Field experiments were conducted in a split plot with three rice varieties. Rice samples were exposed to

sulfur dioxides of different concentrations inside a fumigation chamber. After the measurement of visible and near

infrared radiance of the canopies of rice，leaves of rice canopy were sampled to analyze pH of extracted leaf

fluid and chlorophyll concentration and sulfur content of rice leaves. The correlation between these biophysical-

biochemical indices and the first derivative spectral reflectance shows that the first derivative spectral

characteristics which are significantly correlated with sulfur content, pH and cholorophyll concentration can be

selected to estimate the biophysical-biochemical indices with regression models using 689nm, 584nm and 570nm at

the tiller stage, and using 689nm, 584nm and 585nm at the heading stage respectively. The first derivative

spectral reflectance of 689nm, 584nm and 570nm at the jointing stage, and 689nm, 584nm and 585nm at the filling

stage can be used to test the accuracy of regression models. The correlation coefficients between measured values

of the jointing stage and estimated values of the tiller stage for estimating models are 0.498 (estimated sulfur

content), 0.884 (estimated pH) and 0.63 (estimated chlorophyll concentration) respectively. In addition, the

correlation coefficients between measured values and estimated values are 0.66 (sulfur content), 0.768

(chlorophyll concentration) and 0.50 (pH) respectively.