Scripps Institution of Oceanography

This dissertation examines atmospheric boundary-layer processes.

The equatorial Pacific cold tongue region is examined, with observations showing that monthly anomalies in low-level cloud amount and near-surface atmospheric temperature advection are negatively correlated. In warm advection, soundings show the surface layer is stably stratified, inhibiting the upward mixing of moisture, while cold advection favors a more convective atmospheric boundary layer and greater cloud amount. Two global coupled climate models fail to simulate this, suggesting specific areas for possible improvement.

Climatology and low-level wind variability near three California wind farms are then explored: San Gorgonio Pass and Tehachapi Pass in Southern California, and Solano County further north. Each site has a pronounced annual cycle with highest wind speeds in the warm months. While winter winds depend more on SLP, summertime winds are stronger, more diurnally dependent, and show more topographic influence on direction, though SLP variability is lesser.

Self-organizing maps reveal that oceanic high SLP and continental low SLP synoptic patterns lead to higher wind speeds. SLP gradients at 100km separation are strongly correlated to cotemporaneous site wind speeds. Dynamically downscaled reanalysis data at 10km resolution reveals that a thermally driven flow at the northern site commences at the coast and propagates inland in a distinct packet. The 10km model does not resolve the Gorgonio Pass flow, but another model at 2.5km validates well and shows that the observed winds depend on interactions between solar and lateral boundary forcing.

A statistical downscaling scheme is developed for relating GCM output to site winds. The multilinear regression model integrates weather type information and observational findings to reproduce wind speeds, showing skill for both monthly and daily data. Monthly mean wind speed changes over the 21st century are implied of up to 0.6 m/s in the summertime in three downscaled coupled climate models under greenhouse forcing. However, the discrepancies between models prevent consensus.

Analysis of the discrepancies reveal that in one model the western North American surface heating coincides with a decrease in SLP, while others show increased continental SLP. The coupled models' representations of these regional patterns are discussed.