Gather additional static input variables for baseline LSTM v2

[lekoenig]

We're interested in adding input variables to improve from baseline. From discussion #44, the running list of new input variables includes:

Dynamic

• daily water temperature from PRMS (seg_tave_water in p1_prms_met_data) Holding off on water temperature for now
• daily min/max air temperature in addition to the mean (get from gridmet primary climate variables)
• daily vapor pressure deficit (get from gridmet derived variables); alternatively, min/max relative humidity if VPD not included in output from drb_gridmet_tools
• daily solar radiation (either use seginc_swrad from PRMS outputs, or sw-rad from gridmet primary climate variables)
• daily potential evapotranspiration, PET (from seg_potet in PRMS outputs) Holding off on this for now since we're now working at the NHD-scale

Static

• soil characteristics from STATSGO, including information about texture (percent clay, sand, etc.) and conductivity (e.g. permeability, rainfall-runoff factor, soil thickness, bulk density (~porosity), depth to seasonally high water table, and average water capacity). Get from Wieczorek dataset referenced to NHDPlusv2.
• mean daily or annual precipitation (get from StreamCat?) pulling watershed mean annual precip (mm) based on 800m PRISM data (1971-2000) from Wieczorek dataset.
• Percent impervious surface within local catchment /HRU, NLCD (get from Wieczorek dataset referenced to NHDPlusv2)
• Percent forest within local catchment /HRU , NLCD (get from Wieczorek dataset referenced to NHDplusv2)

We need to:

• gather each of the above input variables aggregated at the scale of the PRMS segment