Synoptic Scale Moisture Variation over the Tropical Pacific Ocean

Abstract

Moisture observing capability is surveyed over the tropical northeast Pacific Ocean. Data are taken from late January 1979 during FGGE. Emphasis is on diagnosis of synoptic scale systems in data sparse areas. The capabilities and limitations of five observing systems are examined: surface observations, satellite cloud imagery, radio- and dropsondes, satellite individual channel brightness temperatures, and model analysis from the European Center for Medium Range Weather Forecasts. Both qualitative and quantitative intercomparisons are made. Surface observations carry almost insignificant moisture information over the tropical oceans. The capability of GOES imagery is well known; however, clouds mask important moisture structure and do not always define moisture patterns well, even at cloud level. Soundings were adequate for synoptic diagnosis, ifthere were enough of them; however, FGGE dropsondes were limited in detail. Satellite channel data provide thorough coverage and show some detail even in nearly overcast regions; ambiguity of interpretation remains a problem. Given the lack of moisture observations for initialization, the ECMWF analysis provides surprisingly realistic moisture patterns. Quantitative intercomparison of data is generally discouraging. Field comparisons of model analysis and satellite observations are poor, with only marginal statistical significance. Both systems, however, clearly define the synoptically active regions in their variability statistics; they both perform better in moist regions, where quantitative estimates of moisture are most important. Comparisons with radiosondes are poor as well. Correspondence of analysis, satellite and radiosondes is good in moist regions, but all three have serious observational problems when radiosonde-observed relative humidity falls below 50%. Each of the five systems describes detail not contained in the other systems. Most importantly, quantitative satellite moisture channel data can be used synoptically, and model analysis provides useful synoptic moisture information, even without initial moisture observations.