7. MIXING DEPTH STRUCTURE AND EVOLUTION ON EPISODE DAYS

The planetary boundary layer is the portion of the atmosphere that is influenced by thermal and frictional forces arising from contact between the air and the earth's surface. These forces in turn cause significant diurnal changes in mixing of the air in the planetary boundary layer, which can range from about 100-300 m at night to 1.5-2.5 km during the day. Mixing depth is defined as the altitude above the surface through which vigorous vertical mixing of heat, moisture, momentum, and pollutants occurs (Holzworth, 1972). Diagnosing the evolution and structure of the planetary boundary layer is important for understanding the dispersion and transport of pollutants such as ozone and ozone precursors. The mixing depth structure and evolution as diagnosed from radar profiler data during July 13-15, 1995 and July 31-August 2, 1995 were described by Dye et al. (1996). In addition, they discussed the atmospheric mechanisms that cause these day-to-day changes in the mixed layer evolution. Their results are summarized here.

The major components of the planetary boundary layer are the nocturnal boundary layer, the daytime convective boundary layer, and the nighttime residual layer, which forms between the top of the nocturnal boundary layer and below the top of the old daytime convective boundary layer. A stable nocturnal boundary layer forms at the surface at night when the temperature of the air near the ground decreases in response to the radiational cooling of the earth's surface. This process produces a shallow inversion and stable conditions, which reduces vertical mixing, thus confining surface-based pollutants to the lowest few hundred meters. Starting shortly after sunrise, the convective boundary layer grows during the day as the ground is heated and thermals vertically mix heat, moisture, momentum, and pollutants. At sunset, these thermals decay and the stable conditions of the nocturnal boundary layer return. Aloft at night, a nighttime residual layer remains and initially has characteristics (in terms of momentum, moisture, and pollutant burden) of the recently decayed convective boundary layer.

The structure of the planetary boundary layer was measured using the radar profilers. Beside measuring winds and temperature, the radar profilers also measured reflectivity, which can be used to compute the refractive index structure parameter (C_n²). This parameter measures the variations in the refractive index of the atmosphere, which are produced when turbulence creates gradients in humidity. Dye et al. (1995) showed that mixing depths estimated from C_n² and RASS during summertime ozone episodes agreed well with mixing depths independently estimated from aircraft pollutant, temperature, and turbulence data.

For this study, hourly mixing depths were estimated from C_n² data during the day using an algorithm developed by Dye et al. (1995). For the night and the early morning hours, RASS virtual temperature (T_v) data were combined with surface data to identify the height of the nocturnal boundary layer and to monitor the growth of the convective boundary layer in the first few hours after sunrise.

The meteorological conditions during the two episodes studied were favorable for formation of high ozone concentrations as high-pressure systems located over the area produced clear skies, weaker winds, hot temperatures, and a subsidence inversion. During the first episode, the highest ozone concentrations were measured on July 14 and July 15. July 13 was considered a "ramp-up" day with 12 ozone exceedances, and on July 14 and 15, 37 and 36 ozone exceedances occurred throughout the OTR. Peak ozone concentrations of 170 ppb and 175 ppb were measured on July 14 at sites in New Jersey and Connecticut.

To illustrate the diurnal changes in the mixed layer structure, Figure 7-1 shows a time series of the mixing depth and the subsidence inversion from July 13-15, 1995 at the New Brunswick, NJ and Gettysburg, PA sites. This figure illustrates several features that are important for understanding the role that mixing depth plays in the air quality of this region:

• Diurnal Evolution. At night, the strong stability in the nocturnal boundary layer limited vertical mixing of momentum and (most likely) surface-based pollutants to a shallow layer (i.e., 100 to 200 m thick) near the ground. Above the nocturnal boundary layer, the residual layer was observed between 200 and about 1500-2000 m agl. After daybreak (0700 EST), the surface warmed, resulting in the growth of the convective boundary layer. The convective boundary layer grew to an afternoon depth of 1500-2000 m agl, but further vertical growth was inhibited by the subsidence inversion. By early evening, the stable conditions associated with the nocturnal boundary layer formed. The residual layer above the nocturnal boundary layer initially had characteristics similar to the recently decayed convective boundary layer. Ozone or ozone precursors mixed upwards in the convective boundary layer during the daytime likely remained in the residual layer, and the strong stability associated with the nocturnal boundary layer "shielded" these pollutants aloft from the titrating effects of fresh NO emissions near the surface. In fact, tethersonde measurements made at Gettysburg on August 1, 1995 (discussed further in Section 8) confirmed that the aloft ozone above the nocturnal boundary layer remained high; whereas within the nocturnal boundary layer it was titrated.

