Closed hmrtn closed 4 years ago
Issue #14 has a discussion of the important difference between subsetting and filtering. My initial assessment is that we should stick with the subset
mentality.
We should be cognizant of the need for subsetting at two distinct levels:
1) While loading data, we may only be interested in data within a particular bounding box. We can use this information at the data load step to minimize the amount of data we load into memory. 2) While working with data that has been loaded, we will want to subset by distance, polygon, etc.
BlueSky models contain a lot of grid cells - too many! For a large amount of analysis, it does not make computational or statistical sense to use cells from irrelevant locations. We need to be able to reduce the number of grids in a number of ways.