High Elevation Precipitation Uncertainty and Inference of Sierra Nevada Precipitation Patterns from Streamflow Observations

As is the case in most mountainous areas, observing precipitation across high-elevation basins of the Sierra Nevada of California is challenging, due to a relatively sparse network of precipitation gauges at these altitudes. Gridded datasets that estimate precipitation across the landscape likewise show larger uncertainties in mountainous areas, often complicating efforts to model hydrological processes. However, numerous observations of streamflow are also made in the Sierra, which offer additional information about the water balance of the basin with which to estimate precipitation. Here, we consider unimpaired streamflow observations from a dataset of >100 current and historic stream gauges across the Sierra from the mid-20th century to the present. We apply a Bayesian methodology for inferring basin-mean precipitation from these streamflow observations. To do this, we calibrate semi-lumped hydrologic models representing each basin, using streamflow to infer multiplicative corrections to gridded precipitation forcing data. We compare the resulting streamflow-based estimates of precipitation against widely-used precipitation gauge-based gridded datasets. From this analysis, we draw the following conclusions regarding Sierra Nevada precipitation patterns: 1) precipitation inferred from streamflow exhibits greater spatial variability than that seen in gridded precipitation datasets, with greater precipitation maxima on windward (southwest) facing basins and larger spatial gradients, 2) year-to-year variability is greater in streamflow-based estimates of precipitation than in gridded datasets, and 3) high-elevation trends in streamflow-based  precipitation datasets differ from trends in gauge-based gridded datasets. These findings suggest that the incorporation of streamflow into precipitation datasets may improve representation of high-elevation hydrologic processes and trends.