The Euro continues to amplify the MJO in the Western Pacific and Western Hemisphere, and is looking more like the GFS every day. Now brings the MJO to nearly 3.5 sigma and even that may not be strong enough.
View attachment 3568
A few important take aways worth bringing up here that I gleaned from reading several pieces of scientific literature last night regarding the ECMWF's handling of MJO events in the Maritime Continent
- The MJO predictability across the Maritime Continent appears to be dependent on the amount of convection and boundary layer warmth in the Western Pacific. Warmer initial SSTs and less convection preceding MJO events that propagate across the Maritime Continent are generally more predictable than those w/ a singular, monopole of deep convection.
- Drier western Pacific helps generate a rossby wave like response that increases meridional flux of moisture which allows the MJO to propagate more easily across the Maritime Continent
- The fractional contribution of surface winds is larger than convection (OLR) in RMM which is a profound weakness in the index
- The European actually has a wet bias over the Maritime Continent due to its poor representation of important topographic features which interfere w/ low-level latent heat fluxes and surface wind exchanges. This wet bias here strengthens the Walker Circulation and creates a corresponding cool bias in the east-central Pacific equatorial cold tongue.
- The stronger diurnal cycle during the northern hemisphere summer and El Nino increases the frequency of MJO blocking by the Maritime Continent because the incident solar radiation is stronger which hastens the diurnal cycle of convection over the archipelago that destructively interferes w/ the MJO
- The diurnal cycle in convection over the Maritime Continent interferes with the MJO by competing for moisture, and impacts to the large-scale circulation through generation of sea breeze circulations, etc.
The following diagram shows RMM MJO bias vectors for the ECMWF, JMA, and NCEP GFS models from Ichikawa and Inatsu (2017) according to each MJO phase. Note here that the Euro and JMA are on average too slow w/ the MJO in general (for example during a Maritime Continent MJO forecast reality may be closer to phase 5 while a Euro/JMA forecast depicts the MJO closer to phase 4, etc.). Also, while the NCEP GFS tends to have an overall oversimplification bias for MJO events as Larry has discussed the past few days, this bias only really occurs in phases 6,7,8, & 1 (western Pacific & Western Hemisphere), while the biases in the Maritime Continent are actually smaller than either the JMA or ECMWF models. This suggests that the GFS is in fact better at forecasting MJO events in the Maritime Continent (phase 4-5) than any other global model, which is where the MJO has been the past week and change.
View attachment 3570