Performance of a simple proxy for U.S. cloud-to-ground lightning

TL;DR

This study evaluates a simple CAPE-precipitation proxy's effectiveness in representing U.S. cloud-to-ground lightning patterns, finding it better at short-term scales and during cooler seasons, with potential for future lightning prediction.

Contribution

It provides an assessment of the CP proxy's performance over the U.S. from 2003 to 2016, highlighting its strengths and limitations across different time scales and seasons.

Findings

Proxy performs better on daily and monthly scales.

Worse performance during warm season (May-October).

Good correlation with negative CG flashes in cool season.

Abstract

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 U.S. (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 ($\sim$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.