Using synoptic-scale land surface moisture indices for hydrologic prediction

Forecasting seasonal runoff is an important challenge in the western U.S. because the timing and volume of summer (dry season) streamflow plays a critical role in managing water supply and delivery systems. Beginning in December of each year, observations of the snow pack begin to provide invaluable information on the amount of runoff expected in late spring and summer. The skill of such water supply forecasts also benefits, in some locations, from the consideration of synoptic climate indices -- e.g., using sea surface temperature anomalies to characterize ENSO or PDO state, which have been shown to have predictable teleconnections to land surface hydrology months in advance. These two sources of forecast information complement each other, with climate/SST state information providing predictability at long lead times (3-9 months), and observed snow and other moisture states of local catchments providing skill at shorter lead times (1-4 months). However, because similar SST anomalies can produce widely different future hydrology outcomes, information at intermediate lead times describing how a particular teleconnection or climate pattern is evolving on a macro-scale level (especially regarding precipitation anomalies) may provide additional insight into future spring and summer runoff. In this study, we explore the use of a long-term land surface hydrology data set to define continentalscale hydrologic indices with potential predictive value, and to assess where in the western U.S. these might enhance predictability already achievable using local observations and climate indices. We find that the west-wide hydrologic indices have potential value in the Pacific Northwest and the Southwest, but yield little improvement in other areas of the western U.S., relative to existing climate and observation-based indices