Comparison of RADIATIVE FLUXES from INSAT and ERBE

Abstract

Accurate knowledge of radiative fluxes at the top of the atmosphere (TOA) is necessary to study the variability of climate. The Indian geostationary satellite INSAT has been measuring radiative fluxes in a narrow spectral band for the past 15 years. This data can be useful for climate studies if it is converted to total radiative fluxes. In this thesis, an accurate algorithm has been developed to convert the longwave (LW) radiance measured by INSAT 1B in the 10.5–12.5 m region to total outgoing longwave radiation (OLR). A radiative transfer (RT) model was used to explore the relationship between narrowband and broadband radiative fluxes. An empirical method was used to obtain a more accurate algorithm for conversion of a narrowband window channel flux to a broadband flux. In this analysis, we have shown that INSAT OLR data calculated using the algorithm developed in this thesis agrees within 5–10 W m ² with the Earth Radiation Budget Experiment (ERBE), which is the accuracy limit of ERBE data. This indicates that the INSAT OLR is as good as ERBE OLR for studies of regional climate variability. Moreover, the high frequency INSAT data can be used to study the diurnal cycle of radiation more accurately, which is not possible using ERBE data. It was found that the error in INSAT OLR calculated using the new algorithm depends on the number of images available for that month. During the period April 1988 to March 1989, the lowest RMS error (~6 W m ²) occurred in March 1989 with 243 images, and the highest RMS error (~12 W m ²) occurred in June 1988 with 154 images. An algorithm was developed to identify clear sky using infrared (IR) data. On a regional mean basis, the difference between INSAT and ERBE clear sky OLR was about 5–10 W m ². A strong diurnal cycle in clear sky OLR was noticed, particularly over land, with the peak occurring a few hours after local noon. Most of the error in OLR was found to be related to the total column water vapour and vertical water vapour distribution. INSAT albedo showed a negative bias over ocean when compared to ERBE, with no such bias over land. This low albedo bias of INSAT was removed by adding a constant term equal to 2% over the ocean. Two clear sky identification algorithms were developed-one using only visible data and the other using both visible and IR data. The second algorithm, using both spectral channels, was more accurate when compared to ERBE. The spatial pattern of LW, shortwave (SW), and net cloud radiative forcing (CRF) from INSAT was similar to that of ERBE for July 1988 and January 1989. INSAT CRF based on the new algorithm was close to ERBE CRF, with an error of less than 10 W m ². INSAT CRF based on the old algorithm was also accurate due to cancellation of biases of the same sign in different terms of the cloud forcing equation. Net radiation was calculated using the OLR and albedo datasets from INSAT. The error in net radiation was within 10 W m ² when compared to ERBE. It was noticed that the contributions of absorbed SW radiation and OLR to the total error in net radiation were equal in magnitude. In some regions, such as the Saudi Arabian desert from December 1988 to January 1989, the RMS error between ERBE and INSAT estimated net radiation (~10 W m ²) was found to be much less than the RMS error between two ERBE satellite observations (~20 W m ²). This thesis demonstrates that operational satellites such as INSAT, equipped with a narrowband sensor, can be used effectively for the study of the radiation budget.