Abstract ID: 030
Impact of Stochastic Kinetic Energy Backscatter Scheme on the Navy Earth System Prediction Capability prediction of the Madden Julian Oscillation during 2017
Lead Author: Marcela Ulate
U.S. Naval Research Laboratory, UCAR, United States of America
Keywords: MJO, ensemble, NAVGEM
Abstract: The Navy Earth System Prediction Capability (Navy ESPC) is used to examine the prediction of the Madden-Julian Oscillation (MJO) with and without the addition of Stochastic Kinetic Energy Backscatter (SKEB) forcing to account for model uncertainty. The formulation of the SKEB forcing in the Navy ESPC System uses a moisture convergence mask which adds random perturbations to the rotational component of the wind. The Navy ESPC is comprised of the Navy Global Environmental Model (NAVGEM) atmospheric model coupled to Hybrid Coordinate Ocean Model (HYCOM) and the Los Alamos Community Ice Code (CICE). These ensembles were run once per week between February 2017 through January 2018 with 5 and 16 members for the SKEB and the Non-SKEB configurations respectively.
We compare the mean state differences between ensembles, the main ensemble statistics, and the MJO predictions using an MJO index. Our results show an improvement in the ensemble spread and decrease in errors for leads times up to 15 days in the forecasts when the SKEB configuration is used. MJO prediction using the Real-Time OLR MJO Index show differences between the SKEB and the Non-SKEB formulations as early as lead time day three, with the greatest differences after days 10-15. We also analyze variables associated with the MJO such as Sea Surface Temperatures (SSTs), precipitation and Outgoing Longwave Radiation (OLR). The SST biases are the largest in the Equatorial East Pacific and Southern Ocean. These biases are of comparable magnitude in the two ensemble systems. The largest SST ensemble spread is concentrated in the western boundary currents. For both SSTs and precipitation the SKEB formulation does not appear to have a clear effect on the spatial distribution of variance, and the spread magnitudes are also similar. The largest errors in precipitation occur close to the Maritime Continent and in the tropical Pacific. The OLR biases indicate overactive convection over the Indian Ocean and tropical West Pacific, which are greatest in the second week of the forecast for both SKEB and Non-SKEB ensemble systems and then slowly decrease at longer lead times.
Co-authors:
Carolyn A. Reynolds (U.S. Naval Research Laboratory)
Matthew A. Janiga (U.S. Naval Research Laboratory)
Justin G. McLay (U.S. Naval Research Laboratory)
William Crawford (U.S. Naval Research Laboratory)
William Komaromi (U.S. Naval Research Laboratory)
