On Air-Sea Interaction in the Gulf of California and its Effect on North American Monsoon

The North American Monsoon (NAM) is a seasonal shift in the large-scale circulation that supplies 60-80% of annual rainfall in northwestern Mexico and 30-40% in the US southwest. A 2012 intercomparison of regional climate models has shown that summer precipitation prediction over North America is poorest in the NAM region. An understanding of the NAM's major governing processes is necessary to improve global and regional climate modeling, including the NAM's remote impacts on the summer circulation, precipitation and drought over North America.

Here, we propose a partial mechanistic understanding of the NAM synthesizing local- and large-scale processes. The local scale mechanism helps explain how low-level moisture from the Gulf of California (GC) fuels NAM rainfall. The inversion at the top of the marine boundary layer (MBL) over the northern Gulf of California (NGC) limits moisture transport into the NAM region. Soundings launched from a ship in the GC during summer reveal that this strong inversion weakens with increasing SSTs and generally disappears once SSTs exceed 29°C, allowing the trapped MBL moisture to mix with free tropospheric air. This generates a deep, moist, well-mixed layer that can be advected inland by favorable low-level jets (LLJs) to form thunderstorms in Arizona and elsewhere. The large-scale mechanism relates tropical surface water, tropospheric moisture and the NAM anticyclone by means of climatologies of satellite SST, outgoing longwave radiation (OLR) and NCEP/NCAR reanalysis of 500 hPa geopotential height from 1983 to 2010. As warm Pacific SSTs propagate northwards up the Mexican coastline, deep convection follows this northward advance, with associated descending air north/northeast of the convection region possibly advancing the position of the anticyclone. This evolution brings mid-level tropical moisture into the NAM region.

A set of carefully designed simulations of WRF is used to investigate the dependence of NAM precipitation, onset and circulation on SSTs along the Mexican coastline and in the GC. North American Monsoon Experiment (NAME) field campaign in summer 2004 provides unique enhanced observational data such as unified rain gauge data (URD) and Multiplatform-Merged (MPM) SST for assimilation and also for evaluation of the model. WRF is able to simulate LLJ parallel to the GC axis during the 2004 monsoon onset. Preliminary WRF simulations show that warmer GC SSTs tend to enhance low-level moisture during this period and as a result more precipitation occurs over the foothills of Sierra Madre Occidental (SMO) and over US southwest. This rainfall enhancement appears to result from a weakening in the MBL inversion. However, predicted inversions are stronger than those observed, and, in contrast to observations, the MBL inversion generally does not disappear when SSTs exceed 29°C. This discrepancy may represent an opportunity to improve WRF performance during summer over North America.