Background

Hydrological drought hazard results from extreme low flows in rivers, reducing water supplies and thus the capacity for abstraction causing water shortages. Currently, the UK’s vulnerability to drought hazard has reached the warning threshold (20%) on the Water Exploitation Index (water abstraction as a percentage of the freshwater resource); it is thus defined by the European Environment Agency as a water-stressed country (EEA, 2019). Recent research (Collet et al., 2018, Visser-Quinn et al., 2019; (Kay et al., 2018, Rudd et al., 2019),suggests that climate change may represent an additional stressor, with hydroclimatological projections indicating an increase in the frequency and intensity of hydrological droughts in the coming decades.

It is clear that drought hazard represents a major threat to water security globally, and the UK is no different. However, these hazards are subject to spatial variation, which may change in the future. With regional variations in these hazards, as well as regional population patterns, it is crucial to study such phenomena at a large geographical scale (e.g. country level).  In addition to spatial considerations, temporal analyses are needed, using flow or precipitation time series. Projections of future river flows (runoff) are the product of a long and complex modelling chain: emissions scenarios force General Circulation Models (GCMs), the outputs of which are downscaled to force hydrological models. Uncertainties in terms of model input, structure, and parameters, cascade through the modelling chain (Clark et al., 2016). Looking forward to future climate projections, any analyses require to consider ensemble projections, thus posing questions around probabilistic analyses.

Aims and Objectives

Droughts can be explored in many ways, one of which is categorising by severity and exploring the transition between severity states. Markov chains are a powerful tool which allow us to explore patterns as an alternative to traditional hydrological models by utilising flows projections either directly from regional climate models (Aitken et al 2022) or through modelling chains such as EFLaG (Aitken et al, 2023). The aim of this PhD is to explore the regional variation in drought severity transitions regionally across the UK. The research will include developing novel ways to establish the probability of different events occurring, as well as grappling with large datasets to detect patterns of transition or occurrence.

Methods

Some initial work by the research group has established the validity of using different ensemble climate projections for analysis such as this. Similarly some tentative exploration has indicated that using Markov Chains can support pattern exploration in drought occurrences in natural catchments. Consequently this PhD will explore hydrological climate projections at the spatial coverage of Great Britain. At this scale the PhD will analyse:

• Flow derived within the regional climate models (RCMs) - the hydrological component is online coupled, allowing greater feedbacks to be captured. Exploring a wide range of GCMs/ESMs captured - across multiple modelling centres. This will consider a wide representation of processes,  and represent a more holistic capture of model structural uncertainties. 
• Use Markov Chains to categorise flow transitions and explore drought progression patterns
• Develop and explore methods to examine and detect patterns across spatial and temporal data