search

dtcenter / MET

Model Evaluation Tools
https://dtcenter.org/community-code/model-evaluation-tools-met
Apache License 2.0
77 stars 24 forks source link

Investigate supporting GeoTIFF files directly in MET. #1018

Open dwfncar opened 6 years ago

dwfncar commented 6 years ago

Multiple people from the 2018 WRF Users's Workshop and the 2018 Unidata User's Workshop suggested that MET be enhanced to read GeoTIFF files directly. GeoTIFF is an image file format to which metadata can be attached.


This task is to consider doing that. [MET-1018] created by johnhg

caroline-owens commented 3 weeks ago

Hi all, is there any update on best practices for using geotiff files as the input (forecast and/or obs) for MET or METplus? thank you!

JohnHalleyGotway commented 3 weeks ago

Hi @caroline-owens, I see you're wondering about the status of this issue. This is currently in our "Backlog of Development Ideas". So while it's been a good idea for a long time, we haven't had the time or funding to prioritize it. We're not actively working on it, and I'm not aware of any upcoming projects for which this would be a deliverable. We have many issues that fall into this category.

That being said, we do sometimes have funding that allows us to prioritize "community" feature requests. So you could comment on this stating why this feature would be of particular use.

In the meantime, I'll note that MET does support the passing of data from Python to MET's C++ tools. So you could write a Python script to extract gridded data and corresponding metadata from a GeoTiff file and pass it via memory to MET. And I could definitely see how such a use case would be really beneficial to add to the METplus repository.

caroline-owens commented 3 weeks ago

Thanks so much @JohnHalleyGotway! I'm looking for this feature in order to use MET (and/or METplus) to evaluate forecasts of harmful algal blooms against satellite products, and potentially to do evaluation of new satellite algorithms for bloom area, density, and movement.

It seems like using Python embedding will be the best path forward. Is there a particular board I should post on to find any existing similar use cases that could provide guidance on what metadata and format is needed? Thanks again!

Caroline