[5pt] Run-time improvements for fimpact processes #1864

Noted significant fim_pipeline.sh runtime for HUC 07120004 (Chicago metro nr O'Hare airport). Looking at the UAT run time history (total runtime):
4_9_14_0 --> 152 mins
4_9_12_1 --> 155 mins
4_9_11_0 --> 78 mins
4_9_10_10 --> 28 mins
4_9_8_0 --> 31 mins

I added runtime counters for the process_roads_fimpact.py and process_buildings_fimpact.py and flagged some significant runtime for the buildings process. This was evident in branch 0 and in some of the GMS branches:

Looking through the code, it appears possible bottlenecks are the polygon-on-polygon intersections (geopandas) and the zonal_stats calls.