Last year, a combination of unprecedented local rainfall intensities and storm surges from Hurricanes Harvey, Irma, and Maria flooded Houston, large parts of Florida, and numerous island nations in the Caribbean. Such hydrologic compound events—when the combination of two or more hazard events or climate variables leads to an extreme impact—have a multiplier effect on the risk to society, the environment, and infrastructure [Zscheischler et al., 2018].
Hydrologic extremes, such as floods and droughts, are among the world’s most dangerous and costly natural hazards. Between 1980 and 2013, flood losses tallied more than $1 trillion and caused more than 220,000 recorded fatalities globally. In 2017 alone, major flood disasters in the United States, Caribbean, and Southeast Asia killed more than 1,000 people, caused damages on the order of hundreds of billions of U.S. dollars, and harmed or destroyed the livelihoods of millions of people.
Droughts, a different type of hydrologic extreme, can lead to widespread food and water crises. In the United States, 23 drought and drought-heatwave events between 1980 and 2016 caused economic losses totaling $216 billion. Severe drought conditions in Somalia between 2015 and 2017 have led more than 700,000 people to leave their homes, and these conditions have left more than 6.2 million people in need of humanitarian assistance.
High-impact compound events make up a large portion of these losses, but they are often overlooked in disaster risk analyses and policy agendas. What factors overlap to cause such extremes? And how does one hydrologic extreme lead to other environmental crises? What do we know about them, and what do we still need to know?
Assessing the Risks
Efforts to reduce the risks posed by hydrologic extremes are essential. Accordingly, risk reduction is at the heart of two recent international agreements: the Sendai Framework for Disaster Risk Reduction and the Warsaw International Mechanism for Loss and Damage Associated with Climate Change Impacts.
However, because compound events emerge from complex processes having multiple causes, they do not conform neatly with traditional categories of extremes or current risk assessment methodologies.
The Intergovernmental Panel on Climate Change defines compound events as (1) two or more extreme events occurring simultaneously or successively, (2) combinations of extreme events with underlying conditions that amplify the impact of the events, or (3) combinations of events that are not themselves extremes but lead to an extreme event or impact when combined [Seneviratne et al., 2012].
In some cases, factors are strongly interdependent (e.g., floods at river confluences) and must be treated accordingly in risk assessments. In other cases, there may be some negligible statistical dependency in the system, small enough to assume physical independence (e.g., tide and surge along some coastline stretches). In yet other cases, all variables involved will be truly independent, but the fact that they occur together aggravates their impacts (e.g., tides and river discharge).
As a result, it is important to understand a system’s behavior, identify all relevant variables that determine the impacts, and carefully assess whether those variables can be treated as independent or if more complex dynamical or statistical models need to be employed to capture the dependence structure between them.
Although the relationships between flood drivers in coastal areas are well known, current risk assessment techniques mainly consider one driver at a time and ignore the dependence between drivers, leading to underestimation or overestimation of the underlying risk. For example, the dependence between hurricane storm tides, existing groundwater levels, and precipitation is not accounted for in the estimation of the 100-year floodplain by the U.S. Federal Emergency Management Agency used for the National Flood Insurance Program.
Each variable contributing to an extreme compound event is not necessarily an extreme event on its own. For example, a moderate storm surge can cause significant flooding by blocking or slowing typical river discharge or gravity-fed drainage through storm water systems. In the United States, Hurricane Harvey brought intense rainfall to the greater Houston area. It also caused a moderate storm surge over 5 days, which reduced the freshwater drainage capacity. Had there not been a storm surge, the flooding would have been less, as the water would have drained faster.
Compounding effects are also apparent when extreme events occur in close succession. A prominent example is the 2013–2014 winter storm season in the United Kingdom, when several storms over a short period of time produced large storm surges, high waves, and strong rainfall [Wadey et al., 2014; Haigh et al., 2016]. In such cases, smaller storms that happen after a large storm may have greater impacts than would normally be expected because of factors such as saturated soils, eroded beaches, and weakened flood defenses that leave communities more vulnerable.
