Mill Creek Watershed Hydrologic Summary & Forecast: Water Year 2021

This summary describes the application of a distributed hydrologic model of the Mill Creek watershed to summarize and forecast hydrologic conditions through September 2021.  The Mill Creek hydrologic model was developed as a Decision Support Tool (DST) to investigate streamflow enhancement options and restoration priorities for the watershed under a grant from the State of California’s Wildlife Conservation Board.  This forecast effort was prepared for Coast Range Watershed Institute by O’Connor Environmental, Inc. to provide timely information to land and water managers and residents regarding the scale and scope of the current drought conditions and to demonstrate the feasibility of using the existing DST as a forecast tool.


The Mill Creek hydrologic model represents all of the major land-based phases of the hydrologic cycle with a series of spatially and temporally distributed Geograhic Information System (GIS) and timeseries inputs, and has been calibrated to available streamflow, groundwater elevation, and wet/dry stream channel mapping data.  Please refer to the links at the end of this summary for additional details on the watershed characterization, model development, and calibration process underpinning the model.  For this forecast, we extended the climate timeseries input to the hydrologic model through 4/15/21 and applied estimates of the average April 15–September 30 rainfall and Potential Evapotranspiration (PET) from available climate monitoring stations at Venado, Healdsburg, and Windsor.  The forecasts account for antecedent hydrologic conditions and the effects of multi-year climate variability by simulating continuous soil moisture, runoff, streamflow, and groundwater conditions beginning in Water Year 2010. 

Water Balance

A mostly complete record of mean annual precipitation is available at Healdsburg from 1894 to present.  Comparison of the total Water Year (WY) 2020 and 2021 rainfall to previous 2-yr totals reveals that the current drought is about as severe as the 1976/1977 drought of record and that this recent 2-yr rainfall is only about 61% of the lowest 2-yr total (2014/2015) during the more recent drought period (Figure 1). 


Figure 1: Comparison of two year total precipitation for Water Years 2020/2021 at Healdsburg with two year totals from previous droughts.  The 2020/2021 total assumes 1.4-inches of additional precipitation will occur between 4/15 and 9/30 (long-term average for this date range).

Annual water balance results indicate that WY 2021 conditions are significantly drier than the driest year previously simulated (2014) in the baseline simulation period (2010-2019).  Total annual streamflow is projected to be about 51% of 2014 conditions and 14% of mean annual conditions (Figure 2).  The Climatic Water Deficit (CWD) is projected to be about 13% higher than 2014 conditions and 64% higher than the long-term average.  Infiltration recharge is particularly sensitive to the low rainfall conditions and is projected to fall to only 9% of 2014 conditions and 2% of mean annual conditions.  The primary processes responsible for maintaining summer streamflow (baseflow and springflow) are more resilient to the short-term reductions in precipitation and are projected to be about 61% and 88% of 2014 conditions respectively (Figure 2).  


Figure 2: Mean annual water balance results simulated with the Mill Creek hydrologic model for Water Year 2021 compared to conditions during the 2010-2019 baseline period. 

Streamflow & Salmonid Habitat

Both spring and summer streamflows are projected to be significantly lower in WY 2021 compared to the driest previously simulated WY of 2014.  Upstream of the confluence of Felta Creek within the bedrock portions of the watershed, streamflows fell below passable levels for smolt and juvenille salmonids in early to mid-June in average water years and in mid-May during the dry year of 2014 (Figure 3).  The projections indicate that flows in the middle and upper portions of the watershed this year have likely fallen below passable levels about 5 weeks earlier than in 2014.  Farther downstream in the alluvial reach (below Felta Creek) flows recede more quickly, falling below fish passage thresholds by early June in average years and by mid-April in 2014.  In 2021, flows are projected to recede to below passable levels about 4 weeks earlier than in 2014 within the lower alluvial reach (Figure 3).


Comparing the timing of the spring flow recession to the timing of salmonid smolt outmigration from available monitoring data collected by California Sea Grant from 2014-2019 indicates that in recent dry years such as 2014, flows became impassable above Dry Creek during the early to middle portion of the primary outmigration period.  The situation in 2021 appears to be significantly worse for outmigration with flows predicted to become impassable from the start of the most critical late-March to mid-May outmigration period (Figure 3). 


Figure 3: Forecasted spring and summer hydrographs for water year 2021 compared to conditions during the 2010-2019 baseline period for locations in the middle/lower bedrock reach (top) and in the lower alluvial reach (bottom); see Figure 4 for locations.  Gray bars represent the total number of salmonid smolts detected to be moving over a given week based on CA Sea Grant monitoring data collected from 2014 to 2019.

