Studentships

Five UK universities are hosting 13 PhD studentships as part of the SHEAR Studentship Cohort (SSC)

The SHEAR Studentship Cohort (SSC) is a cohort of 13 PhD studentships hosted at 5 different UK Universities (Imperial College London, King’s College London, University of Birmingham, University of Sussex and University of Reading) and aims to complement and integrate further the current SHEAR-funded projects, with a particular focus on providing the students with excellent opportunities to gain transferable skills, to conduct original and scientifically excellent research, and to contribute to impact creation beyond the existing SHEAR projects. The SSC is hosted by Imperial College London and the Programme Director is Dr Wouter Buytaert. If you have any questions about the SSC please contact the Project Manager.

The studentships have now been allocated with projects commencing during the 2017/2018 academic year; indicative project descriptions are included below.

Interactive visualization and communication of hydrological information to support multi-level natural hazard reduction and resilience building

Student: Anna Twomlow, Imperial College London

Supervisor: Dr Wouter Buytaert

Leveraging interactive web technologies such as infographics and social networks to enhance the flow of information between different actors in a complex development context

The ubiquitous presence of internet-connected computers and mobile phones has revolutionized the way that people communicate and share information. This has resulted in an exponential rise in the use of social computer networks, and the availability of unprecedented amounts of information.

This PhD project aims to leverage this evolution to decrease human vulnerability and promote resilience building in the context of natural hazards such as floods and droughts.

The societal impact of natural hazards is increasing in many parts of the world, particularly in developing countries. However, the scarcity and low quality of existing environmental information (such as rainfall and river flow records), as well as suboptimal processing and communication of that information to the people who need it, is a major challenge.

The use of citizen and science and other participatory approaches present an opportunity to alleviate this data scarcity. Examples of such initiatives include community-based flood early warning systems, low-cost weather stations, and participatory monitoring of ecosystem services.

This project will study how new types of collaborative decision-support systems can be developed to support grassroots initiatives for the collection and communication of environmental data that can increase local resilience to natural hazards. The project will leverage new technological developments in ICT such as cloud computing and social networks, and will build upon the most recent insights in the creation of dynamic visualization and infographics. In particular, the project will study how the visualization of information (including inherent properties such as uncertainties) needs to be tailored and adapted to be accessible for a wide range of actors, and in particular disadvantaged groups and people with low technical and scientific background.

The project will use two or more case studies in remote mountainous environments: one case study in Western Nepal (Karnali River) and a case study in Tanzania (Kilombero River).

Low-cost sensor networks for flood and drought monitoring and forecasting

Student: Neeraj Sah, Imperial College London

Supervisor: Dr Wouter Buytaert

Developing the next generation of real-time sensor networks to support participatory monitoring of water resources and water-related risks in a development context

Technological developments, such as wireless communication and the miniaturization of electronics for computing and data storage, provide great opportunities for environmental sensing. This is of global importance: large parts of the world and developing regions in particular are still very data scarce in terms of environmental variables such as precipitation and streamflow.

At the same time, stressors such as environmental degradation, climate change, and population growth put increasing pressures on water resources and aggravate water-related risks.

This PhD project will study the potential of low-cost, real-time pervasive sensor networks to monitor drought and flood risk and to support mitigation and resilience-building actions.

A first part project will focus on the development and installation of prototypes of real-time low-cost water level sensors to monitor storm surges in rivers, as well as water availability in groundwater wells and reservoirs. For this, the project can draw upon previous expertise at Imperial College London in the use of a variety of technologies for water level sensing, with current applications in Peru, Nepal, Somalia, and Tanzania.

Subsequently, new hydraulic and hydrological models will be developed to process and analyse the data, and integrate them in end-user oriented solutions such as flood early warning systems, drought monitors, and derived products such as forecast-based financing.

Lastly, working closely with end-users, specific solutions and roadmaps for their integration in operational systems will be developed in collaboration with commercial partners such as local small and medium-sized enterprises (SMEs).

