Anticipating Glacier Lake Outburst Floods (GLOFs): an impact-based forecasting framework for managing GLOF risks in Nepal.

英语. Glacier lake outburst floods (GLOFs) are an increasingly documented threat across the Himalayan region, wherein Nepal is situated. GLOFs involve a rapid discharge of water from a lake situated at the side, front, within, beneath, or on the surface of a glacier. Forecasting the impacts associated with such events could facilitate the development of proactive risk management approaches. Impact-based forecasting (IbF) and anticipatory action (AA) are proactive strategies which are of emerging interest among humanitarian actors and disaster risk management authorities both in Nepal and globally. To assess what a proactive approach could look like for addressing GLOF risks in Nepal, this thesis takes inspiration from existing GLOF risk models and applications of anticipatory risk management approaches for other hazard types. The research questions addressed are as follows: 1. How are glacier lake outburst flood risks to Nepalese communities presented in academic literature? 2. How could impact-based forecasting address glacier lake outburst flood risks in Nepal? A systematic literature review of potentially dangerous glacier lake (PDGL) risk models enabled the assessment of existing capacities to anticipate GLOF threats. The research also collates insights from experiences in applications of IbF through six semi-structured interviews and a non-systematic content analysis of IbF and AA documents. Both research components ultimately inform a framework showing how GLOF risks could be addressed through IbF. Part 1 of the thesis looks at the state of risk data for GLOFs in Nepal, presenting an overview of PDGLs and associated outburst models. The findings show that of the 47 classified PDGLs posing a risk to Nepal, only four are assessed to have outburst models quantifying projected impacts using physically-based hydrodynamic models. Although, there is high uncertainty and discrepancies in projected inundation zones and lead times across studies for all assessed lakes. Most studies additionally do not assess household-level vulnerabilities, serving as a key knowledge gap in capacities to anticipate impacts in a way that is meaningful for holistically risk-informed activities. Part 2 discusses the results of the study around the proposed framework, showing how GLOF threats in Nepal could be addressed through IbF. First, capacities to predict GLOF events are discussed under two categories, the longer-term readiness triggers, and shorter-term activation triggers. The GLOF risk assessments studied model an array of trigger and outburst mechanisms that could theoretically be observed at each PDGL to assess the likelihood of imminent outburst and aligned with anticipated outburst scenarios. Some examples of assessed triggers include mass movements, high temperatures or precipitation, and earthquakes. A discussion of the lead times available for implementing risk-mitigating early actions upon recognition of these triggers follows this. Next the two overarching phases of the framework, the readiness phase, and the activation phase, are expanded upon, discussing the proposed activities within each. The readiness phase involves the long-term anticipation of GLOF-related impacts to enable the implementation of risk-informed preparedness activities. The overarching stages of this phase include identifying PDGLs, developing impact-based forecasts, and developing early action protocols. A related theme drawn from the interview respondents regards the importance of IbF-informed preparedness activities such as developing community capacities to operationalise forecasts and implement early actions. The activation phase is the period for refining IbF scenarios and implementing predefined early actions. An issue emerging from the interviews was that of working with dynamic risk data in anticipatory action. It is suggested this could be mitigated in the GLOF context by enabling scenarios and associated protocols to be refined according to emerging risk information. This thesis ultimately provides an assessment of the risk data associated with GLOFs in Nepal and proposes an IbF framework to address these risks, within the realms of assessed forecasting capacities. By leveraging IbF and AA approaches, supported with investments in hazard modelling, the proposed framework could enable timely risk-reducing actions in the face of anticipated GLOF threats.

英语. Impact-based forecasting is like envisioning future disaster scenarios via a crystal ball; except, instead of a crystal ball, we have hazard models and predictions of risk. The potential for proactive risk management approaches to anticipate the impacts associated with future disaster events and guide risk reducing efforts is increasingly being recognised. This motivates an investigation into its possible applications to a hitherto unconsidered hazard type within the field of anticipatory action – glacier lake outburst floods. Glacier lake outburst floods involve a rapid discharge of water from a lake situated at the side, front, within, beneath, or on the surface of a glacier. These hazard events can have devastating consequences for downstream communities, claiming lives, destroying houses and infrastructure, and occasionally driving secondary hazards in the form of landslides or debris flows. This thesis seeks to provide insight into the complex arena of anticipating such events. The study assesses what a proactive approach to addressing the risks associated with flood events from glacier lakes could look like in Nepal. It collates existing risk models for this hazard with insights from discussions held with actors applying anticipatory risk management approaches to other hazard types. The result is an impact-based forecasting framework for glacier lake outburst floods. This approach draws on the idea that there are learnings to be made across hazard types within the field of disaster risk management. Good quality impact data is seemingly imperative to successful applications of anticipatory risk management approaches. Knowing what could happen in a future disaster event can help improve capacities to reduce loss of life, property, and other assets. For impacts to be accurately predicted and losses subsequently reduced, risk data needs to be effectively and continually gathered and evaluated. The findings here suggest the present inconsistencies and uncertainties in assessed GLOF models (relating to forecast lead times and the classification of exposed areas) may ultimately compromise impact-based forecasting capacities for this hazard type. GLOF risk models are shown to be limited in Nepal, with conflicting assessments of the risks posed by ‘potentially-dangerous’ lakes having noted implications for applications of anticipatory risk management approaches. Despite these observed constraints, this study proports that shifting decision making processes in advance of outburst - by assessing the range of potential early actions, identifying household-level vulnerabilities, or assigning responsibilities based off predicted scenarios – could help practitioners and populations increase the time available to reduce exposure to risk. The framework presented in this study illustrates how long- and short-term forecasts could be utilised in a proactive manner to reduce the risk among exposed populations. The proposed framework shows how forecasting disaster events and associated impacts could facilitate the systematic and informed development and implementation of plans and activities to prepare communities exposed to GLOF risks. This framework could be an important step towards assessing how anticipatory risk management approaches could be applied to more sudden-onset hazard types, since presently considered hazards often work with longer lead times for action. It is my hope this preliminary analysis of the subject area can be utilised by humanitarian actors and disaster risk management practitioners to drive an increase in the production and proactive utilisation of risk assessments and outburst models for identified potentially dangerous glacier lakes in Nepal, and beyond.