Photograph of a LoRa-cellular gateway antenna tower near Quincy California, in the Feather River watershed.
Monitoring and modeling soil moisture to improve runoff forecasting and drought characterization in the Feather River Basin
Drought, along with recent wildfires, have had enormous impacts downstream on water quantity and quality. These conditions have greatly impacted the agriculture and the residents of the nearby Central Valley.
For decades, snowpack measurements have been used as a fairly reliable metric to forecast runoff into California reservoirs which store water for later use during the dry summer months. However, in recent years, especially after multi-year droughts, runoff predictions have fallen short. Dry soils may be absorbing some of the snowmelt that would have historically drained into rivers and reservoirs (Lapides et al, 2022). To better understand these processes and to improve water availability predictions, the USGS in cooperation with the California Department of Water Resources (DWR) has initiated a new study to monitor soil moisture and to incorporate those data into a hydrologic model.
The upper Feather River watershed was selected for the study location because it is one of California’s highest precipitation regions and supplies the largest portion of water delivered by the State Water Project. The watershed drains an estimated 3,200 square miles and provides an average 3.2 million acre-feet annually to the Feather River system and feeding into Lake Oroville, which is the largest reservoir in the State Water Project. Changes in water availability due to drought and wildfire have enormous impacts downstream on water rights allocations for the $58 billion agricultural industry of the Central Valley and the cities that depend on surface water delivery. Soil moisture has been recognized as one of the major data gaps in hydrologic monitoring in this area. At the start of this project, there were no soil moisture stations in the Feather River watershed.
Expanding monitoring and modeling of both surface water and soil moisture in the Feather River watershed would provide insight into how much snowmelt is being absorbed into the soil. The data gathered will also feed into a watershed model to enhance early drought warning and improve forecasts of runoff into reservoirs. If successful, the sensor network and watershed model in this research could be expanded into other parts of California.
Watershed Modeling
A USGS water balance model will inform two aspects of this study; 1) developing an initial monitoring installation strategy, and 2) to use data from this study to calibrate the model and extend its predictive capacity in the extend the monitoring locations spatially and temporally for the Feather River watershed, including unmonitored subbasins.
The Basin Characterization Model (BCM: Flint et al., 2021a) is a monthly, fine-scale (270 meters), spatially distributed water balance model that has been rigorously calibrated to streamflow, actual evapotranspiration, snowpack, and potential evapotranspiration data across California. The statewide BCM (Flint et al., 2021b) will be extracted to the Feather River watershed, and locally refined and calibrated to improve the fine-scale characterization of local water balance processes. Soil moisture and streamflow collected in this study will be used to calibrate the refined BCM model and extend the drought and flood prediction capabilities spatially for the entire watershed and temporally for the historical period (1896-present).
Soil Moisture Monitoring
An important initial step in this study was determining soil moisture monitoring station locations that best represent the variability of the watershed. Their locations were determined by analyzing the hydrology, geology, elevation, and other factors to group areas together based on similar soil moisture responses called “soil moisture response units” (SRMUs; following methods of Curtis and others, 2019). The Feather River Watershed was divided into 15 SRMUs with 1-2 soil moisture monitoring stations planned for each SMRU. Figure 1 shows a map of the Feather River Watershed with the SMRUs and proposed locations for soil moisture monitoring stations.
Figure 1. Upper Feather River watershed showing 15 different soil moisture response units (SMRU) proposed sites of soil moisture monitoring stations (2 proposed sites for each SMRU).
Soil moisture sensors will be installed at up to 5 depths (5, 10, 20, 50, and 100 cm) at each soil moisture monitoring station. Figure 2 shows an example of soil moisture sensors that are being used in this study. Soil moisture data will be telemetered in real-time using either LoRa or GOES satellite telemetry. LoRa telemetry uses low-power, long-range radios to transmit data to local uplink sites, called gateways. The gateways relay data packets through cellular, to the USGS servers (see Grimsley et al, 2022 for additional information). Figure 4 shows an example of one of the LoRa-cellular gateway antenna towers which was constructed as part of this project.
Figure 2. Soil moisture sensors installed in a soil moisture monitoring pit as part of a USGS monitoring station for real-time soil moisture monitoring.
Figure 4. Photograph of a LoRa-cellular gateway antenna tower near Quincy California, in the Feather River watershed. (Photo by Patrick Haluska.)
