Creating a land-sea mask for gridded data - just like magic!

This video shows how you can create a mask for your gridded netcdf or grib data so you can calculate spatial statistics such as the mean just over land or ocean. The method uses cdo and needs just four simple steps.

The example uses the two metre temperature, so that you can show the trends in global mean land temperatures compared to global mean ocean temperatures. The first graphic was created using the methods outlined in the video and show that global land temperatures have already increased by 2 degrees since 1950 according to the ERA-5 reanalysis!

Someone pointed out that this mask will chop a part of Holland off, for being below sea level - this is true and the underlying topography available in cdo is stored anyway at low resolution (0.5 degrees I think) and this method is only meant to be rough and ready at your fingertips. If you want to have a more accurate map you can also download a ready made mask from here:
https://ldas.gsfc.nasa.gov/gldas/vege...

or at high resolution here is a "distance to coastline" dataset available at 0.01 degrees https://data.noaa.gov/dataset/dataset...
you can convert to a mask using the logical functions outlined in my functions video.

I hope you enjoy my latest effort, and I look forward to seeing you back here on climate unboxed soon.