UC San Diego

Santa Ana Winds of Southern California and Northern Baja California (SoCal) are the primary weather drivers of wildfires that frequently and infamously ravage this topographically and demographically complex region. As the available wind observations are scarce, Santa Ana Winds (SAWs) have not been adequately studied on climate timescales. Yet, wildfire behavior has been changing wildly even during my dissertation work. For example, the largest wildfire in California’s recorded history occurred in December of 2017, well outside the unique fall season for the largest wildfires typical of this region. This dissertation presents results of my efforts to understand Santa Ana wind behavior on climate timescales and in our changing climate. In Chapter 1, I developed and analyzed the longest and most complete record of hourly SAWs heretofore available. These results provide a robust perspective on both high- and low-frequency SAW behavior, uncovering previously unknown climate influences on SAW activity, and laying the groundwork for eventual studies into seasonal SAW predictability. Notably, the 65-year record of hourly SAWs did not manifest clear long-term trends. In Chapter 2, using a dynamically downscaled training data set, I developed an approach to statistically downscale coarsely resolved winds from a global Reanalysis with spatial resolution typical of global climate models (GCMs). The result was an efficient and skillful downscaling of coarse daily surface wind vectors onto a fine grid over the SoCal domain spanning 70 years. From these daily wind fields, I derived SAWs and validated them against my previously validated SAWs derived in Chapter 1 from our dynamical training data set. A capacity to statistically downscale daily winds from reanalyses and global climate models was thus developed and applied, in Chapter 3, to a set of eight GCMs yielding an ensemble of daily 10X10 km wind data sets covering SoCal and spanning a historical and future projected time period from 1950 to 2100. Analyses of these data yielded physically meaningful, robust and clear projections of SAW activity gradually decreasing and constricting around its traditional seasonal peak centered on December. These results provide robustness, nuance and meteorological context to expectations of future SoCal wildfire activity.