Hello, Peter gave some ideas, but I just want to add some more discussion. The general technique is masking, where you somehow create a mask variable whose value is 1 in the region where you want to do something, and missing elsewhere. Another example is in this FAQ, Averaging/integrating over irregular regions in Ferret at http://ferret.pmel.noaa.gov/FERRET_17sep07/FAQ/analysis/avg_integ_irreg.html We don't have a a general way to create a mask that picks out the points inside a national boundary or other general curve. You could start with a topography dataset and regrid that to the grid of your data, and create a land-only mask in the general region of China. Here is how to make first a land mask over the whole globe, then a rectangular mask that contains China: yes? use etopo20Now you can see that what you want is a mask which is 1 at each grid point inside the boundary of China. This is not something that we have a method for though. Ansley xhbhyq11 wrote: Hi : |