Actionable knowledge for disaster risk reduction: understanding social vulnerability to enhance community resilience to natural hazards

Student: Caroline Russell, University of Birmingham

Supervisor: Dr Julian Clark

Developing new forms of collaborative governance for disaster risk reduction in low and middle income countries (India, Nepal, Mozambique)

Effective disaster risk reduction depends on improving coordination between state agencies and multiple civil society actors, and enhancing community resilience to hazards by addressing specific local knowledge gaps.

However, little systematic data exists on these issues. Such coordination challenges and knowledge gaps are particularly acute in low and middle income countries in East Africa (e.g. Mozambique) and South East Asia (e.g. Nepal and India), which face manifest governance challenges across geographic and political-administrative scales.

Mitigation of disaster risk here requires

  1. improved coordination between government, NGOs and affected (often remote rural) communities
  2. developing a greater understanding of the under-explored social vulnerabilities to hazards in affected communities.

Both challenges offer invaluable opportunities to generate actionable knowledge to enhance community resilience to disaster risks: that is, knowledge that directly improves community resilience capabilities and helps shape the development of grassroots strategies for practical preparedness and response.

To address the first of these issues, this PhD project will focus on collaborative governance of drought, flooding or landslide risk to address four key research questions:

  1. What regional and national structures and procedures currently exist to provide advice and information for communities at risk of drought/flooding/landslides?
  2. At community scale, what institutional structures, norms and beliefs condition local disaster risk mitigation/adaptation strategies, and how do these interact with formal advice provision from state agencies and civil society groups?
  3. What new forms of collaborative engagement (including ‘bridging mechanisms’, cross-scale policy initiatives and citizen science approaches) might be developed between national and local scales to reduce disaster risk and increase community resilience?
  4. to explore the possibilities for sharing of best practice on collaborative governance of disaster risk management between the focal countries among the SHEAR project consortium.

Understanding water-related multi-hazards in a sustainable development context

Student: Julia Docherty, University of Birmingham

Supervisor: Prof. David Hannah

Assessing hydrometeorological drivers, river basin controls and local impacts of hydrologically-induced landslides and floods in a multi-hazard framework

Hydrologically-induced landslides and floods are among the most destructive natural hazards globally. Water-related research in many developing countries is often hindered by a variety of issues, such as topographic complexity (especially in mountain regions), intensive and poorly understood climate patterns (e.g., the South Asian monsoon). This makes data collection problematic and limits transferability of findings from elsewhere to data-scarce regions. Consequently, there is a paucity of critical knowledge about landslides and floods space-time distribution, controls and local impacts.

Often research on natural hazards adopts a single hazard approach. But in many regions, and mountains in particular, there is a direct connection between landslides and other hazards — notably flooding. Thus, hazards overlap and interactions (or cascades) between hazards take place. By omitting these inter-relationships, there is potential to underestimate risk or increase vulnerability.

Water-related hazards are associated typically with intense and/ or prolonged rainfall events that alter material strength to cause landslides, or generate rapid runoff resulting in floods. Although meteorology is known to be a key driver of these hazards, the response to rainfall inputs is modified spatially and temporally by river basin properties (including relief, relative position within the basin, geology, soils, land use change and infrastructure development).

This PhD will assess hydrometeorological drivers, river basin controls and local impacts of hydrologically-induced landslides and floods in a multi-hazard framework. It will use as the main case study the Middle Mountains of Nepal, which are a global hotspot of water-related hazards, but will aim to include other cases for comparative analysis and testing of transferability of knowledge and methods.

Prediction and early warning of cascading, multi-hazards is vital to increase human preparedness and, thus, to increase potential resilience of people and infrastructure. This studentship will yield information of direct practical relevance with potential to support decision making/ early warning systems for water-related natural disasters risk reduction.

Cascading vulnerability in the Eastern Himalayas: disasters, politics and poverty

Student: Peter McGowran, King’s College London

Supervisors: Dr Amy Donovan; Dr George Adamson

Developing understanding of vulnerability transfer related to multiple hazards and social, economic and political factors

Vulnerability to environmental disasters may be increased or decreased by development and economic factors, often in complex ways.

