Based on snowpack and baseflow conditions as of April 1, predicted natural streamflows for the upcoming April 1 - September 30 time period are:
- Henry's Lake: 95% of average
- Henry's Lake to Island Park Dam: 85% of average
- Henry's Fork upstream of Ashton: 89% of average
- Fall River: 115% of average
- Teton River at St. Anthony: 125% of average
Full results are shown in the table below.
In last week's blog, I discussed the importance of long-term April 1 snow-water-equivalent (SWE) data for predicting streamflows for the upcoming spring and summer. I also showed that the long-term mean (or arithmetic average) of SWE is greater than the current 1981-2010 median value to which current SWE values are compared. The primary observation to make was that although April 1 SWE at many sites in the upper Henry's Fork watershed was above the 1981-2010 median, it was actually below the long-term mean at most sites, leading to below-average predictions of spring/summer streamflow. The percentages shown above are relative to the 1972-2016 mean for the Henry's Fork and the 1981-2016 mean for Fall River and the Teton River.
The table above provides predictions of April 1 through September 30 natural streamflow in the Henry's Fork watershed, based on April 1 SWE values and on baseflows from last winter. Baseflow is defined as streamflow during the winter, when streamflow is primarily maintained by water stored in soils and aquifers. Methods and interpretation are given in my March 10 blog, where I predicted April-June streamflows based on early March conditions. As percent of average, the April-September predictions shown in the table above are not much different from the April-June predictions made back in March. Henry's Lake is a notable exception. The prediction here is quite a bit lower than that made in March, primarily because baseflows calculated in March were based on preliminary streamflow data. Updating the baseflow calculations with recently published streamflow data approved by USGS reduced the estimate of baseflows and hence the prediction. The prediction for the Teton River is lower than it was back in March because a substantial amount of low- and mid-elevation snow melted during the month of March, contributing to record streamflows in the Teton River during March but subtracting from streamflow that would have occurred later in the spring and summer had that snow not melted.
In the table above, remember that the 90% exceedance values are those that will be exceeded with 90% probability, given the SWE and baseflow conditions observed on April 1. In particular, there is a 90% chance that streamflow this spring and summer will be greater than last year, except at Henry's Lake.
The predictive models of streamflow based on April 1 data are quite a bit better in their predictive power, as measured by the R2 value, which is defined as the percent of total variability in the data explained by the predictive model. This is discussed more fully in the March 10 blog. This is intuitive, since SWE on April 1 is very close to its peak value for the season, whereas March 1 SWE values represent only a fraction of the season-total snow accumulation. R2 values for the March predictions ranged from 56.6% at Henry's Lake to 74.0% in the Henry's Fork between Henry's Lake and Island Park. Compare those to the R2 values in the table below for the April predictions.
To translate these natural streamflow predictions into predictions for actual streamflows, we need to incorporate the effects of reservoir storage and delivery, irrigation demand, and timing of runoff into the equation. We are working on a comprehnsive model of reservoir operations and irrigation delivery for the whole watershed and will have some predictions for the regulated system within a week or two.