There is no universal definition of drought. It has been defined in many ways by various disciplines because the characteristics of drought and its impacts reflect the climate and societal characteristics of the region affected. Conceptually, drought is a deficiency of precipitation from expected or “normal” conditions that extends over a season or longer period of time and where water is thus insufficient to meet the needs of human activities and the environment. Four types of drought are generally recognized: meteorological, agricultural, hydrological, and socioeconomic. Drought risk is a product of a region’s exposure to natural hazards and its vulnerability to extended periods of water shortage. Drought is different from many other natural hazards in that it does not begin and end swiftly. Its onset is gradual, and as intensity and duration increase, the effects correspondingly become more widespread. This creeping phenomenon is expressed through multiple indicators because impacts are nonstructural and spread over large areas.
All meteorological droughts begin with a deficiency of precipitation over time. But nature is not the only cause of droughts. Droughts can be socially constructed or amplified when there is insufficient precipitation to meet needs (e.g., needs may be greater than “normal” supply if unsustainable development has occurred). Therefore, droughts are caused by changes in both supply of and demand for water, and both are dynamic. Impacts of drought are multifaceted. For instance, for agriculture, lack of precipitation affects the critical variables of soil moisture and evapotranspiration, and these become the vectors of adverse changes for society. For water supply for human uses, as the intensity, duration, and spatial extent of drought increase, the natural dimension decreases in
significance and a high priority is placed on water resources management. Some water shortages can be the result of overappropriation of supply so that “drought” occurs even in years of “normal” precipitation.
On average approximately 15 percent of the United States is affected by drought each year, based on the historical record from 1895 to now (National Climatic Data Center/NOAA, 2005). Figure 2-1 illustrates recent drought trends. Figure 2-2 illustrates longer-term trends for one location and shows a number of periods noteworthy for their duration, severity, and spatial extent.
A still ongoing drought began in 1996 for large parts of the country. Approximately 35 to 40 percent of the country has been affected at one time or another by severe to extreme drought during this period, and for some regions drought conditions have persisted for five or more years. For example, parts of the Southeast, particularly Georgia, North Carolina, South Carolina, and Florida, experienced three to four consecutive years of drought between 1999 and 2002. In the West much of the Southwest experienced five consecutive years of drought between 2001 and 2004, while much of Montana, Idaho, and surrounding states have experienced severe drought for as many as seven consecutive years since 1999. The most recent drought is particularly notable in that it was hotter than a similar drought of the 1950s. In general, western North America has seen significant warming over the last 100 years, particularly in the last couple of decades.
The recent period of unprecedented population increase in the western United States coincided with one of the wettest periods on record. During the
1990s, for example, the population of Nevada increased by 66 percent, Arizona by 40 percent, Colorado by 30 percent, Idaho by 28 percent, and New Mexico by 20 percent. The rapidly expanding population in the region is exacerbating water conflicts between multiple and competing users. It is also worth noting that the most important management agreement in the West (the Colorado River Compact of 1922) was based on an overestimation of the reliable average annual supply of water (estimated at 16.4 million acre feet [MAF] when the true long-term average flow of the Colorado River is estimated at between 12 and 16 MAF, with the current best estimate at 13.5 MAF) because the allocation was based on a short observational record that represented one of the region’s wettest periods.1
The paleoclimate record, seen through proxy records such as tree rings, coral, and boreholes in ice, allows scientists to estimate precipitation amounts back much farther than the past few hundred years of instrument records, and it is clear that significant droughts are not solely a modern phenomenon (Cook et al., 2004; Woodhouse and Overpeck, 1998). Droughts comparable to those of the 1930s, 1950s, or early 2000s have occurred in general once or twice per century over the last 2000 years.2 Research by Clark et al. (2002) using lake sediment records from
the northern Great Plains shows pronounced 100- to 130-year drought cycles back at least 8,000 years.
