Development of a Convection Risk Index to forecast severe weather, and application to predict maximum wind speeds

Abstract:

We have developed a tool, the Convection Risk Index (CRI), to represent the severity, timing and location of convection for select geographic areas. The CRI is calculated from the Convection Risk Matrix (CRM), a tabulation of numerous meteorological parameters which are categorized into four broad factors that contribute to convection (surface and lower level moisture, atmospheric instability, vertical wind shear, and lift); each of these factors have historically been utilized by meteorologists to predict the likelihood for development of thunderstorms. The CRM ascribes a specific threshold value to each parameter in such a way that it creates a unique tool used to calculate the risk for seeing the development of thunderstorms. The parameters were combined using a weighted formula and which when calculated, yields the Convection Risk Index 1 to 4 scale, with 4 being the highest risk for seeing strong convection. In addition, we also evaluated the performance of the parameters in the CRM and CRI for predicting the maximum wind speed in areas where we calculated the CRI using nonparametric tree-based model, Bayesian additive trees (BART). The use of the CRI and the predicted wind speeds from BART can be used to better inform emergency preparedness efforts in government and industry.We have developed a tool, the Convection Risk Index (CRI), to represent the severity, timing and location of convection for select geographic areas. The CRI is calculated from the Convection Risk Matrix (CRM), a tabulation of numerous meteorological parameters which are categorized into four broad factors that contribute to convection (surface and lower level moisture, atmospheric instability, vertical wind shear, and lift); each of these factors have historically been utilized by meteorologists to predict the likelihood for development of thunderstorms. The CRM ascribes a specific threshold value to each parameter in such a way that it creates a unique tool used to calculate the risk for seeing the development of thunderstorms. The parameters were combined using a weighted formula and which when calculated, yields the Convection Risk Index 1 to 4 scale, with 4 being the highest risk for seeing strong convection. In addition, we also evaluated the performance of the parameters in the CRM and CRI for predicting the maximum wind speed in areas where we calculated the CRI using nonparametric tree-based model, Bayesian additive trees (BART). The use of the CRI and the predicted wind speeds from BART can be used to better inform emergency preparedness efforts in government and industry.