Assessment of biases in MODIS surface reflectance due to Lambertian approximation

摘要: Using MODIS data and the AERONET-based Surface Reflectance Validation Network (ASRVN), this work studies errors of atmospheric correction caused by Lambertian approximation. On one hand, approximation greatly simplifies radiative transfer model, reduces size look-up tables, makes operational algorithm faster. other uncompensated scattering model systematically biases results. For example, for a typical bowl-shaped bidirectional reflectance distribution function (BRDF), derived is underestimated at high solar or view zenith angles, where BRDF high, overestimated low angles low. The magnitude grows with amount in atmosphere, i.e., shorter wavelengths higher aerosol concentration. slope regression surface vs. ASRVN factor (BRF) about 0.85 red 0.6 green bands. This error propagates into BRDF/albedo algorithm, slightly reducing overall anisotropy BRDF. results small negative bias spectral albedo. An assessment GSFC (Greenbelt, USA) validation site shows albedo reduction 0.004 near infrared, 0.005 red, 0.008