Modeling of Mountain-Valley Wind Fields in the Southern San Joaquin Valley, California

Abstract

A dry three-dimensional mesoscale model was used to study the diurnal cycle of mountain-valley winds in the southern San Joaquin Valley during a summer day. A scheme for interpolating potential temperature was developed to provide hourly temperature fields to initialize and force the dynamically predicted wind fields. A simplified modeling approach was used to produce steady state solutions that are dynamically consistent with the momentum equation and supplied temperature fields. Model performance was evaluated by comparing observed and predicted surface winds. Some features of the wind field flow aloft were qualitatively examined with regard to their importance in air quality studies. The morning drainage-upslope transition and the evening reversal of upslope flow were realistically simulated throughout most of the valley. The variation of wind speeds throughout the valley and over the course of the day were simulated with an average bias of 9% of the average wind. Wind directions were simulated with an overall average bias of 5° and midday hourly correlation coefficients of typically r = 0.8. Model performance was below average during the morning and evening transition periods, when thermal forcing is at a minimum and valley winds are light and variable. At midday, the model produces strong upward vertical motions near the ridge crests and divergence-driven subsidences at the foot of the mountains typical of observations made in mountain-valley systems. During the morning, modeled drainage flow down the mountains results in a convergence zone in the southern and narrowest part of the valley, resulting in rising motions; down-valley flow, sometimes observed in mountain-valley systems, also occurs. The model is best suited for applications in mountain-valley regions for which wind observations are sparse and do not adequately reflect thermally driven circulation.