Support for complex terrain #52
Description
Milos Rompar has been using PyDDA to retrieve winds from a radar/lidar network where the domain includes a 1400 m hill where the radar is. Since we assume flat terrain in our retrieval, there would definitely be some inaccuracy in the winds since we do not consider the terrain.
I think adding a complex terrain constraint would be as simple as constraining the gradient of the cost function to be zero at and below the ground. I'm looking for people to work with on implementing such terrain maps so that we can test out such a feature.