Investigation of morning PBL rise and impact on ozone production in regulatory simulations used in the Houston, TX SIP

A large region in southeast Texas, including the Houston metropolitan area, is in violation of the federal standard for ozone (O3). To demonstrate attainment of the O3 standard the Texas Commission on Environmental Quality (TCEQ) is developing CAMx simulations for 2005, 2006, and 2018 to support a 2010 State Implementation Plan (SIP) submission. The credibility of the simulations depends on accurate reproduction of both concentrations, and concentration sensitivity to reduction strategies. This analysis focuses on one component of the relationship between meteorology and ozone production, the influence of the time evolution of the simulated Planetary Boundary Layer (PBL).

For the SIP attainment demonstration, the EPA Guidelines recommends the use of a modeling based metric for every monitor called the relative reduction factor (RRF). This metric is a ratio of future (2018) to baseline (2005 and 2006) “qualified” 8-h ozone model predictions. The ozone predictions are qualified if they are within 15 km of the monitor, and above 85 ppb. This analysis focused on the 63 simulated days with “qualified” predictions that were used by the TCEQ for their RRF calculations. In addition, this analysis focused on the most photochemically active region within the modeling domain. This region, which we have labeled as Central Houston (CHOU), covers downtown Houston in the west and the highly industrialized Houston Ship Channel in the east. For this sub-domain, we used the 63 modeling days to quantify the sensitivity of ozone production to changes in precursor concentrations due to morning PBL rise. Throughout the analysis, Process Analysis (PA) techniques are relied upon to quantify the change in physical and chemical processes that lead to differences in ozone production.

Our analysis of SIP relevant modeling days revealed rapid and slow categories of morning PBL rise. Simulated days are classified as rapid or slow based on the CHOU sub-domain wide average PBL rise. Simulated days are “rapid risers” if there is a change of greater than 700 m/hr, between the hours of 6-11 LST, and they are “slow risers” if the change is less than 700 m/hr. Slow riser PBL results in relatively high morning volatile organic carbon (VOC) and NOx concentrations. PA data show that this results in a large fraction of OH radicals reacting with NO2 to produce HNO3, with virtually no H2O2 production. Slow riser days also exhibited relatively small amount of new OH radicals to initiate the chemistry necessary to oxidize the large amounts of NOx. Thus, slow riser days predicted high ozone concentrations with limited radical sources, under NOx inhibited conditions. Rapid riser high ozone days had a much different VOC/NOx ratio. Rapidly rising PBL increased dilution rates of the morning NOx by a factor of five, and also entrained more VOCs from aloft. The larger dilution and entrainment rates of these precursors had a significant impact on the chemical processes and resulted in a higher ozone production rate in the morning, and an earlier ozone peak. PA results show that a smaller fraction of OH reacting in termination processes, and a shift in the distribution of VOC that reacted with OH. Rapid riser simulation days become NOx limited and radicals begin to self terminate in the late afternoon producing H2O2.

These results suggest that PBL rise can determine whether ozone was produced under NOx limited, or NOx inhibited conditions. This was confirmed using an indicator based on the ratio of the two primary radical termination reactions; the reaction of OH with NO2 that produces nitric acid, and the reaction of HO2 with itself to produce peroxides. When NOx is abundant, the ratio of peroxide production to nitric acid production is low and ozone production is considered NOx inhibited. When NOx is scarce, the ratio of peroxide production to nitric acid production is high and the ozone production is considered NOx limited. We calculated the ratio of peroxide to nitric acid production, and found that the fast riser PBL day becomes NOx limited much earlier in the day than the slow riser, restraining ozone production to NOx availability. These results suggest that ozone was produced in a NOx limited environment, and would respond differently to precursor controls. If this simulation is deemed reasonable, then it has large implications for control policies in Houston.

The final step in the analysis was to quantify whether the response to controls on each of these type of morning rises is different. This was accomplished by calculating the EPA attainment metric, the RRF. An RRF was calculated for every monitor in Houston using all modeling days, and then calculated again using only rapid riser or only slow riser days. At any given monitor, up to 51% of days used for the RRF were rapid rise PBL days. The RRF results indicate that the all modeling day RRF is dependent on location. The western side of Houston saw the largest reductions in ozone with an RRF of 0.84 and the eastern side had very little change with a RRF near 0.92. At every monitor the rapid riser days were more responsive to precursor emissions than a slow riser days with a difference in RRF of as much as 0.07.

These data suggest that Houston is a complex environment, and that different parts of Houston would respond differently to controls. In addition, the morning PBL height influence on ozone production will require different targets of control. The RRF as stated in the EPA Guidance is an average of both these type of phenomena. This averaging results in an artificial response where, in many cases, is less responsive than actual conditions.