Lower Boundary Forcing related to Occurrence of Rain in the Tropical Western Pacific
A four year timeseries of satellite estimated rainfall and daily SST are employed. We investigate three closely related topics: 1) the influence of lower boundary forcings on the distribution of deep moist convection, as evidenced by the onset of rainfall; 2) the correlation of total rainfall with these lower boundary forcings in the MJO passband; 3) characteristics of the coupled ocean-atmosphere responses associated with propagation of the MJO.
A statistical association is revealed between mesoscale SST structure and rainfall, including the cooperative roles of SST gradients and SST itself. Preferred locations for rainfall onset are in the mid-range of the background SST distribution, centered at the mode of SST and associated with gradients. The warmest SST locations are not favored (i.e. neutral) and coolest are disfavored, primarily a consequence of seasonal influences.
There are phase relationships between SST structure at short-range, fully consistent with coupled ocean-atmosphere responses at the local weather scale. For the MJO, it appears that an SST positive anomaly always leads MJO rainfall, essentially irrespective of the MJO phase, wet or dry. Interestingly, the -Laplacian of SST is in quadrature with the MJO, specific to it's active phase, and leads MJO rainfall by ~10-14 days. This relationship is evident across the eastern hemisphere from the east coast of Africa to the dateline.
We conclude that the primary relationship of SST to tropical oceanic rainfall is conditional, associated with production of moist static energy …i.e. if there is rain, then a warm anomaly will produce more rain. We conclude the primary role of mesoscale SST structure is as a systematic trigger function. This cooperative relationship between SST and its derivative fields, if properly represented in global models, potentially could reduce systematic rainfall prediction/projection errors. Recent progress at ECMWFappears to exhibit consistency with this hypothesis.