Drop stats "summary" from `/conus_hydrology/stats/` and `/arctic_hydrology/stats/` endpoints

[Joshdpaul]

The summary is only needed for the app, and should be included in the hidden /conus_hydrology/hydroviz/ endpoint exclusively. Since we now have different "source" options (original GCM, GCM diff applied to historical baseline) the origin of the values in the summary dictionary is unclear.

For reference:

Summary dictionary for CONUS is calculated here and the source is:

• Projected: source=2 (gcm_diff_applied_to_maurer) · landcover=1 (static) · scenario=3 (rcp60) · era=2 (2046-2075)
• Historical: source=0 (original_gcm) · landcover=1 · model=8 (Maurer) · scenario=0 · era=0 (1976-2005)
• rcp60 models: ACCESS1-0 (0) and BNU-ESM (2) excluded

Summary dictionary for Arctic is calculated here and the source is:

• eras: historical = 1990-2021, projected = 2034-2065
• C2LE models: source=2 (gcm_diff_applied_to_cheng: Cheng ratio pre-applied)
• PGW models: source=0 (original_gcm: no historical run, ratio cannot be applied)
• Historical baseline: source=0 (original_gcm)
• Ensemble min/avg/max computed across all 6 projected models combined