[SOLVED] For the same longitude or latitude raster data use the function based on terra

For the same longitude or latitude raster data use the function based on terra

Assuming that I need to calculate the variance of the raster data for the same longitude, Is using the sapp function the right way to go? But I failed.

I can now use the more complex method of converting to data.frame and grouping to calculate the variance:

library(terra)
library(tidyverse)

r <- rast(nrows = 5, ncols = 10, vals = c(1:50))
plot(r)


r |> 
  terra::as.data.frame(xy= TRUE) |> 
  group_by(x) |> 
  summarise(lonsd = sd(lyr.1))

Is there a more straightforward function in terra? Please tell me if you know.

There are also more complex scenarios where this needs to be done on multiple layers of raster data at the same time, how should this be done?

Solution

You can use aggregate to compute something for each column (or row) in a raster

Example data

library(terra)
set.seed(1)
r <- rast(nrows = 5, ncols = 10, vals = sample(100), nlyr=2)

Solution

a <- aggregate(r, c(nrow(r), 1), var, na.rm=TRUE)
data.frame(lon=xFromCol(r), values(a))
#    lon  lyr.1  lyr.2
#1  -162  468.2 1364.3
#2  -126 1007.3  493.0
#3   -90  557.5  822.2
#4   -54  358.3 1089.7
#5   -18 1224.3  665.7
#6    18 1125.3  260.3
#7    54 1180.3  207.2
#8    90  631.3  433.0
#9   126 1564.7 1669.7
#10  162  881.7  752.3

a <- as.data.frame(r, xy=TRUE)
aggregate(a[,3:4], a["x"], var, na.rm=TRUE)