When a drought coincides with a heat wave, the impacts are also much greater than when they occur in isolation. The California drought of 2012–2016, for example, is characterized by both below-average precipitation and sustained high temperatures [AghaKouchak et al., 2014]. The precipitation deficits are not the most extreme on record, but record temperatures and multiple extreme heat waves exacerbated the impacts of the hydrological drought.
The Human Dimension
The changing nature of human activities is a challenging aspect of compound events. For example, the availability of freshwater resources in a given location is driven by locally generated runoff as well as discharge from upstream areas. The amount of water coming from upstream is often strongly influenced by human interventions, for example, water withdrawal for irrigation, electricity production, domestic use, and reservoir and land management [Veldkamp et al., 2017; Mehran et al., 2017].
In many regions, water consumption already exceeds the available renewable water supplies, making alternative resources, such as water diversion, desalination, and deep wells, necessary to satisfy the demand.
The effects of meteorological droughts can be exacerbated by increased human water use. These “anthropogenic droughts” are broadly defined as water stress caused or intensified by such human activities as increased demand, mismanagement, and anthropogenic greenhouse gas emissions [AghaKouchak et al., 2015].
Not only are anthropogenic influences a feature of most systems, but they can also act to worsen the effects of compound events. In Brisbane, Australia, during the 2011 Queensland floods, dam engineers focusing on maintaining an adequate water supply after years of drought failed to anticipate the unusual volume of rainfall that eventually caused the reservoir to overflow the Wivenhoe Dam, causing extensive flood damage downstream [Van den Honert and McAneney, 2011]. In a similar context, lack of updated operating rules likely led to the damaging of the Oroville Dam in California in 2017, when an extreme wet season followed a 5-year record-setting drought [Vahedifard et al., 2017]
The Role of Climate Change and Variability
Unraveling the potential influence of climate change and natural variability on compound events is a challenging task because different mechanisms affect the variables involved. Even if the driving variables are stationary, their dependence structure may change, which changes the likelihood that they will coincide.
Recent research has already revealed changes in the frequency of the concurrent occurrence of storm surges and rainfall [Wahl et al., 2015] and drought and heat waves [Mazdiyasni and AghaKouchak, 2015] in the United States. Given the expected rise in global temperatures due to human activities, the chance of concurrent droughts and high temperatures will likely increase in the future. Higher sea levels will likewise increase the chances of compound flooding in coastal regions [Moftakhari et al., 2017].
Although the popular press sometimes attributes extreme weather events to climate change, establishing the most likely causes of a given event is an emerging field and particularly challenging for compound events. A recent report by the National Academies of Sciences, Engineering, and Medicine  takes an important step toward benchmarking scientific approaches that characterize uncertainty and articulation of event attribution. However, compound events quickly illuminate the challenge of understanding and attributing interlinked phenomena. The propagation of errors and uncertainties inherent in creating models of these phenomena potentially imposes limits on usable knowledge for climate adaptation concerns.
The Way Forward
Experts who study and simulate extreme events and their impacts are best able to pinpoint knowledge gaps. In this vein, future efforts to address hydrologic compound events will benefit from efforts on several fronts.
System stress tests can be exploited to articulate the relevant attributes of the impact system [e.g., Bevacqua et al., 2017]. These approaches include understanding how human (ir)rational behavior may trigger, create, or alleviate the occurrence and impact of compound events.
Additional key variables and event combinations that require further analysis must be identified [e.g., Liu et al., 2016]. Appropriate statistical methods must be used to simulate dependence in time and space [Haigh et al., 2016] and across multiple variables [e.g., Wahl et al., 2015]. Data and model requirements must be identified for documenting, understanding, simulating, and attributing compound events. Last, compound events must be incorporated into impact assessments and disaster risk mitigation works [e.g., Zheng et al., 2017].
These efforts will be enhanced by close collaboration and communication between scientists from various fields, including natural sciences, engineering, and social sciences, as well as stakeholders and policy makers. Research in this area is still in its infancy. A community effort, including promotion by major scientific associations, will help improve methods of detection, modeling, and risk assessment of compound extremes. For example, a recently established European Union COST Action Network on “Understanding and modeling compound climate and weather events (DAMOCLES)” will facilitate international collaboration on this important topic.