HydraProbe used to improve water supply forecast from 3 SNOTEL sites in Idaho over traditional antecedent conditions using the principle component forecast model
Much of the water in the Western United States used for irrigation, municipal supplies and hydro-electric production originates as winter snow pack at higher elevations. Since the early 1900s, the correlation between stream flow and the snow water equivalent (SWE) has been used to forecast the available water for beneficial uses and assess flood potential.
A network of nearly 850 SNOTEL sites in addition to over 1100 manually measured snow courses is used in the statistical based forecasting of streamflow. Because the soil under the snowpack represents a significant storage reservoir for snow melt water, soil sensors began to be installed at SNOTEL sites starting in the late 1990s. In this network, some of the longest records of soil moisture are at three Idaho SNOTEL sites.
This study was conducted to determine if 14 years of soil moisture data at 3 SNOTEL sites measured with the Stevens HydraProbe could statistically improve the stream flow forecasts at a river gage operated by the US Geological Survey (USGS). A parameter call the Soil Moisture Deficit Index (θdi) was calculated from the average soil moisture and the soil’s water content at 333 hPa, was used as an attenuation coefficient in the stream flow forecasts. The forecasted stream flow was compared to the actual stream flow recorded by the USGS and a correlation faction (R) was developed to compare the accuracy of the forecasts. Preliminary work shows that soil moisture data improved the water supply forecast and showed a positive correlation between stream flow and soil moisture deficit index. The impetus for an improved forecast are many, as water resources continue to be limited, especially in light of increasing demand in an environment where there is evidence that climate change is changing seasonal snowpack and snowmelt timing.