The highest correlative relations for air pollution levels are often with meteorological variables such as temperature and wind speed. Today, sophisticated gridded high-resolution meteorological models are used to produce meteorological fields that drive chemical transport models for air quality management. Errors in specification of the physical atmosphere such as temperature, clouds and winds can affect the air quality predictions. Additionally, the efficiency and efficacy of emission control strategies can be compromised by errors in the meteorological fields. In this paper, the role of meteorology in air quality behavior, primarily from the viewpoint of regional ozone modeling as carried out in the U.S., is reviewed. Particular attention is given to physics and new techniques for improving meteorological model performance. Uncertainties in model turbulent mixing in the nighttime boundary layer, where large model differences exist, are examined. The role of spatial mesoscale features such as topography and land/water systems in models are discussed. The nocturnal low level jet, a mesoscale temporal and spatial feature, and its impact on air quality are examined. Traditional air quality concerns have focused on synoptic conditions at the center of high-pressure systems. However, high ozone levels have also been associated with stationary fronts. The ability of models to capture mesoscale structure and yet retain synoptic structure and its timing is challenging. Data assimilation and its ability to improve model performance are examined. Particular attention is given to vertical nudging strategies that can affect formation of the nocturnal low-level jets. Finally, clouds can have a major impact on air quality since insolation impacts temperature, biogenic emissions and photolysis rates and extremes in stability. Traditional techniques, which attempt to insert cloud water where there is not dynamical support, can lead to additional errors. New dynamical approaches for improving model cloud performance are discussed.

High mixing ratios of ozone along the shores of Lake Michigan have been a recurring theme over the last 40 years. Models continue to have difficulty in replicating ozone behavior in the region. Although emissions and chemistry may play a role in model performance, the complex meteorological setting of the relatively cold lake in the summer ozone season and the ability of the physical model to replicate this environment may contribute to air quality modeling errors. In this paper, several aspects of the physical atmosphere that may affect air quality, along with potential paths to improve the physical simulations, are broadly examined. The first topic is the consistent overwater overprediction of ozone. Although overwater measurements are scarce, special boat and ferry ozone measurements over the last 15 years have indicated consistent overprediction by models. The roles of model mixing and lake surface temperatures are examined in terms of changing stability over the lake. From an analysis of a 2009 case, it is tentatively concluded that excessive mixing in the meteorological model may lead to an underestimate of mixing in offline chemical models when different boundary layer mixing schemes are used. This is because the stable boundary layer shear, which is removed by mixing in the meteorological model, can no longer produce mixing when mixing is rediagnosed in the offline chemistry model. Second, air temperature has an important role in directly affecting chemistry and emissions. Land–water temperature contrasts are critical to lake and land breezes, which have an impact on mixing and transport. Here, satellite-derived skin temperatures are employed as a path to improve model temperature performance. It is concluded that land surface schemes that adjust moisture based on surface energetics are important in reducing temperature errors.

A lightning nitrogen oxides (LNOx) emissions model using satellite‐observed lightning optical energy is introduced for utilization in Air Quality modeling systems. The effort supports assessments of air‐quality/climate coupling as related to the influence of LNOx on atmospheric chemistry. The Geostationary Lightning Mapper (GLM), International Space Station Lightning Imaging Sensor (ISS‐LIS), and the Tropical Rainfall Measuring Mission (TRMM) LIS data are used to examine the efficacy of the method, extend the previously derived LNOx record, and demonstrate a path for using ISS‐LIS observations to cross‐calibrate regional LNOx estimates from the future global constellation of geostationary lightning observations. A detailed evaluation of the GLM data set is provided to establish the robustness of observations for LNOx estimates and to make preliminary assessments of the LNOx emissions model. Seasonal and geographical variation, land/ocean contrast, and annual fluctuation in the GLM observed lightning activity and flash optical energy are provided. GLM detection substantially degrades with the increase in the field of view, resulting in 44% more flashes and 40% less optical energy observation by GLM‐16 (compared to GLM‐17) to the east of the middle‐longitude between the two mappers (106.2°W). Regular horizontal striations are found in the optical energy product. On average, GLM flashes matched to the cloud‐to‐ground flashes have ∼30% longer duration, 50%–70% more extension, and ≥100% higher optical energy compared to the unmatched flashes (assumed to be intra‐cloud). The results from summer‐long chemical transport simulations using LNOx generated from the emission model agrees with previous studies and shows consistency across the GLM/LIS data sets.