ECMWF and GFS model forecast verification during DYNAMO: Multiscale variability in MJO initiation over the equatorial Indian Ocean

Abstract Improving understanding and prediction of the initiation of the Madden-Julian Oscillation (MJO) over the equatorial Indian Ocean (IO) was a main objective of the Dynamics of the MJO (DYNAMO) field campaign. This study evaluates the European Centre for Medium-Range Weather Forecasts (ECMWF) and the National Centers for Environmental Prediction Global Forecast System (GFS) operational global model forecasts of MJO-2 in DYNAMO (November–December 2011). The model rainfall and 850 hPa zonal winds are compared with the Tropical Rainfall Measurement Mission (TRMM) rainfall and operational analysis in the equatorial IO. The ECMWF forecasts of the total equatorial rainfall were more consistent with TRMM observations. The GFS did not accurately predict the onset of equatorial convection beyond 5 day lead time. Instead, the GFS rainfall was mainly off the Equator in the Intertropical Convergence Zone (ITCZ). The GFS did not capture synoptic scale systems that bring in dry air intrusions from the subtropics and disrupt convection in the ITCZ which is favorable for equatorial convection associated with the MJO initiation. DYNAMO observations suggest that multiscale interactions among convection, synoptic systems, and the MJO large-scale circulations were important for the initiation of MJO-2. The model forecasts and analysis of 850 hPa zonal wind are decomposed into mesoscale, synoptic, and large-scale flows. Results show that (1) both models have good initial conditions and 1–2day forecasts and (2) the ECMWF has significantly better skill than the GFS in terms of forecasting equatorial convection, synoptic systems, and the large-scale MJO circulation at 5–15day lead times. 1