Chapter 2 - Key drivers of flood risk change

Extreme floods are one of the most damaging natural hazards, accounting for the majority of all economic losses from natural hazards worldwide. Several intertwined natural and anthropogenic drivers influence flood risk and its change: global warming, precipitation patterns, flood triggering processes, river morphology, river engineering works, population and values at risk, and flood risk reduction strategies. Sustainable flood risk management requires understanding all aspects of flood risk and its change in space and time. Thus, flood risks must be analyzed from a dynamic rather than a static perspective. However, methods to analyze and quantify environmental and socio-economic changes related to flood risk, both in space and time, are nearly not existent. Within this cumulative habilitation thesis methods are examined and developed that allow the analysis of past and future changes in both the natural and human environment with a spatially explicit perspective, and methods that allow disentangling the different drivers of change that are mostly interwoven and have opposing effects on flood risk evolution. The habilitation extends the frontiers of research on flood risk changes with three main methodological approaches: (1) data-driven analyses of environmental and socio-economic change, (2) development of models for specific aspects of flood risk, and (3) model coupling. Coupled component models provide an interesting approach for analyzing flood risk change, for modelling feedback mechanisms between human activities and the natural environment, and for the regionalization of global environmental and socio-economic changes. The habilitation thesis gives an outlook for enabling coupled model frameworks to predict and evaluate the effects of different adaptation strategies on flood risk evolution. Finally a modelling framework that couples specialist models toward whole-system models offers the potential for obtaining an universalist view and unifying several approaches in geography. Such a holistic approach is supporting the search for sustainable solutions for the complex and interconnected problems we are facing today.

Flood risk (d)evolution: Disentangling key drivers of flood risk change with a retro-model experiment

The economic stress and damage from natural hazards are escalating at an alarming rate, calling for anticipatory risk management. Yet few studies have projected flood and drought risk, owing to large uncertainties, strong non‐linearities, and complex spatial‐temporal dynamics. Here, we develop an integrative global risk analysis framework encapsulating future changes in flood and drought hazards as well as associated exposure and vulnerability dimensions. Flood characteristics are quantified by fitting a generalized extreme value distribution (GEV) to the annual flow maxima time series, while drought properties are characterized by the standardized precipitation evapotranspiration index (SPEI) and the standardized precipitation index (SPI). The drivers of drought and flood risk changes at the global and regional scales are explored, and the wide cascade of uncertainties in the risk assessment is decomposed. We find a substantial increase in both flood and drought risk towards the end of the century over most of the globe, driven by compounding changes in exposure, vulnerability, and hazard. A shift from a fossil‐fueled development to a sustainable one decreases the global area facing a risk doubling from 61% to 33% for flood and from 41% to 23% for drought. South America and Africa are identified as hotspot regions where a concomitant, large increase in both flood and drought risk are projected. The hazard quantification method is ubiquitously the dominant uncertainty source for drought risk changes, while the contribution of uncertainty sources for flood risk changes is highly variable in space.

Our understanding of the key drivers of change in flood risk due to climate change remains incomplete. Here, to understand and quantify the key drivers of change in flood risk, we present a framework to undertake a ‘bottom-up’ (‘scenario-neutral’) climate change impact study on flood risk using an event-based flood model that considers non-stationarity in rainfall extremes and catchment wetness. A key advantage of this approach is that by using an event-based model, which explicitly represents key flood drivers, we can directly understand how changes in these drivers will influence changes in flood peaks. The utility of this modelling framework is demonstrated by applying it to one temperate and one tropical catchment in Australia, focusing on the sensitivity of frequent (1 in 5 annual exceedance probability, AEP) and rare (1 in 50 AEP) flood peaks to changes in thermodynamic and dynamic drivers of rainfall extremes, and changes in catchment processes, which are quantified by changes in rainfall losses.Response surfaces of the catchment sensitivity are first produced and demonstrate that increases in rainfall intensity drive increases in the flood peak. Smaller flood peaks in drier catchments are most sensitive to changes in rainfall losses, with the modulating impact of changes in losses decreasing as the runoff ratio increases. Climate model projections of these drivers are then superimposed on these surfaces to identify plausible shifts in flood risk under climate change. Several climate model ensemble members imply a decrease in future flood peaks, most notably 14 % of ensemble members suggest future decreases for the 1 in 5 AEP event for the temperate catchment. Despite large agreement on the direction of change, we found large variability in the expected magnitude of change for both frequent and rare events with flood peaks projected to increase between 0 % and 45 % for the tropical catchment and between −10 % to more than 50 % for the temperate catchment under RCP8.5 in 2085, highlighting the deep uncertainty associated with projecting flood risk under climate change. We believe this study acts as proof of concept for how a bottom-up climate change assessment can be undertaken within an event-based flood modelling framework, and provides insights to help us better understand and communicate the key drivers of changes to flood risk into the future.

