Tropical forests are disproportionately important to the global carbon budget, and large tree mortality influences tropical forest carbon stocks. Large tree mortality also influences population structure as uncommon large trees contribute a substantial proportion of total population fecundity. However, information about the causes of large tree mortality is limited. We combined surveys of lightning strikes from a real-time monitoring system with large-scale forest inventories and satellite data to quantify the landscape and intercontinental distribution of lightning-caused disturbance in the tropics. Specifically, we test two hypotheses: 1) the likelihood of being struck by lightning varies predictably with tree structural characteristics and 2) rates of lightning-caused disturbance, estimated as satellite-detected flashes, differ among ecosystems and increase with greater woody vegetation complexity.
Results/Conclusions
The probability of being directly struck by lightning increases with greater tree crown area and relative height, and the likelihood of secondary lightning damage increases with shorter inter-crown distances. Combining these empirical models, we constructed a lightning risk model for forest trees that indicated rates of lightning damage differ among species and are concentrated in areas of high carbon storage. Lightning was most frequent in tropical forests, although it varies substantially among regions, and we estimate that 1.38 billion tropical trees are damaged by lightning annually. Consequently, within and among forest patterns of lightning damage likely shape tropical forest dynamics and carbon stocks. Apart from forests, our findings indicate that lightning-caused disturbance is similarly important in savannas and has substantial undocumented effects in other ecosystems.