Sub-seasonal Predictability of a Genesis Potential Index

Sub-seasonal Predictability of a Genesis Potential Index

Rodrigo J. Bombardi ¹

Laurie Trenary ²

Kerry Emanuel ³

¹Dept. of Geography, Texas A&M University

²Dept. of Atmospheric, Oceanic, and Earth Sciences, George Mason University

³Dept or Earth, Atmospheric, and Planetary Sciences. Massachusetts Institute of Technology

Tropical cyclone genesis has important implications to society and the environment. Recent studies have investigated the variability of tropical cyclone genesis through the study of Genesis Potential Indices. The objective of this work is to evaluate the predictability of a Genesis Potential Index (GPI) on sub-seasonal timescales. We analyze retrospective forecasts (hindcasts) from four operational models included in the Sub-seasonal to Seasonal (S2S) project. The hindcasts initializations vary from model to model and span the period of 1994 through 2014. Our results show that some models (Fig. 1) show skill in predicting monthly GPI on lead-times starting on week 3 (i.e. comprising lead-times between weeks 3 and 6). We also investigate the predictability of large-scale GPI patterns over the North Atlantic Ocean using Predictable Component Analysis (PrCA). Large-scale GPI patterns are also predictable in some operational models at sub-seasonal timescales.

Fig 1. GPI Anomaly correlation (compared against MERRA2) for lead-times between weeks 3 and 6 and four different models. The hindcasts were initialized in July.