Biophysical processes and people's livelihoods largely depend on ecosystem productivity; thus, land degradation is harmful for local sustainable development. Most previous assessments of land degradation used a comprehensive index system or trend analysis without localized benchmarks; thus, they were not fully objective. Here, we assessed the land degradation in the Heihe River Basin (HRB), China using the local net primary production (NPP) scaling (LNS) method. We improved the method by dividing land capability classes (LCC) using remote sensing data, and determining the reference NPP for each LCC to represent the productivity under optimal environmental conditions. To assess the extent, severity and trend of land degradation, we obtained LNS percent values which was defined as the ratio of present NPP to reference NPP. We categorized our identified land degradation hotspots, determined their causes and proposed potential strategies for land conservation. Our results showed that nearly one quarter of the area was degraded land, mainly located in the upper stream. The average NPP was 35.75% lower than the reference NPP, with a reduction amount of 0.82 GgC yr −1 . Only the lower stream showed a significant decreasing trend in LNS percent values. The areas with a positive trend in LNS percent values were primarily observed in the middle stream croplands. We found the degradation condition deteriorated first and recovered later in the HRB. Land use change contributed more than climate change to the local NPP of the degraded areas. Our research provides a more general framework for assessing land productivity reduction induced by anthropogenic and natural processes in the drylands.
K E Y W O R D Sdrylands, land capability classes, land degradation, local net primary production scaling, NPP, the Heihe River Basin