PSM model improvements

[alanalutz]

Description

Improvements to the propensity score model:

• Adds three more covariates: tree cover in 2000, travel time to nearest city, and population density
• Enforces a minimum distance between sample points to avoid spatial autocorrelation
• Applies a land mask to avoid sampling large permanent water bodies
• Stratifies samples by biome

Created, updated, or replaced Files and/or tables

• ‎notebooks/ps_model.ipynb renamed to notebooks/match_cells.ipynb
• notebooks/match_cells.ipynb: Code updated to apply the changes described above.

Notes for reviewers

This PR does not require close review, as the propensity score model is no longer used in the current TPARA methodology. This PR mostly serves to preserve the development process in the commit history.
However, there are a couple of elements here that are still in use that I would like review on:

• The three additional covariates, and the way they are resampled and assembled into a 1km-resolution stack
• Use of a land mask to avoid including large, permanent bodies of water in the analysis

_{Stack created with GitHub Stacks CLI • Give Feedback 💬}

[cjthomas730]

Everything looks good to me; a minor nitpick on the variables.py which would be nice to address but isn't needed for approval. Resampling and land masking look good