Mean summer baseflows in 2021 are projected to fall to about 82% of 2014 levels with periodic disconnection of pool habitats expected in portions of the bedrock areas of Mill Creek (i.e. upstream of Felta Creek) and extended long-term disconnection in the lower alluvial reaches (Figure 4).  The extent of predicted disconnection in Mill Creek upstream of the alluvial reach expands somewhat in 2021 compared to 2014 and riffle depths remain low throughout the watershed.  Most of the reach from the Palmer Creek confluence to about 1-mile upstream of the Wallace Creek confluence maintains pool connection with riffle depths above 0.1-ft and appears to provide the most persistent summer habitat (with respect to streamflow) for salmonids during periods of extreme drought (Figure 4).  Fisheries monitoring has consistently documented utilization by juvenile coho in portions of this reach, further highlighting its critical importance in providing drought refugia for salmonids.  Summer rearing habitat in most of the other high utilization reaches is expected to be severely limited by critically low summer streamflow conditions (Figure 4).       


Figure 4: Forecasted summer riffle depths and extents of flow disconnection for water year 2021 compared to conditions simulated for 2014.  Also shown are the reaches where CA Sea Grant monitoring data indicates consistent summer rearing habitat utilization by juvenile coho salmon and the locations of the hydrographs shown in Figure 3.   

Wildfire Risk & Recovery

A major limitation of the present modeling analysis is that the model was constrcuted prior to the August 2020 Walbridge Fire which burned through approximately 58% of the watershed.  In the first few years following the fire, the expectation is that summer streamflows may be increased owing to reduced evapotranspiration (ET) by damaged and destroyed vegetation.  If this effect is pronounced, the fire may provide some mitigating benefits to the current drought conditions and the current model predictions may be expected to under-predict spring and summer streamflows.  The post-fire increases in streamflow, to the extent they occur, may be expected to be a relatively short-term phenomenon and the long-term effects (which could potentially include increased ET and reduced streamflow) will depend on the effects of the fire on stand age compositions and the degree and types of post-fire forest management activities that are implemented.

The Climatic Water Deficit (CWD) provides a measure of the seasonal moisture stress experienced by vegetation and may be indicative of the component of wildfire hazard associated with vegetation moisture levels.  CWD in 2021 is expected to be about 13% higher than in the recent dry year of 2014 and exhibits substaintial spatial variability depending on aspect, soil, and vegetation conditions (Figure 5).  To better understand where in the watershed forests and other vegetation may be experiencing particularly high moisture stress, we calculated the percent increase in CWD for 2021 relative to 2014.  Much of the upland portions of the watershed exhibit CWD anomalies of less than 10%, whereas certain areas (particularly north facing slopes along the main-stem of Mill Creek) exhibit CWD anomalies in excess of 40%.  Vegetation in these areas has likely adapted to the generally low CWDs and may be especially vulnerable to the increased CWD posed by the ongoing drought (Figure 5).     


Figure 5: Forecasted annual Climatic Water Deficit (CWD) for water year 2021 (left) and the percent increase in CWD for 2021 relative to 2014 (right).  


Current drought conditions in the Mill Creek watershed appear to be about as severe as the historic 1976/1977 drought of record.  Streamflow conditions are likely to be significantly drier in spring and summer 2021 relative to the recent drought year of 2014.  Of particular significance for salmonids is the very early predicted development of impassable streamflow conditions in lower Mill Creek which may be very problematic for outmigrating smolts.  Juvenile rearing habitat is also expected to be impaired, however the reach of Mill Creek between Palmer Creek and about 1-mile upstream of Wallace Creek is likely to retain surface flows between adjacent pool habitats and provide a key refugia for over-summering salmonids.  Elevated wildfire risk may also be expected owing to the increased Cimatic Water Deficit and associated vegetation moisture stress relative to even the dry conditions of 2014.  Heightened vegetation stress and fire risk may be concentrated in north-facing slopes in the central portions of the watershed where the change in CWD relative to past droughts such as 2014 is simulated to be most extreme.


The current analysis does not include the effects of the Walbridge Fire which may be expected to result in short-term increases in streamflow and to some degree modulate the severity of the current drought with respect to streamflow.  The hydrologic predictions presented here contain uncertainty from a variety of sources and represent approximations of reality based on model simulations developed using best available information.  Depending on the utility of the forecast products presented here, CRWI may expand the forecast effort next year to provide forcasts for other watersheds in the lower Russian River with existing distibuted hydrologic models including the Mark West Creek, Green Valley/Atascadero Creek, and Dutch Bill Creek watersheds.  Although one may expect susbtantial variability in drought response between watersheds, the general patterns of hydrologic response and the severity of the current drought conditions relative to past droughts revealed by this analysis of the Mill Creek watershed are likely generally representative of other coastal salmonids watersheds in the northern California Coast Range.  

Final Report - Coming Soon