Droughts of the 20th century were eclipsed by past droughts in terms of annual maximum severity, duration of drought, and geographic extent of drought. The paleoclimatic record indicates that droughts longer than a decade (i.e., “megadroughts”) were not rare, and that droughts affecting much of the western United States have lasted as long as a century or more within the last 2,000 years (Gray et al., 2004). The true range of “natural” drought variability is thus substantially larger and more complex than suggested by the last century, when there are accurate records of drought variations provided by modern instrumentation.
In general, most of the large droughts of the western United States have affected more than one major river basin at a time, and some (e.g., a megadrought in the late 16th century) apparently affected the United States from coast to coast and from northern to southern borders. Another key aspect of drought variability illuminated by the paleoclimatic record is that decades- to centuries-long hydrological “regimes” (e.g., characterized by rare/short or frequent/longer droughts) have begun and ended abruptly; transitions between drought regimes can take place over years to decades, whereas the regimes themselves can be significantly longer.3
Great strides have been made in recent years with respect to understanding the proximal cause of drought in North America. Drought in the southwestern United States (e.g., 1950s and the recent drought) is known to be connected with the El Niño/Southern Oscillation (ENSO), and dry winters are favored in La Niña years.4 More recently, studies have confirmed that anomalous sea surface temperature (SST) patterns, particularly in the tropics and the Indo-Pacific, can explain both 20th-century and earlier droughts, and research has identified strong statistical associations between decadal modes of Pacific and Atlantic variability with decadal patterns of wet and dry conditions over North America (McCabe et al., 2004; Hoerling and Kumar, 2003). Of course, the major challenge is to explain the causes of the anomalous—and persistent—ocean conditions that lead to North American drought.5
There is increasing consensus that anthropogenic forcing will likely increase the probability of drought in central and western North America.6 Exacerbating this likelihood is the fact that temperatures are already rising significantly in the American West, and snowpack is retreating in the same region. A lesson of the paleoclimatic record is that anthropogenic forcing could trigger an abrupt transition into a more drought-prone climatic regime, thus increasing the frequency and duration of drought. Given that these possibilities could materialize with or with-
out significant future human-induced climate change, it makes sense to consider no-regrets strategies to reduce vulnerability to drought in either case.7
CONTEXT AND IMPACTS
As noted earlier, drought was selected for examination as a case study because it both is driven by multiple environmental stresses and leads to multiple stresses across a wide range of scales. Drought is, at its simplest, an imbalance between water supply and water demand. Yet in reality, so many variables are present in both sides of that equation that it is clearly a classic example of a multiple-stresses scenario, where many factors combine in ways that are not entirely predictable. Drought typically evolves slowly, and as it progresses the impacts accumulate and expand in scope, extent, and intensity. From any one impact, there can be cascading impacts. This is clearly a challenge for decision makers who must plan for and react to the accumulating impacts.
Both climate and society are dynamic, as are the relationships between them (i.e., impacts and vulnerabilities). The different influences occur on varying scales (both time and space) and thus are challenging to predict with any accuracy. Drought is indeed a normal component of climate variability, but as a generator of multiple environmental stresses, it is the longer droughts—multiyear, multidecadal, and even centennial—that are of greatest concern. The longer the drought, the wider the range of ecosystem and societal effects that cascade from primary to tertiary impacts as time lengthens and spatial scales widen.
Terrestrial ecosystems respond not only to temperature but particularly to decreasing soil moisture that in turn induces woody plant mortality, rapid canopy change, and increased soil erosion.8 Drought-induced tree mortality results in heightened vulnerability to fires. Note that canopy fires are the largest type of wildfire events in forests (Strauss et al., 1989). Increased temperature over long periods of time can also facilitate larger infestations of pests that magnify tree mortality and thereby expand the spatial scale of forest fires. Ecotones, or zones of transition between ecosystems, appear to be most susceptible to rapid vegetation change under stress (Allen and Breshears, 1998). As reductions in herbaceous groundcover increase, the distribution of near-ground energy is altered and can affect a wide range of ecosystem processes dependent on that energy. The threat of nonlinear increases in soil erosion rises significantly as a consequence of that shift in near-ground energy.
