Question

i want to select June, July and August data from the 3D array (ssta_sst, 360*180*362).

the loop works but the output of ssta_winter has the identical values for ssta_winter[,,i].seen below. I have set it up as an array of (360,180,29).

I think the problem is the variable temp, i want to define it as an array first but i don’t know what size it should be (360,180,3) or (360,180,3*29) and how to keep a loop counter in temp when it passes down to finding mean stage?

ssta_winter = array( data=NA, dim = c(360,180,29))
temp = array( data=NA, dim = c(360,180,3))

for (yr in 1982:2010) {
    temp <- ssta_sst[,,year_sst==yr & (month_sst>=6 & month_sst<=8)]
    for (i in 1:360) {
        for (j in 1:180) {
            ssta_winter[i,j,] <- mean(temp[i,j,])
        }
    }
}

> for (i in 1:29){
+ print(ssta_winter[180,166,i])
+ }
[1] 0.2222583
[1] 0.2222583
[1] 0.2222583
[1] 0.2222583
[1] 0.2222583
[1] 0.2222583
[1] 0.2222583
[1] 0.2222583
[1] 0.2222583
[1] 0.2222583
[1] 0.2222583
[1] 0.2222583
[1] 0.2222583
[1] 0.2222583
[1] 0.2222583
[1] 0.2222583
[1] 0.2222583
[1] 0.2222583
[1] 0.2222583
[1] 0.2222583
[1] 0.2222583
[1] 0.2222583
[1] 0.2222583
[1] 0.2222583
[1] 0.2222583
[1] 0.2222583
[1] 0.2222583
[1] 0.2222583
[1] 0.2222583

Answer 1

This : mean(temp[i,j,]) will only produce a single value, so the LHS of that assignment should be ssta_winter[i,j], and that means your dimensions for the means of winter (in the Southern Hemisphere) months should be only c(360,180). (And that explains the pattern you are puzzled by.)

If you want a single value for the “year” index, then you need to use that seq_along applied to the index inside your loop on the LHS of the assingment:

for (yr in seq_along(1982:2010 ) ) {
   temp <- ssta_sst[,,year_sst==(1982:2010)[yr] & (month_sst>=6 & month_sst<=8)]
   for (i in 1:360) {
     for (j in 1:180) {
        ssta_winter[i,j,yr] <- mean(temp[i,j,])
                                   }
                     }
                      }