Skip to contents

Mathematical operations on two or more GRasters

Source: R/06_GRaster_functions_acrossLayers.r
functions.Rd

These functions can be applied to a "stack" of GRasters with two or more layers. They return a single-layered GRaster. If you want to summarize across cells in a raster (e.g., calculate the mean value of all cells on a raster), use global(). Options include:

  • Numeration: count() (number of non-NA cells), sum().

  • Central tendency: mean(), mmode() (mode), median().

  • Extremes: min(), max(), which.min() (index of raster with the minimum value), which.max() (index of the raster with the maximum value)

  • Dispersion: range(), stdev() (standard deviation), var() (sample variance), varpop() (population variance), nunique() (number of unique values), quantile() (use argument probs), skewness(), and kurtosis().

  • Regression: Assuming we calculate a linear regression for each set of cells through all values of the cells, we can calculate its slope(), intercept(), r2(), and tvalue().

Usage

# S4 method for class 'GRaster'
mean(x, na.rm = FALSE)

# S4 method for class 'GRaster'
mmode(x, na.rm = FALSE)

# S4 method for class 'GRaster'
median(x, na.rm = FALSE)

# S4 method for class 'GRaster'
count(x)

# S4 method for class 'GRaster'
sum(x, na.rm = FALSE)

# S4 method for class 'GRaster'
min(x, na.rm = FALSE)

# S4 method for class 'GRaster'
max(x, na.rm = FALSE)

# S4 method for class 'GRaster'
which.min(x)

# S4 method for class 'GRaster'
which.max(x)

# S4 method for class 'numeric'
sdpop(x, na.rm = FALSE)

# S4 method for class 'GRaster'
varpop(x, na.rm = FALSE)

# S4 method for class 'numeric'
varpop(x, na.rm = FALSE)

# S4 method for class 'GRaster'
stdev(x, pop = TRUE, na.rm = FALSE)

# S4 method for class 'GRaster'
var(x, na.rm = FALSE)

# S4 method for class 'GRaster'
nunique(x, na.rm = FALSE)

# S4 method for class 'GRaster'
skewness(x, na.rm = FALSE)

# S4 method for class 'GRaster'
kurtosis(x, na.rm = FALSE)

# S4 method for class 'GRaster'
slope(x, na.rm = FALSE)

# S4 method for class 'GRaster'
intercept(x, na.rm = FALSE)

# S4 method for class 'GRaster'
r2(x, na.rm = FALSE)

# S4 method for class 'GRaster'
tvalue(x, na.rm = FALSE)

# S4 method for class 'GRaster'
range(x, na.rm = FALSE)

# S4 method for class 'GRaster'
quantile(x, prob, na.rm = FALSE)

Arguments

x

A GRaster. Typically, this raster will have two or more layers. Values will be calculated within cells across rasters.

na.rm

Logical: If FALSE (default), of one cell value has an NA, the result will be NA. If TRUE, NAs are ignored.

pop

Logical (for stdev()): If TRUE (default), calculate the population standard deviation across layers. If FALSE, calculate the sample standard deviation.

prob

Numeric: Quantile to calculate. Used for quantile().

Value

A GRaster.

Examples

if (grassStarted()) {

# Setup
library(sf)
library(terra)

# Example data
madElev <- fastData("madElev")

# Convert a SpatRaster to a GRaster
elev <- fast(madElev)
elevs <- c(elev, elev, log10(elev) - 1, sqrt(elev))
names(elevs) <- c("elev1", "elev2", "log_elev", "sqrt_elev")

elev
elevs

# do some math
elev + 100
elev - 100
elev * 100
elev / 100
elev ^ 2
elev %/% 100 # divide then round down
elev %% 100 # modulus

100 + elev
100 %/% elev
100 %% elev

elevs + 100
100 + elevs

# math with logicals
elev + TRUE
elev - TRUE
elev * TRUE
elev / TRUE
elev ^ TRUE
elev %/% TRUE # divide then round down
elev %% TRUE # modulus

elevs + TRUE
TRUE + elevs

# Raster interacting with raster(s):
elev + elev
elev - elev
elev * elev
elev / elev
elev ^ log(elev)
elev %/% sqrt(elev) # divide then round down
elev %% sqrt(elev) # modulus

elevs + elev
elev * elevs

# sign
abs(-1 * elev)
abs(elevs)

# powers
sqrt(elevs)

# trigonometry
sin(elev)
cos(elev)
tan(elev)

asin(elev)
acos(elev)
atan(elev)

atan(elevs)
atan2(elev, elev^1.2)
atan2(elevs, elev^1.2)
atan2(elev, elevs^1.2)
atan2(elevs, elevs^1.2)

# logarithms
exp(elev)
log(elev)
ln(elev)
log2(elev)
log1p(elev)
log10(elev)
log10p(elev)
log(elev, 3)

log(elevs)

# rounding
round(elev + 0.5)
floor(elev + 0.5)
ceiling(elev + 0.5)
trunc(elev + 0.5)

# comparison
elev < 100
elev <= 100
elev == 100
elev != 100
elev > 100
elev >= 100

elev + 100 < 2 * elev

elevs > 10
10 > elevs

# logic
elev < 10 | elev > 200
elev < 10 | cos(elev) > 0.9

elev < 10 | TRUE
TRUE | elev > 200

elev < 10 | FALSE
FALSE | elev > 200

elev < 10 & cos(elev) > 0.9

elev < 10 & TRUE
TRUE & elev > 200

elev < 10 & FALSE
FALSE & elev > 200

# Mathematical functions on GRasters with >= 2 layers:
mean(elevs)
mmode(elevs)
median(elevs)
nunique(elevs)

sum(elevs)
count(elevs)
min(elevs)
max(elevs)
range(elevs)
skewness(elevs)
kurtosis(elevs)

which.min(elevs)
which.max(elevs)

slope(elevs)
intercept(elevs)
r2(elevs)
tvalue(elevs)

stdev(elevs)
stdev(elevs, pop = FALSE)
var(elevs)
varpop(elevs)

# Note: To get quantiles for each layer, use
# global(x, "quantile", probs = 0.2).
quantile(elevs, 0.1)

}