Our work shows that the onset time of this threshold SST may be predicted 1-2 months in advance from the sea surface topography in the eastern tropical Pacific. By defining different ocean regions, sea surface height anomalies (SHA) in May were evaluated for each region for 10 years of TOPEX/POSEIDON satellite data. Combining these measurements into a SHA Index and regressing this index against the N.-GC 29.5 C onset time, the resulting correlation coefficient r=-0.94, explaining 88% of the variance. The forecasted time window for the onset of heavy rainfall has an uncertainty (std. dev.) of ±3.4 days. Although SHA were evaluated by careful visual inspection of satellite imagery, in the near future we hope to refine this result using the digital data.
This forecast of the N.-GC SST onset may also be used to forecast whether Arizona rainfall during June-August will be dry-to-average or wet-to-average. The latter condition indicates dry-to-average conditions in the Midwest. This is based on a 20 year SST data set valid for the GC (weekly temporal &18 km spatial resolution) and statewide AZ precipitation records, and on a strong anti-correlation between Arizona and Midwest summer rainfall amounts.
This relationship between the N.-GC SST onset and summer rainfall amounts may derive from two factors: (1) longer monsoon seasons tend to have more rainfall events, and (2) the positioning of the mid-level anticyclone over North America during summer may depend on the poleward propagation of warm water up the west coast of Mexico. As these higher SSTs advance poleward out of the eastern Pacific warm pool, deep convection commences over the adjacent land, and subsiding air to the north of the convection positions the center of the poleward advancing anticyclone. Hence the latitude of the anticyclone (affecting moisture transport) may depend on the advance of the warmest GC SSTs. Observational and modeling evidence will be presented that supports this hypothesis. Note that the summertime circulation over North America is largely defined by the position of this anticyclone.
The mechanism behind the above SHA-SST relationship is not well understood at present, but apparently involves an intrusion of high SST from the warm pool into the GC in May-June. This intrusion appears related to the behavior of a "dome" in sea surface height off the coast of southern Mexico. It is observed that in May-June, the California Current and associated winds slacken, the dome diminishes and the SST intrusion occurs. Ocean topography from satellite indicates that warm pool water during March-April circulates clockwise around this dome (decoupled from the coast). As the dome begins to collapse in late May-early June, it bridges with the coast near Cabo Corrientes, with geostrophic currents directed toward the Cape and into the GC. A dramatic reconfiguration of the 28.5oC isotherm occurs during June, perhaps in response to this change in surface currents. During the first week of June, this isotherm extends far to the west of southern Mexico but south of 18N, whereas during the last week of June, this isotherm typically describes a narrow band "hugging" the coast, extending deep into the GC.
Finally, based on 18 years of reliable GC SST data, time series of GC SSTs show a clear warming trend (r=0.736 for July in the central GC), with linear regression showing more than a 1 degree C SST increase over this period. A similar warming trend is observed in the eastern Pacific warm pool. If GC SSTs influence the timing and amounts of NAM rainfall, this SST warming trend suggests that the behavior of the NAM may be rapidly changing.
Supplementary URL: