Russian River Watershed Overview

The Russian River in northern California is a 1,485 square mile drainage approximately eighty miles long and thirty two miles across at the widest point, sitting between the mountains of the Mendocino Range to the west and the Mayacamas mountains to the east.  The land cover within the watershed is largely forested in the higher elevation portions with agricultural and urbanized areas within the lower elevation portions.  The hydrology of the watershed is dominated by atmospheric river driven rain events, which can deliver significant amounts of water in a few days during the wet season of November through April, with very little rainfall occurring during the drier summer and fall months.  Flooding from these events typically occurs one to two days after the beginning of the storm and recede within two to three days.

Two reservoirs in the watershed, Lake Mendocino and Lake Sonoma are operated by USACE for flood protection, water supply, and environmental flow objectives, managed in coordination with the Sonoma County Water Agency that delivers water to municipal users in the watershed.  Major population centers impacted by flooding include Hopland, Healdsburg, and Guerneville.  

Background: HEC-WAT for Forecast Informed Reservoir Operations

Lake Mendocino, impounded by Coyote Valley Dam, was the first reservoir evaluated for Forecast Informed Reservoir Operations (FIRO) with USACE.  The FIRO program seeks to better incorporate meteorologic and hydrologic forecasts into reservoir release decision, improving project performance for flood risk management objectives as well as conservation-based objectives such as water availability and ecosystem operations.  In order to evaluate this proposal under the systems-based approach with risk and uncertainty required by USACE guidance, a HEC-WAT model was developed with components for HEC-HMS to model rainfall-runoff processes, HEC-ResSim to model reservoir operations, HEC-RAS to model hydraulics and flood inundation, and HEC-FIA for evaluation of economic consequences from flood damages. As a result of the FIRO pilot study, leading to a FIRO Viability Assessment study and eventual water control manual (WCM) update study, Lake Mendocino regularly operates under Forecast Informed Reservoir Operations procedures allowing up to 11,650 acre-feet of storage to be retained during the flood season and used to meet water availability demands in the following year.

The development of the Russian River HEC-WAT model for FIRO is described in detail in HEC Publication PR-100.  This model has been simplified and updated to the latest HEC-WAT version for this Quick Start Guide, with only the Existing Conditions version of the model provided to reduce complexity of the example.   Note: The example models for the Russian River in this Quick Start Guide are provided for the sole-purpose of learning about HEC-WAT and demonstrating the software's features and capabilities.  These models should not used for any purpose beyond this Quick Start Guide.


 



HEC-WAT Quick Start Guide Example for Russian River Watershed

Hydrologic Sampler and Meteorologic Data

The HEC-WAT Hydrologic Sampler that generates storm events during the HEC-WAT Flood Risk Analysis (FRA) compute used to evaluate uncertainty for flood risk impacts in the watershed was developed with precipitation frequency and storm shape data derrived from the National Climatic Data Center (NCDC) records for 28 hourly precipitation gages within the watershed, as well as average annual precipitation derrived from the 30 year period-of-record from the PRISM (Parameter-elevation Regressions on Independent Slopes) model available at the time.

HEC-HMS model

The HEC-HMS model for the Russian River FIRO study was initially developed with HEC-HMS version 4.2 and has since been updated to HEC-HMS version 4.13 that is included with HEC-WAT 1.1.  This HEC-HMS model delineated the watershed into eighteen sub-basins, with basin parameters developed through the use of various geospatial data sources including USGS 10-meter DEM terrain and soil properties from the NRCS Soil Survey Geographic (SSURGO) database.  Channel routing parameters were initially estimated with the DEM data, calibrated to observed flows, and then replaced with parameters derived from the HEC-RAS model.   The HEC-HMS model was calibrated with the sixty year analysis period for which observed runoff hydrographs and seasonal volumes were available at the key streamgages within the watershed.  From this calibration process, a single set of model parameters to represent both dry conditions and flood events was developed.  Results for antecedent soil moisture conditions prior to flood events form this period-of-record simulation provided the basis for a statistical distribution of the initial soil moisture deficits used in the HEC-WAT FRA Monte Carlo simulations.

HEC-ResSim model

The HEC-ResSim model for the Russian River FIRO study was initially developed with HEC-ResSim 3.3 and has been updated to HEC-ResSim version 3.5 that is included with HEC-WAT 1.1.  The model was initially developed under prior studies in coordination with USACE San Francisco District and Sonoma County Water Authority staff.  The model was modified to match the topology of the HEC-HMS model with a small number of changes to support additional locations required for reservoir operations rules at downstream control points.  Reservoir evaporation data from the HEC-HMS model was used to develop a seasonal pattern applied to each reservoir, while data from Sonoma County Water Authority was used to define consumptive diversions from the river.

HEC-RAS model

The HEC-RAS model that was developed for the Russian River FIRO study was initially developed with HEC-RAS version 5.0.3 and has been updated to HEC-RAS version 6.5 that is compatible with HEC-WAT 1.1.  The HEC-RAS model was developed from an existing model for the USACE Corps Water Management system, used to estimate flood damages during real-time forecasting proceedures.  Detailed hydraulic modeling of the watershed was limited to the maximum annual events within each year of the period-of-record to reduce compute time.  This HEC-RAS model was updated with cross-sections developed using the USGS 10-meter DEM dataset and assumed bathymetry where not available.  Calibration of the model emphasized matching flood conditions and was not considered accurate for lower flow conditions.