Hydrohistorian RiverWare bundle (control-file DMI)
=================================================

control.ctl     one line per series: Object.Slot: file=~/... units=... scale=1.0 import=resize
streamflow/     one file per gage, daily mean cfs, slot Gage Inflow.
                DWR_<abbrev> and USGS_<site> are often the SAME gage recorded by
                two agencies (gages.csv same_site_as pairs them): import one.
diversions/     one file per structure (WDID_<wdid>), DWR's daily Total (Diversion), cfs.
reservoirs/     Reclamation HydroData per reservoir: Storage (acre-ft), Pool Elevation
                (ft), Outflow and Inflow (cfs), Evaporation (acre-ft per daily timestep).
                "Unregulated Inflow" slots are not standard reservoir slots; add a
                custom slot or drop the line. Values Reclamation publishes that are
                physically implausible are written as NaN and listed in each file's
                header (reservoir_daily.flag in the release has them all).
reservoirs/*_elevation_volume_<year>.txt, *_elevation_area_<year>.txt
                published area-capacity tables (Navajo 2019 only): two tab-separated
                columns, no header -- elevation (ft) then volume (acre-ft) or area
                (acre) -- strictly increasing, for the Elevation Volume / Elevation
                Area Table slots via the slot dialog's File > Import. RiverWare
                assumes imported values are in the units the slot dialog shows, so
                set ft / acre-ft / acre there first. RiverWare interpolates linearly
                between rows; Reclamation's own rule between whole feet is
                V + A*x + C*x^M (area_capacity.csv has A, C and M). Elevations are
                on the table's datum (Navajo: RPVD, 2.97 ft below NAVD88).
network.csv     every structure, gage and confluence on DWR's stream network, each with
                downstream_node_id and riverware_object (the name used in control.ctl).
streams.csv     each stream, its parent and the parent mile where it joins.

Every object name is a placeholder: rename to your model's objects and keep
the file= paths. Dates: each value is stamped at the END of its day (24:00),
the RiverWare and TSTool convention. Missing days are NaN, never zero.
Source tables, data dictionary and checksums: https://hydrohistorian.com/data
