Improve wx model-ninjafoam coupling

[nwagenbrenner]

Right now we just use the average of the wx model grid to specify the inlet boundary of the ninjafoam domain. We should look into more sophisticated coupling options. We could look into specifying the lateral boundary conditions similar to:

https://www.researchgate.net/profile/Engin_Leblebici/publication/305987888_Wind_Power_Estimations_using_OpenFoam_Coupled_with_WRF/links/57b581c608aeddbf36e7d8be.pdf

and perhaps nudging within the domain.

[nwagenbrenner]

Info on WRF lateral BCs:

http://www2.mmm.ucar.edu/wrf/users/docs/arw_v3.pdf

and observation nudging:

http://www2.mmm.ucar.edu/wrf/users/workshops/WS2009/abstracts/1-09.pdf

http://journals.ametsoc.org/doi/pdf/10.1175/1520-0450%281994%29033%3C0416%3AMFDDA%3E2.0.CO%3B2

http://www2.mmm.ucar.edu/wrf/users/docs/ObsNudgingGuide.pdf

[jforthofer]

One big difference between us and WRF is that our solution is a steady state solution, so no time variable. So we wouldn't be able to directly use the publications above, but maybe we can use the same idea... That a source term is added to the momentum equations that is a function of the difference from the measurement to the current (at this iteration) solution. The function should go to zero when the difference is zero and not get too big, like not bigger than any of the other terms in the momentum equation.