This PhD project addresses the vulnerability aspects of multiple hazard events with different impacts across space and time.

Nepal and India are both prone to flooding, landslides and earthquakes, which often occur within a short period of time. This leads to compounded vulnerability that can be exacerbated by ongoing hazards, but also by complex sociopolitical and cultural reactions to such hazardous events.

For example, an earthquake triggers the loss of infrastructure, residences and social networks, which leads to a loss in social capital and an increase in temporary accommodation. It also alters supply chains and disrupts local businesses. A subsequent intense monsoon or even an unseasonal extreme event like cloudbursts with accompanying landslides is then more destructive because of the increase in vulnerability and the lack of capacity in local businesses and infrastructure to respond.

The PhD student will spend time in north east India at the LANDSLIP case study locations in Darjeeling and/or East Sikkim conducting detailed interviews with key informants, focus group discussions and survey work to understand how the population have coped with successive disasters (including civil unrest), and how they feel that this has affected their vulnerability to future disasters. This would focus on political issues and challenges around poverty alleviation and empowerment.

The research will be supplemented by work in Nepal in collaboration with Practical Action Nepal, who have been working on flooding as well as landslides. The student will also use documentary records (including media reports) and census data to triangulate with interviews (elites and residents) and household surveys. They will be encouraged to map the evolution of vulnerability in small areas through a sequence of disasters over time, thus examining the extent of cascading vulnerability and its nature and relationship (if any) to cascading hazards.

Understanding transboundary impacts of multi-hazard early-warning systems and their cultural context

Student: Gaurab Dawadi, King’s College London

Supervisors: Dr Amy Donovan; Dr George Adamson

Identifying cultural influences on interpretation and response to early warnings

Early warnings are understood in different ways by different people, and this can affect whether or not they are acted upon — and how quickly (Eiser et al., 2015).

This project will use examples from India and Nepal to understand the dynamics of multi-hazard early warning across borders (transboundary).

The PhD project will draw on the work done in the SHEAR projects, and also assess early warning systems (EWS) for meteorological hazards, and public perceptions of their usefulness.

In particular, it will examine differences between Nepal and India, and between two Indian states, both in terms of cultural use of EWS and in terms of their effectiveness. The student will use surveys, interviews and focus groups to gain a deep understanding of the cultural dynamics that affect EWS.

Elite interviews will be conducted to gain understanding of some of the geopolitical issues around early warning, including how India (government, institutions, communities) perceive the systems in Nepal and vice versa. One of the challenges in border regions is to produce consistent messages — if one side evacuates and the other does not, then trust issues may arise.

This project will therefore look both at the state level and at the national level to assess transboundary EWS challenges. Darjeeling (West Bengal State) in India has a substantial Nepali-speaking population and a developed separatist movement that campaigns for a separate Indian state, while Sikkim State is located on a disputed border with China, and also borders both Nepal and (in a small area) Bhutan. It has a very different cultural and political demographic to Darjeeling. These two areas thus share a national government but widely different cultural and geopolitical contexts.

Nepal, meanwhile, has a different government with different levels of investment in early warning and different approaches to India. While a short distance apart, the three areas provide diverse comparative sites for a study of early warning in complex cultural contexts.

Examples of cultural issues with warning might include reactions to colour codes, ways of interpreting maps or numerical information, linguistic nuances, levels of trust in warning agencies and scientific institutions.

Understanding these issues is important for early warning studies more broadly, even if the results of this study are locally specific: there are critical challenges in the uptake of early warning, and ensuring that knowledge is catalysed into action. Some of these relate to understanding, but others relate to more complex ideas around culture, beliefs, trust, and ideology.

Bayesian networks for multi-hazard risk assessment in the Himalaya as a way of improving early warnings

Student: Shreyasi Choudhury, King's College London

Supervisors: Prof. Bruce Malamud; Dr Amy Donovan

Investigating the use of Bayesian Belief Networks for multi-hazard (vs. single hazard) risk reduction in the Himalaya

A multi-hazard cascade occurs when a single natural hazard triggers another hazard, for example a large rain storm triggering a landslide which blocks a river causing a flood. A multi-hazard coincident example would be an earthquake and heavy rain occurring near each other in time and space, and the two together resulting in landslides.

