LANDFIRE

[HeatherSavoy-USDA]

This dataset was requested in the 2022 geospatial WG workshop.
Variables of interest: Existing Vegetation Type, National Vegetation Classification, Vegetation Height and Cover

Data available from: https://www.landfire.gov/getdata.php
It looks like there is an API we could use for this to be a remote dataset.

The tricky part of this dataset is the versions: not all versions/updates have all variables. Do we use the latest major version, 2.0.0 from 2016, or do we support multiple versions?

[Hedrick-ARS]

We use LANDFIRE products in our snow modeling framework and each subsequent release brings new challenges for incorporation into our pipeline. The latest version should be the most accurate, but I could envision some scenarios where users may need older versions (e.g. historical reanalysis). Perhaps develop remote access to the latest version through the API, then investigate the older versions as needed?

[HeatherSavoy-USDA]

Thanks, @Hedrick-ARS! I'll request the latest version available per variable, at least for now.

[HeatherSavoy-USDA]

Another complication: Some variables are temporal, and some are not. The disturbance variable has values for every year in 1999-2020 which spans 4 out of 5 versions. But the vegetation-related variables from the original request are only associated with versions and not years. This is similar to some of the issues with temporal coverage differences between variables in #86.

[HeatherSavoy-USDA]

This dataset is fun. Here are some more complications:

1. The API by default returns the data in a custom CRS which has coordinates with an origin in the middle of the area of interest, i.e., the CRS changes depending on the area of interest. There is an option to specify an output CRS, so we could specify the CRS of the user's request. However, that complicates some of the rules/assumptions we have for user CRSs.

By default, the service will return the product bundle in a localized Albers projection intended to keep the output product aligned with North and to keep the cell sizes constant.

2. This is more of an annoyance for the user, but the raster attribute tables don't seem to be returning for categorical raster requests in my initial API tests.