The San Joaquin Valley (SJV) of California is categorized by the U.S. Environmental Protection Agency (EPA) as being in serious non-attainment of the National Ambient Air Quality Standard (NAAQS) for 8-hr ozone. While the SJV as a whole experiences regionally high ozone levels, the urban and rural areas around Fresno, CA experience NAAQS exceedances at a particularly high rate. A favorable interaction of local emissions with a mesoscale meteorological phenomenon known as the “Fresno Eddy” allows for an ozone hotspot around Fresno. On most days of the ozone season, a horizontally oriented eddy forms aloft in the mixed layer during the overnight hours; ground level conditions in the immediate vicinity of the Fresno urban area are affected during the morning, allowing for recirculation of the morning rush hour emissions. Enhanced understanding of the Fresno Eddy is required to further the knowledge of SJV ozone buildup processes. Furthermore, identification of representative regimes of eddy formation is prerequisite for performing meaningful simulations to aid with regulatory strategy development.

Several previous researchers have studied the Fresno Eddy using simulation approaches. The results of such simulations are insightful for understanding Fresno eddy dynamics. Such simulation studies are inherently limited for air quality applications, however, because they only consider a relatively small number of days for which the representativeness is unknown. Here, we consider statistical methods for the investigation of extended records from surface meteorological monitors sampling the eddy at an hourly rate. Though the surface measurements do not fully capture the 3-dimensional eddy, they provide sufficient information to infer characteristics such as degree of eddy development and strength of the resulting recirculation on a given day. Analysis of data spanning 9 ozone seasons allows for the identification of various regimes of eddy formation and the frequency at which these regimes are realized.

A cluster analysis technique is applied to identify groups of days having similar diurnal (24-hr) cycles for the surface wind field measurements. The cluster analysis considers each day from the months of May through October of 1996-2004, or 1656 days total. The clusters of days with similar surface flow pattern evolution generally capture synoptic influences, indicating the effect of the upper atmospheric conditions to drive or inhibit eddy formation. Anticyclonic regimes tend to favor eddy development due to the shallow boundary layer and limited degree of marine ventilation. The large day to night temperature swings associated with these sunny and hot days additionally contribute to the marked upslope/downslope cycle on the SJV rims that is intimately connected to Fresno Eddy formation. More ventilated conditions with deeper boundary layer depths and higher Froude numbers are less likely to have significant eddy formation.

The synoptic features associated with the various clusters indicate the relative propensities of upper atmospheric conditions to result in strong eddy formation. The eddy itself is a mesoscale phenomenon, however, and exhibits considerable variability within the above identified synoptic regimes. As such, regional ozone levels are strongly affected by synoptic conditions, but in turn they ultimately depend on the mesoscale conditions resulting from the eddy development on a particular day. Thus, to better determine the nature of Fresno area ozone buildup mechanisms, each synoptic regime is explored for mesoscale variability relating to regional ozone levels. The strength of the eddy within a synoptic regime can be inferred by gauging the degree to which the eddy results in a reversal of the ground level winds around Fresno. The eddy affects surface wind monitors at 6 different monitoring locations sequentially (in time) from south to north. Stronger eddy development results in a region of eddy influence extending further northward. Additionally, stronger eddy development results in a full reversal (180 degree shift) of the surface wind direction at the affected monitor, whereas weaker eddy development may result in a lesser shift in direction.

Using the above indicators of Fresno Eddy strength, it is possible to determine days having stronger and weaker eddy development within each synoptic regime. Ozone levels within each cluster generally increase with increasing eddy strength, however very strong eddy flows may also allow for some pollutant dispersion. Within each cluster, increasing degree of marine ventilation is noted to hinder the eddy development and thus result in reduced ozone levels. Mesoscale conditions under different synoptic regimes are more difficult to compare directly. The clusters are associated with distinct, synoptic characteristics in terms of boundary layer height, atmospheric stability, and degree of (or lack of) subsidence. Thus, for different clusters, seemingly similar degrees of eddy development, as measured by the effects of the eddy on the ground level winds, may result in very different ozone levels. Alternatively, similar ozone levels may result under days with different synoptic influences because of different effects of the eddy recirculation. The distribution of eddy strength is determined independently for each synoptic regime, and there are distinct responses of ozone levels to the degree of eddy strength under each synoptic regime.

In summary, the Fresno Eddy is a mesoscale flow phenomenon that affects airborne pollutant recirculation and allows for high ozone levels in SJV. Synoptic conditions largely determine the probability that the eddy will form, however mesoscale characteristics of the eddy under a given synoptic influence may vary considerably. The resulting ozone levels are affected by both the synoptic and eddy influences. Thus, understanding of SJV ozone dynamics requires a combined knowledge of synoptic and mesoscale conditions affecting California.