Resampling Xarray Dataset to higher spatial resolution?

Question

This feels like it should be an easy problem with a known solution but struggling to figure it out.

I have a 2D xarray dataset (with latitude/longitude) on a regular grid. I want to increase the resolution of this data by sampling it on to a higher resolution grid, e.g. going from 1 degree to 0.1 degree.

I thought this should have been as easy as using the .interp method and providing it with the new grid but this causes issues at the boundaries where we lose data points. What I think is happening is that the "half" of the grid box at the edge of the domain ends up as a NaN as that's the nearest neighbour to that data point.

From what I can tell, none of the interpolation methods give the desired effect and it feels like some data manipulation might be required to "hack" this in to working correctly.

Thanks

Answer 1

rioxarray's reproject method would be the right tool you are looking for. You can find an example here and the documentation for the function here.