The Influence of Patterned Warming and CO_{2<\sub> on North Atlantic Hurricane Frequency}

Modeling annual North Atlantic hurricane frequency and predicting its response to climate change are of great societal interest. Prior research has found that low-order Poisson regression models using North Atlantic (local) and tropical-average (remote) sea surface temperatures (SSTs) can explain North Atlantic hurricane frequencies. Remote SSTs are thought to influence hurricane genesis by affecting free-tropospheric temperatures. However, the main mechanism by which the surface can affect free-tropospheric temperatures is through deep convection, which is collocated with just the warmest tropical SSTs. Here, we propose and show evidence of a physical mechanism through which these warmest SSTs influence hurricane frequencies, with warmer warmest SSTs reducing hurricane frequency. Using thousands of years of tropical cyclone–permitting atmospheric model simulations with various uniformly and nonuniformly perturbed SST fields, we train statistical models to predict annual North Atlantic hurricane counts. We show that an index characterizing the warmest tropical waters is a superior predictor of North Atlantic hurricane activity compared to tropical-average SSTs. Models based on tropical averages fail to predict even the sign of frequency changes under strong patterned warming scenarios. Additionally, higher CO_{2<\sub> concentrations independent of SSTs decrease hurricane frequencies. We apply this statistical model to project hurricane frequency changes through the late twenty-first century using Coupled Model Intercomparison Project phase 6 (CMIP6) SST projections, identifying a decreasing trend through 2100. Using tropical-average SSTs instead of an index for the warmest waters yields different predictions. Our findings emphasize that patterned warming significantly influences North Atlantic hurricane frequencies. This patterned warming is especially important insofar as the warmest tropical SSTs warm disproportionately relative to local SSTs.}