Flood damage has increased significantly and is expected to rise further in many parts of the world. For assessing potential changes in flood risk, this paper presents an integrated model chain quantifying flood hazards and losses while considering climate and land use changes. In the case study region, risk estimates for the present and the near future illustrate that changes in flood risk by 2030 are relatively low compared to historic periods. While the impact of climate change on the flood hazard and risk by 2030 is slight or negligible, strong urbanisation associated with economic growth contributes to a remarkable increase in flood risk. Therefore, it is recommended to frequently consider land use scenarios and economic developments when assessing future flood risks. Further, an adapted and sustainable risk management is necessary to encounter rising flood losses, in which non-structural measures are becoming more and more important. The case study demonstrates that adaptation by non-structural measures such as stricter land use regulations or enhancement of private precaution is capable of reducing flood risk by around 30 %. Ignoring flood risks, in contrast, always leads to further increasing losses—with our assumptions by 17 %. These findings underline that private precaution and land use regulation could be taken into account as low cost adaptation strategies to global climate change in many flood prone areas. Since such measures reduce flood risk regardless of climate or land use changes, they can also be recommended as no-regret measures.

Confronting complexity

Using precipitation and temperature data for the 20th century in combination with a macroscale hydrologic model, we evaluate changes in flood risk in the western U.S. associated both with century‐scale warming and interannual climate variations. In addition, we examine the implications of apparent increases in precipitation variability over the region since the mid‐1970s. We use detrended temperature data representing early and late 20th century climate to force the variable infiltration capacity hydrologic model and show that spatially homogeneous temperature changes over the western U.S. in the 20th century on the order of +1°C per century have resulted in substantial changes in flood risks over much of the region. Although changes specific to particular geographic areas are apparent in some cases, the overall changes due to observed warming trends are well categorized by midwinter temperature regimes in each watershed. Cold river basins where snow processes dominate the annual hydrologic cycle (<−6°C average in midwinter) typically show reductions in flood risk due to overall reductions in spring snowpack. Relatively warm rain‐dominant basins (>5°C average in midwinter) show little systematic change. Intermediate or transient basins show a wide range of effects depending on competing factors such as the relative role of antecedent snow and contributing basin area during storms that cause flooding. Warmer transient basins along the coast in Washington, Oregon, and California, in particular, tend to show increased flood risk. While the absolute value of simulated changes in flood risk is affected by basin scale, the nature of the relationship of flood risk to basin temperatures in midwinter is largely scale‐independent. Climate variations associated with Pacific Decadal Oscillation (PDO) and El Niño Southern Oscillation (ENSO) also have strong effects on flood risks. In contrast to the effects associated with 20th century warming, the climate variability signal is characterized by regional scale patterns related to the geographic distribution of cool season precipitation also identified in many previous studies. In general, the largest changes in simulated flood risks are associated with years when PDO and ENSO are “in phase,” particularly in the southwest. Changes in the variability of cool season precipitation after about 1973, the causes of which are uncertain, are shown to result in increased flood risk over much of the western U.S. in the simulations.

Managing flood risks in a changing climate

Quantified scenarios analysis of drivers and impacts of changing flood risk in England and Wales: 2030–2100

Floodplains, as seen from the flood risk management perspective, are composed of co-evolving natural and human systems. Both flood processes (that is, the hazard) and the values at risk (that is, settlements and infrastructure built in hazardous areas) are dynamically changing over time and influence each other. These changes influence future risk pathways. The co-evolution of all of these drivers for changes in flood risk could lead to emergent behavior. Hence, complexity theory and systems science can provide a sound theoretical framework for flood risk management in the 21st century. This review aims at providing an entry point for modelers in flood risk research to consider floodplains as complex adaptive systems. For the systems science community, the actual problems and approaches in the flood risk research community are summarized. Finally, an outlook is given on potential future coupled component modeling approaches that aims at bringing together both disciplines.

