Evaluation of real-time hydrometeorological ensemble prediction on hydrologic scales in Northern California

•Integrated Forecast and Reservoir Management (INFORM) demonstration project.•Dynamic downscaling of NCEP GEFS and CFS ensemble forecasts (6h to 41.5days).•Precipitation, temperature and reservoir inflow ensemble forecasts.•Good performance for lead times of 1–2weeks or longer (in some cases).•Adjustment of bias in the ensemble precipitation forecasts improves performance. The paper presents an evaluation of real time ensemble forecasts produced during 2010–2012 by the demonstration project INFORM (Integrated Forecast and Reservoir Management) in Northern California. In addition, the innovative elements of the forecast component of the INFORM project are highlighted. The forecast component is designed to dynamically downscale operational multi-lead ensemble forecasts from the Global Ensemble Forecast System (GEFS) and the Climate Forecast system (CFS) of the National Centers of Environmental Prediction (NCEP), and to use adaptations of the operational hydrologic models of the US National Weather Service California Nevada River Forecast Center to provide ensemble reservoir inflow forecasts in real time. A full-physics 10-km resolution (10km on the side) mesoscale model was implemented for the ensemble prediction of surface precipitation and temperature over the domain of Northern California with lead times out to 16days with 6-hourly temporal resolution. An intermediate complexity regional model with a 10km resolution was implemented to downscale the NCEP CFS ensemble forecasts for lead times out to 41.5days. Methodologies for precipitation and temperature model forecast adjustment to comply with the corresponding observations were formulated and tested as regards their effectiveness for improving the ensemble predictions of these two variables and also for improving reservoir inflow forecasts. The evaluation is done using the real time databases of INFORM and concerns the snow accumulation and melt seasons. Performance is measured by metrics that range from those that use forecast means to those that use the entire forecast ensemble. The results show very good skill in forecasting precipitation and temperature over the subcatchments of the INFORM domain out to a week in advance for all basins, models and seasons. For temperature, in some cases, non-negligible skill has been obtained out to four weeks for the melt season. Reservoir inflow forecasts exhibit also good skill for the shorter lead-times out to a week or so, and provide a good quantitative basis