The workshop presentations and discussions spanned an array of issues that coalesced around several major topics and research needs. These were identified primarily in the breakout sessions, based on discussions with speakers and workshop participants. In the committee’s view, several key findings that address three major categories emerged from the workshop deliberations. The first category focuses on the current state-of-the-art in observation of atmospheric dynamics and the propagation of these dynamics into the land-based hydrologic realm. The second category assesses the status of translating the scientific knowledge base across the climate and hydrologic science communities. The third category identifies opportunities for progress to better unite the scientific perspectives and increase their usefulness in the water resource planning and management arenas.
Although its focus is on the United States, this report necessarily considers a large body of research on the global climate and water system. Research at the global scale is relevant here in terms of its contributions to our general understanding of climate dynamics. But the global perspective also becomes relevant by providing a context for U.S. hydrologic extremes.
Characterizing the Conventional Wisdom
Observational evidence shows that the nation’s hydrology is changing with respect to water cycle variables, yet uncertainties in the sources and characterization of this change persist.
One way by which scientists measure the presence of change in extreme hydrologic events is to represent the events through statistical distributions, from which they can assess changes in, for example, mean or median values and percentiles. Workshop participants noted that the weight of observational evidence shows an ongoing acceleration of the water cycle as the climate warms, with a broad spectrum of atmospheric and land surface variables associated with these changes (also see Huntington, 2006, 2010). Recent assessment of a broad array of water cycle variables for the United States corroborates this finding (Karl et al., 2009) (Figure1).
For the United States, changes in the upper percentiles of the precipitation distributions indicate that much of the nation has become generally wetter over the past century (IPCC, 2001 and others). These changes are manifested as increases in, for example, the number of days per year with precipitation exceeding fixed thresholds, such as ~50 mm per day (Karl and Knight,
Figure 1 Century-scale changes in a broad array of water cycle variables contribute to the scientific evidence for a detectable greenhouse warming signal. Taken together, these variables indicate acceleration of the hydrologic cycle and the non-uniform spatial distribution of these changes. SOURCE: From the U.S. Global Change Research Program’s national assessment of climate impacts. Reprinted, with the permission of Cambridge University Press, from Karl et al. (2009). © 2009 by University Corporation of Atmospheric Research.
1998). But for the associated hydrologic variables, results are mixed using standard hydrologic measures. Analysis of flood occurrence (i.e., the annual maxima series) shows essentially no trends at a set of U.S. Geological Survey (USGS) stream gages that were carefully selected to minimize any influences of water management (USGS, 2005). This phenomenon was also noted at a 2008 workshop hosted by the COHS, both by speakers citing the same USGS report as well as by participants citing other research (NRC, 2008). Yet, evidence of changes in U.S. drought characteristics is mixed. Across much of the eastern and central United States, trends in increasing precipitation appear to have resulted in reductions in drought severity and length. In contrast, in parts of the West a general warming appears to have increased evaporative demand more rapidly than precipitation, with the result that these areas tend toward more, longer, and more severe droughts (Groisman et al., 2004). These results point again to difficulties in interpreting climate- and weather-oriented extremes in a hydrologic context.
The breakout discussions noted that major uncertainties have presented themselves but have yet to be reconciled. Why have continental U.S. streamflow changes over the past 50-60
years been evidenced primarily in low flows and not in flood flows? Are there inconsistencies in observed precipitation extremes and streamflow extremes, and if so, why? What are the causes of pervasive increases in occurrence of low flows, for example, in the upper Midwest? To what extent is land cover, as contrasted with climate change, the cause of observed streamflow changes?
Assumptions on the occurrence of major hydrologic events to analyze extremes are based on the notion of stationarity, yet observational evidence increasingly shows that this assumption is untenable.
Stationarity represents the idea that hydrologic systems fluctuate in an unchanging envelope of variability (i.e., the mean and the degree of variability of hydrologic time series do not change over time). Water management systems have been traditionally designed based on this assumption. Therefore, it is critical to the protection of life and property to understand if and how these assumptions are being violated (Milly et al., 2008). From a scientific standpoint, fluctuations in stage heights and flood flows over the historical past constitute a natural experiment, with particular realizations that have in some cases been unexpected and changing over time. A good example is the American River in California, where over the past ~100 years the 5 largest three-day peak flood volumes all occurred in the second half of the record, as had 10 of the largest 13 (see also NRC, 1999). In other words, this hydrologic system no longer operates within its expected unchanging envelope of variability. The statistical distribution of flood volumes that represent this system has become non-stationary.
Bulletin 17B of the Interagency Advisory Committee on Water Data (IACWD, 1982), titled Guidelines for Determining Flood Flow Frequency, details a set of data-based methods that allows one to define flood potential. This document is the current standard in the United States, but it has not been updated since the early 1980s. The workshop participants broadly agreed that although Bulletin 17B was concerned about non-stationarity, a remedy is not well addressed in the document. Because the available evidence (at that time) indicated that major climate-induced changes occur on the scale of thousands of years, Bulletin 17B assumed that floods are unaffected by the shorter-term changes that have been documented in the context of anthropogenically induced climate change. Participants discussed the United States’ unmet need for new flood-frequency guidelines that draw on advances in hydrologic and climate science over the past 25 years, an observation that is supported by presentations and agreement at a previous COHS workshop (NRC, 2008). Regular revision of the Bulletin 17B guidelines as modeling and understanding of relevant phenomena improves would also be valuable. Regardless, continuing to use the assumption of stationarity in designing water management systems is no longer practical or defensible.