Abstract. Coastal managers face the task of assessing and managing flood risk. This requires knowledge of the area of land, the number of people, properties and other infrastructure potentially affected by floods. Such analyses are usually static; i.e. they only consider a snapshot of the current situation. This misses the opportunity to learn about the role of key drivers of historical changes in flood risk, such as development and population rise in the coastal flood plain, as well as sea-level rise. In this paper, we develop and apply a method to analyse the temporal evolution of residential population exposure to coastal flooding. It uses readily available data in a GIS environment. We examine how population and sea-level change have modified exposure over two centuries in two neighbouring coastal sites: Portsea and Hayling Islands on the UK south coast. The analysis shows that flood exposure changes as a result of increases in population, changes in coastal population density and sea level rise. The results indicate that to date, population change is the dominant driver of the increase in exposure to flooding in the study sites, but climate change may outweigh this in the future. A full analysis of changing flood risk is not possible as data on historic defences and wider vulnerability are not available. Hence, the historic evolution of flood exposure is as close as we can get to a historic evolution of flood risk. The method is applicable anywhere that suitable floodplain geometry, sea level and population data sets are available and could be widely applied, and will help inform coastal managers of the time evolution in coastal flood drivers.

Drivers of future fluvial flood risk change for residential buildings in Europe

We report on a regional flood and earthquake risk assessment for 33 countries in Eastern Europe and Central Asia. Flood and earthquake risk were defined in terms of affected population and affected gross domestic product (GDP). Earthquake risk was also quantified in terms of fatalities and capital loss. Estimates of future population and GDP affected by earthquakes vary significantly among five shared socioeconomic pathways that are used to represent population and GDP in 2030 and 2080. There is a linear relationship between the future relative change in a nation's exposure (population or GDP) and its future relative change in annual average population or GDP affected by earthquakes. The evolution of flood hazard was quantified using a flood model with boundary conditions derived from five different general circulation models and two representative concentration pathways, and changes in population and GDP were quantified using two shared socioeconomic pathways. There is a nonlinear relationship between the future relative change in a nation's exposure (population or GDP) and its future relative change in its annual average population or GDP affected by floods. Six regions can be defined for positive and negative relative change in population that designate whether climate change can temper, counter, or reinforce relative changes in flood risk produced by changes in population or exposure. The departure from the one‐to‐one relationship between a relative change in a nation's population or GDP and its relative change in flood risk could be used to inform further efforts at flood mitigation and adaptation.

<p>In the coming decades, climate change will likely become a complex issue affecting hydrological regimes and flood hazard conditions. According to the IPCC reports, significant changes in atmospheric temperature, precipitation, humidity, and circulation are expected which may lead to extreme events including flood, droughts, heatwaves, heavy precipitation, and more intense cyclones. Although the effects of climate change on flood hazard indices is subject to large uncertainty, the evaluation of high-flows plays a crucial role in flood risk planning and extreme event management. With the advent of the Coupled Model Intercomparison Project Phase 6 (CMIP6), flood managers are interested to know how changes in catchment flood risk are expected to alter relative to previous assessments. Here we examine catchment based projected changes in flood quantiles and extreme high flow events for Irish catchments, selected to be representative of the range of hydrological conditions on the island. Conceptual hydrological models, together with different downscaling techniques are used to examine changes in flood risk projected from the CMIP6 archive for mid and end of century. Results will inform the range of plausible changes expected for policy relevant flood indices, the sensitivity of findings to use of different climate model ensembles and inform the tailoring of adaptation plans to account for the new generation of climate model outputs.</p>

There is growing evidence of coherent, global patterns of change in annual precipitation and runoff with high latitudes experiencing increases consistent with climate model projections. This paper describes a methodology for estimating detection times for changes in seasonal precipitation extremes. The approach is illustrated using changes in UK precipitation projected by the European Union PRUDENCE climate model ensemble. We show that because of high variability from year to year and confounding factors, detection of anthropogenic climate change at regional scales is not generally expected for decades to come. Overall, the earliest detection times were found for 10 day winter precipitation totals with 10 year return period in SW England. In this case, formal detection could be possible within a decade from now if the climate model projections are realized. The outlook for changes in summer flash flood risk is highly uncertain. Our analysis further demonstrates that existing precautionary allowances for climate change used for flood management may not be sufficiently robust in NE England and east Scotland. These findings imply that for certain types of flood mechanism, adaptation decisions might have to be taken in advance of formally detected changes in flood risk. This reinforces the case for long‐term environmental monitoring and reporting of climate change indices at “sentinel” locations.