Hi Onyango Augustine Omondi,

The CDO operators have the function "gtc" (greater than a constant) which makes a grid with 1s for cells greater than a given value, i.e.

> cdo gtc,0.1 input.nc output.nc 

'output.nc' is a file with 1s in cells where the condition is true.

See CDO manual, Section 2.5.2. COMPC - Comparison of a field with a constant

Hope it helps,

Augusto

cdo gtc,0.1 input.nc output.nc

You could do the same in Grads. For instance, check the manual page for the const() and maskout() functions:

http://cola.gmu.edu/grads/gadoc/gradfuncconst.html

They have exactly that example. To calculate a daily timeseries of the fraction of the globe covered by precipitation greater than 10mm/day:

ga-> set lon 0 360

ga-> set lat -90 90

ga-> set t 1 last

ga-> define ones = const(const(maskout(p,p-10),1),0,-u)

ga-> set x 1

ga-> set y 1

ga-> display tloop(aave(ones,lon=0,lon=360,lat=0,lat=360))

In Matlab, as in CDO, it is much simpler. You can just create a new logical variable (0 and 1) with the result of your conditional testing. For instance:

>> data=rand(10,10); % some random precip data from 0-1

>> mask=data>0.7; % create a matrix of 0 and 1 with =1 when data > 0.7

>> mean(data(mask)) % average just the masked values

ans =

0.8502

>> mean(data(~mask)) % average the un-masked values

ans =

0.3325

Thanks all.