Partialgwmodel

[LaurelineJ]

Better dealing of groundwater data, especially for state groundwater model.

[LaurelineJ]

This allows us to save partial gw model so that when we want to focus on a given state, we do not need to read the contus data.

This is in conflict with the state level model handling of data.

[jrising]

Hi Laureline, can you please merge in the bystate branch into this branch and update the code?

[LaurelineJ]

I'll work on it next week.

[LaurelineJ]

The main initial goal of this branch was to save a state specific groundwater model for state-specific analysis. The idea was to make the process faster and not load all the groundwater data. I was also to shorten the Groundwater.jl file to move all data reading lines to a groundwaterdata.jl script. All vector type data have been aggregated in one csv file, while the connexion and lateral conductivity matrix are kept in separate csv file. In the future, this should be improved by having only one sparse matrix.

I have moved the data in the correct place (I hope), it is now in data/counties/gwmodel. I have also added something to warn that the state-level groundwater model does not exist yet. Though I think I can construct something quickly by making bold assumptions on storage coefficient. Updates on this to come in the near future.

There are a few additional changes in there that are not related to the branch per say. I am not sure if they are useful or not. For instance:

• Caching: here cached_fallback(), cached_store() and the test_caching script all follow the cachepath.
• I have temporally removed the variation of groundwater cost in space until we have depth to groundwater information (rather than thickness of groundwater layer)
• simulate.jl does not necessarily read a config file if one is already loaded

[jrising]

Seems great. Want to merge it in?