The interplay between natural site conditions and farming raises erosion by water above geological background levels. We examined the hypothesis that farmers take erosion into account in their farming decisions and switch to farming practices with lower erosion risk the higher the site-specific hazard becomes. Erosion since the last tillage was observed from aerial orthorectified photographs for 8100 fields belonging to 1879 farmers distributed across Bavaria (South Germany) and it was modeled by the Universal Soil Loss Equation using highly detailed input data (e.g., digital terrain model with 5 × 5 m² resolution, rain data with 1 × 1 km² and 5 min resolution, crop and cropping method from annual field-specific data from incentive schemes). Observed and predicted soil loss correlated closely, demonstrating the accuracy of this method. The close correlation also indicted that the farmers could easily observe the degree of recent erosion on their fields, even without modelling. Farmers clearly did not consider erosion in their decisions. When natural risk increased, e.g. due to steeper slopes, they neither grew crops with lower erosion potential, nor reduced field size, nor used contouring. In addition, they did not compensate for the cultivation of crops with higher erosion potential by using conservation techniques like mulch tillage or contouring, or by reducing field size. Only subsidized measures, like mulch tillage or organic farming, were applied but only at the absolute minimum that was necessary to obtain subsidies. However, this did not achieve the reduction in erosion that would be possible if these measures had been fully applied. We conclude that subsidies may be an appropriate method of reducing erosion but the present weak supervision, which assumes that farmers themselves will take erosion into account and that subsidies are only needed to compensate for any disadvantages caused by erosion-reducing measures, is clearly not justified.

中文翻译：

---

农民的耕作方式反映了其对水土流失的行为

自然场地条件与耕作之间的相互作用加剧了高于地质本底水平的水的侵蚀。我们检验了以下假设：农民在其耕种决策中考虑到侵蚀，并且转向特定耕作方式时，特定地点的危害越大，侵蚀风险越低。自上次耕作以来的侵蚀，是从分布在巴伐利亚（德国南部）的1879个农民的8100个田地的正射照片上观察到的，并使用通用的土壤流失方程使用高度详细的输入数据（例如，具有5×5的数字地形模型）对侵蚀进行了建模。 m ²分辨率，降雨数据为1×1 km ²和5分钟的分辨率，作物和耕种方法（来自激励计划的年度田间特定数据）。观测到的土壤流失与预测的土壤流失密切相关，证明了该方法的准确性。密切的相关性还表明，即使不进行建模，农民也可以轻松观察其田地最近的侵蚀程度。农民显然在他们的决定中没有考虑侵蚀。当自然风险增加时，例如由于陡峭的斜坡，他们既不会种植具有较低侵蚀潜力的作物，也不会缩小耕地的面积，也不会使用等高线。此外，他们没有通过采用覆盖耕作或等高线等保护性技术，或通过缩小田地面积来补偿具有较高侵蚀潜力的农作物的种植。只有补贴措施，例如覆盖耕作或有机耕作，仅适用于获得补贴所必需的绝对最低限度。但是，如果完全采用这些措施，就无法减少侵蚀。我们得出的结论是，补贴可能是减少侵蚀的一种适当方法，但是目前薄弱的监督假设农民自己将考虑侵蚀，并且仅需要补贴以补偿因减少侵蚀而造成的不利条件，显然是没有道理的。