Wildfire and Changes in Hydrology
The Feather River watershed has also been severely impacted by wildfire in recent years. Between 2020 and 2021, two wildfires, the Dixie Fire and the North Complex Fire, burned nearly half the watershed upstream of Lake Oroville. Wildfire can have a huge impact on hydrology due to the loss of trees. Fewer trees will likely lead to lower overall evapotranspiration.
One limitation of current BCM applications is the inability to dynamically change vegetation and land use through time. Relationships between climate and vegetation are assumed to be constant for future simulations, and the current spatial pattern of vegetation types is limited to a static representation based on currently available published vegetation data. Updates to vegetation datasets can take many years and therefore are not suitable for modeling post-fire response in real-time. Modifying the vegetation layer in an area following a major wildfire will allow the BCM model to better represent reduced evapotranspiration in an area that was burned. Vegetation is also expected to recover over a period of several years depending on the climate and vegetation type.
The incorporation of the wildfire-caused vegetation changes was not part of the original scope of work for this project but was added through additional support from the USGS Post-Fire Hazards Impacts to Resources and Ecosystems (PHIRE) program. More information on related PHIRE studies can be found here: https://www.usgs.gov/special-topics/wildland-fire-science/science/post-fire-hazards-impacts-resources-and-ecosystems.
Figure 3. Map of the upper Feather River watershed showing outlines of two of the largest wildfires in the basin from recent years: North Complex Fire in 2020 and Dixie Fire in 2021.
Relevance and Benefits
This project provides benefit to the cooperator with the goal of improving the understanding of soil moisture in the Feather River basin and how drought affects soil moisture and subsequently water availability. This knowledge can improve runoff forecasts based on snowpack (SWE) measurements alone or watershed modeling forecasts that do not include soil moisture observations.
Previous research has shown that soil-moisture measurements can improve water availability forecasts in both snow-dominated (Harpold et al, 2016) and rainfall-dominated basins (Wyatt, et al, 2020). The Feather River watershed above Lake Oroville is prone to mixed rain-snow events and rain-on-snow events, and the rain-snow transition zone (elevation) is expected to increase due to climate change (Avansi, et al, 2018). A combination of in-situ measurements in targeted sub-basins and basin-wide modeling could help provide more accurate forecasts of water availability for the year and reduce some of the uncertainty related to water resource management. The challenge is balancing the right density and location for additional in-situ monitoring that provides the most benefit for a basin-wide modeling approach.
References
Avanzi, F. , Maurer, T.P., Malek, S.A., Glaser, S.D. , Bales, R.C., Conklin, M.H., 2018, Feather river hydrologic observatory: improving snowpack forecasting for hydropower generation using intelligent information systems. Report for California Energy Commission, available at: energy.ca.gov/sites/default/files/2019-11/Energy_CCCA4-CEC-2018-001_ADA.pdf
Curtis, J.A., Flint, L.E. and Stern, M.A., 2019. A multi‐scale soil moisture monitoring strategy for California: Design and validation. JAWRA Journal of the American Water Resources Association, 55(3), pp.740-758.
Flint, L.E., Flint, A.L., and Stern, M.A., 2021a, The basin characterization model—A regional water balance software package: U.S. Geological Survey Techniques and Methods 6–H1, 85 p., https://doi.org/10.3133/tm6H1.
Flint, L.E., Flint, A.L., and Stern, M.A., 2021b, The Basin Characterization Model - A regional water balance software package (BCMv8) data release and model archive for hydrologic California, water years 1896-2020, U.S. Geological Survey data release, https://doi.org/10.5066/P9PT36UI.
Grimsley, R., Marineau, M.D., Iannucci, R., 2022, Experiences in LP-IoT: EnviSense deployment of remotely reprogrammable environmental sensors, Proceedings of LP-IoT ’21, https://doi.org/10.1145/3477085.3478988
Harpold, A.A., Sutcliffe, K, Clayton, J., Goodbody, A., Vazquez, S., 2016, Does Including Soil Moisture Observations Improve Operational Streamflow Forecasts in Snow-Dominated Watersheds? 53(1), pp 179-196, https://doi.org/10.1111/1752-1688.12490
Lapides, D.A., Hahm, W.J., Rempe, D.M., Whiting, J., Dralle, D.N., 2022, Causes of Missing Snowmelt Following Drought, Geophysical Research Letters, Vol 49, Issue 19, https://doi.org/10.1029/2022GL100505
Wyatt, B. M., Ochsner, T.E., Krueger, E.S., Jones, E.T., 2020, In-situ soil moisture data improve seasonal streamflow forecast accuracy in rainfall-dominated watersheds, Journal of Hydrology, vol 590, November 2020, 125404, https://doi.org/10.1016/j.jhydrol.2020.125404.
