The Relative Merits of Narrowband Channels for Estimating Broadband Albedos

Abstract

Observations made with the current and proposed narrowband shortwave channels aboard the NOAA series of satellites were simulated for a number of different surfaces (ocean, vegetative land, desert, cloud and snow) using the ATRAD radiation model to study the relative merit of each channel and, in various combinations to predict the broadband albedo. Solar zenith angles were varied over the range from 0 to 60 degrees. The results indicated that for all of the surfaces considered there would be no significant difference in predicting the broadband albedo with either the current (0.58?0.68 ?n) or proposed (0.58?0.68 ?m) channel 1 of the AVHRR. The proposed narrower channel 2(0.84?0.87 ?m), however, would be a better predictor than the current wider channel 2(0.725?1.0 ?m). Channel 1 is better than channel 2 for surfaces of low or moderate reflectivity, while over snow, the error in using channel 2 would be less than half of that for channel 1. Combining channels 1 and 2 would reduce the error by about 50% for vegetation, ocean and snow. Adding the proposed channel 3A (1.58?1.64 ?m) to channels 1 and 2 would further improve the prediction of the broadband albedo. Channel 20(0.65?0.73 ?m) of the HIRS instrument was similarly studied to ascertain how well the broadband albedo would be predicted if the spectral filter was removed to widen the bandpass. Two different detectors (Si and InGaAs) with the current and a modified beamsplitter were considered. The results indicated that the modified beamsplitter was preferred. The use of this beamsplitter with the Si detector (0.46?1.04 ?m) gave the best prediction for ocean, vegetative land, and desert scenes, while the InGaAs detector (0.64?1.74 ?m) was best for cloud and snow scenes. Although the use of a widened channel 20 was shown to be less successful than the combination of channels 1, 2 and 3A of the AVHRR, flattening the response curve for the InGaAs detector using a compensating filter was comparable to using the AVHRR channels.