataset is named SoCal_Veg_Topo_Facets.tif. It is a raster GIS file in geotiff format with 30 meter resolution. The projection for it is California Albers, NAD83 datum (EPSG:3310). Dimensions of the raster are 22748 rows x 15223 columns, and the bounding box x,y coordinates in the California Albers projection are -235440.000, -156780.000 (upper left) and 447000.000, -613470.000 (lower right). The raster is encoded as bytes with a minimum value of 0, a maximum value of 11, and a nodata value of 255.
The values in the raster represent individual facet types that are numbered 0 to 11. The additional file SoCal_Veg_Topo_Facets.tsv is a tab-delimited lookup table giving the class names associated with each raster value.
The study region of the dataset consists of the area of 385 HUC12 USGS watersheds encompassing the four National Forests in Southern California (Angeles, Cleveland, Los Padres, and San Bernardino). A map of the study area is provided with the attached file collection (SoCal_Veg_Topo_Facets_Studyregion.png).
The original use of this dataset is as an input to a geospatial software tool to assist in the prioritization of areas for shrubland restoration postfire (see https://github.com/adhollander/postfire). Other intended users of this dataset might include resource managers, researchers who are carrying out biogeographic studies, and people needing to make estimates of patterns of shrublands and the functional groups of shrubland species across this landscape (e.g., estimating biomass or biomass recovery postfire).
This dataset is made available under a CC0 license.
U.S. Forest Service
National Fish and Wildlife Foundation