Landscape Evolution Models (LEMs) simulate the movement of water and sediment over the landscape. Although much progress has been made in the development of LEMs, few have been tested in rivers subject to anthropogenic impacts that produce high energy flows, transporting large amounts of sediment and causing significant geomorphic changes. As such, it remains uncertain if LEMs are useful and stable under relatively short term ‘extreme’ geomorphic conditions. To shed light on this topic we use a LEM (CAESAR-Lisflood) and historical documents to develop a detailed reach scale model of the Kander river (Switzerland). This model was used to simulate the unintended impacts of engineering works, occurring in 1714, that deviated the Kander river into a lake and resulted in a large decrease in base level of the river. In 10 years, the model simulates knickpoint propagation that rapidly erodes 2.5 million m³ of sediment and produces a remarkable 27 m of channel erosion. Simultaneously, the model develops the formation of a delta via frequent avulsions. Model testing is performed by comparing model predictions against historical observations of channel incision, knickpoint location, and delta spatial extent. Overall, model error is low and the model remained stable as results do not contain erratic erosion or deposition. Importantly the model suggests that downstream processes occurring at and near the delta have an effect on upstream channel erosion. We also recommend that studies replicating historic landscape changes with LEMs reduce uncertainty in hydrological inputs.

景观演化模型（LEM）可模拟水和沉积物在景观上的运动。尽管LEM的开发已经取得了很大进展，但是在受到人为影响而产生高能流，运输大量沉积物并引起重大地貌变化的河流中，很少进行测试。因此，仍然不确定LEM在相对短期的“极端”地貌条件下是否有用和稳定。为了阐明这一主题，我们使用LEM（CAESAR-Lisflood）和历史文献来开发Kander河（瑞士）的详细河段规模模型。该模型用于模拟1714年发生的工程设计的意外影响，该工程使Kander河转向湖泊，并导致河底水位大大降低。10年后³的沉积物，产生显着的27 m的河道侵蚀。同时，该模型通过频繁撕脱形成三角洲。通过将模型预测与通道切口，拐点位置和增量空间范围的历史观测值进行比较，来执行模型测试。总体而言，模型误差低，并且模型保持稳定，因为结果不包含不稳定的侵蚀或沉积。重要的是，该模型表明，在三角洲及其附近发生的下游过程对上游河道侵蚀有影响。我们还建议使用LEM复制历史景观变化的研究，以减少水文输入的不确定性。