option to assign different resample methods to different requested gridded products

[AmyHudson]

If grabbing classification layer (e.g. crops) and continous layer (e.g. NDVI), would likely want 'nearest' for classification and 'bilinear' for classification.

[HeatherSavoy-USDA]

Good point! We have an open issue on the GeoCDL repo for this. The current plan is to keep 'nearest' method the default for all variables, but let the user specify one method or two, one for categorical variables and one for continuous. We will add an error if a non-applicable method for categorical variables is requested.

I'll leave this open until the GeoCDL issue is resolved and we have the documentation in rgeocdl updated for it.

[HeatherSavoy-USDA]

This is partially implemented on GeoCDL