One of the research priority areas in the GEWEX Americas Prediction Project (GAPP) science plan is to determine the extent of land surface memory processes and their contribution to precipitation predictability at seasonal to interannual timescales.
This relates to improving the predictability of the hydrological cycle with special regards to the land surface and the role of predictions for water resources management. There is the potential to utilize NCEP's Climate Forecast System (CFS) seasonal forecast model, using the North American Land Data Assimilation System (NLDAS) to provide initial soil moisture states, that should lead to improved summertime precipitation forecasts. Currently CFS is initialized only using tropical ocean sea surface temperatures (SST). Related to these issues, especially from a water management perspective, is the predictability of the extent and persistence of drought. But any improvement in precipitation prediction skill at seasonal to interannual timescales requires that the following unresolved problems be addressed:
- What is the linkage between the patterns of soil moisture anomalies and precipitation anomalies both in space and time as predicted from the NCEP Eta model, and are they different from analyses made from soil moisture observation proxies (i.e. from off-line observationally-forced LSMs) or other seasonal climate models?
- How can the strength of any `linkage' between soil moisture and precipitation anomalies be estimated, and how does it change for different seasons and regions?
- Can we understand the mechanism and the pathways that connect soil moisture anomalies toprecipitation anomalies at the relevant time lags, and for different regions?
- Do improved predictions in precipitation, if found, lead to improved streamflow predictions, and can these improvements be quantified?
These questions are critical in understanding precipitation predictability and are central in developing improved precipitation predictions at different timescales
The project objectives are as follows:
- To search for possible teleconnection patterns between soil moisture and precipitation over the continental U.S. by using a suite of observational and modeling products that include NLDAS off-line land surface model output, observed precipitation and the new NCEP regional reanalysis model output.
- To identify "sensitive" areas and periods of soil moisture anomalies that can be used as an indicator or predictor for precipitation forecast.
- To investigate the physical processes related to these teleconnection patterns, and linkage between soil moisture anomalies and precipitation predictability, through detailed diagnostic analyses of water vapor fluxes, recycling processes, temperature and humidity fields and similar variables from the regional reanalysis output.
- To quantify the improvements in river discharge forecast skill from enhanced precipitation forecasting through teleconnections between soil moisture and precipitation.