Fig. 5: Relationships between SWE and potential improvements to water supply prediction when expanding measurements in different station networks.
These 2-d histograms summarize the relative frequency of pixels from all study basins (n = 390) and forecast dates (n = 4) by percentile values for ΔR2 versus SWE for two scenarios for expanded SWE monitoring: a no existing stations, and b multiple existing stations. For each axis, percentile values are computed relative to the spatial distribution for each basin and forecast date. The histograms are normalized by basin size to avoid placing greater weight on larger basins. We define hotspots as having ΔR2 in the 99th percentile; these generally have higher SWE values relative to other locations in a basin for both network scenarios.
