Aggregating in arrays

Question

The netCDF4 file I want to work with in R is too large. I want to write a loop that will read in a chunk of the data and summarize it.

The variable I wish to read in has 4 dimensions; 'lat', 'lon', 'member' and 'time'. The time has a monthly resolution and the member contains 60 ensemble runs from a climate model.

Using the ncvar_get command I have extracted 12 time slices from the netCDF, leaving me with a 4 dimension array.

num[1:144,1:69:1:60,1:12]

How would I aggregate this so that I would have annual data.

Answer 1

I am assuming that your fourth dimension, which has a length of 12 is the number of months and you would like to aggregate over this dimension and return an array of dimension c(144, 69, 60) .

reproducible data (an array of the same dimensions, with all 1s)

myArray <- array(1, dim =c(144, 69, 60, 12))

Here is a method using apply :

mySumArray <- apply(myArray, c(1,2,3), sum)

This returns an array with the following dimensions:

dim(mySumArray)
[1] 144  69  60

and where the first three elements are:

mySumArray[1:3]
[1] 12 12 12

If you wanted the mean or some other function, just replace sum with you desired function.

An optimized version of summing and calculating the mean is rowSums and rowMeans .

mySumArray <- rowSums(myArray, dims=3)

returns the same result as above MUCH faster.