Surface Heat Fluxes Using Satellite Observations: A Case Study in the Northwest Pacific

Abstract

Satellite-observed surface-level meteorological parameters are used to derive 10-day mean surface heat fluxes of solar, longwave, sensible, and latent heat fluxes in the northwest Pacific. Cloud amount in tenths were taken from Geostationary Meteorological Satellite cloud data, published by the Japan Meteorological Agency. Sea surface temperature (SST) was derived from the NOAA Advanced Very High Resolution Radiometer observations using the cross-product SST. Surface wind speed was obtained from the Special Sensor Microwave/Imager (SSM/I). Air temperature and surface level humidity were derived statistically from the SSM/I precipitable water. Synoptic flux estimates using satellite observations are compared with earlier estimates for this region (same season) derived from in situ data and were found to be reasonable. While the Kuroshio regime shows large net heat loss toward the south, it decreases and has heat surplus. For an evaluation of the satellite-derived heat fluxes, these are compared with flux estimates from in situ observations at Ocean Weather Station Tango and were found to be comparable within 10 W m?2, except for latent heat flux for which the discrepancy is 30 W m?2. This high discrepancy between in situ and satellite-derived latent heat flux is attributed to the higher uncertainties involved in satellite-derived latent heat flux compared to other heat flux components. To find out the validity and applicability of satellite-derived heat fluxes an extensive error analysis is provided. Uncertainties in the flux estimates, resulting from the use of bulk method and remotely sensed data are worked out and are presented for individual and total fluxes. These uncertainties in satellite-derived fluxes are further compared with uncertainties for the same flux values resulting from climatological ship observations. For net satellite-derived heat flux varying from 0 to 300 W m?2 the uncertainties were found to be of the order of 50?90 W m?2. For the same range of flux values the possible uncertainties in the ground-based climatological estimate were from 40 to 80 W m?2. This increase in uncertainties for satellite-derived fluxes was found to be due to remote sensing error. By incorporating the satellite-measured radiance and atmospheric radiative transfer model in the estimation of the radiation budget, the total error bar can be reduced to 30?80 W m?2 from 50?90 W m?2, showing that satellite-derived heat fluxes are having accuracies better than (or equal to) those of climatological estimates. In spite of these uncertainties, satellite-derived heat fluxes can be used in upper-ocean heat balance studies. The above conclusion is based on two facts. 1) Near-real-time 10-day mean satellite fluxes will be more informative in finding the natural variabilities than the climatological mean fluxes. 2) Though the uncertainties are substantial, when we study the interannual variabilities or anomalies, a larger part of the uncertainty corresponding to the systematic errors in observations and bulk parameterizations will be nullified.