Relating the Frequency of Tropical Cyclogenesis Events to the Scale of Precursor Disturbances in Aquaplanet Simulations with Different Horizontal Grid Spacings

Climate projections indicate uncertainty regarding how the frequency of tropical cyclones (TCs) might change amid a warmer climate. Previous studies have further noted substantial variability in projections of TC frequency as a function of the horizontal grid spacing in general circulation models. The frequency of TCs is posited to be inextricably connected to variability among the multi-scale processes that constitute cyclogenesis. Consequently, projections of TC frequency derived from explicit simulations of TCs will depend on how models represent variability among the multi-scale processes that constitute genesis. As a first step towards understanding variability among genesis events in the context of the Earth’s warming climate system, intercomparisons are presented between simulations of genesis by varying the horizontal grid spacing in an aquaplanet modeling framework using the Model for Prediction Across Scales-Atmosphere (MPAS–A). Four simulations are constructed with quasi-uniform horizontal grid spacings—120 km, 60 km, 30 km, and 15 km—and each simulation is integrated for three years. Tropical precursor disturbances to genesis are identified and objectively tracked with the TRACK algorithm.

Both the number of viable precursor disturbances to genesis and the number of those disturbances that undergo genesis decrease monotonically with increasing horizontal grid spacing, such that the ratio of genesis events among the four simulations is approximately conserved. The sensitivity of this relationship is examined by varying the time duration required for a disturbance to exhibit vertical coherence and a warm core in TRACK. Shear-relative composite-mean analyses are created to assess structural differences between disturbances in the simulations. Contrasting genesis events between the 120-km and 15-km simulations reveals that genesis events in the 120-km simulation are characterized by weaker boundary layer convergence; weaker updraft mass fluxes throughout much of the troposphere; and broader circulations. To further quantify variability in the scale of the precursor disturbances undergoing genesis among the four simulations, distributions of vortex size are analyzed. Results indicate that the vortex size distribution monotonically shifts towards larger sizes and less (positive) skew with increasing grid spacing. A simple argument is explored relating the variability in the number of genesis events to the vortex size distributions. That is, the minimum vortex size is primarily constrained by the model horizontal grid spacing as the scale of resolved features within a circulation diminishes with coarser grid spacing. The maximum vortex size is universally limited by wave–vortex dynamics that contract vortices towards their intrinsic Rhines scale. Therefore, the simple argument would indicate fewer TCs with coarser grid spacings by virtue of these simulations resolving fewer relatively small TCs.