As drought intensifies over time, its spatial scale expands and its societal ramifications deepen (Wilhite et al., 2005). Infrastructures are affected, as is the supply of social services, and there emerge serious distributive consequences for the less well off members of society. Trends in the spatial distribution of water,
population size, demand, and competing uses all gain heightened significance and increased stress at times of increasing water scarcity.
Rates of change in the timing and supply of water, as well as sequencing of wet and dry years, can aggravate the impacts of drought. When wet and dry years can alternate frequently, human and ecological systems adapt to the extremes of variability; multiyear variability, on the other hand, presents the illusion of stability and the human impacts can be greater because people plan inadequately. Not only do human expectations differ based on timing and sequencing, but insects, pathogens, and other pests also respond in nonlinear fashion to some hydrological trends. Sequencing of water availability in the second half of the 20th century and into the first decade of the 21st century is of particular significance to the American West. For instance, snowpack provides up to 80 percent of the runoff, and since the 1950s there has been a long-term trend in increasing temperature and decreasing snowpack (Hamlet et al., 2005). The onset of spring and growing season evapotranspiration, the timing of snowmelt and snowmelt discharge, and the amount of recharge as the proportion of precipitation shifts from snow to rain at critical elevations will all have effects (Stewart et al., 2005; Cayan et al., 2001; Knowles et al., 2006). Increasing temperature, even if precipitation remains constant, has the potential to dramatically alter the hydrology of river basins and the severity of drought episodes. These types of trends are likely to lead to multiple stresses on both ecosystems and human social systems, thereby exacerbating competition between the two. Given these challenges, there is an acute need for further development of our capacity to predict the onset of drought. We now recognize that the true range of natural drought variability is substantially larger and more complex than is suggested by the 20th-century instrumental record. Research (Hoerling and Kumar, 2003) shows that while land surface feedbacks with the atmosphere can be important to amplify or dampen drought in some locations or seasons, these feedbacks are not always dominant in driving temporal drought variability, and patterns in coupled atmosphere-ocean variability also play an important role.
If the United States and other nations are to make progress in reducing the serious consequences of drought, an improved understanding of the hazard and its prediction and the full range of factors that influence vulnerability is needed. Enhancing our knowledge of the hazard will require a complex, integrated early warning system that incorporates climate, soil, and water supply factors such as precipitation, temperature, soil moisture, snowpack, reservoir and lake levels, groundwater levels, and streamflow. The implementation of the National Integrated Drought Information System (NIDIS), currently underway within the National Oceanic and Atmospheric Association (NOAA), represents a multiagency and multiorganizational effort and was viewed by workshop participants as an important step in the development of an improved decision-support system for the country.
Because of the slow-onset nature of drought, the ability of resource managers to adapt and/or impose alternative management practices in a timely fashion would be greatly enhanced by more reliable seasonal climate forecasts. Unfortunately, little skill currently exists to reliably predict the onset, severity, duration, spatial extent, or end of a drought event a season or more in advance. This is a critical research need. In addition to improved seasonal forecasts, there is need for improved methods of probability risk assessments that rely on reconstructions of past climates and account for climatic nonstationarity to calculate the occurrence and return probabilities of drought (Enfield and Cid-Serrano, forthcoming). Improving seasonal forecasts and their application will also require collaboration between scientists at research institutes, universities, and federal agencies. The primary end users of these forecasts need to be involved in this process so researchers understand their needs.