Next Generation Stream Gaging
New Technologies for Mapping Surface Soil Moisture Over Wildfire-Prone Landscapes
Soil moisture datasets at five sites in the central Sierra Nevada and northern Coast Ranges, California
Photograph of a LoRa-cellular gateway antenna tower near Quincy California, in the Feather River watershed.
Figure 1. Upper Feather River watershed showing 15 different soil moisture response units (SMRU) proposed sites of soil moisture monitoring stations (2 proposed sites for each SMRU).
Figure 1. Upper Feather River watershed showing 15 different soil moisture response units (SMRU) proposed sites of soil moisture monitoring stations (2 proposed sites for each SMRU).
Figure 4. Map of the upper Feather River watershed showing outlines of two of the largest wildfires in the basin from recent years: North Complex Fire in 2020 and Dixie Fire in 2021.
Figure 4. Map of the upper Feather River watershed showing outlines of two of the largest wildfires in the basin from recent years: North Complex Fire in 2020 and Dixie Fire in 2021.
Soil moisture sensors installed in a soil moisture monitoring pit as part of a USGS monitoring station for real-time soil moisture monitoring. Photo credit: Michelle Stern, USGS
linkFigure 5. Soil moisture sensors installed in a soil moisture monitoring pit as part of a USGS monitoring station for real-time soil moisture monitoring. Photo credit: Michelle Stern, USGS.
Soil moisture sensors installed in a soil moisture monitoring pit as part of a USGS monitoring station for real-time soil moisture monitoring. Photo credit: Michelle Stern, USGS
linkFigure 5. Soil moisture sensors installed in a soil moisture monitoring pit as part of a USGS monitoring station for real-time soil moisture monitoring. Photo credit: Michelle Stern, USGS.
Experiences in LP-IoT: EnviSense deployment of remotely reprogrammable environmental sensors
The basin characterization model—A regional water balance software package
A multi-scale soil moisture monitoring strategy for California: Design and validation
Runoff Estimates for California
Streams, rivers, lakes and reservoirs are important natural resources for irrigation, public supply, wetlands and wildlife. Excess precipitation that flows into these sources is called runoff, and it's an important drought indicator. The California Water Science Center tracks both monthly and annual runoff.
Drought, along with recent wildfires, have had enormous impacts downstream on water quantity and quality. These conditions have greatly impacted the agriculture and the residents of the nearby Central Valley.
For decades, snowpack measurements have been used as a fairly reliable metric to forecast runoff into California reservoirs which store water for later use during the dry summer months. However, in recent years, especially after multi-year droughts, runoff predictions have fallen short. Dry soils may be absorbing some of the snowmelt that would have historically drained into rivers and reservoirs (Lapides et al, 2022). To better understand these processes and to improve water availability predictions, the USGS in cooperation with the California Department of Water Resources (DWR) has initiated a new study to monitor soil moisture and to incorporate those data into a hydrologic model.
The upper Feather River watershed was selected for the study location because it is one of California’s highest precipitation regions and supplies the largest portion of water delivered by the State Water Project. The watershed drains an estimated 3,200 square miles and provides an average 3.2 million acre-feet annually to the Feather River system and feeding into Lake Oroville, which is the largest reservoir in the State Water Project. Changes in water availability due to drought and wildfire have enormous impacts downstream on water rights allocations for the $58 billion agricultural industry of the Central Valley and the cities that depend on surface water delivery. Soil moisture has been recognized as one of the major data gaps in hydrologic monitoring in this area. At the start of this project, there were no soil moisture stations in the Feather River watershed.
Expanding monitoring and modeling of both surface water and soil moisture in the Feather River watershed would provide insight into how much snowmelt is being absorbed into the soil. The data gathered will also feed into a watershed model to enhance early drought warning and improve forecasts of runoff into reservoirs. If successful, the sensor network and watershed model in this research could be expanded into other parts of California.
Watershed Modeling
A USGS water balance model will inform two aspects of this study; 1) developing an initial monitoring installation strategy, and 2) to use data from this study to calibrate the model and extend its predictive capacity in the extend the monitoring locations spatially and temporally for the Feather River watershed, including unmonitored subbasins.
