使用另一个栅格的单元格值按索引对栅格堆栈进行子集化

Question

I have a 12-layer raster stack consisting of temperature data - one layer for each month. I also have one raster each for planting and harvest dates of a certain crop. I would like to make some calculations of various climate parameters (eg temperature standard deviation), but only within the cropping period between the planting/harvest dates, excluding months outside of this period. All months within the temperature raster stack that do not fall within the cropping period I would like to assign as NA for a given cell. More specifically, I would like to use the cell values within my planting/harvest rasters to subset the cropping months by their index (ie layers 1 through 12) and NOT by their cell values. I am at my wit's end with this one. Below is the code I attempted to use on an example dataset, but does not give me the output I want:

set.seed(150)
t1<- raster(ncol=10, nrow=5) #Create rasters
values(t1) = round(runif(ncell(t1),1,20))
t2<- raster(ncol=10, nrow=5)
values(t2) = round(runif(ncell(t2),1,20))
t3<- raster(ncol=10, nrow=5)
values(t3) = round(runif(ncell(t3),1,20))
t4<- raster(ncol=10, nrow=5)
values(t4) = round(runif(ncell(t4),1,20))
t5<- raster(ncol=10, nrow=5)
values(t5) = round(runif(ncell(t5),1,20))
t6<- raster(ncol=10, nrow=5)
values(t6) = round(runif(ncell(t6),1,20))

Planting<- raster(ncol=10, nrow=5) #Create Planting date raster
values(Planting) = round(runif(ncell(Planting),3,5)) # All planting dates are between 3 and 5

t.stack<-stack(t1,t2,t3,t4,t5,t6) #Create raster stack
layer.names<-c(1:6) # Rename raster stack layers from 1 to 6
names(t.stack)<-as.numeric(as.character(layer.names)) # Attempt to coerce raster stack layer names to numeric
names(t.stack) #View new raster stack layer names (they don't seem to be numberic)

t.stack[t.stack[["X1"]]< Planting] <- NA #Attempt to use Planting raster cell values to coerce raster layer index X1 (1st layer) to NA because it lies outside of the range of 3 to 5

head(t.stack[["X1"]]) #View new values of raster layer X1

The output from the 1st raster stack layer ( X1 ) is:

#   1  2  3  4  5  6  7  8  9 10
# 1 NA  9 10 17  6  8  8  8  9 17
# 2 12 19 16  9 14 19 12  6  5  7
# 3 10 NA NA NA 10 15  6 15  6 12
# 4 20 16 12  5 18 13 13  5 NA 10
# 5 13 14 18 10 16  8 10 NA 20  7

As you can see, I attempted to change the layer names to numeric values and use conditional statements to query them against the cell values in t.stack . Instead, what happened was that specific cell values within the X1 layer were replaced if they were less than the corresponding cell values in the Planting raster. However, I want the whole X1 layer in this dataset to be designated NA because the planting dates only span from 3 to 5, and do not include 1 (ie the first layer). Any thoughts would be appreciated.

Answer 1

Not sure, if I understood you correctly, maybe this helps to get you started:

for (i in 1:nlayers(t.stack)) {
  tt <- t.stack[[i]][]
  tt[i < Planting[]] <-  NA
  t.stack[[i]][] <- tt
}

Answer 2

Here is a more formal (safer, memory wise) approach:

set.seed(150)
r <- raster(ncol=10, nrow=5)
tt <- sapply(1:6, function(x) setValues(r,  round(runif(ncell(r),1,20))))
t.stack <- stack(tt)
Planting <- setValues(r, round(runif(ncell(r), 3, 5)))

f <- function(p) {
    x <- matrix(0, nrow=length(p), ncol=12)
    x[cbind(1:nrow(x), p)] <- 1
    x <- t(apply(x, 1, cumsum))
    x[x==0] <- NA
    x
}

p <- calc(Planting, f, filename='planting12.grd')

tp <- t.stack * p