Quantifying the sensitivity of maximum, limiting, and potential tropical cyclone intensity to SST: Observations versus the FSU/COAPS global climate model

Abstract

Previous research quantified the sensitivity of limiting intensity to SST for observed tropical cyclones (TCs) and for TCs generated by two global climate models (GCMs). On average, a 1 C increase in sea surface temperature (SST) is associated with a 7.9 m s 21 increase in the statistical upper limit of observed intensity. Conversely, a 1 C increase in SST does not significantly affect the limiting intensity of GCM-generated TCs. The study presented here builds on previous research in two ways: (1) A comparison is made between the statistically defined limiting intensity and the physically defined potential intensity, and (2) a test is performed on the ability of a 0.94 resolution GCM to reproduce the observed statistical relationship between potential intensity and SST. Data from NASA's Modern Era Reanalysis are used to approximate the observed sensitivity of potential intensity to SST for the 1982-2008 time period. Results indicate that the sensitivity of potential intensity to SST is not statistically different from the sensitivity of observed maximum or limiting intensity to SST. This result links the statistically defined sensitivity to the physically based theory of hurricanes. Potential intensity is also estimated from the FSU/COAPS GCM. Although the FSU/COAPS model does not capture the observed sensitivity of TC maximum or limiting intensity to SST, the model reproduces the observed sensitivity of potential intensity to SST. The model generates suitable atmospheric conditions for the development of strong TCs, however strong TCs do not develop, possibly as a result of insufficient resolution.

The influence of model resolution on the simulated sensitivity of North Atlantic tropical cyclone maximum intensity to sea surface temperature

“…The FSU-COAPS model successfully simulates the thermodynamic environment needed to sustain intense TCs, but does not actually generate intense TCs. Strazzo et al [2015] speculate that this occurs partly because model grid spacing is too coarse to resolve the inner core and thus the maximum wind speed of simulated TCs. Strazzo et al [2015] also find that the sensitivity of observed maximum TC intensity to SST is not statistically different from the sensitivity of observed potential intensity to SST.…”

“…Strazzo et al [2015] speculate that this occurs partly because model grid spacing is too coarse to resolve the inner core and thus the maximum wind speed of simulated TCs. Strazzo et al [2015] also find that the sensitivity of observed maximum TC intensity to SST is not statistically different from the sensitivity of observed potential intensity to SST. If model resolution explains the inability of the FSU-COAPS model to simulate the sensitivity of TC maximum intensity to SST, we might expect that the sensitivity of simulated maximum intensity to SST will approach the sensitivity of simulated potential intensity to SST as model resolution increases.…”

“…One of the primary objectives of this research is to understand whether GCMs that generate TCs with intensities exceeding 60 m s 21 are better able to reproduce the sensitivity of maximum TC intensity to SST compared to GCMs that do not generate strong TCs. As in and Strazzo et al [2013bStrazzo et al [ , 2015, we use a spatial tessellation approach to spatially bin TC and SST data for statistical analysis. We begin by tessellating the North Atlantic basin into equal-area hexagons-hereafter referred to as ''regions''-and subsequently overlaying all TC track and SST data onto this set of regions.…”

“…This allows us to test whether the ability of a GCM to generate strong TCs influences the sensitivity of simulated TC intensity to SST. Similarly, we extend the results of Strazzo et al [2015] to determine whether the difference between the sensitivity of potential and maximum intensity to SST decreases as model resolution improves. Finally, we address concerns about uncertainty by estimating the error associated with the spatial tessellation methodology.…”

“…High-resolution models (i.e., models with horizontal grid spacing near 0.258) may be able to simulate TCs with intensity distributions that resemble what we observe, but do the strongest model-generated TCs form over the warmest SSTs, as is the case for observed TCs [Elsner et al, 2012b]? Strazzo et al [2015] demonstrate the ability of the Florida State University-Center for Ocean-Atmospheric Prediction Studies (FSU-COAPS) model (0.948) to reproduce the relationship between the theoretically defined potential intensity and SST despite that model's inability to simulate the observed sensitivity of maximum intensity to SST. The FSU-COAPS model successfully simulates the thermodynamic environment needed to sustain intense TCs, but does not actually generate intense TCs.…”