oss.texture.r: Use particle size analysis SpatRaster data to create a...

In newdale/onsoilsurvey: Making PDSM in Ontario Better

Use particle size analysis SpatRaster data to create a texture class SpatRaster based on the the Canadian System of Soil Classification or USDA. If sand fraction data is provided, modifiers (coarse, fine, very fine) will be assigned to the sands, loamy sands and sandy loams.

Description

Use particle size analysis SpatRaster data to create a texture class SpatRaster based on the the Canadian System of Soil Classification or USDA. If sand fraction data is provided, modifiers (coarse, fine, very fine) will be assigned to the sands, loamy sands and sandy loams.

Usage

oss.texture.r(
  sand,
  silt,
  clay,
  vcs = NULL,
  cs = NULL,
  ms = NULL,
  fs = NULL,
  vfs = NULL,
  tri = "CSSC"
)

Arguments

sand	RasterLayer, or SpatRaster of sand (0.5 - 2 mm) content either in decimal or percentage (e.g. 0.25 or 25)
silt	RasterLayer, or SpatRaster of silt (0.002 - 0.05 mm) content either in decimal or percentage (e.g. 0.25 or 25)
clay	RasterLayer, or SpatRaster of clay (<0.002 mm) content either in decimal or percentage (e.g. 0.25 or 25)
vcs	RasterLayer, or SpatRaster of very coarse sand (1 - 2 mm) content expressed as percentage of the sum of the sand fractions or portion of total sand
cs	RasterLayer, or SpatRaster of coarse sand (0.5 - 1 mm) content expressed as percentage of the sum of the sand fractions or portion of total sand
ms	RasterLayer, or SpatRaster of medium sand (0.25 - 0.50 mm) content expressed as percentage of the sum of the sand fractions or portion of total sand
fs	RasterLayer, or SpatRaster of fine sand (0.10 - 0.25 mm) content expressed as percentage of the sum of the sand fractions or portion of total sand
vfs	RasterLayer, or SpatRaster of very fine sand (0.05 - 0.10 mm) content expressed as percentage of the sum of the sand fractions or portion of total sand
tri	character, current choices are "CSSC" for Canadian System of Soil Classification (default), or "USDA" for United States Department of Agriculture

Value

When the input data is SpatRaster with only one layer, returns a list of two objects: texture_raster SpatRaster of soil texture class as per the selected soil classification system

legend data frame containing the factor levels for the soil texture class raster

When the input data is a SpatRaster with more than one layer (i.e., stack), the function will return a list of lists. Each outer list item will represent the list of outputs for each layer of the SpatRaster. The inner list will contain the 'texture_raster' and 'legend' objects as described above.

Examples

# create sample data which includes all combinations of sand, silt and clay
dat<- data.frame(expand.grid(sand=seq(0,100,1), silt=seq(0,100,1), clay=seq(0,100,1)))
dat$sum<- dat$sand+dat$silt+dat$clay
dat<- dat[dat$sum==100,]
dat$x<- dat$sand
dat$y<- dat$clay
dat<- dat[,c(5,6,1,2,3)]

sand<- terra::rast(dat[,c(1:3)], type="xyz")
silt<- terra::rast(dat[,c(1,2,4)], type="xyz")
clay<- terra::rast(dat[,c(1,2,5)], type="xyz")

# Create a texture class raster without sand fractions
tex<- oss.texture.r(sand,silt,clay)

# And we can visualize
texture.map<- terra::as.factor(tex[[1]])
rat<- data.frame(terra::levels(texture.map))
rat[["Texture"]]<- tex[[2]]$Class[match(rat$ID,tex$legend$Code)]
rat<- rat[,c(1,3)]
levels(texture.map)<- rat
coltb<- data.frame(value=rat$ID, col=colorRampPalette(RColorBrewer::brewer.pal(12, "Set3"))(nrow(rat)))
terra::coltab(texture.map)<- coltb
terra::plot(texture.map)

# Create a texture class raster with sand fractions

# generate random values for sand, silt and clay, normalize to sum 100, assign to SpatRaster
sand<- sample(seq(0,100,1),10000,replace=TRUE)
silt<- sample(seq(0,100,1),10000,replace=TRUE)
clay<- sample(seq(0,100,1),10000,replace=TRUE)

vals<- data.frame(sand=(sand/(sand+silt+clay))*100,
                  silt=(silt/(sand+silt+clay))*100,
                  clay=(clay/(sand+silt+clay))*100)

sand<- terra::rast(ncol=100, nrow=100, vals=vals$sand, xmin=0, xmax=100, ymin=0, ymax=100)
silt<- terra::rast(ncol=100, nrow=100, vals=vals$silt, xmin=0, xmax=100, ymin=0, ymax=100)
clay<- terra::rast(ncol=100, nrow=100, vals=vals$clay, xmin=0, xmax=100, ymin=0, ymax=100)

# generate random values for sand fractions, normalize to sum 100, assing to SpatRaster
vfs<- sample(seq(0,100,1),10000,replace=TRUE)
fs<- sample(seq(0,100,1),10000,replace=TRUE)
ms<- sample(seq(0,100,1),10000,replace=TRUE)
cs<- sample(seq(0,100,1),10000,replace=TRUE)
vcs<- sample(seq(0,100,1),10000,replace=TRUE)

vals<- data.frame(vfs=(vfs/(vfs+fs+ms+cs+vcs))*100,
                  fs=(fs/(vfs+fs+ms+cs+vcs))*100,
                  ms=(ms/(vfs+fs+ms+cs+vcs))*100,
                  cs=(cs/(vfs+fs+ms+cs+vcs))*100,
                  vcs=(vcs/(vfs+fs+ms+cs+vcs))*100)

vfs<- terra::rast(ncol=100, nrow=100, vals=vals$vfs, xmin=0, xmax=100, ymin=0, ymax=100)
fs<- terra::rast(ncol=100, nrow=100, vals=vals$fs, xmin=0, xmax=100, ymin=0, ymax=100)
ms<- terra::rast(ncol=100, nrow=100, vals=vals$ms, xmin=0, xmax=100, ymin=0, ymax=100)
cs<- terra::rast(ncol=100, nrow=100, vals=vals$cs, xmin=0, xmax=100, ymin=0, ymax=100)
vcs<- terra::rast(ncol=100, nrow=100, vals=vals$vcs, xmin=0, xmax=100, ymin=0, ymax=100)

# Create a texture class raster without sand fractions
tex_fractions<- oss.texture.r(sand,silt,clay, vcs, cs, ms, fs, vfs)

# And we can visualize
texture.map<- terra::as.factor(tex_fractions[[1]])
rat<- data.frame(terra::levels(texture.map))
rat[["Texture"]]<- tex_fractions[[2]]$Class[match(rat$ID,tex_fractions$legend$Code)]
rat<- rat[,c(1,3)]
levels(texture.map)<- rat
coltb<- data.frame(value=rat$ID, col=colorRampPalette(RColorBrewer::brewer.pal(12, "Set3"))(nrow(rat)))
terra::coltab(texture.map)<- coltb
terra::plot(texture.map)

newdale/onsoilsurvey documentation built on Jan. 5, 2024, 1:35 a.m.