1Department of Earth, Atmospheric and Planetary Sciences, Massachusetts Institute of Technology, USA
2Centro de Ciencias de la Atmósfera, Universidad Nacional Autónoma de México, Mexico
3General Direction of the National Center for Environmental Research and Training (CENICA), National Institute of Ecology (INE), Mexico
4Secretaría del Medio Ambiente, Gobierno del Distrito Federal, Mexico
Abstract. MCMA-2003 was a major field campaign investigating the atmospheric chemistry of the Mexico City Metropolitan Area (MCMA) in April of 2003. This paper describes the wind circulation patterns during the campaign both within the Mexico City basin and on the regional scale. ''Time roses'' are introduced to concisely analyze the diurnal wind patterns. Three episode types were identified that explain the conditions encountered: ''O3-South'', ''Cold Surge'' and ''O3-North''. These can be diagnosed from a combination of synoptic and basin observations based on whether the day was predominantly cloudy, or whether the O3 peak was in the north or south of the basin. O3-South days have weak synoptic forcing due to an anti-cyclone over the eastern Pacific. Strong solar heating leads to northerly flows in the basin and an evening shift due to a gap flow from the south-east. Peak ozone concentrations are in the convergence zone in the south of the city. Cold Surge days are associated with ''El Norte'' events, with strong surface northerlies bringing cold moist air and rain. Stable conditions lead to high concentrations of primary pollutants and peak ozone in the city center. O3-North days occur when the sub-tropical jet is closer to Mexico City. With strong westerlies aloft, the circulation pattern is the same as O3-South days except for a wind shift in the mid-afternoon leading to ozone peaks in the north of the city. This classification is proposed as a means of understanding pollutant transport in the Mexico City basin and as a basis for future meteorological and chemical analysis. Furthermore, model evaluation and design of policy recommendations will need to take into account the three episode types.