Logo: to the web site of Uppsala University

Floods and droughts often worsen socioeconomic inequalities, as marginalized groups face disproportionate impacts and have fewer resources to recover. Inequalities, in turn, limit their ability to recover, and amplify the impacts of future hydrological extremes, creating vicious cycles of worsening disasters. Yet such floods and drought events can also serve as opportunities for transformative change, addressing underlying vulnerabilities and reducing inequalities. Progress in sociohydrology is essential for better understanding these two-way interactions, requiring case studies and models of human–water systems that account for uneven risk distribution, along with global comparative analyses using emerging datasets. We propose an interdisciplinary research agenda that combines sociohydrological modelling, critical social science perspectives on power and inequality, and the innovative, but carefully scrutinized, use of artificial intelligence (AI) tools, such as text mining. This integrated approach can reveal context-specific dynamics, and help break vicious cycles of disasters and inequalities.

Global data have served an integral role in characterizing large-scale groundwater systems, identifying their sustainability challenges, and informing on socioeconomic and ecological dimensions of groundwater. These insights have revealed groundwater as a dynamic component of the water cycle and social–ecological systems, leading to an expansion in groundwater science that increasingly focuses on groundwater’s interactions with ecological, socioeconomic, and Earth systems. This shift presents many opportunities that are conditional on broader, more interdisciplinary system conceptualizations, models, and methods that require the integration of a greater diversity of data in contrast to conventional hydrogeological investigations. Here, we catalogue 144 global open access datasets and dataset collections relevant to groundwater science that span elements of the hydrosphere, biosphere, atmosphere, lithosphere, food systems, governance, management, and other socioeconomic system dimensions. The assembled catalogue offers a reference of available data for use in interdisciplinary assessments, and we summarize these data across their primary system, spatial resolution, temporal range, data type, generation method, level of groundwater representation, and institutional location of lead authorship. The catalogue includes 15 groundwater datasets, 23 datasets derived in relation to groundwater, and 106 datasets associated with groundwater. We find the majority of datasets are temporally static and that temporally dynamic data peak in availability during the 2000–2010 decade. Only a small fraction of temporally dynamic data is derived with any direct representation of groundwater, highlighting the need for greater incorporation of groundwater in Earth system models and data collection initiatives across socioeconomic, governance, and environmental science research communities. A small number of countries, led by the USA, Germany, the Netherlands, and Canada, generate most global groundwater data, reflecting a global North bias in the institutional leadership of these data generation activities. We raise three priority themes for future global groundwater data initiatives, which include: data improvements through prioritizing observed and temporally dynamic data; elevating regional and local scale data and perspectives to address challenges relating to equity and bias; and advancing data sharing initiatives founded on reciprocal benefits between global initiatives and data providers.

To better understand the increasing human impact on the water cycle and the feedbacks between hydrology and society, the International Association of Hydrological Sciences (IAHS) organized the scientific decade “Panta Rhei – Everything Flows: Change in hydrology and society” (2013–2022). A key finding is the need to use integrated approaches to assess the co-evolution of human–water systems in order to avoid unintended consequences of human interventions over long periods of time. Additionally, substantial progress has been made in leveraging new data sources on human behaviour, e.g. through text mining of social media posts. Much has been learned about detecting hydrological changes and attributing them to their drivers, e.g. quantifying climate effects on floods. To achieve further progress, we recommend broadening the understanding, the discipline and training activities, while at the same time pursuing synthesis by focusing on key themes, developing innovative approaches and finding sustainable solutions to the world’s water problems.

International databases of disaster impacts are crucial for advancing disaster risk research, particularly as climate change intensifies the frequency and intensity of many natural hazards – including temperature extremes. However, many widely-used disaster impact databases lack information on the physical dimension of the hazards associated with an impact, and on the exposure to such hazards. This hinders analysing drivers of severe disaster outcomes. To bridge this knowledge gap, we present SHEDIS-Temperature, a dataset that provides Subnational Hazard and Exposure information for temperature-related DISaster impact records (https://doi.org/10.7910/DVN/WNOTTC; Lindersson and Messori, 2025). This open-access dataset links temperature-related impact records from the Emergency Events Database (EM-DAT) with subnational data on their locations, associated meteorological time series, and population maps. SHEDIS-Temperature provides hazard and exposure data for 2835 subnational locations associated with 382 disaster records from 1979–2018 in 71 countries. Detailed hazard metrics, derived from 0.1° 3 hourly data, encompass absolute indicators, such as the heat stress measure apparent temperature accounting for humidity and wind speed, as well as percentile-based indicators of when and where temperatures exceeded local thresholds. Population exposure data include annual population figures for impacted subnational administrative units and person-days of exposure to threshold-exceeding temperatures. Outputs are available at grid-point level as well as zonally aggregated to administrative subdivision units, and disaster-record levels. Technical validation against a station-based dataset indicated minor systematic biases – slightly overestimated minimum and underestimated maximum temperatures – but confirmed high consistency between datasets, with correlation coefficients ≥0.9 and mean absolute errors ≤2 °C. By providing comprehensive attributes across the hazard-exposure spectrum, SHEDIS-Temperature supports interdisciplinary research on past temperature-related disasters, offering valuable insights for future risk mitigation and resilience strategies.

