There is growing consensus that land degradation assessments should occur at local to regional scales, as global assessments fail to capture the complexity of land degradation processes, and how they manifest in different social‐ecological contexts. We have tested a statistically rigorous method for regional land degradation assessment in drylands, using fuzzy classification theory and techniques, that results in the creation of a habitat condition archetype for the Namaqualand Hardeveld bioregion of the Succulent Karoo biome in South Africa. The methodology initially groups field data into classes based on known indicators of land degradation, before predicting observed field cover with 16 remote sensing variables, derived from both LANDSAT 8 and SENTINEL‐2A satellite data. Using fuzzy classification techniques, projections of field cover from the resulting model are combined with variables representing a potential change in cover to produce a map of the habitat condition of the bioregion. Approximately 17% of the bioregion had habitat condition values less than one standard deviation smaller than the mean and could therefore be considered degraded. To provide context, this map is presented with potential drivers of change for the bioregion. The map, and accompanying drivers of change, is ultimately presented in a Google Earth Engine app for easy interpretation and use.

中文翻译：

---

干旱地区环境土地退化评估：以多肉植物Karoo生物群落的Namaqualand Hardeveld生物区为案例研究

越来越多的共识认为，土地退化评估应在地方到区域范围内进行，因为全球评估未能反映出土地退化过程的复杂性及其在不同社会生态环境中的表现方式。我们已经使用模糊分类理论和技术对干旱地区的土地退化评估进行了统计学上严格的测试，结果为多肉植物Karoo的Namaqualand Hardeveld生物区创建了栖息地条件原型南非的生物群落。该方法最初根据已知的土地退化指标将野外数据分组，然后根据LANDSAT 8和SENTINEL-2A卫星数据得出的16个遥感变量预测实地覆盖率。使用模糊分类技术，将所得模型的田间覆盖投影与代表覆盖潜在变化的变量组合起来，以绘制生物区域栖息地状况的地图。大约17％的生物区域的栖息地条件值小于一个小于平均值的标准差，因此可以认为其退化。为了提供背景信息，该地图提供了生物区域变化的潜在驱动因素。地图，以及随之而来的变化驱动力，