Diagnosing MJO hindcast biases in NCAR CAM3 using nudging during the DYNAMO field campaign

This study evaluates the Madden–Julian Oscillation (MJO) hindcast skill and investigates the hindcast biases in the dynamic and thermodynamic fields of the National Center for Atmospheric Research Community Atmosphere Model version 3. The analysis is based on the October 2011 MJO event observed during the Dynamics of the Madden–Julian Oscillation field campaign. The model captures the MJO initiation but, compared to the observations, the hindcast has a faster MJO phase speed, a dry relative humidity bias, a stronger zonal wind shear, and a weaker MJO peak amplitude. The MJO hindcast is then nudged toward the European Centre for Medium‐Range Weather Forecast Reanalysis fields of temperature, specific humidity, horizontal winds, and surface pressure. The nudging tendencies highlight the model physics parameterization biases, such as not enough convective diabatic heating during the MJO initiation, not enough upper tropospheric stratiform condensation, and lower tropospheric reevaporation during the mature and decay phases and a strong zonal wind shear during the MJO evolution. To determine the role of temperature, specific humidity, and horizontal winds in the model physics parameterization errors, six additional nudging experiments are carried out, with either one or two of the fields allowed to evolve freely while the others are nudged. Results show that convection and precipitation increase when temperature or specific humidity are unconstrained and decrease when horizontal winds evolve freely or temperature alone is constrained to reanalysis. Budget analysis of moist static energy shows that the nudging tendency compensates for different process biases during different MJO phases. The diagnosis of such nudging tendencies provides a unique objective way to identify model physics biases, which usefully guides the model physics parameterization development.