Advances in numerical algorithms, improvement of computational power and progress in remote sensing have led to the development of global flood models (GFMs), which promise to be a useful tool for large-scale flood risk management. However, performance and reliability of GFMs, especially in data-scarce regions, is still uncertain, as they are difficult to validate. Here we aim at contributing to develop alternative, more flexible, and consistent methods for GFM validation by applying a change detection analysis on synthetic aperture radar (CD-SAR) imagery obtained from the Sentinel-1 imagery, on a cloud-based geospatial analysis platform. The study addresses two main objectives. First, to validate four widely adopted GFMs with flood maps generated through the proposed CD-SAR approach. This exercise was conducted for eight different large river basins on four continents, to account for a diverse range of hydro-climatic environments. Second, to compare CD-SAR-derived flood maps with those obtained from alternative remote sensing sources. These comparative results offer valuable insights into the reliability of CD-SAR data as a validation tool, more specifically how it stacks up against flood maps generated by other remote sensing techniques.

Managing climate change-related risks requires robust and actionable insights into future climates. Here we develop the plural climate storylines framework to complement existing physical climate storylines, which have strengthened the usability of climate projections yet struggled to generate action for just climate futures. By taking urban adaptation as a case in point, we illustrate the plural climate storylines framework through four complementary methodological schools that bring together multiple knowledges on complex social and climatic processes: power-sensitive storylines, decolonizing storylines, co-producing storylines and aspirational storylines. Our framework generates storylines with the potential to advance transformative policies and new pathways towards climate-just futures.

The human consequences of flood and drought disasters are widespread and detrimental. Large-scale studies, drawing on global geodata products and international databases, can systematically examine how anthropogenic megatrends shape disaster risk and test the generalisability of findings from other scientific methodologies. However, the top-down lens of these global studies often misses the pivotal role that human societies play in shaping disaster risk, including how water management influences physical hazards and how political factors shape social vulnerability. It is precisely this tension – characterised by the need for global perspectives alongside the need to incorporate human influences in the study of disaster risk – that motivates my research.

This thesis specifically examines how observations from global data can leverage our understanding of how humans shape hydrological disaster risk, in terms of the hazard, human exposure and social vulnerability. To this end, the thesis draws on multiple methodologies across four individual studies, including one scoping review and three quantitative geospatial studies. The findings of this thesis provide insights into 1) how the landscape of global data shapes disaster studies and 2) how human societies shape disaster risk.

For the former, my thesis shows that key data opportunities and challenges vary across disaster types and risk dimensions. Addressing each of these limitations is important because of the interrelated nature of disaster risk. The thesis also underlines how the pursuit of transforming fragmented disaster knowledge into holistic and useful information would encounter fewer obstacles if the global datasets were more integrated or, at the very least, more compatible. Databases recording past disaster losses serve as a natural place for such an integration.

For the latter, this thesis brings to light the heterogeneous impact that large-scale infrastructure projects can have on disaster risk, by showing that river regulation does not serve as a universal solution for reducing long-term drought risk. The thesis also highlights the central role of human exposure and economic inequality in shaping human losses during severe flood disasters. Taken together, this underlines the importance of addressing root causes of vulnerability to reduce fatalities during disasters.

Economic inequality is rising within many countries globally, and this can significantly influence the social vulnerability to natural hazards. We analysed income inequality and flood disasters in 67 middle- and high-income countries between 1990 and 2018 and found that unequal countries tend to suffer more flood fatalities. This study integrates geocoded mortality records from 573 major flood disasters with population and economic data to perform generalized linear mixed regression modelling. Our results show that the significant association between income inequality and flood mortality persists after accounting for the per-capita real gross domestic product, population size in flood-affected regions and other potentially confounding variables. The protective effect of increasing gross domestic product disappeared when accounting for income inequality and population size in flood-affected regions. On the basis of our results, we argue that the increasingly uneven distribution of wealth deserves more attention within international disaster-risk research and policy arenas.

This study presents a global explanatory analysis of the interplay between the severity of flood losses and human presence in floodplain areas. In particular, we relate economic losses and fatalities caused by floods during 1990-2000, with changes in human population and built-up areas in floodplains during 2000-2015 by exploiting global archives. We found that population and built-up areas in floodplains increased in the period 2000-2015 for the majority of the analyzed countries, albeit frequent flood losses in the previous period 1990-2000. In some countries, however, population in floodplains decreased in the period 2000-2015, following more severe floods losses that occurred in the period 1975-2000. Our analysis shows that (i) in low-income countries, population in floodplains increased after a period of high flood fatalities; while (ii) in upper-middle and high-income countries, built-up areas increased after a period of frequent economic losses. In this study, we also provide a general framework to advance knowledge of human-flood interactions and support the development of sustainable policies and measures for flood risk management and disaster risk reduction.

Riverine flood risk studies often require the identification of areas prone to potential flooding. This modelling process can be based on either (hydrologically derived) flood hazard maps or (topography-based) hydrogeomorphic floodplain maps. In this paper, we derive and compare riverine flood exposure from three global products: a hydrogeomorphic floodplain map (GFPLAIN250m, hereinafter GFPLAIN) and two flood hazard maps (Flood Hazard Map of the World by the European Commission's Joint Research Centre, hereinafter JRC, and the flood hazard maps produced for the Global Assessment Report on Disaster Risk Reduction 2015, hereinafter GAR). We find an average spatial agreement between these maps of around 30 % at the river basin level on a global scale. This agreement is highly variable across model combinations and geographic conditions, influenced by climatic humidity, river volume, topography, and coastal proximity. Contrary to expectations, the agreement between the two flood hazard maps is lower compared to their agreement with the hydrogeomorphic floodplain map. We also map riverine flood exposure for 26 countries across the global south by intersecting these maps with three human population maps (Global Human Settlement population grid, hereinafter GHS; High Resolution Settlement Layer, hereinafter HRSL; and WorldPop). The findings of this study indicate that hydrogeomorphic floodplain maps can be a valuable way of producing high-resolution maps of flood-prone zones to support riverine flood risk studies, but caution should be taken in regions that are dry, steep, very flat, or near the coast.