Impact of resolution and downscaling technique in simulating Atlantic tropical cyclone activity with GEM-Climate

In this presentation, we discuss the ability of GEM, a Canadian Atmospherics Climate Model, to run in a number of different configurations, to simulate the observed variability in Atlantic Tropical Cyclones over the recent past. Specifically, we have run GEM in a number of configurations, using observed Sea Surface Temperatures (SSTs) and sea-ice surface boundary conditions interpolated from the Atmospheric Model Intercomparison Project v2 (AMIP2) for the period 1979-2006. The different configurations include (i) a 2º Global, regular resolution version of GEM, ii) a 1º Global, regular resolution version of GEM, (iii) a Global Variable resolution version of GEM, where the area of high resolution (0.3º) covers the tropical Atlantic and all of North Africa and the remaining part of the Globe is covered at 2º resolution. (iv) 2 Limited-Area (LAM) domains of GEM, one covering only the tropical Atlantic and the other both the tropical Atlantic and the entire North Africa. These two LAM domains are both run at 0.3º resolution and are forced by results from integration (i) and (ii), namely output from the Global regular resolution integrations. (iv) 2 LAM versions of GEM, employing the exact same resolution and domain coverage as the LAMs in (iv), but employing ECMWF Reanalysis data (ERA-40 and ERA-Interim) as lateral boundary forcing, rather than output from the 2º GEM-Global run.

By intercomparing the ability of each of these configurations to simulate the observed number, geographical distribution, intensity distribution, seasonal and interannual variability of Atlantic tropical cyclones as well as large-scale fields know to influence tropical cyclone activity, we can begin to understand the importance of simulating Atlantic Tropical Cyclones of factors such as: (i) Model resolution local to the Atlantic Basin, (ii) Treatment of LAM lateral boundary conditions compared to the open-boundary, Variable Global Model approach, (iii) Importance of the inclusion of a high-resolution representation of the upstream African Easterly Wave (AEW) track in LAM configurations, versus AEWs being represented on the LAM boundaries as derived from the 2º GCM run (iv) the importance of employing ERA40 versus GEM lateral boundary conditions.

In this presentation, we will present an overview of the performance of the Global Variable Resolution version of CRCM5 in simulating the climatology and variability of Atlantic Tropical Cyclones. The range of other model configurations will then be compared and contrasted to address the list of factors outlined above.