Abstract: Because some of the desertification evaluation indices cannot be directly obtained by remote sensing, and simple visual interpretation or direct classification evaluation methods have strong subjectivity, a comprehensive evaluation method of the desertification status based on remote sensing in combination with principal component analysis was put forward. Five desertification evaluation indices were selected following the representative, comprehensive, predominant and operable principles. These indices were directly obtained from inversed ETM⁺ data. Comprehensive evaluation model was set up based on principal component analysis of priori data and the determination of indices weight. The present desertification status of the whole study area was evaluated by the model. Results indicate that the model is more objective than that method of analysis and is more suitable for the desertification evaluation at regional scale. And the evaluation effects of the model reflected the superiority of the remote sensing method at large environmental scale.

Key words: desertification evaluation, index inversion, principle component analysis