Subseasonal to Seasonal Prediction of Extratropical Storm Track Activity Over the U.S. Using NMME Data

The extratropical storm tracks are regions over which mid-latitude cyclones tend to frequent. Cyclones propagating along the storm tracks provide much of the high impact weather over the mid-latitudes, including heavy precipitation, high wind, and extreme cold events, hence skillful forecasts of storm track activity will provide policy makers, emergency management, and stakeholders advanced warning to prepare for mitigation measures. The North American Multi-Model Ensemble system (NMME) is an experimental but operational multimodel climate prediction ensemble system made up of 8 state-of-the-art climate prediction models, and is supported by various U.S. and Canadian agencies including the Department of Energy (DOE). Previous applications of the NMME have focused on seasonal forecasts of the large scale circulation since only monthly mean data were available, but with phase 2 of NMME, high frequency (daily) data are becoming available and storm track activity predicted by the models can be derived from the high frequency model output. This project proposes to develop a new product for subseasonal to seasonal prediction of storm track activity over the U.S. and its vicinity using NMME and reanalysis data, and to assess the skill of this product. Preliminary assessment of storm track prediction by the PI’s research group using output from the National Center for Environmental Prediction (NCEP) Climate Forecasting System version 2 (CFSv2, one of the 8 models participating in NMME) reforecast data have suggested that the model has substantial skills in predicting storm track variability. These preliminary results also suggest that aggregating the CFS forecasts from different initial start times to form ensemble averages provide significant enhancement to the forecast skill. These results suggest that aggregating forecasts from different models from the same initial time may further enhance the skill of the forecasts, hence our hypothesis is that NMME ensemble forecasts should be more skillful than forecasts based only on CFSv2 alone. In addition, bias correction models will be developed using predicted and analyzed storm track anomaly fields during the training period using reforecast data to further enhance the forecast skill. Efforts will also be made to further investigate how variations in storm track activity modulate the frequency of occurrence of extreme weather events, and to understand the physical processes that give rise to the predictability of storm track variability. This project assesses the skill of NMME prediction of storm track activity, and develops a novel application of NMME data to provide subseasonal to seasonal prediction of storm tracks, hence this proposal contributes to both area A (evaluation of NMME system predictions) and area B (exploration of new applications of NMME system predictions) of the competition. The proposed research is clearly relevant to the goals of the Climate and Environmental Sciences Division (CESD) of the DOE’s Office of Biological and Environmental Research to “Synthesize new process knowledge and innovative computational methods advancing next-generation, integrated models of the human-Earth system”, and “Develop, test, and simulate process-level understanding of atmospheric systems and terrestrial ecosystems, extending from bedrock to the top of the vegetative canopy”.  More specifically, the proposed research is clearly relevant to the goals of the Regional and Global Climate Modelling Program under CESD including “diagnosing and analyzing state-of-the-science coupled climate and earth system models to understand climate variability and change at regional and global scales”, and to “develop metrics for model validation, to conduct climate analysis research in order to inform the model development strategies of ESM”.