Exploratory Usage of Global WRF for Ensemble Tropical Storm Simulations

Atlantic tropical cyclones have the potential to wreck havoc along the east coast of the United States. Accurate prediction of their occurrences is crucial for our preparedness against their impending arrivals. In 2013, based on various model predictions, we expected the Atlantic hurricane season to be highly active. However, the observed Atlantic tropical cyclone activity was actually well below average, with a total of only 14 named storms.

In this study, we explore the usage of GlobalWRF to provide an initial understanding of why the seasonal forecast differs significantly from reality in 2013. In particular, we investigate the effects of sea surface temperature (SST) on the amount of Atlantic tropical cyclone activity. Experiments were performed using 2 ensemble seasonal runs of GlobalWRF (initialized on 1 August 2013 using the Global Forecast System analyses) with a nested domain over the Atlantic Ocean Basin. The global SST is fixed in one ensemble run and variable in the other. Each ensemble run is made up of 20 members, each initialized with random atmospheric disturbances. For each ensemble member, vortex tracking was performed over the Atlantic Ocean Basin to detect the locus where the lowest pressure, maximum wind speed, and maximum vorticity near the surface occur concurrently. Assumed to be the center of the tropical cyclone, this locus is then tracked in space and time. Base on the tracked information, the total number of tropical cyclones and storm intensities are found for that specific ensemble member.

For the varying SST ensemble run, the probability distribution of the number of cyclones reveal that the most probable number of tropical cyclones to be 10-14. While this result is consistent with the observed 14 named storms of the 2013 hurricane season, it is somewhat sensitive to the criteria imposed on the longevity of the tropical cyclone lifecycle in the tracking process. Currently, the fixed SST ensemble run is being finalized. We will compare the resulting probability distribution of the number of cyclones with the varying SST run. We are also in the process of obtaining the probability distribution of tropical cyclone strength for each ensemble run.