Abstract

The passive remote sensing of green leaf area index (GLAI) has utilized measurements of red (R) and near-infrared (NIR) radiance. Increased availability of mid-infrared (MIR) radiance data (e.g., 1.55 to 1.75 microns, from Landsat TM band 5) and the correlation between MIR radiance and GLAI has encouraged the use of these wavelengths for GLAI estimation. The aim of this paper is twofold; first, to refine the methodology used for such GLAI estimation, and second, to assess the effect of incorporating MIR radiance data into this methodology. Empirical models based upon airborne multispectral scanner measurements of R, NIR, and MIR radiance were inverted, and the GLAI of grassland was estimated to an accuracy of 18 to 58 percent (95 percent confidence level) for five levels of GLAI and + or -0.48 to 0.75 GLAI for a point. Refinement of the methodology increased the accuracy to 60 to 85 percent (95 percent confidence level) for five levels of GLAI and + or - 0.09 to + or - 0.12 GLAI for a point. The empirical models provided significantly similar estimates of GLAI regardless of the contribution of MIR radiance. It was concluded that measurements of MIR radiance added little extra information to measurements of R and NIR for the purposes of GLAI estimation.