Multi-hazard cascades and coincident hazards are much more difficult to calculate probabilities for, making early warning more difficult. Bayesian Belief Networks (BBNs) have been developed in several hazard domains, including volcanology (e.g., see Donovan et al., 2012) and tsunami science, as a means of aggregating complex information to produce a probability of a hazard.

While they are subjective in that they require considerable expert judgement to set up and design, they have the benefit of being able to combine multiple varieties of information, including qualitative or binary information.

In the absence of straightforward equations for the probability of hazard events — especially in a multi-hazard context — BBNs present one way to try to produce useful forecasts.

This PhD project has two strands:

  1. The design and implementation of a simple BBN for a multi-hazard problem in the Himalaya, for example, earthquake → landslide → flood multi-hazard cascades. The student would investigate the necessary information and the likely potential sources of information (including the sources of information that the LANDSLIP project is developing), and then set up a BBN to manage this information and produce a risk assessment. This would involve expert elicitation to design the details of the BBN and to estimate uncertainties.
  2. At the same time, the student will conduct a meta-study involving interviews of expert participants and project members on their feelings and concerns about the use of subjective probabilistic methods in risk assessment.

Together, these two strands would produce a robust and innovative analysis of a potential method for multi-hazard risk assessment.

A model for impact-based flood early warning in Uganda

Student: University of Reading

Supervisors: Prof. Rosalind Cornforth; Dr Elisabeth Stephens

Linking data on hazard, vulnerability and exposure to develop impact-based early warning in Uganda

The initial Forecast-based Financing pilot project in North Eastern Uganda required considerable human resource to establish danger thresholds for which flood forecasting systems would be required to forecast (Coughlan de Perez et al. 2016). A new approach is therefore needed to establish country-wide danger thresholds for scaling out Forecast-based Financing to all areas of Uganda which have sufficient forecast skill (as being evaluated by the FATHUM postdoc). This research will involve the development of measures of vulnerability and resilience in Uganda, for example, based on an exploration of crop yield and livelihoods data.

This PhD project will work towards a model for national-scale impact-based forecasting of flood risk by answering the following questions:

  1. What are the key indicators of flood vulnerability across different parts of Uganda?
    • Working alongside the Uganda Red Cross Society (URCS) and the FATHUM project team to determine the criteria for flood vulnerability which can be addressed using anticipatory measures, to include the Vulnerability and Capacity Assessment (VCA) carried out by URCS and analysis of historical disaster impacts (e.g. through Red Cross Disaster Relief Emergency Fund (DREF) reports).
  2. How can a GIS tool be developed that integrates all components of risk: hazard, exposure, and vulnerability, which would support Forecast-based Financing action planning?
    • Identify the criteria for spatially explicit flood vulnerability and capacity assessment
    • Identifying key datasets for mapping flood vulnerability across Uganda
    • Building national-scale flood hazard maps from locally and globally available resources
    • Linking flood inundation mapping with population and settlement information, administrative areas and vulnerability classifications to delineate (an appropriate number of) possible action areas
  3. How can this GIS analysis be linked to forecasts in real-time to enable decision-relevant impact-based probabilistic flood forecasting?
    • GloFAS, or local hydrological forecasting systems, will be combined with the GIS tool to create a probabilistic decision support tool for Forecast-based Action by the Uganda Red Cross. This tool will highlight in real time the areas that could see impact, and in which early actions would be worth taking.

The student will undertake a placement period with Makerere University, linking with the Uganda Red Cross and Irene Amuron of the Red Cross Red Crescent Climate Centre.

Project Advisors: Shuaib Lwasa (Makerere U.), Erin Coughlan de Perez (Red Cross Red Crescent Climate Centre), Dai Clegg (Evidence for Development) Dr John Seaman (Evidence for Development) with support from Dr Celia Petty (Evidence for Development).

