Landslides most commonly occur in the rainy season, resulting in damage to infrastructure and human lives. An early prediction framework for landslides would clearly help to mitigate damage. In this work, a prediction framework for shallow landslide initiation was developed and validated with a real case study. To assess the reliability of the prediction framework, back-analysis of a landslide that occurred in Malin village, Maharashtra, India on July 2014 was performed. Relations of landslide stability with soil saturation were established through a physically based approach using GeoStudio software. A leaky barrel algorithm was developed for the study location to monitor the effect of rainfall' on soil saturation. Simulation results of landslide stability were compared with the rainfall–soil saturation algorithm based on the leaky barrel. The presented framework was found to have good predictability of shallow landslide occurrence and is therefore recommended for real-time monitoring of landslide-prone locations.

中文翻译：

---

通过饱和度监测和预测推进滑坡预警系统

山体滑坡最常发生在雨季，导致基础设施和人员生命受损。山体滑坡的早期预测框架显然有助于减轻损失。在这项工作中，开发了浅层滑坡发生的预测框架，并通过实际案例研究进行了验证。为了评估预测框架的可靠性，对 2014 年 7 月印度马哈拉施特拉邦马林村发生的山体滑坡进行了回溯分析。使用 GeoStudio 软件通过基于物理的方法建立了滑坡稳定性与土壤饱和度的关系。为研究地点开发了漏桶算法，以监测降雨对土壤饱和度的影响。将滑坡稳定性模拟结果与基于漏桶的降雨-土壤饱和度算法进行了比较。研究发现，所提出的框架对浅层滑坡发生具有良好的预测性，因此建议用于滑坡易发位置的实时监测。