Informal Seminar: On the relationship between intraseasonal and interannual teleconnections from Indo-Pacific heating
by
Many studies on the extratropical teleconnections of the Madden-Julian Oscillation (MJO) in boreal-winter have associated an increased frequency of positive North Atlantic Oscillation (NAO) anomalies with the occurrence of MJO phase-3 10 to 15 days earlier, when increased convection is located in the eastern Indian Ocean. Connections between the NAO and the MJO phase-2 (when increased convection is located further west) appear to be weaker in intra-seasonal diagnostics. On the other hand, on the interannual time scale, teleconnections computed from 2-month or 3-month means indicate a stronger connection of positive NAO with convection in the western-to-central part of the tropical Indian Ocean.
In the first part of this study, diabatic heating computed from ERA-5 data (specifically, as a residual between total and adiabatic temperature tendencies) in a 36-year period, and 200-hPa geopotential height from the same re-analysis, are used to investigate the relationship between inter-annual and intra-seasonal teleconnections of MJO-like heating anomalies over the Indian Ocean and the Maritime Continents. Results show that the discrepancy between results on the two timescales are due, to a large extent, to the impact of ENSO on the inter-annual component of Indian Ocean circulation and heating anomalies. This implies that the way in which the seasonal-mean signal is (or is not) filtered out may significantly affect the estimation of intra-seasonal teleconnections. Since the impact of ENSO on both Indian Ocean convection and the NAO varies between early and late winter, results are also dependent on the specific period of the year considered.
In the second part, results from a set of sub-seasonal ensembles run with the ECMWF coupled model, and initialized on 1 Jan and 1 Nov 1981 to 2016, will be compared with ERA-5 results for the Nov-Dec and Jan-Feb periods. In these ensembles, stochastic physics is only applied over the tropical Indian Ocean, and is the only source of spread among different ensemble members. Confirming earlier findings on ECMWF sub-seasonal and seasonal forecasts, a good agreement is found on the patterns of tropical heating variability, but the amplitude of their extratropical teleconnections is significantly under-estimated. It is also found that the lagged correlation between the two leading (MJO-like) modes of diabatic heating variability is weaker in the model than in ERA-5, suggesting that MJO propagation is less coherent in the model than in reality.