Citizen science for landslide risk reduction and disaster resilience building in mountain regions

山区是自然灾害的多发地区，特别是与山体滑坡有关的灾害。与此同时，由于自然环境的复杂性和数据收集面临的困难，对造成这些灾害的自然过程的科学认识落后。这些灾害对社会的影响非常大，特别是因为山区往往拥有欠发达的基础设施和弱势人口。因此，迫切需要提高我们对山区自然灾害如何发生、如何减轻灾害以及如何使处于危险中的人们更具抵御能力的认识。该提案将利用最近在环境数据收集、处理和通信方面的技术和概念突破，在数据匮乏和贫困的南亚山区社区实现跨越式恢复能力建设。我们特别指出了三种具有巨大前景的趋同进化。首先，传感器网络和数据管理的技术发展使参与性和基层数据收集和公民科学成为可能。其次，基于网络和云的信息通信技术使建立更强大的分析和预测系统成为可能，吸收异构数据源并跟踪不确定性。最后，这为知识共同创造和决策提供了更有针对性的信息流。这些演变是一种趋势的一部分，这种趋势趋向于采用更自下而上和参与性的方法来产生支持环境过程决策的科学证据，这通常被称为“公民科学”。我们认为，公民科学方法在偏远山区环境中特别有前景，因为在这些地区提高抗灾能力和人道主义反应本质上是多中心活动：广泛的行为体参与到相关信息和科学证据的生成、决策和政策制定以及灾害发生期间和前后的行动实施中。因此，最重要的是加强这些活动中心之间的信息流动，最好地利用现有知识，查明主要的知识差距，并分配资源以消除这些差距。我们将以尼泊尔西部的卡纳利盆地作为试点研究。卡纳利盆地是一个偏远且研究不足的盆地，遭受自然灾害的复杂相互作用，包括水文引起的山体滑坡和洪水等级联灾害。在过去几年中，这些灾害对当地基础设施（如道路、灌溉渠、房屋、桥梁）造成了严重破坏，并影响了生计（例如，在2014年8月的洪水中，34760个家庭受到影响）。利用高性价比的传感器技术，实施基层降水、河流流量、土壤湿度和地貌监测。我们将利用这些数据分析极端气象事件及其对滑坡风险时空格局的影响。通过将这些数据与卫星图像等其他数据源合并，我们的目标是以前所未有的分辨率生成滑坡风险地图。与此同时，我们的参与式公民科学方法将使我们能够设计和实施一个基于知识共同产生和共享的自下而上和多中心社区抗灾能力框架。最后，我们将在合作伙伴尼泊尔实际行动（Practical Action Nepal）现有的以社区为基础的洪水预警系统的基础上，创建一个全面的多灾种预警系统和知识交流平台。为此，我们将利用基于开放标准、分散的数据处理和知识传播方面的最新发展，例如移动电话和网络接口。

Mountains are hotspot of natural disasters, in particular those related to landslides. At the same time, scientific understanding about the natural processes that cause these disasters is lagging behind, because of the complexity of the physical environment and the difficulties facing data collection. The impact of these disasters on society is very high, especially because mountain regions often host less developed infrastructure and vulnerable populations. As a result, there is an urgent need to improve our understanding about how natural disasters in mountain regions occur, how they can be mitigated, and how people at risk can be made more resilient.This proposal will leverage recent technological and conceptual breakthroughs in environmental data collection, processing and communication to leapfrog resilience building in data-scarce and poor mountain communities in South Asia. In particular, we identify three convergent evolutions that hold great promise. First, technological developments in sensor networks and data management allow for participatory and grass-roots data collection and citizen science. Second, web- and cloud based ICT makes it possible to build more powerful analysis and prediction systems, assimilating heterogeneous data sources and tracking uncertainties. Lastly, this enables a more tailored and targeted flow of information for knowledge co-creation and decision-making. These evolutions are part of a trend towards more bottom-up and participatory approaches to the generation of scientific evidence that supports decision making on environmental processes, which is often referred to as "citizen science". We believe that a citizen science approach is particularly promising in remote mountain environments, because improving resilience and humanitarian response in these regions are inherently polycentric activities: a wide range of actors is involved in generating relevant information and scientific evidence, in decision-making and policy building, and in implementing actions both during a hazard and before and after. It is therefore paramount to strengthen the flow of information between these centres of activity, to make best use of existing knowledge, to identify the major knowledge gaps, and to allocate resources to eliminate these gaps.We will use the Karnali basin in Western Nepal as a pilot study. The Karnali basin is a remote and understudied basin that suffers from a complex interplay of natural hazards, including hydrologically-induced landslides and cascading hazards such as flooding. Over the last years, these hazards have caused serious damage to local infrastructure (e.g., roads, irrigation canals, houses, bridges) and affected livelihoods (e.g., 34760 families in the August 2014 floods).Using cost-effective sensor technologies, we will implement grass-roots monitoring of precipitation, river flow, soil moisture, and geomorphology. We will use those data to analyse meteorological extremes, and their impact on spatiotemporal patterns of landslide risk. By merging these data will other data sources such as satellite imagery, we aim to generate landslide risk maps at unprecedented resolution.At the same time, our participatory citizen science approach will enable us to design and implement a framework for bottom-up and polycentric community disaster resilience, based upon knowledge co-generation and sharing.Lastly, we will build upon the existing community-based flood early warning system implemented by our partner Practical Action Nepal, to create a comprehensive multi-hazard early warning system and knowledge exchange platform. For this, we will leverage recent developments in open-standards based, decentralized data processing and knowledge dissemination, such as mobile phones and web-interfaces.