UNDERSTANDING VULNERABILITY AND RESPONSE STRATEGIES
Vulnerability to drought is dynamic and is influenced by a multitude of factors, including increasing population, regional population shifts, urbanization, technology, government policies, land use and other natural resource management practices, desertification or land degradation processes, water use trends, and changes in environmental values (e.g., protection of wetlands or endangered species). Therefore, the magnitude of drought impacts may increase in the future as a result of an increased frequency of occurrence of the natural event (i.e., meteorological drought), changes in the factors that affect vulnerability, or a combination of these elements. The development of a national drought policy and preparedness plans at all levels of government that place emphasis on risk management rather than following the traditional approach of crisis management would be a prudent step for the United States to take. Crisis management decreases self-reliance and increases dependence on government and donors.
The impacts of drought in recent years have been increasing, although there is no systematic national assessment of drought impacts such as exists with other natural disasters. The Federal Emergency Management Agency (FEMA) (1995) estimated annual losses in the United States because of drought at $6 to $8 billion, making drought the most costly natural disaster in the country until Hurricane Katrina in 2005. Losses from the 1988 drought have been estimated at more than $39 billion. The National Drought Mitigation Center (Hayes et al., 2004) has estimated that the losses associated with the 2002 drought exceeded $10 billion, based on estimates made by 10 states. If these losses are extrapolated to include all drought-affected states, they would be significantly higher. It is important to note that these are estimates for a single drought year; major drought events often occur over a series of years (e.g., 1999-2005).
The impacts of drought have also been growing in complexity. Historically, the most significant impacts associated with drought have occurred in the agricul-
tural sector (i.e., crop and livestock production). There has been an expansion of impacts in other sectors, particularly energy production, recreation and tourism, transportation, forest and wildland fires, urban water supply, environment, and human health. The recent drought years in the western United States, for example, have resulted in financial impacts in nonagricultural sectors that have likely exceeded those in agriculture. In addition to the direct impacts of drought, there are significant indirect impacts that, in many cases, exceed in value the direct losses associated with drought episodes. In addition to these human-focused impacts, there are the effects on nonhuman systems that are even more difficult to quantify.
DROUGHT POLICY AND PREPAREDNESS
In the past decade or so drought policy and preparedness have received increasing attention from governments, international and regional organizations, and nongovernmental organizations. Simply stated, a national drought policy is a way to establish a clear set of principles or operating guidelines to govern the management of drought and its impacts. The ideal policy is consistent and equitable for all regions, population groups, and economic sectors and is consistent with the goals of sustainable development and the wise stewardship of natural resources. The overriding principle of drought policy is an emphasis on risk management through the application of preparedness and mitigation measures. Preparedness refers to predisaster activities designed to increase the level of readiness or improve operational and institutional capabilities for responding to a drought episode. Mitigation is short- and long-term actions, programs, or policies implemented during and in advance of drought that reduce the degree of risk to human life, property, and productive capacity. These actions are most effective if done before the event. Emergency response will always be a part of drought management because it is unlikely that government and others can anticipate, avoid, or reduce all potential impacts through mitigation programs. A future drought event may also exceed the capacity of a region to respond. However, emergency response is best used sparingly and only if it is consistent with longer-term drought policy goals and objectives.
A key component of a national drought policy is to reduce risk by developing better awareness and understanding of the drought hazard and the underlying causes of societal vulnerability (Hayes et al., 2004). The principles of risk management can be promoted by encouraging the improvement and application of seasonal and shorter-term forecasts, developing integrated monitoring and drought early warning systems and information delivery systems, developing preparedness plans at various levels of government, adopting mitigation actions and programs, and creating a safety net of emergency response programs that ensure timely and targeted relief. The delivery of information in a timely manner
is essential so that informed decisions can be made by resource managers and others.
The traditional approach to drought management has been reactive, relying largely on crisis management. This approach has been ineffective because response is untimely, poorly coordinated, and poorly targeted to drought-stricken groups or areas. In addition, drought response is post-impact and relief tends to reinforce existing resource management practices (i.e., it rewards poor resource management and the lack of preparedness planning). Many governments are striving to learn how to employ proper risk management techniques to reduce societal vulnerability to drought and therefore lessen the impacts associated with future drought events.