• Growth Rates. As shown in Figure 7-1, the structure and evolution of the mixed layer for these days were basically similar, with one important difference: the growth of the convective boundary layer was faster on the ramp-up day (July 13) than on the major exceedance days (July 14 and 15). For example, at the New Brunswick site on the ramp-up day, the convective boundary layer grew from about 300-2250 m agl between 0700-1200 EST, which translates to a growth rate of 325 m/hr. On the exceedance days (July 14 and July 15), the growth rate was 180 m/hr and 230 m/hr during the same time period. This slower growth rate likely confined emissions near the surface for a longer time, and thus photochemistry occurred in the presence of higher precursor concentrations. This slower growth rate was perhaps one key factor in the production of higher ozone concentrations on July 14 and July 15.

• Descending Inversion. Another important factor for understanding air quality was the descending subsidence inversion observed at the Gettysburg, PA profiler site during July 13-15, 1995. Notice that the subsidence inversion descended from about 2000 m agl on July 13 to 1000 m agl on July 15. This lowering of the inversion was also observed in rawinsonde data collected at the Mid-Atlantic sites. The descending inversion reduced the vertical extend of mixing almost in half on July 15, which resulted in less dilution of ozone and ozone precursors.

The same type of mixing depth analyses was performed on the July 31-August 2, 1995 ozone episode. On July 31, 12 ozone exceedances were confined to New Jersey and Pennsylvania. On August 1 and 2, ozone exceedances were more numerous and wide spread.

Figure 7-2 shows a time series of the mixing depth and the subsidence inversion during this period at the New Brunswick, NJ and Gettysburg, PA profiler sites. The mixing depths in this figure show several similarities to those discussed above, but also indicate some differences.

The similarities between these two episodes include:

• Strong changes in the diurnal mixing depth, which was shallowest at night and deepest during the daytime.

• The mixing depth tended to grow faster on the ramp-up day (July 31, 1995) than on the major exceedance days (August 1-2, 1995). For example, at the Gettysburg site, the mixed layer grew rapidly on July 31 from 200 m to 1900 m between 0700 and 1200 EST, which translates to a growth rate of 280 m/hr. On the following two days, the convective boundary layer grew at less than half the rate of July 31 (110-115 m/hr). Again, it appears that slower growing boundary layers were associated with days with higher ozone concentrations.

The major difference between these two episodes was:

• The subsidence inversion ascended during the second period. At the Gettysburg site, it ascended from 1750 m agl on July 31 to 2400 m agl on August 2. The difference between the two episodes may indicate that the maximum mixing height reached during the day is not as important for ozone formation as the rate of rise of the mixing height in the morning.

We have not examined the processes that might cause the slower rise in mixing height to result in higher ozone concentrations onshore. However, it is clear that emissions transported offshore in the morning on a slow-rise day will be confined to a thinner layer than on a day with more rapid growth. This confinement should result in higher precursor concentrations in the layer that will generate higher ozone concentrations, which might subsequently impact shoreline sites.

Another feature that we observed during these episodes was a warming and stabilization of the aloft air on the high ozone days. This warming was observed in both the rawinsonde and RASS temperature data. Figure 7-3 shows rawinsonde profiles of potential temperature at Aberdeen Proving Ground, Maryland from July 31 to August 2, 1995 at 1100 EST. The temperature profiles show that aloft air warmed 1-2°K each day. In addition, the changing slope of the potential temperature profile indicates the stabilizing of the aloft air. This warming and stabilization of the aloft air was likely caused by two atmospheric phenomena:

• Sinking air in the high-pressure system that adiabatically warmed.

• Warm-air advection that occurred when aloft westerly to northwesterly winds on the back side of the high-pressure system transported warmer air into the region.

The effect of the warming and stabilization of the aloft air on mixing depth evolution was that more surface heat was needed to mix air to greater altitudes on the major exceedance days (August 1 and 2). For example, on July 31 only 600 °K m of heat is needed to mix the air to 1700 m msl. However, on August 1 and 2, it would take more than six times (3800 °K m) the heat to mix to this same altitude. In other words, on these two exceedance days it took longer for the surface to generate enough heat to create thermals that would overcome the aloft stability; hence, the convective boundary layer grew at a slower rate. To accurately resolve these day-to-day differences in the mixing depth evolution, it is necessary that the models have the "physics" to reproduce this aloft warming.