[en] Estimation of trends in ozone air quality is a complex problem because of the strong influence of meteorology on ozone and its precursors. In this paper we present a physically-consistent methodology to link changes in ozone to changes in precursor emissions in the presence of meteorological fluctuations. Time series ozone concentrations (O₃), nitrogen oxides (NO_x) and non-methane hydrocarbons (NMHC) measured from 1985 to 1993 in the Los Angeles Air Basin are first separated into different time scales using the Kolmogorov-Zurbenko (KZ) filter. Linear regression analysis is performed on meteorological variables and ozone and its precursors on each time scale. The downward trend in the daily hourly maxima of O₃ is about twice that in the daily hourly mean O₃. The results demonstrate that emission control programs implemented over Southern California were very effective in improving ambient ozone levels during a period of robust economic growth. (Author)