Seasonal atmospheric transitions in the Caribbean basin and Central America

Abstract

The dates of the transition between winter and summer (W2S) and between summer and winter (S2W) regional-scale atmospheric regimes have been defined using daily weather types above and around the Caribbean basin from 1979 to 2017. The uncertainties due to either the use of two different reanalyses (i.e., NCEP-DOE and ERA-Interim) or the parametrization used for the definition of the transition dates have typically a small impact on the interannual variability of the seasonal transitions. When both reanalyses are considered together, the average W2S transition date occurs, on average, on May 13 (with a standard deviation of 9 days) while the S2W transition date occurs, on average, on October 26 (with a standard deviation of 12 days). The atmospheric characteristics associated with both transitions reveal asymmetries in the annual cycle. The W2S transition is rather abrupt and near-synchronous to a rather sharp increase of rainfall, propagating from Central America to the NE of the Caribbean basin, and a weakening of the Caribbean Low Level Jet. The W2S transition is also not preceded by any significant sea surface temperature (SST) anomalies either in the tropical North Atlantic or the Eastern Pacific. On the other hand, the S2W transition is overall smoother, and anomalously warm (cold) SST over the Caribbean Sea and Gulf of Mexico (Eastern Pacific) during the boreal summer are usually related to a delayed transition (and vice versa). The interannual variations of S2W and W2S transitions are mostly independent to each other. The potential and real-time predictability of the W2S transition is explored using a subseasonal-to-seasonal prediction ensemble (11 runs from 1998 to 2017) from the ECMWF model. Its skill is close to zero with a lead time longer than 15–20 days, confirming the weak impact of the antecedent SST upon the W2S transition. The skill suddenly increases from late April, 2–3 weeks only before the mean W2S transition date. It suggests that some atmospheric forcing, operating from synoptic to intra-seasonal time scale, plays a role, but it seems barely related to any occurrence, or sequence, of specific weather types.