Statistical Downscaling and Forecast of Global Warming Reverse-Reaction Coastal Cooling in the California South Coast Air Basin

Summer coastal cooling along the California coast has been reported by Lebassi et al. (2009) while analyzing summer surface mean monthly maximum air temperature trends between 1950-2005 for two California air basins: San Francisco Bay Area (SFBA) and South Coast Air Basin (SoCAB) using daily data from 273 National Weather Service (NWS) Coop Sites. Furthermore, Sequera et al. (2012) reported other foci of coastal cooling in the California Coast, in particular, at the San Diego Air Basin where the maximum coastal cooling reached a level of -1.3°C/decade. The spatial distribution of the observed JJA max temperatures showed a complex pattern in which cooling trends were found at low elevation coastal areas open to marine air penetration and warming trends at inland and high elevation coastal areas. As a consequence of an increased gradient of the concurrent sea breeze potential for the same period, the authors suggested that this increased sea breeze activity was responsible for the observed coastal cooling. Other recent reports have detected coastal cooling along the South America Pacific coast and the Persian Gulf.

In this report, the evolution in time of the summer coastal cooling in the SoCAB is tested against global records and global circulation models (GCMs) in anticipation of forecasting future trends. Previous efforts in using GCMs to detect summer coastal cooling in California have reflected coastal-inland thermal gradients, however, failed to detect the magnitude of the decadal temperature decrease, attributed to lack of specificity of the local physical processes caused by local topography. For this study, a statistical downscaling approach is first used from the World Climate Research Programme's (WCRP's) Coupled Model Intercomparison Project phase 3 (CMIP3) multi-model dataset under present climate conditions (1960-2010) and validated against present data. The statistical downscaling techniques are based on linear regression between temperature and large scale atmospheric forcing in order to obtain forecasted average surface temperatures at resolutions of 1/8 degrees from the 2 degree resolution of General Circulation Models (GCM). A Thin Plate Smoothing Spline interpolation technique is also used to improve the regression models. Results with the proposed approach show good agreement with present tendencies. For future conditions (2010-2050) it was found that summer coastal cooling has an increasing tendency consequence of the increased pressure gradient caused by the faster heating rate of inland regions in contrast to the slower heating rate of the sea surface temperatures. These future tendencies are modulated by the Pacific Decadal Oscillation, which controls the coastal upwelling.