Annual, Seasonal, and Interannual Variability of Air–Sea Heat Fluxes in the Indian Ocean

• DOI: 10.1175/jcli4163.1
• Publication Date: 2007-06-29T19:02:35Z
• Authors: Lisan Yu; Xiangze Jin; Robert A. Weller
• Keywords: Shortwave; Outgoing longwave radiation; Climate Forecast System

Abstract This study investigated the accuracy and physical representation of air–sea surface heat flux estimates for Indian Ocean on annual, seasonal, interannual time scales. Six products were analyzed, including newly developed latent sensible fluxes from Objectively Analyzed Air–Sea Heat Fluxes (OAFlux) project net shortwave longwave radiation results International Satellite Cloud Climatology Project (ISCCP), analysis Southampton Oceanography Centre (SOC), National Centers Environmental Prediction reanalysis 1 (NCEP1) reanalysis-2 (NCEP2) datasets, European Medium-Range Weather Forecasts operational (ECMWF-OP) 40-yr Re-Analysis (ERA-40) products. paper presents six in depicting mean, seasonal cycle, variability into ocean. Two series situ measurements, one taken a 1-yr Arabian Sea Experiment field program other 1-month Joint Monsoon Interaction (JASMINE) Bay Bengal used to evaluate statistical properties over measurement periods. The consistency between scales was using standard deviation physically based correlation analysis. has three findings. First all, large differences exist mean value Part may be attributable bias numerical weather prediction (NWP) models that underestimates Ocean. Along JASMINE ship tracks, four NWP modeled all have sign opposite observations, with NCEP1 being underestimated by 53 W m−2 (the least biased) ECMWF-OP 108 most biased). At buoy site, also an underestimation bias, smallest 26 largest 69 (NCEP1). On hand, OAFlux+ISCCP best comparison at both sites. Second, effect changes scale. Despite fact is biased significantly, there no major cycle except ECMWF-OP. latter does not fixed due frequent updates model platform. Finally, among (OAFlux+ISCCP, ERA-40, NCEP1, NCEP2) can studying variability, ERA-40 Qnet good as judged measures. shows broad agreement two products, varying details. By comparison, NCEP2 representative variabilities basin