There are four key components in an effective drought risk reduction strategy (O’Meagher et al., 2000): (1) the availability of timely and reliable information on which to base decisions; (2) policies and institutional arrangements that encourage assessment, communication, and application of that information; (3) a suite of appropriate risk management measures for decision makers; and (4) actions by decision makers that are effective and consistent. In 1992 Australia adopted a national drought policy that applied these components through three objectives: (1) to encourage primary producers and other sections of rural Australia to adopt self-reliant approaches to managing for climatic variability; (2) to maintain and protect Australia’s agricultural and environmental resource base during periods of extreme climate stress; and (3) to ensure early recovery of agricultural and rural industries, consistent with long-term sustainable goals (O’Meagher et al., 2000). Australia’s national drought policy is widely known and its philosophy adaptable in other settings (Botterill, 2005).
In the United States there has been some progress in the development of preparedness plans. The most noticeable progress has been at the state level, where the number of states with drought plans has increased dramatically during the past two decades. In 1982 only three states had drought plans in place; in 2005, 38 states had developed plans. The basic goal of state drought plans should be to improve the effectiveness of preparedness and response efforts by enhancing monitoring and early warning, risk and impact assessment, and mitigation and response. Plans should also contain provisions (i.e., an organizational structure or framework) to improve coordination within agencies of state government and between local and federal government. Initially, state drought plans largely focused on response efforts aimed at improving coordination and shortening response time; today the trend is for states to place greater emphasis on mitigation as the fundamental element of a drought plan. Thus, some plans are now more proactive, adopting a more risk management approach to drought management.
Drought mitigation plans have three essential components, regardless of whether they are developed at the state, national, regional, or local scale.
A comprehensive monitoring and early warning system provides the basis for many of the decisions that must be made by a wide range of decision makers as drought conditions evolve and become more severe. Equally important, early warning systems need to be coupled to an effective delivery system that disseminates timely and reliable information. As drought plans incorporate more mitigation actions, it is imperative that these actions be linked to thresholds (e.g., reservoir levels, climate index values) that can serve as triggers for mitigation and emergency response actions.
A critical step in the development of a mitigation plan is the conduct of a risk assessment of vulnerable population groups, economic sectors, and regions (Knutson et al., 1998). The purpose of the risk assessment is to determine who and what is at risk and why. This is accomplished through an analysis of historical and recent impacts associated with drought events.
After impacts have been identified and prioritized, the next step is to identify appropriate mitigation actions that can help to reduce the risk of each impact for future drought events.
One existing need is quantifying the advantages of a risk-based drought mitigation planning effort over the crisis management approach so policy makers see the advantages of committing resources up front to develop and implement mitigation actions rather than waiting to deal with impacts during a crisis. In most cases the costs associated with mitigation actions are minimal when compared with the costs of drought, which often are in the billions of dollars.
The U.S. Congress has considered actions that could be taken in response to recommendations issued in May 2000 by the National Drought Policy Commission (NDPC) but has not moved on these discussions. One of the NDPC’s recommendations strongly endorses drought planning at all levels of government. Legislation has been introduced in the U.S. Congress that could lead to the creation of a more permanent national drought council and a national drought policy. Key components of this bill included an emphasis on risk management, preparedness planning, and improvement of the nation’s drought monitoring system and forecasting capabilities. A project, the National Integrated Drought Information System (NIDIS),9 is intended to provide the foundation for development of an improved drought monitoring system (Western Governors’ Association, 2004). NIDIS is one of the components of the National Drought Preparedness Act, and authorization for NIDIS has been included in two bills introduced in Congress (HR 5136, S2751) in spring 2006.
