Isochrones of frontal position are determined subjectively through examination of time series of surface reports. These time series are constructed from all available regular hourly and special between hourly surface observations from the National Weather Service's Automation of Field Operations and Services network (AFOS) archived at the National Climatic Data Center. Objective analyses of observed surface temperature and winds are prepared in order to calculate diagnostic quantities. The role of mesoscale potential vorticity (PV) anomalies in the upper troposphere and their associated shortwave troughs on the surface fronts is examined using European Centre for Medium Range Weather Forecasting (ECMWF) global analyses archived at NCAR. Satellite observations every three hours from the McIDAS archive at the University of Wisconsin-Madison are also used in the analysis.

Three distinct phases of frontal evolution have been identified. The first phase, 1500 UTC 24 March to 1200 UTC 25 March, consists of the interaction of a decaying maritime cyclone and its associated front with the coastal mountains and Sierra Nevadas in California. The second phase, 1500 UTC 25 March to 1200 UTC 26 March 1991, involves the rapid development of a very strong cold front associated with a Nevada lee cyclone that moves through the intermountain region. The third phase, 1200 UTC 26 March to 0300 UTC 27 March, also involves rapid development of a second strong cold front that causes extremely high winds in the desert Southwest. In all three phases, the strength and location of the upper tropospheric PV anomaly and its associated low level convergence and upward vertical motions is highly correlated with the strength and location of the fronts at the surface. Differential diabatic heating and deformation of flow by topogrphy are two mesoscale processes found to significantly modify frontal characteristics. The strength of the potential temperature gradient associated with the front is strongly modulated by differential sensible heating across the front. An estimate of the contribution to frontogenesis from differential diabatic heating from phase III shows that it is several times greater than the contribution from the surface winds alone. In phase I and phase II, the fronts are highly deformed by the complex topography of the western United States. Ridges oriented perpendicular to the direction of frontal movement act to block the front. In addition, the fronts are observed to accelerate in valleys and parallel to the rigdes. Lastly, frontal evolution is not well described by the Norwegian Cyclone Model because the fronts are not associated with a synoptic scale cyclone. Rather, the fronts are strongly affected by the mesoscale upper level, topographic and differential diabatic processes such that their temporal and spatial evolutions are mesoscale.