降雨滑坡多指标监测预警方法研究

降雨滑坡是发生最频繁,损失最严重的地质灾害,但如何实现其准确预警仍是目前所面临的主要挑战。而探究降雨滑坡的物理机制是开展预警的关键突破口。结合典型震后滑坡5个水文年的长时间序列实时监测数据和非饱和水-力耦合分析方法,开展都江堰银洞子沟滑坡现场实测数据的物理机制研究及长时间序列边坡稳定性分析。基于非饱和边坡稳定性分析方法,结合长时间序列实测大数据的实时计算和分析,提出基于降雨强度-概率(I-P)、饱和度、基质吸力、地表倾斜角度和稳定性实时计算分析的多参数指标预警方法体系,基于多参数预警体系和实时监测数据,分别于2017年8月和2018年6月在银洞子沟两次实现成功临灾预警,成功临灾预警实例验证了多参数阈值预警模型的可靠性和实用性,研究旨在为降雨型滑坡的准确预警提供新的参考和预警模式。

Research of monitoring and early warning methods for rainfall-induced landslides based on multivariate thresholds

Rainfall-induced landslides are the most frequent and catastrophic geological disasters,while the accurately early warning of precipitate landslides in response to intense rainfall is still challenging.Understanding the rainfall-induced mechanism of landslides is the prerequisite for the early warning.In the study,a typical post-earthquake landslide in Dujiangyan County was selected as a real-time monitoring site,and the stability and failure mechanism of the slope were investigated in a long period by combining the real-time monitoring data of five hydrological years and hydro-mechanical coupling theory.A landslide warning model with multi-parameter thresholds including rainfall intensity-probability threshold(I-P),saturation,matric suction,inclination and slope stability,was proposed based on the unsaturated slope stability calculation and the long-term monitoring data analysis.The proposed method was successfully implemented in early warning at Yindongzi gulley in Aug.2017 and Jun.2018 and proved reliable and practicable.The research provides a reference for early warning of rainfall-induced landslides.