Evaluation of land degradation for resource management
Neural network modelling was applied to evaluate land degradation hazard of a marginal agricultural system using a resource map database. Degradation degree and extent as supervisors, elevation, slope, direction, soil and vegetation as explanatory factors, were selected for the neural network modelling. Precisions of the modelling for the evaluation of the degree and extent of degradation were 86% and 79%, respectively. The evaluation of the degradation hazard was found valid in the northern foothills and plain during the field verification, where hazardously evaluated areas corresponded to sites with frequent gully erosion.
Japan International Research Center for Agricultural Sciences Animal Production and Grassland Division
International Center for Agricultural Research In the Dry Areas
- Term of research
- Responsible researcher
YAMAMOTO Yukiyo ( Animal Production and Grassland Division )
FUJITA Haruhiro ( Animal Production and Grassland Division )
GINTZBURGER Gustav ( International Center for Agricultural Research In the Dry Areas )
- Publication, etc.
- Japanese PDF