The Basin Characterization Model (BCM: Flint et al., 2021a) is a monthly, fine-scale (270 meters), spatially distributed water balance model that has been rigorously calibrated to streamflow, actual evapotranspiration, snowpack, and potential evapotranspiration data across California. The statewide BCM (Flint et al., 2021b) will be extracted to the Feather River watershed, and locally refined and calibrated to improve the fine-scale characterization of local water balance processes. Soil moisture and streamflow collected in this study will be used to calibrate the refined BCM model and extend the drought and flood prediction capabilities spatially for the entire watershed and temporally for the historical period (1896-present).
Soil Moisture Monitoring
An important initial step in this study was determining soil moisture monitoring station locations that best represent the variability of the watershed. Their locations were determined by analyzing the hydrology, geology, elevation, and other factors to group areas together based on similar soil moisture responses called “soil moisture response units” (SRMUs; following methods of Curtis and others, 2019). The Feather River Watershed was divided into 15 SRMUs with 1-2 soil moisture monitoring stations planned for each SMRU. Figure 1 shows a map of the Feather River Watershed with the SMRUs and proposed locations for soil moisture monitoring stations.
Figure 1. Upper Feather River watershed showing 15 different soil moisture response units (SMRU) proposed sites of soil moisture monitoring stations (2 proposed sites for each SMRU).
Soil moisture sensors will be installed at up to 5 depths (5, 10, 20, 50, and 100 cm) at each soil moisture monitoring station. Figure 2 shows an example of soil moisture sensors that are being used in this study. Soil moisture data will be telemetered in real-time using either LoRa or GOES satellite telemetry. LoRa telemetry uses low-power, long-range radios to transmit data to local uplink sites, called gateways. The gateways relay data packets through cellular, to the USGS servers (see Grimsley et al, 2022 for additional information). Figure 4 shows an example of one of the LoRa-cellular gateway antenna towers which was constructed as part of this project.
Figure 2. Soil moisture sensors installed in a soil moisture monitoring pit as part of a USGS monitoring station for real-time soil moisture monitoring.
Figure 4. Photograph of a LoRa-cellular gateway antenna tower near Quincy California, in the Feather River watershed. (Photo by Patrick Haluska.)
Wildfire and Changes in Hydrology
The Feather River watershed has also been severely impacted by wildfire in recent years. Between 2020 and 2021, two wildfires, the Dixie Fire and the North Complex Fire, burned nearly half the watershed upstream of Lake Oroville. Wildfire can have a huge impact on hydrology due to the loss of trees. Fewer trees will likely lead to lower overall evapotranspiration.
One limitation of current BCM applications is the inability to dynamically change vegetation and land use through time. Relationships between climate and vegetation are assumed to be constant for future simulations, and the current spatial pattern of vegetation types is limited to a static representation based on currently available published vegetation data. Updates to vegetation datasets can take many years and therefore are not suitable for modeling post-fire response in real-time. Modifying the vegetation layer in an area following a major wildfire will allow the BCM model to better represent reduced evapotranspiration in an area that was burned. Vegetation is also expected to recover over a period of several years depending on the climate and vegetation type.
The incorporation of the wildfire-caused vegetation changes was not part of the original scope of work for this project but was added through additional support from the USGS Post-Fire Hazards Impacts to Resources and Ecosystems (PHIRE) program. More information on related PHIRE studies can be found here: https://www.usgs.gov/special-topics/wildland-fire-science/science/post-fire-hazards-impacts-resources-and-ecosystems.
Figure 3. Map of the upper Feather River watershed showing outlines of two of the largest wildfires in the basin from recent years: North Complex Fire in 2020 and Dixie Fire in 2021.
Relevance and Benefits
This project provides benefit to the cooperator with the goal of improving the understanding of soil moisture in the Feather River basin and how drought affects soil moisture and subsequently water availability. This knowledge can improve runoff forecasts based on snowpack (SWE) measurements alone or watershed modeling forecasts that do not include soil moisture observations.
Previous research has shown that soil-moisture measurements can improve water availability forecasts in both snow-dominated (Harpold et al, 2016) and rainfall-dominated basins (Wyatt, et al, 2020). The Feather River watershed above Lake Oroville is prone to mixed rain-snow events and rain-on-snow events, and the rain-snow transition zone (elevation) is expected to increase due to climate change (Avansi, et al, 2018). A combination of in-situ measurements in targeted sub-basins and basin-wide modeling could help provide more accurate forecasts of water availability for the year and reduce some of the uncertainty related to water resource management. The challenge is balancing the right density and location for additional in-situ monitoring that provides the most benefit for a basin-wide modeling approach.
