Observing requirements for geostationary satellites to enable ozone air quality prediction

Abstract. We conduct a variety of analyses to support mission planning for geostationary satellite measurements of atmospheric composition. We carry out a simplified observing system simulation experiment (OSSE) using a photochemical box model and its adjoint integrated with a Lagrangian 4-D-variational data assimilation system. Using this framework in conjunction with pseudo observational constraints we estimate surface emissions and assess the improvement in ozone air quality forecasting and prediction. We use an analytical model as our principle method of conducting uncertainty analyses, which is the primary focus of this work. We investigate the impacts of changing the observed species (e.g., ozone, carbon monoxide (CO), nitrogen dioxide (NO₂), and formaldehyde (HCHO)), observation frequency and quality upon the ability to predict the magnitude of summertime peak ozone events, characterize the uncertainties of those predictions, and the performance of the assimilation system. We use three observed species scenarios: CO and NO₂; ozone, CO, and NO₂; and HCHO, CO and NO₂. These scenarios are designed to test the effects of adding observations of either ozone or HCHO to an existing CO and NO₂ observing system. The studies were conducted using the photochemical model setup to simulate a range of summertime polluted environments spanning NO_x limited to volatile organic compound (VOC) limited conditions. As the photochemical regime changes the relative importance of trace gas observations to constrain emission estimates and subsequent ozone forecasts varies. For example, adding ozone observations to an NO₂ and CO observing system is found to decrease ozone prediction error under NO_x and VOC limited regimes, and complimenting the NO₂ and CO system with HCHO observations would improve ozone prediction in the transitional regime and under VOC limited conditions.