During the workshop, once the formal presentations were complete, the participants brainstormed a variety of possible research questions and needs. It was noted that the ability to make projections about the future is what makes knowledge powerful, and that in drought this means that much of the work is site specific and situation dependent. Because investment strategies related to water management and drought mitigation will differ depending on the size, cost, and service life of the strategy or facility, such projections are particularly important to policy makers. Integrated analysis of multiple stresses needs to replace cause-effect-type analysis. Much needs to be done to link large and small scales so that broad-based knowledge actually has practical application on the ground. There are still major gaps in understanding and communicating to and from the user community, and this inherently includes education and outreach. There is great value to studying the paleo record both because this gives us a basis for understanding megadrought and because looking at the past gives a greater time span over which relationships between duration, frequency, severity, extent, and location can be defined and quantified at the continental scale. There is also a need to consider the impacts on the services provided by ecosystems.
Finally, participants discussed possible steps to improve our capability to integrate science knowledge so that we are better able to deal with multiple stresses in decision support, with a focus on research needs. For the drought case study, the participants listed the following as steps that could advance our understanding of the multiple-stresses components and interactions for drought:
Implement the National Integrated Drought Information System, emphasizing a partnership between federal and nonfederal agencies and organizations. This would improve monitoring and early warning systems to provide better and more timely and reliable information to decision makers; address data gaps in drought monitoring and enhance networks, particularly for soil moisture, snowpack, and groundwater; and develop new monitoring and assessment tools/ products that will provide resource managers at all levels with proper decision-support tools at higher resolution.
Help the scientific and policy communities and resource managers better understand drought as a complex natural hazard. For example, more effort could be made to communicate information on probabilities of single- and multiple-year drought events to natural resource managers and planners, policy makers, and the public in ways and formats that make sense to different audiences. Paleoclimate and historical climate research could be done to better understand past droughts at the regional scale. Additional research could be done on nonlinear processes and thresholds to provide improved prediction capability about the connections between tree mortality, energy and water budget shifts, and soil erosion.
Improve the reliability of seasonal climate forecasts and train end users on how to apply this information to improve resource management decisions
with the goal of reducing drought risk. One part of this might be to develop more competitive research grant programs to fund research on drought prediction. In particular, there is a need for enhanced observations and research on both the paleoclimate record and the drought-related dynamics of ocean-atmosphere coupling. Another idea might be to form a consortium of scientists to encourage collaboration on drought prediction. New funding mechanisms might be needed that explicitly encourage multidisciplinary research bridging the gap between physical and biological science and human needs. Finally, it might be useful to develop a network of scientists and end users to assess the practical needs of end users and how forecast information can be communicated more effectively to the user community to maximize its application.
Assess the economic, social, and environmental impacts associated with drought. Unless we improve our understanding of human behavior, the best intentioned plans will continue to produce less than desired results. The inadequate assessment of drought costs continues to be a significant problem in communicating the importance of drought mitigation to the management and policy communities. More accurate assessments of the true impacts of drought would provide greater justification for investments in mitigation actions at the local, state, and regional levels. Finally, work could be done to improve early assessments of drought impacts through the application of appropriate models (i.e., crop, hydrological).
Assess the science and technology needs for improving drought planning, mitigation, and response at the local, state, tribal, regional, and national levels. To do this, it might be necessary to evaluate current drought planning models available to governments and other authorities for developing drought mitigation plans at the state and local levels of government and require plans to follow proposed standards or guidelines. Efforts could be made to identify improved triggers (i.e., links between climate/water supply indicators/indices and impacts) for the phase-in and phase-out of drought mitigation and response programs and actions during drought events. Work could be done to develop vulnerability profiles for various economic sectors, population groups, and regions and to identify appropriate mitigation actions for reducing vulnerability to drought for critical sectors.
Increase awareness of drought, its impacts, trends in societal vulnerability, and the need for improved drought management. This might include initiating K-12 drought/water awareness programs/curricula or launching public awareness campaigns for adult audiences, directed at water conservation and the wise stewardship of natural resources.
Design more focused and systematic education and outreach programs for stakeholders based on information derived from periodic surveys of their interests. From the results of such surveys, design workshops tailored to the specific interests of different combinations of stakeholders with the objective of producing decision-support tools on a continuing basis.