Abstract ID: 237
Uncertainty in Diabatic Heating within MJO phases 3-4 and Consequences for Mid-Latitude Predictability: Large Ensemble Studies with the ECMWF Model
Lead Author: David Martin Straus
George Mason University, United States of America
Keywords: MJO, diabatic heating, Rossby Wave Source, Predictability, Tropical-Extratropical Interactions
Abstract: The substantial amount of research carried out on the Madden-Julian Oscillation (MJO), its evolution and predictability, and the consequences for its extra-tropical teleconnections, have all relied almost entirely on the Wheeler and Hendon multi-variate EOF framework. Much less studied is the effect on mid-latitude predictability of the variability of diabatic heating within a given MJO phase. This substantial variability is due to the high degree of intermittency in time and space that characterizes heating.
This study will report on a large suite of unique reforecasts carried out with the ECMWF coupled model. The 60-day reforecasts span 13 initial conditions residing within MJO phases 3 and 4 (heating in the Indian Ocean), with each forecast set having 51 ensemble members. The uniqueness of this set lies in the setting of the stochastic parameterization configuration: each of the 51 ensemble members sees the same identical initial condition, and is perturbed only in the tropical Indian – West Pacific Ocean regions in which the MJO heating is active initially. ALL subsequent spread is due only to the uncertainty in heating in this MJO region.
We find that while the total intra-ensemble error variance in heating exceeds the signal in heating anomaly, the planetary wave tropical diabatic heating (200 hPa divergence) ensemble spread and error variance has not saturated by forecast day 30 (day 40), indicating predictability of the heating on the monthly time scale. Defining a predictability time based on the internal error variance reaching half its saturation value, the planetary wave tropical heating has a predictability time of 24 days, while the corresponding Rossby wave source predictability time is only 14 days.
In mid-latitudes, while the heating error (indicative of the storm tracks) has spread into the Northern Hemisphere ocean basins by day 15, the large scale error variance in 200 hPa vorticity saturates only between days 30 and 40, indicating potential long-time scale predictability. The role of the stratosphere in enhancing the error growth is limited to the latter part of the reforecasts, when stratospheric error induced by the earlier spread in upward propagating tropospheric heat flux begins to descend.
This study highlights the role of the uncertainty in the Rossby wave source associated with MJO heating in limiting mid-latitude predictability.
Sarah-Jane Lock (ECMWF)
Franco Molteni (ECMWF)
Daniela Domeisen (ETH Zurich)
Priyanka Yadav (ETH Zurich)