Secondary criteria functions: average, closest

[nocollier]

Write a series of secondary criteria functions. Given the neighborhood, what do we do with it? These functions need to resolve the neighborhood so that it matches (in terms of dimensions/coordinates) the site data. They should take the form:

neighborhood_...(ds_neighborhood: xr.Dataset, ds_point: xr.Dataset) -> xr.Dataset`

We will need (names flexible):

• neighborhood_average: take the average of the neighborhood and also encode the standard deviation as an ancillary variable so it may be used as uncertainty if desired.
• neighborhood_closest: I am not sure we need this function as it would be a "do nothing" combined with a neighborhood defined by nlayer=0. Perhaps we do write the function and make sure we find the nearest non-masked/nan cell. This could be useful for near-ocean sites or in combination with other functions.