Abstract
Satellite remote sensing provides the only practical means of measuring and monitoring the total albedo of the earth's surface at regional and global scales. However, satellite sensors measure the spectral bidirectional reflectance of the earth's surface (plus atmospheric effects). Thus algorithms must be developed to compute, first, the spectral hemispherical reflectance from the bidirectional reflectance for discrete wavebands and, second, to estimate the total albedo (0.3–4.0 μm) from these spectral hemispheric reflectances.
In this paper we present results of a field experiment in which the nearly complete bidirectional reflectance distribution function is measured by the PARABOLA instrument for Russian steppe grassland sites. The spectral hemispheric reflectances were computed by angular integration of these measurements for three wavebands: red (650–670 nm), near infrared (810–840 nm), and shortwave infrared (1620–1690nm). Estimates of total albedo were then made by weighting the spectral hemispheric reflectances by the total solar irradiance in three broadband spectral regions (300–700 nm, 700–1300 nm, and 1300–4000 nm) which were assumed to be represented by the narrowband measurements. These calculations resulted in albedo estimates with a mean relative error of 14% as compared to pyranometer measured albedo. Since vegetation reflectance varies significantly over each of the three broadband regions additional reflectance weighting factors were computed from a combination of high spectral resolution green canopy reflectance data and corresponding computed spectral solar irradiance. This additional reflectance weighting resulted in a reduction in the mean relative error to 3.9% relative to pyranometer measured albedo.
It is noted that the three spectral bands of the PARABOLA instrument data reported here are similar to those of the spectral wavebands which are planned for future AVHRR sensors on NOAA satellites. Therefore, the results and techniques presented here may be useful for future global albedo estimation utilizing AVHRR sensors.
Keywords:
Notes
Corresponding author.