Extended range hydrometerological forecasting for improved flood early warning in Bangladesh

Student: Sazzad Hossein, University of Reading

Supervisors: Dr Elisabeth Stephens; Prof. Hannah Cloake

Supporting earlier flood preparedness in Bangladesh

Bangladesh is located downstream of three major world rivers, with flooding from these rivers affecting millions of people directly, as well as causing huge damage to agricultural lands. Currently the Flood Forecasting and Warning Centre (FFWC) in Bangladesh provides deterministic flood forecasts out to 5 days, with experimental probabilistic forecasts out to 10 days using forecasts from the European Centre for Medium-Range Weather Forecasts (ECMWF).

This project will tackle the key scientific questions that will enable the FFWC to provide warnings further head of time. Working together with FFWC, the student will carry out research to understand the main drivers of uncertainty involved in probabilistic sub-seasonal to seasonal river flow predictions in Bangladesh for improved early warning and early action to floods.

The PhD student will use reanalysis and reforecast archives from the Global Flood Awareness System held at the ECMWF, and compare with river flow records held by the FFWC and other relevant flood observational datasets. The Global Flood Awareness System is an operational forecasting system funded by the European Commission’s Copernicus Emergency Management Service.

Topics for consideration include, but are not limited to:

  • The limit of predictability of the South Asian monsoon
  • Land surface controls on floods flows including soil moisture, snow melt and evaporation
  • Parameterisation and physics of river routing models, particularly related to backwater effects at major river confluences
  • Improving statistical characteristics of probabilistic forecasts

The student will undertake a placement period within Bangladesh with the Red Cross Red Crescent Climate Centre (RCCC) to inform the implementation of the Forecast-based Financing pilot project in the Bogra district of Bangladesh.

Project Advisors: Christel Prudhomme (ECMWF), Hassan Ahmadul (RCCC).

Flash floods and landslide risk in South East Asia: assessing coincidence, predictability and early warning potential using global scale forecasting systems

Siobhan Dolan, University of Reading

Supervisors: Dr Elisabeth Stephens; Prof. Hannah Cloake

Assessing coincidence, predictability and early warning potential using global scale forecasting systems

Flooding from intense rainfall regularly affects several thousands of people across SE Asia. In August 2017 at least 175 people died, and thousands fled their homes as floods swept across Nepal, India and Bangladesh. These devastating floods are often accompanied by destructive landslides which adds to the loss of lives and livelihoods. The European Centre for Medium-Range Weather Forecasts (ECMWF) is developing a flash flood forecasting system as part of the Global Flood Awareness System, which uses global ensemble numerical weather forecasts to provide early warning of intense rainfall.

This project will benefit from support from both flooding and landslide experts from the SHEAR consortium of projects, meteorological experts from ECMWF and the Met Office as well as working alongside partners in India and Nepal, for example the Medium term forecasting centre in India and Practical Action in Nepal.

The PhD student will:

  • Develop and evaluate long term datasets of flash flood indices such as those based on the Extreme Forecast Index from ensemble reforecasts and new reanalysis products such as ERA5 held at ECMWF.
  • Evaluate the potential skill of the flash flood forecasts using these long term datasets and flow archives held by authorities in India/Nepal and remotely sensed data sources (where available).
  • Identify coincidence and timings of flash floods and landslides, and the factors influencing predictability in selected study areas.
  • Develop post-processing tools to improve communication of landslide risk from flash flooding, which can be adopted by project partners in India/Nepal.

The student will undertake a series of UK and overseas placements in the partner organisations, including a period working with ECMWF on the flash flood forecasting, with the BGS and Met Office on landsliding, and 2 overseas placement periods with partners in India/Nepal including Practical Action and the Medium Term Forecasting Centre in India.

Project Advisors: Helen Reeves (BGS), Calum Baugh (ECMWF), Bruce Malamud (KCL), Jo Robbins / Rutger Dankers (Met Office).

