B. Drought Prediction
a. Drought prediction based on forecasted Standardized Precipitation Indices
Many international experts and representatives from national meteorological and hydrological services around the world have identified the SPI as a universal meteorological drought index for effective drought monitoring and climate risk management. A simple method has been developed at the NCEP CPC to operationally forecast three month and six month Standardized Precipitation Indices (SPIs) for the prediction of meteorological drought over the contiguous United States. The methodology is based on precipitation (P) seasonal forecasts from the NCEP Climate Forecast System, version 2 (CFSv2). Before predicting SPI, the P forecasts from the coarse resolution CFS global model are corrected for biases and downscaled to higher resolution, based on the probability distribution function method, to a regional grid of 50 km. The downscaled and bias-corrected CFSv2 P forecasts, out to 9 months, are appended to the P analyses to form an extended P data set. The SPIs are then calculated from this new time series. The skill is regionally and seasonally dependent, however overall, the six month SPI is skillful out to three or months.
NCEP CPC operationally implemented the SPI three month (SPI3) and six month (SPI6) forecasts in April 2011. Figure 3 shows the forecasts of SPI3 and SPI6 for July-September 2012. It shows that drought over the western interior states and the area from Mississippi basin to Illinois is likely to continue for two to three months.
Figure 3: SPI forecasts initialized on July 2 and 3 2012 for (a) SPI6 one month lead, (b) two month lead, (c) three month lead, (d) SPI3 one-month lead. Contours are indicated by the color bar.
b. National Multi-Model Prediction Experiment for drought prediction.
Research has shown that climate prediction based on an ensemble of multiple climate models is generally more accurate that that from any single model. Hence, as part of its Climate Test Bed research efforts, NOAA's MAPP program has long invested in research to develop methodologies that optimally combine predictions for multiple models. These efforts have been accompanied by similarly important research efforts to fundamentally improve NOAA operational climate forecast system (manifested in the June 2010 release of the latest version of the Climate Forecast System).
More recently, the NOAA MAPP Program, in partnership with the Climate Test Bed, and with contributions from NSF, DOE, and NASA has initiated a research project, named the National Multi-Model Ensemble (NMME) Experiment to seek further improvements in intraseasonal to seasonal climate and drought prediction. The NMME project is utilizing an ensemble of leading national climate models in near-operational mode to provide climate forecasts for research purposes since August 2011. The NMME project includes a comprehensive research investigation regarding the optimal design and added value of this multi-model predictive system for climate and drought prediction. Figure 4 shows preliminary results from this multi-model system for precipitation and surface air temperature compared to results from single systems. In agreement with previous research, the NMME has predictive ability than that of any individual model. Although the NMME system is still an experiment, NCEP seasonal forecasters are already using its predictions as one of the inputs to producing official operational seasonal climate forecasts.
Figure 4: Percentages of skillful forecasts (Continuous Ranked Probability Skill Score: CRPSS>0) for seasonal mean precipitation and surface air temperature anomalies over global land areas. N1-N6 are individual models from NMME, and E1-E5 are from ENSEMBLES. (Yuan and Wood, submitted)
--- Content provided by Kingtse Mo, NCEP/CPC ---