option to run with GRU forcings when using multiple HRUs per GRU

[andywood]

This would be a useful mode of application, similar to the operation of VIC with veg/snowband tiles. Rather than develop individual forcing inputs for all the HRUs in a GRU, allow SUMMA to estimate the HRU forcings internally based on the GRU-level forcings. HRU-level forcings files can be much larger than GRU-level forcings files, making explicit HRU forcings cumbersome; and they also require a preparation strategy, making the overall workflow more difficult than just using GRU forcings.

Several strategies are possible, eg:

1. read input maps of scaling/adjustment from GRU to HRU for each forcing variable
2. read global options for scaling (eg a temperature lapse rate with elevation, precip adjustment with elevation and aspect)

[wknoben]

Following from our discussion in (now closed) #496 :

A slightly more general approach would be to make all remapping (i.e. native forcing resolution to HRUs) happen internally in SUMMA, which avoids the need to specify GRU-level forcing outside SUMMA. This is particularly helpful in cases where each GRU contains 1 HRU, but there are still many GRUs to a given forcing input cell/polygon.

The mizuRoute source code may provide a good example of how to incorporate remapping from arbitrary forcing shape (grid, polygon, possibly also points?) to HRUs:

• https://github.com/ESCOMP/mizuRoute/blob/main/route/build/src/remap.f90
• https://mizuroute.readthedocs.io/en/latest/Input_data.html#runoff-mapping-data