Conditional mean of rasterstack

Question

I am trying to calculate the mean raster from a large RasterStack, with the following (using package raster):

raster_mean <- calc(raster_stack, conditonalmean)

However, I do not want a standard mean, but I wish to add some conditions to the computation. I have written the following function (where three conditions appear, one is to remove null values and the others are numerical):

conditonalmean <- function(x, na.rm){
v = var(x,na.rm = na.rm)
mean(v[v>=-10 & v<=10])

}

Which however returns the following error:

Error in .calcTest(x[1:5], fun, na.rm, forcefun, forceapply) : 

cannot use this function

Any suggestion?

Answer 1

Its failing because there's no default for na.rm in your function, and calc is calling it without a na.rm set. Try running your function on an arbitrary vector and you'll see what I mean:

> conditonalmean(1:10)
Error in var(x, na.rm = na.rm) : 
  argument "na.rm" is missing, with no default

If you add a value for na.rm in the call to calc this will work:

> cm = calc(s, conditonalmean, na.rm=TRUE)
> cm
class       : RasterLayer 
dimensions  : 4, 4, 16  (nrow, ncol, ncell)

because the na.rm value to calc is passed through to calls to conditonalmean and so provide a value.

You can also fix it better by giving a default in your definition:

 conditonalmean =
  function(x, na.rm=TRUE){
   v = var(x,na.rm = na.rm)
   mean(v[v>=-10 & v<=10])
  }

Also, you should probably check the spelling! Your function is "conditonal" mean!