How hydrologic extremes are intertwined with other anthropogenic effects is poorly understood.
The nature of climate-induced hydrologic extremes is currently confounded by other, engineering-based, anthropogenic effects. Interacting anthropogenic factors such as land cover change and the operation of water engineering facilities including dams, irrigation works, wells drawing on and drawing down aquifers, and interbasin transfers define the 21st century landscape and hydrologic dynamics of the continents. Additional confounding arises from the changing nature of climate variability, including the trajectories and patterns of storm tracks. Consequently, exploration of the climate extremes issue requires a better understanding of all of these factors and how they interact. Such an understanding would greatly benefit the applications community, which requires highly region-specific and, in many cases, site-specific information. This is precisely the scale at which the current state of the art in climate modeling is least robust and least certain—and would create opportunities to harmonize traditional hydrologic field research carried out on more local domains with next-generation, high-resolution atmospheric modeling.
Translating the Science of Hydrologic Extremes to the Policy and Management Sectors
Management and mission-oriented agencies with public-sector responsibilities have been provided with marginally useful scientific information about the likely manifestations of future climate change.
Future increases in flood extremes have often been inferred by climate modelers from extreme rainfall projections generalized in many cases within a global context. Yet, according to workshop participants, floods of interest to the user community occur locally in a magnitude and frequency context that is not the same as that implied by global models. It was noted during breakout sessions that the long time horizons, substantial uncertainty bounds, and relatively coarse-scale spatial resolutions (despite progress in downscaling) limit the usefulness of global climate model output for most hydrologic applications (e.g., water resources planning, floodplain management; see also, NRC, 2008). Although research from the climate modeling community enhances understanding of atmospheric dynamics, it generates outputs that are not directly comparable to hydrologic extremes and are even less applicable to operational needs (e.g., water regulation at dams, designing flood protection infrastructure) (NRC, 2007a).
Furthermore, smaller-scale regional climate models are not yet sophisticated enough to add significant value to this endeavor. Higher resolution regional climate models are now available that better resolve the effects of topography and may provide better estimates of precipitation extremes in areas where floods are mostly associated with large-scale storms (e.g., the western United States). However, the ability of these models to reproduce observed extremes remains to be demonstrated, and their resolution is insufficient to resolve the processes that control extreme precipitation in warm seasons, which dominate most of the United States outside of the West. In addition, distinguishing signal versus noise is a challenging issue in the prediction
and assessment of hydrologic extremes. This challenge is exacerbated for regional climate models because variability from daily to interannual timescales is greater for small regions. Therefore, articulating regional results at a long enough timescale to tease out signal versus noise is difficult. Often this requires a time-frame that exceeds the infrastructure’s useful life. Planning and operations for water management and design of new projects require high-quality information in a site-specific context that the current generation of climate models cannot yet deliver to respond to realities dictated by regulation, current public policy, and other factors.
There are insufficient interactions and knowledge exchange between climate scientists, water scientists, and engineers and practitioners to solve these challenges.
This contention is well illustrated by the use of terminology in the different communities involved with climatic and hydrologic extremes. For example, terms such as “extreme flood” and “change in extremes” are applied differently by climate modelers, agency hydrologists, and academic hydrologists. In a hydrologic context, “extremes” are usually connected with risk analysis, and at least indirectly to infrastructure design criteria. For instance, the “100-year floodplain” is widely used in land use planning, and the design discharge for culverts crossed by certain classes of roads is the “50-year event.” Hydrologists usually consider events of this general magnitude “extreme.” On the other hand, the climate literature often describes much more frequent events as “extreme.” These varied metrics could lead to miscommunication regarding the degree and location of the exposure of the nation’s infrastructure to flood risk (see also below).
If the scientific and practitioner communities can communicate and plan for extreme events now, then the results of their work will provide for improved preparation for events in the future. Close cooperation between these communities to better assist each other is critical. A common vocabulary or understanding of the meaning behind various types of “extreme events” would facilitate collaboration. In the absence of a common language, the different uses of important terms should be clearly defined and accepted, and each community should be more flexible and adaptable with respect to how the other uses the terms.
Risk to the nation’s infrastructure from water-related extremes depends on not only the changing probabilities of climatic means or extremes but also the exposure of assets to these extremes and their value in economic terms.
