Analyses of FORMOSAT‐3/COSMIC humidity retrievals and comparisons with AIRS retrievals and NCEP/NCAR reanalyses

Abstract

The specific humidity retrieved from the FORMOSAT‐3/Constellation Observation System for the Meteorology, Ionosphere, and Climate (COSMIC) (FC) occultation observations of the GPS radiowaves were analyzed and compared with radiosonde observations, aircraft dropsonde observations, Atmospheric Infrared Sounder (AIRS) retrievals, and National Centers for Environmental Prediction/National Center for Atmospheric Research (NCEP/NCAR) reanalyses. For the limited number of radio soundings at the sites of small islands near Taiwan coincident in space and time with FC observations, the difference of humidity in individual cases is <40% below ∼7 km without obvious bias. Above this level, the humidity of radiosonde observations is systematically lower than that of FC observations, which is consistent with previous studies showing the dry bias of certain radiosonde observations. The humidity of FC observations differs significantly throughout the troposphere with dropsonde observations surrounding a strong typhoon. The difference reaches 100% in the middle and upper troposphere. The cause of the large discrepancy is likely to be the large horizontal variation of humidity in regions of an active weather system as shown in closely spaced dropsonde observations, where FC observations represent humidity averaged over humid and dry regions in the long path of GPS radiowaves, whereas dropsonde observations are a point measurement. Global distributions of specific humidity were compared between FC retrievals and the AIRS retrievals and NCEP/NCAR reanalyses. There is a good agreement between AIRS and FC humidity retrievals. The discrepancy is generally within 15%. The discrepancy between the NCEP/NCAR and FC humidity is significant, especially in the upper troposphere. For the layer 300–400 hPa, the NCEP/NCAR humidity is higher than the FC humidity by >30% over most of oceanic regions.