PNNL

Abstract

This study assesses the 40-day 3.25-kilometer global simulation of Simple Cloud-Resolving E3SM Model (SCREAMv0) using high-resolution observations from the Atmospheric Radiation Measurement (ARM) Green Ocean Amazon (GoAmazon) field campaign. SCREAMv0 can reasonably capture the diurnal cycle of boundary layer clouds and early-afternoon precipitation, however, lacks the development of mid-level congestus. SCREAMv0 does not form enough precipitation clusters larger than 200 km, with a higher frequency of smaller clusters than the observations by the scanning precipitation radar. This precipitation cluster size bias suggests an inhibition of convection upscale growth from isolated deep convective clusters into larger mesoscale propagating systems, which possibly due to misrepresentations in land-atmosphere coupling. This study highlights the great potential of high-resolution ground-based observations in diagnosing convective processes in global storm resolving model simulations, identifying key model deficiencies, and providing guidance on the detailed process-oriented sensitivity tests and analyses in the future.