Assessing the predictability of Tropical Storm Erika (2009) through an analysis of environmental factors

In this study, the predictability of Tropical Storm Erika is evaluated by analyzing a 60-member ensemble WRF simulation. As in Sippel and Zhang (2009), an Ensemble Kalman Filter (EnKF) is utilized to introduce perturbations into the initial model fields, which leads to a fairly large spread in the final intensities of the 60 ensemble members. Tropical Storm Erika was forecast to intensify into a hurricane by most operational models, but in reality never exceeded 50 kt. An investigation into which factors prevented intensification of the weaker ensemble members provides insight that may aid in the forecasting of the intensity of future tropical cyclones under similar conditions.

A variety of environmental factors are examined and the most important factors are determined based on correlation analysis. It appears that mid-level relative humidity, the distribution of convection relative to the storm center, and the deep layer shear all play a role in determining whether a given ensemble member intensifies or not. It is further shown that many of the ensemble members that do not intensify fail to do so because of apparent dry air intrusions, particularly in the 700-500 mb layer, that wrap around the centers of the storms. Because of the moderate shear, this dry air (which appears to be both environmental and shear-induced in origin) is able to penetrate the core of the cyclones.

A relationship between the direction of the shear vector, the direction of the center tilt vector, and the distribution of convection is also identified. Consistent with previous studies, deep convection is concentrated in the downshear left quadrant in a large percentage of ensemble members. Additionally, the tilt vector is generally oriented about ninety degrees to the left of the shear vector, consistent with past studies on this subject.