Spectral retrieval of latent heating profiles from TRMM PR Data: Comparisons of lookup tables from two- and three-dimensional simulations

The Spectral Latent Heating (SLH) algorithm has been developed to estimate Q1 and Q2 profiles for the TRMM PR (Shige, Takayabu et al. 2004, 2007a, 2007b). The method uses PR information [precipitation top height (PTH), precipitation rates at the surface and melting level, and rain type] to select Q1 and Q2 profiles from lookup tables. Lookup tables for the three rain types—convective, shallow stratiform, and anvil rain (deep stratiform with a melting level)—were derived from numerical simulations of tropical cloud systems from the Tropical Ocean and Global Atmosphere Coupled Ocean–Atmosphere Response Experiment (TOGA COARE) utilizing a cloud-resolving model (CRM). The SLH-retrieved Q1 and Q2 profiles for the SCSMEX NESA are in good agreement with sounding-based ones from Johnson and Ciesielski (2002). Differences of Q1 and Q2 profiles between western Pacific and Atlantic estimated by the SLH algorithm are also consistent with those from the budget study (e.g., Thompson et al. 1979).

The two-dimensional (“2D”) version of the Goddard Cumulus Ensemble (GCE) model was used in the previous studies. Real clouds and cloud systems are three-dimensional. The availability of exponentially increasing computer capabilities has resulted in three-dimensional (“3D”) CRM simulations for multiday periods with large horizontal domains becoming increasing prevalent. We are now using the 3D GCE model in order to compare look-up tables from the 2D and 3D simulations. The lookup table (LUT) from 3D with a horizontal resolution of 2 km produces worse agreement between the SLH-retrieved and sounding-based Q1 profiles for the SCSMEX NESA than the LUT from 2D does. This is explained by differences in the vertical profile of convective heating of a given depth between 2D and 3D. While convection in 2D has latent heating concentrated below the freezing level, consistent with TOGA-COARE radar observation (DeMott and Rutledge 1998), convection in 3D has stronger latent heating above the freezing level. Preliminary results indicate that more realistic convective heating profiles in 3D are achieved by lowering the horizontal resolution from 2 km to 0.5 km.