On the observation of cumulus convection using infrared satellite data

In the late 60's, Jim Holton described to me some lectures that Ooyama had given concerning the parameterization of cumulus convection. It was clear to me that there was much concerning this matter that had not previously occurred to me, and I began following developments more closely. After a burst of interest over the next decade, it seems to me that progress has been slow and uncertain. In particular, what knowledge we do have seems to have little impact on attempts to measure convective activity from space. In particular, knowledge of scales, life history and the explicit role of cumulonimbus convection in precipitation do not seem to have impacted the relevant choices of spatial and temporal resolution used in observational analyses of satellite data. The present lecture uses work with my student, Roberto Rondanelli, wherein we examine infrared and visible radiance from geostationary data at spatial resolutions of 5 km (or less) and temporal resolutions of one hour. The results illustrate the reasons for a number of problems using coarser resolutions, and lead to improved analyses which are tested against land based radar observations form Kwajelein. The results suggest promising improvements in a number of matters including the measurement of tropical rainfall using geostationary data, and the proper normalization of observations of the iris effect.