Predicting Climate Change impacts on Shallow Landslide Risk at regional scales

1336911 (Istanbulluoglu)。山体滑坡破坏水生栖息地，破坏基础设施（如道路、公用事业、水坝）。随着气候变化，西部滑坡灾害预计会增加，但迄今为止，地质滑坡研究通常独立于水文气候研究之外进行。需要将这两方面的研究统一起来，为资源管理和气候适应战略提供区域尺度的滑坡预测。华盛顿喀斯喀特山脉经历了各种气候、植被和地形的滑坡，因此，这里所做的工作与全球各地的山区有关。与州和联邦机构合作伙伴合作，该项目团队将开发一个区域尺度的分布式数值模型，结合50多年来在华盛顿瀑布的滑坡观测，回答以下研究问题：1)位置（地质、地形、坡度）与气候（降水、温度、融雪和补给率）对滑坡频率的相对作用是什么？2)结合地质技术、地质学和水文气候预测方法的一种新的变革性模型对过去滑坡的时空模式和频率的再现效果如何？具体来说，a)概率方法更可靠还是确定性方法更可靠？b)在哪种情况下，需要更精细、更具体的地点数据（例如，山坡上的积雪提供额外的载荷和滑坡的关键触发因素，或改变土壤稳定性的冻融循环）才能获得准确的模型性能。3)气候变化将如何影响滑坡的位置和频率？具体来说，a)最大的预期危害是什么（例如，道路冲刷、泥沙通量、涵洞破坏）？b)在这些预测下，需要什么样的决策支持工具来对复杂地形上的景观和溪流进行可持续管理？近几十年来，气候变化对水资源和河流温度的影响已经通过数值模型进行了广泛的研究，但对滑坡输沙量的研究有限。这些研究要么使用经验降雨阈值和区域尺度的地质易感性图来识别滑坡，要么关注几米范围内的详细水文。第一种方法排除了基本的物理，而第二种方法不能在大面积上使用。该项目将整合这两种思路，并有可能改变区域滑坡研究的方式。该模型将结合5500米分辨率陆地表面模型（VIC）和10米分辨率概率斜坡稳定性模型（SINMAP）的地下水流补给和地表径流，开创了一种创新的数值模型设计。这项工作的目标是直接影响资源管理，并将纳入K-12、本科和研究生教育。该研究计划是与州和联邦机构密切合作制定的，这些机构将立即将研究结果用于土地管理。在项目的第三年，我们将举办一个研讨会，对资源管理人员进行模型培训，并与他们分享我们最终的建模框架。与教育专业人士一起，将为高中学生开发一个教育模块。该模块将通过华盛顿大学的“课堂上的大学项目”在西雅图的选定学校进行展示，在该项目中，学生将获得高水平工作的大学学分。该项目的研究结果将有助于华盛顿大学本科和研究生水平的教学。

1336911 (Istanbulluoglu). Landslides disrupt aquatic habitats and damage infrastructure (e.g., roads, utilities, dams). Landslide hazards in the west are expected to grow with climate change, but to date, geologic landslide research has been typically conducted independently from hydroclimate research. There is need for unifying these two lines of research to provide regional scale landslide prediction for resource management and climate adaptation strategies. Washington Cascade Mountains experience landslides across a wide a range of climates, vegetation, and topography, and thus, work done here is relevant to mountain areas across the globe. Working with state and federal agency partners, this project team will develop a regional-scale distributed numerical model in conjunction with 50+ years of landslide observations across the Washington Cascades to answer the following research questions: 1) What are the relative roles of location (geology, topography, slope) vs. climate (precipitation, temperature, snowmelt, and recharge rates) on landslide frequencies? 2) How well does a new, transformative model combining the methods of geotechnology, geology, and hydroclimate prediction reproduce past spatial-temporal patterns and frequencies of landslides? Specifically, a) Are probabilistic or deterministic methods more reliable? and b) In which cases are finer, more site-specific data needed (e.g., snow pack on mountainsides providing extra load and a key trigger for landslides, or freeze-thaw cycles changing soil stability) for accurate model performance. 3) How will climate change likely impact landslide locations and frequencies? Specifically, a) What are the largest expected hazards (e.g., road wash-out, sediment flux, culvert failure)? and b) Under these predictions, what decision support tools are needed for sustainable management of landscapes and streams over complex terrain? While the effects of climate change on water resources and stream temperatures have been extensively studied using numerical models in recent decades, only limited studies focused on landslide sediment delivery. These studies either use empirical rainfall thresholds and geologic susceptibility maps at regional scales to identify landslides, or focus on detailed hydrology over several meters. The first method excludes essential physics, while the second cannot be used over large areas. This project will integrate these two lines of thought and potentially transform how regional landslide research is done. The proposed model will pioneer an innovative numerical model design by combining subsurface flow recharge and surface runoff from an 5,500-m resolution land surface model (VIC) with a 10-m resolution probabilistic slope stability model (SINMAP). The work is targetedto directly impact resource management and will be incorporated into K-12, undergraduate, and graduate education. The research plan has been developed in close collaboration with state and federal agencies, who will immediately put results to use in land management. A workshop will be conducted in year 3 of the project to give model training to resource managers and share our final modeling framework with them. In conjunction with education professionals, an education module will be developed for high school students. This module will be presented at select schools in Seattle through the University of Washington's "University in the Classroom project," wherein students receive university credit for high-level work. The findings of the project will contribute to undergraduate and graduate level teaching at the University of Washington.