Predicting soil albedo from soil color and spectral reflectance data

The albedo of earth surface features, such as soil, is an important compone nt of models that define land-surface meteorological processes. If land sur faces have no vegetative cover, soil properties determine the amount of sol ar radiation absorbed or reflected. We evaluated the influence of two soil properties, soil color and soil moisture, on soil albedo. Two soil moisture conditions were studied, air dry and wet, defined as the condition when th e water films are absorbed by the soil and no water glistens on the soil su rface. The albedos for 26 U.S. soils were measured with an Eppley pyranomet er, which integrates radiant energy in wavelengths between 0.3 to 2.8 mu m. Soil colors were measured with a Minolta Chroma Meter and spectral reflect ance curves from 0.45 to 0.9 mu m (measured in 0.1-mu m increments) were de termined with a multispectral radiometer. All measurements were made on <2- mm smooth soil surfaces, and the dry and wet data were combined for statist ical analyses. Soil albedos were significantly correlated with Munsell soil color value (r(2) = 0.93), blue (r(2) = 0.82), green (r(2) = 0.90), red (r (2) = 0.93), near infrared (NIR), (r(2) = 0.95), and sum of the four bands (r(2) = 0.94); however, the slopes and intercepts for these relationships w ere different. The 52 spectral curves yielded nine cluster groups, which mo stly related to the Munsell soil color value and soil albedo soil character istics. The 0.3- to 2.8-mu m albedos of smoothed soils fan be accurately es timated using the regression relationship: soil albedo (0.3-2.8 mu m) = 0.0 69 (color value) - 0.114. Using the regression equations presented here, sp ectral reflectance data in selected visible and NIR bands can also be used to predict albedo.