The Relationship Between Tropical SST Structure and Systematic Rainfall Errors in Global Models

A strong association between ordinary mesoscale SST gradients and the occurrence of tropical rainfall has been observed from GHRSST and CMORPH data in the eastern hemisphere (Li and Carbone, 2012). Under slow wind speed conditions, lower boundary forcing results from the convergent Laplacian of SST (LSSTc), which induces ascent in the marine atmospheric boundary layer, MABL. When such forcing exceeds convective inhibition (CIN) in the lower troposphere, moist convection can be triggered, the depth and intensity of which is dependent on deep tropospheric convective available potential energy, CAPE. Averaged over long timeseries and large areas of ocean, LSSTc is equally divided between convergent and divergent mesoscale dipoles, yet 3 of 4 rainfall events have their origin over LSSTc in the tropical western Pacific, and largely independent of SST itself. Over tropical oceans, this correlation with rainfall event frequency and cumulative rainfall is especially prominent in the active phase of MJOs and the eastern hemisphere more generally, as revealed by analyses of CMORPH and GHRSST timeseries (Carbone and Li, 2015). Parameterizations of tropical convection in global models are heavily influenced by SST. However, the influence of mesoscale LSSTc forcing is rarely, if ever, represented in highly parameterized global models, potentially leading to systematic rainfall errors. We will present preliminary results, which attempt to determine if there is a statistical relationship between known systematic errors in the NCAR CESM model and observed regional/seasonal variability in mesoscale LSSTc forcing.