Performance of a simple reanalysis proxy for US cloud-to-ground lightning

The product of convective available potential energy (CAPE) and precipitation rate has previously been used as a proxy for cloud-to-ground (CG) lightning flash counts in climate change applications. Here, the ability of this proxy, denoted CP, to represent the climatology and variability of CG lightning flash counts over the contiguous United States (CONUS) during the period 2003-2016 is assessed. CP values computed using the North American Regional Reanalysis are compared with negative and positive polarity CG flash counts from the National Lightning Detection Network. Overall, the proxy performs better on shorter time scales (daily and monthly) than on longer time scales (annual and semi-annual). Proxy performance tends to be worse during the warm season (May-October), when most lightning occurs, and better during the cool season (November-April). The correlation of annually accumulated CONUS CP with CG flash counts is not statistically significant because of poor warm-season performance. Cool season negative CG flash counts are well correlated with CONUS CP values. Positive CG flash counts (similar to 7% of all CG flashes) are well correlated with annual values of CONUS CP. The relatively strong relations between CP and CG flash counts in some regions and times of the year at daily resolution provide a benchmark for more complex proxies and suggest that proxy-based extended- and long-range prediction of lightning activity may be feasible to the extent that precipitation rate and CAPE can be predicted.