Sub-Seasonal Forecasting of Precipitation and Soil Moisture in CFSv2 Near Gulf Coast Tropical Cyclone Landfall Locations

Abstract

Landfalling tropical cyclones (TCs) produce intense winds, heavy precipitation, and damaging storm surges along the Gulf and East coast of the U.S. There is growing evidence that high antecedent soil moisture (SM) conditions preceding landfall can maintain or even intensify TCs over land, sustaining strong winds and precipitation further inland. In this study, we investigate the ability of the 9-month Climate Forecast System Version 2 (CFSv2) operational forecast to represent near surface SM at forecast lead times (LTs) from hours to weeks over the Gulf coast TC landfalling region. Our results show that the decrease in the predictability of SM and associated precipitation with LTs not only appears at relatively short timescales (e.g., weekly or shorter events), but also evident in the interannual variability and long term (7-year) mean climatology. Moreover, the spatial biases increase with the LT at about one to 2 weeks and act to damp the spatial variability. Nevertheless, even at shorter LTs (i.e., hours to days), the forecasted precipitation fails to capture the more intense events, which contributes to biases in the SM at all LTs. Our results assess the relationship between the predictability of the SM and the related precipitation with LTs. The results suggest that CFSv2 is unlikely to capture the influence of antecedent SM variability associated with preceding convective events on subsequent events (e.g., evaluation of TCs after landfall) on LTs longer than a week, which has implications for overall subseasonal-to-seasonal predictability.