References
Avanzi, F. , Maurer, T.P., Malek, S.A., Glaser, S.D. , Bales, R.C., Conklin, M.H., 2018, Feather river hydrologic observatory: improving snowpack forecasting for hydropower generation using intelligent information systems. Report for California Energy Commission, available at: energy.ca.gov/sites/default/files/2019-11/Energy_CCCA4-CEC-2018-001_ADA.pdf
Curtis, J.A., Flint, L.E. and Stern, M.A., 2019. A multi‐scale soil moisture monitoring strategy for California: Design and validation. JAWRA Journal of the American Water Resources Association, 55(3), pp.740-758.
Flint, L.E., Flint, A.L., and Stern, M.A., 2021a, The basin characterization model—A regional water balance software package: U.S. Geological Survey Techniques and Methods 6–H1, 85 p., https://doi.org/10.3133/tm6H1.
Flint, L.E., Flint, A.L., and Stern, M.A., 2021b, The Basin Characterization Model - A regional water balance software package (BCMv8) data release and model archive for hydrologic California, water years 1896-2020, U.S. Geological Survey data release, https://doi.org/10.5066/P9PT36UI.
Grimsley, R., Marineau, M.D., Iannucci, R., 2022, Experiences in LP-IoT: EnviSense deployment of remotely reprogrammable environmental sensors, Proceedings of LP-IoT ’21, https://doi.org/10.1145/3477085.3478988
Harpold, A.A., Sutcliffe, K, Clayton, J., Goodbody, A., Vazquez, S., 2016, Does Including Soil Moisture Observations Improve Operational Streamflow Forecasts in Snow-Dominated Watersheds? 53(1), pp 179-196, https://doi.org/10.1111/1752-1688.12490
Lapides, D.A., Hahm, W.J., Rempe, D.M., Whiting, J., Dralle, D.N., 2022, Causes of Missing Snowmelt Following Drought, Geophysical Research Letters, Vol 49, Issue 19, https://doi.org/10.1029/2022GL100505
Wyatt, B. M., Ochsner, T.E., Krueger, E.S., Jones, E.T., 2020, In-situ soil moisture data improve seasonal streamflow forecast accuracy in rainfall-dominated watersheds, Journal of Hydrology, vol 590, November 2020, 125404, https://doi.org/10.1016/j.jhydrol.2020.125404.
Next Generation Stream Gaging
New Technologies for Mapping Surface Soil Moisture Over Wildfire-Prone Landscapes
Soil moisture datasets at five sites in the central Sierra Nevada and northern Coast Ranges, California
Photograph of a LoRa-cellular gateway antenna tower near Quincy California, in the Feather River watershed.
Photograph of a LoRa-cellular gateway antenna tower near Quincy California, in the Feather River watershed.
Figure 1. Upper Feather River watershed showing 15 different soil moisture response units (SMRU) proposed sites of soil moisture monitoring stations (2 proposed sites for each SMRU).
Figure 1. Upper Feather River watershed showing 15 different soil moisture response units (SMRU) proposed sites of soil moisture monitoring stations (2 proposed sites for each SMRU).
Figure 4. Map of the upper Feather River watershed showing outlines of two of the largest wildfires in the basin from recent years: North Complex Fire in 2020 and Dixie Fire in 2021.
Figure 4. Map of the upper Feather River watershed showing outlines of two of the largest wildfires in the basin from recent years: North Complex Fire in 2020 and Dixie Fire in 2021.
Soil moisture sensors installed in a soil moisture monitoring pit as part of a USGS monitoring station for real-time soil moisture monitoring. Photo credit: Michelle Stern, USGS
linkFigure 5. Soil moisture sensors installed in a soil moisture monitoring pit as part of a USGS monitoring station for real-time soil moisture monitoring. Photo credit: Michelle Stern, USGS.
Soil moisture sensors installed in a soil moisture monitoring pit as part of a USGS monitoring station for real-time soil moisture monitoring. Photo credit: Michelle Stern, USGS
linkFigure 5. Soil moisture sensors installed in a soil moisture monitoring pit as part of a USGS monitoring station for real-time soil moisture monitoring. Photo credit: Michelle Stern, USGS.
Experiences in LP-IoT: EnviSense deployment of remotely reprogrammable environmental sensors
The basin characterization model—A regional water balance software package
A multi-scale soil moisture monitoring strategy for California: Design and validation
Runoff Estimates for California
Streams, rivers, lakes and reservoirs are important natural resources for irrigation, public supply, wetlands and wildlife. Excess precipitation that flows into these sources is called runoff, and it's an important drought indicator. The California Water Science Center tracks both monthly and annual runoff.