A method for estimating the amount of dead grass fuel based on spectral reflectance characteristics

Abstract

Estimation of the amount of dead grass fuel is essential for the assessment of risk of grassland fires. This paper develops a method to estimate the amount of dead grass fuel based on spectral reflectance. Samples of soil and dead grass were collected in the Songliao Plain, China. The spectral reflectance of these samples at different densities and at various wavelengths (350–2500 nm) was measured in the laboratory. A new spectral index for dead grass was designed based on the equivalent bands of a Moderate Resolution Imaging Spectroradiometer (MODIS) satellite image. In the short-wave infrared region of the electromagnetic spectrum, an absorption feature associated with cellulose and lignin was observed at ~2100 nm in the spectra of dead grass. This absorption feature was not present in the spectra of soil. This observation provides a basis for discriminating between dead grass and soil. The dead grass fuel index created using bands 6 and 7 of MODIS correlated significantly with the field measurements of the mass of dead grass fuel (R2 = 0.84). Hence, the dead grass fuel index can be used to produce an estimate of the amount of dead grass fuel, via the regression function identified above. Such methods of estimating the amount of dead grass fuel can contribute to studies of grassland fire hazard.