This study introduces two new approaches for determining tropospheric column ozone from satellite data. In the first method, stratospheric column ozone is derived by combining Upper Atmosphere Research Satellite (UARS) halogen occultation experiment (HALOE) and microwave limb sounder (MLS) ozone measurements. Tropospheric column ozone is then obtained by subtracting these stratospheric amounts from the total column. Total column ozone in this study include retrievals from Nimbus 7 (November 1978 to May 1993) and Earth probe (July 1996 to present) total ozone mapping spectrometer (TOMS). Data from HALOE are used in this first method to extend the vertical span of MLS (highest pressure level 46 hPa) using simple regression. This assimilation enables high-resolution daily maps of tropospheric and stratospheric ozone which is not possible from solar occultation measurements alone, such as from HALOE or Stratospheric Aerosols and Gas Experiment (SAGE). We also examine another new and promising technique that yields tropospheric column ozone directly from TOMS high-density footprint measurements in regions of high convective clouds. We define this method as the convective cloud differential (CCD) technique. The CCD method is shown to provide long time series (essentially late 1978 to the present) to tropospheric ozone in regions dominated by persistent high tropopause-level clouds, such as the maritime tropical Pacific and within or near midlatitude continental landmasses. In this our first study of the CCD and MLS/HALOE methods we limit analyses to tropical latitudes. Separation of stratospheric from tropospheric column ozone in the eastern Pacific tropics for January 1979 to December 1997 shows that the dominant interannual variability of stratospheric ozone is the quasi-biennial oscillation (QBO), whereas for tropospheric ozone it is driven by El Ni¿o events. For validation purposes, both the CCD and assimilated UARS MLS/HALOE results are compared with ozonesonde data from several southern tropical stations. Despite all three measurements being distinctly different in sampling and technique, all three show good qualitative agreement. ¿ 1998 American Geophysical Union