Resilience and pastoralism: satellite-based decision-support systems for pastures

Student: James Mumima, University of Sussex

Supervisors: Pedram Rowhani; Alexander Antonarakis

Exploring Earth observation data to map and predict useful and accessible biomass for pastoralist communities

Livestock accounts for 37.5% of Kenya’s land area, 12% of its GDP and 40% of its agricultural sector, but is susceptible to frequent droughts and degradation due to overgrazing. In this context, this Sussex University-led work will assess the potential of new earth observation datasets (e.g., Sentinel 1 and 2) to deliver near real-time monitoring and prediction of useful and accessible biomass for pastoralism.

The accessibility of pasture areas will be mapped using a participatory approach with local stakeholders to identify issues related to land tenure, conservation requirements, migration patterns, water, and other socio-cultural factors.

Monitoring of useful biomass will rely on mapping major plant functional types (PFTs), and then tracking biomass dynamics, plant health, and phenological cycles of these PFTs. PFTs in pastures include annual and perennial grasses, and deciduous and evergreen shrubs. This classification will be achieved using spectral mixture analysis or machine learning techniques, taking advantage of spectral and temporal differences in Sentinel’s spectra.

The outcome of this research will support pastoralist communities in Kenya to decide the suitability and location of pastureland for their various livestock through:

  1. improved understanding of spatio-temporal distribution of pastures
  2. improved understanding of ecological changes and resilience of pastures
  3. near-future predictions of pasture suitability.

This will enhance their livelihood resilience in the wake of large and extensive droughts, overgrazing, and land cover change.

Seamless seasonal to subseasonal forecasts of flood risk for the River Tana, Kenya

Student: Augustine Kiptum, University of Sussex

Supervisor: Prof. Martin Todd

Assessing the potential for seamless prediction of flood risk at the river basin scale

This project will utilise the global Flood Awareness system (GLOFAS, Alfieri et al., 2013) which links a global hydrological model (Lisflood) to state-of-the-art weather/climate forecasts from the ECMWF ensemble forecast system.

The GLOFAS model system produces forecasts of river discharge at 10 km resolution from which probabilities of exceeding various critical thresholds are derived. Forecast lead times for the standard system extend to out to 45 days (including 15 days driven directly from the weather forecast) and for out to a number of months in new ‘GLOFAS-seasonal’ (Emerton et al., 2017).

The project will aim to evaluate the predictability of flood events at seamless lead times of days to months, and to explore the potential application of this scientific information in potential Forecast based Action (FbA) for flood preparedness initiatives.

The political economy of Forecast-Based Action in the context of the 'new humanitarianism' paradigm: a multi-scalar analysis of anticipatory intervention

Student: Olivia Taylor, University of Sussex

Supervisors: Prof. Dominic Kniveton; Prof. Peter Newell

Understanding the operationalization of Forecast-Based Action in Kenya

Donors, governments and the NGO community often do not do enough to prepare for climate and ‘natural’ hazards. Traditionally, the bulk of funding has been supplied post-disaster (Kellet and Caravani, 2013), as was the case with the response to the Greater Horn of Africa 2011 drought, where despite early warnings the international community failed to respond in advance (Lautze et al, 2012). However, recent years have seen a suite of new approaches aimed at acting in advance of hazards. This approach is typified by integrating early warning systems into humanitarian actions, such as Forecast-Based Financing (FbF) and Forecast-Based Action (FbA); a shift to ‘on demand’ climate information services; and by novel insurance tools such as those promised by the London Centre for Global Disaster Protection.

The following research questions reflect the proposed multi-scalar approach:

  1. At a national level, how does the paradigm shift outlined above align with pre-existing development priorities and policy processes?
  2. At an institutional level, how constrained are institutions by the perceived risk of acting in vain (Coughlan de Perez, 2015), or by the delicate positionality of humanitarian organisations within diplomatic spaces (Marin and Naess, 2017: 26)?
  3. At a local level, initial research suggests that increasing attention to resilience and adaptation among humanitarian actors may not lead to reduced vulnerability because resources tend to be captured through existing power structures (Mosberg et al., 2017). Is this replicated through the implementation of FbA?