Abstract Estimating landslide volume based on pre- and post-event elevation models without detailed field investigations is still a major challenge due to hidden failure surfaces. A MATLAB model named “SLIPFITTER” was developed to model failure surfaces using polynomial surface to fit the geometry of exposed scarps extracted from Digital Elevation Models (DEMs). To test our model, a series of analog experiments were designed to acquire data for model testing in a controlled environment. The test indicates the only option to simulate failure surfaces with only remotely sensed data, is to assume that the failure surface was outcropping at the toe of slopes. It is not possible to model landslides that do not reach the bottom of slopes without survey data. Using the assumption, the model was tested on real landslides, showing its capability of modeling for multiple landslides over an area in a single simulation, and the possibility to predict potential failure surfaces from a reported crack. The major limitations of the model include the high requirement for the input DEM quality, the incapability to consider underlying bedrock structures, and the difficulty of simulating landslides with complex movements.

中文翻译：

---

通过多项式表面拟合模拟滑坡破坏面

摘要 由于隐藏的故障面，在没有详细现场调查的情况下，基于事前和事后高程模型估算滑坡体积仍然是一个主要挑战。开发了一个名为“SLIPFITTER”的 MATLAB 模型，以使用多项式曲面对失效表面进行建模，以拟合从数字高程模型 (DEM) 中提取的暴露陡坡的几何形状。为了测试我们的模型，我们设计了一系列模拟实验来获取数据，以便在受控环境中进行模型测试。该测试表明，仅使用遥感数据模拟破坏面的唯一选择是假设破坏面在坡脚处露头。如果没有调查数据，就不可能对未到达斜坡底部的滑坡进行建模。使用该假设，该模型在真实的滑坡上进行了测试，展示了它在单个模拟中对一个区域内的多个滑坡进行建模的能力，以及从报告的裂缝中预测潜在故障面的可能性。该模型的主要局限性包括对输入 DEM 质量要求高、无法考虑下伏基岩结构以及难以模拟复杂运动的滑坡。