Merge multiple rasters into one final raster by averaging for all points

user_3374236 0
Viewed 357

I'm working on many rasters to get one final raster to represent the average level.

I've 365 continuous rasters: each raster represents the daily tempeature in a region. One raster to be generated: to average daily values of 365 rasters to reflect the temperature level within a year.

I'm not sure if the code below is correct or not?

temperature_average<- calc(h, mean) ## h is a stack of 365 rasters (daily temperature) in 2020

Is there a way to reaching this purpose? Much appreciated for any guidance.

raster r
1 Answers

there are different ways for doing raster algebra, with packages raster and terra (which is the recommended package for raster calculation and manipulation). See the benchmark below

library(raster)
library(terra)
#toy data
r <- raster(ncol=36, nrow=18, vals=1:(18*36))
s <- stack(r,r*0.5,r/3,sin(r))
#convert raster to SpatRaster from terra package
t <- rast(s)

#raster::calc
temperature_average<- calc(s, mean)
#raster::mean
temperature_average2 <- mean(s)
#terra::app
temperature_average3 <- app(t,mean)
#terra::mean
temperature_average4 <- mean(t)
#see which is the fastes
library(microbenchmark)
microbenchmark("calc"=temperature_average<- calc(s, mean),
               "raster::mean"=temperature_average2 <- mean(s),
               "app"=temperature_average3 <- app(t,mean),
               "terra::mean"=temperature_average5 <- mean(t))

Unit: milliseconds
         expr    min      lq     mean  median      uq     max neval cld
         calc 2.8666 2.95185 3.336269 3.02475 3.10360 17.6321   100   c
 raster::mean 2.3398 2.39725 2.842573 2.44095 2.49595 32.4901   100  bc
          app 1.5573 1.59200 1.859306 1.63535 1.71005 20.6778   100 a  
  terra::mean 1.8386 1.88065 2.243889 1.94090 2.02975 21.0170   100 ab

As you can see terra::app is the fastest, followed by terra::mean

user_107826740
Related