Lake head flow model with multiprocessor run

[Shyamal Karmakar]

Hello Everyone, 

I am trying to run a large-domain MODFLOW GWF simulation in the Flopy environment using a multiprocessor run. When splitting the model domain (optimize_splitting_mask function), hundreds of lakes had issues with connection indexing and slicing, even though the code implements a forced integration of the lake parts so that it does not split the lake into different segments, and the MPI simulation failed because of the connection indexing error. 

Can anybody suggest a better/anyother approach to handle this? I can not consider the lake as GHB since they are dynamically connected to the aquifer. 

Thanks

Shyamal