Abstract: Hyperspectral reflectances of Populus euphratica and P. pruinosa leaves were obtained using field spectrometer in the upper reaches of Tarim River. Leaf nitrogen content was also determined. The bands of 657 nm and 689 nm could be used as sensitive bands to estimate leaf nitrogen content of P. euphratica and P. pruinosa, respectively. The determination coefficient of the linear model between reflectance and leaf nitrogen content of P. euphratica was higher than that of P. pruinosa. Highly significant linear relationships were also found between leaf nitrogen content and hyperspectral characteristic variables, vegetation index and red edge position, respectively. The determination coefficients of linear models between leaf nitrogen content and vegetation index were higher than those of other models. Therefore, inversion models based on vegetation index have good predictability. Significant relationships also existed between leaf nitrogen content and spectral areas. Therefore, spectral areas also can be used to estimate leaf nitrogen content. It was feasible to estimate nitrogen content of P. euphratica and P. pruinosa leaves based on hyperspectral reflectance and parameters transformed from hyperspectral reflectance.

Key words: water use efficiency, grain yield, dry matter accumulation and allocation, tillage practices, photosynthetic characteristics