I have a climatic dataset of a region where I am calculating time series (for drought index). The dimension of the dataset is 164*153*1416 (1416 months in the dataset and the others are the location). How I can average this over space and time?

Average= nanmean(Box_SPEI(:,:,:))

Answer 1

[Chad Greene]

To get a map of average SPEI,

SPEI_mean = mean(SPEI,3,'omitnan'); 

and that calculates the mean along the third dimension (time).

[Md Muktadir]

Hi Chad! Thanks. But I am still stucked as, when I average it along the third dimension, it converts into 164*153 (two dimension). I want to extract/plot the time series for average SPEI. In the dataset, the dimension of time is 1*1416. How I can do that as the length is different for average average SPEI and time?

[Chad Greene]

By default, the mean function calculates the mean across the first nonsingleton dimension. So if you calculate

SPEI_mean = mean(SPEI,'omitnan'); 

the result will be 1x153x1416, because it's taken the mean along the first dimension. If you take the mean of the mean like this:

SPEI_mean = mean(mean(SPEI,'omitnan'),'omitnan'); 

then the answer will be 1x1x1416. That's the time series you want, but you'll probably want to squeeze it like this:

SPEI_mean = squeeze(mean(mean(SPEI,'omitnan'),'omitnan')); 

which turns it into a 1416x1 vector that you can plot via

plot(t,SPEI_mean) 

If you want to calculate the mean within a specific region rather than the entire dataset, like the mean SPEI within the boundaries of a desert or something, try the local function in the Climate Data Toolbox.

[Md Muktadir]

It works @Chad Greene. Thanks a lot. The cdt is really helpful and a great piece of good work.

[Md Muktadir]

@Chad Greene, If i have a dataset (360*720*1416) where first two indicate latitude and longitude and 1416 is the time step. If I need finer resolution (0.25 degree spatial resolution) to make it (720*1440*1416), how I can do that?