evaluation of model complexity from a model files perspective

[slawler]

Maybe there should be more specificity with respect to model description. For example, if we assume that models will be in more specific categories that Water Model. A few examples that come to mind are

• Hydraulic Model (e.g. HEC-RAS)
• Hydrologic Model (e.g. HEC-HMS)
• Coastal Model (e.g. Delft3D)

There may even need to be refinement beyond this however, to differentiate hydraulic models for pipe flow vs surface flow. Or Hydraulic models coupled with hydrodynamic models, or doing combined riverine-coastal (Delft suite, Mike Suite, RAS suite etc.).

https://github.com/TexasDIS/metadata/blob/87c36a9a072eeb0e3b2fe64f1b8c1c2314173868/examples_and_templates/metadata_twdb_example_data_artifact_level_2021_10_06.yaml#L29

Also thinking about the broader issues, beyond classification of models using the data_artifact_title line, there may need to be clear differentiation between model input and output files. https://github.com/TexasDIS/metadata/blob/87c36a9a072eeb0e3b2fe64f1b8c1c2314173868/examples_and_templates/metadata_twdb_example_data_artifact_level_2021_10_06.yaml#L2

For most of the models there are input and file output types that are identifiable via extensions. Often there will be model input files available, but not model outputs.

[ajdabrowski]

Thank you for this comment. Entirely agreed and this is prescient.

As you noticed, in the current structure "models" are grouped within "Artifact". We're aware that this doesn't account for appropriate model-specific description. So far, we've refined and pulled out metadata fields for geospatial data descriptions into "Layer". We aim to further develop model-specific descriptive fields as a next step, including the need you've expressed for supporting a "Model Type" classification.

Your feedback is helpful and I would like to follow-up with you to further discuss model classification and establishing relations between models, and their input and output files.

[slawler]

Sounds good, looking forward to collaborating!

Some preliminary work we took on for tackling this issue can be found on the github.com/USACE page at the mcat-ras and mcat-hms repositories.

[ajdabrowski]

Thanks! We'll take a look at those repos. We've taken a first pass at generic model metadata fields with an update to the documentation published here. I'll be reaching out to you soon to schedule a discussion, and hope the update here can become a talking point. For now, I'll close this issue and we can follow-up with new ones as we see that further things need to be accounted for.