Matt Tankersley

reconfigured the ice_vel call in PR #101. Not sure how this should work with bedmap2 or bedmachine since they have many combinations of different layers and reference frames. Should we...

Discovered that you can't use `os.remove(fname)` since pooch uses the originally download file names (unzipped) in each call, even if preprocessing is already applied.

Two more thinks to consider: * look into using [dask-geopandas](https://dask-geopandas.readthedocs.io/en/latest/) to speed up reading large shapefiles. * maybe switch to [geopandas-base](https://github.com/conda-forge/geopandas-feedstock) to reduce dependencies.

Look's like it still requires GeoPandas to read a .shp 😑, so I think I'd still need to include GeoPandas in the .toml dependencies, which won't work on Windows, unless...

Amazing! Thanks for that 👍 I originally tried `pip install pyogrio[geopandas]` which wasn't working, but `pip install --pre pyogrio[geopandas]` did the trick. I'll add that in my .toml to save...

I've tried to replace `geopandas` with `pyogrio` in poetry, but am still having issues. It worked fine to just add `pyogrio`, but to read a shapefile, `geopandas` still needs to...

We now use Geopandas for more than just reading shapefiles, so for now will close this as it is a core dependency.

How would this work for calculating the IGRF for a dataset with a wide range of dates? Would users need to chunk the dataset into chunks of similar dates (e.g....

Yeah that sounds like a good approach! Great that you got this implemented, I'll be excited to test it out!

On top of @ckohnke comments, here is an issue in `pyIGRF` which describes the various options available now: https://github.com/ciaranbe/pyIGRF/issues/2 Once this is implemented, might be good to leave a comment...