Risk is generally defined as the probability of hazard occurrence multiplied by some measure of hazard consequence or capacity to be harmed given a particular level of hazard (i.e., the vulnerability) (NRC, 2010a). In the past, considerable emphasis was placed on the probability estimates for a particular hazardous process. This is illustrated by the traditional civil engineering approach that uses probabilistic entities (e.g., the “100-year” flood) as measures of the hazard. Flood and drought vulnerabilities are a consequence of human planning and actions. Humans have a propensity to settle in hydrologically dangerous settings, such as floodplains or drought-prone arid and semi-arid regions. Less emphasis has generally been placed on developing well-defined measures of vulnerability, which in this context is the varying level of
susceptibility to and inability to cope with losses given exposure to extreme events of varying intensity and frequency (IPCC, 2007b). Vulnerability depends, in part, on social factors, and it continually changes in response to factors that differ from those measured in probabilistic analyses. Thus, the construction of dams and levees may decrease the probabilistic aspects of risk to “protected areas.” Yet vulnerability may thereby be increased when an associated false sense of security results in the construction of more infrastructure in the at-risk zones.
Although one of the primary goals of research on extreme events such as floods and droughts is risk reduction, the current emphasis of global climate science using models and observations addresses the probabilistic hazard component of risk (NRC, 2010a). Efforts aimed more at the vulnerability component of risk (including placement of infrastructure) are much less well developed. Examples of agency interests include risk-based design, changes to the probable maximum flood or in the distributions employed in flood-frequency analysis, revisions of economic procedures for evaluating future projects, increasing of the skill for intermediate-term (6 to 60 days) projections, and the incorporation of more real-world data (not just model predictions). This latter issue specifically includes paleohydrological and paleoclimatological data to provide broad historical context for hydrologic variability.
A coherent set of research objectives has yet to be fully developed that simultaneously addresses the science needed to better understand the hydrologic events under climate change and their social dimensions and to provide policy that mitigates the consequences of these events. Until these definitive links are established, the public discourse will remain unclear.
A Way Forward
Hydrologists occupy a useful “nexus” between climate change scientists and practitioners, promoting the translation of critical research findings into better informed planning and applications.
Hydrology evolved in the service of water resource management with roots deep within civil engineering (NRC, 1991). Because emphasis has recently shifted toward addressing water-related scientific uncertainties rather than toward planning and management, strong linkages between the hydrologic sciences and the water management community are currently eroding (Loucks, 2007). Frustration abounds, as was observed during the workshop, when the climate science community provides research results that are not in a form that is easily translatable into management decisions (see also NRC, 2007a). Building better interactions between all relevant communities is necessary, and hydrologists serve a central and essential role in these interactions. Hydrologists can fill a critical niche at the interface between the climate science and engineering applications communities by translating research on climate extremes for the applications community.
One key issue to resolve is how the nature of extremes in the atmospheric phase of the hydrologic cycle translates into extreme hydrologic events. In the Southwest, where floods are generated by multiple mechanisms such as the El Niño-Southern Oscillation and snowmelt, understanding of the sources of floods has been advanced by analyses of the atmospheric
conditions that accompanied observed hydrologic extremes (e.g., blocking, fronts, position of jet streams, sea surface temperature, storm tracks). Explicit analysis of the linkages between observed hydrologic data and atmospheric conditions and phenomena can help to develop a basis for flood risk estimation in a changing climate (NRC, 2007b) and to catalyze an interaction between the atmospheric and hydrologic science communities.
Another important topic addressed at the workshop was the translation of the science regarding hydrologic extremes into agency-relevant and policy-actionable knowledge. It was noted during a breakout session that the Harvard Water Program of the 1960s provided an important model for both cross-disciplinary and applications-oriented hydrology (Reuss, 2003; Lettenmaier, 2008; Milly et al., 2008). Thus, precedent exists for bridging the divide between hydrologic research and the operational community. To accommodate water planning and management in the context of an accelerating hydrologic cycle, workshop participants discussed a modern-era version of this program that would emphasize decision-based frameworks that incorporate risk and uncertainty from climate variability as well as other aspects of hydrologic change such as land management and hydraulic engineering impacts, including their associated uncertainties. Results from new environmental surveillance technologies that detect changes in extremes—for example, satellite remote sensing of groundwater fluctuations to micro-sensor arrays for soil moisture, data assimilation, econometric and coupled water resource system decision support tools—will arm policy-makers and managers with up-to-date monitoring capabilities and thus will better inform their decision-making processes. However, caution should be employed to select the more robust of these new technologies for operational needs.
Interim, practical approaches to cope with the possibility of elevated risks of climate and hydrologic extremes on infrastructure will remain essential as the scientific basis for improved methods to analyze the sources and consequences of such extremes continue to evolve.
Workshop participants articulated that irrespective of the status of all these factors—from research aimed at disentangling climate from land-based contributions to extreme events to a lack of robust statistical procedures to define risk—the nation will continue to make major investments in civil and private infrastructure. The design of water-related engineering facilities, such as dams and levees, and ecosystem restoration projects has obvious links to the climate extremes question. The specter of climate change is leading U.S. agencies to contemplate how they will deal with hydrologic extremes, with or without scientific certainty. But life, economic security, and property will all be placed at risk should there be more frequent and severe droughts and floods.
From a planning standpoint, increasingly uncertain flood or drought frequencies cause major problems with projects having long lifespans (e.g., sewers, levees, and dams). Infrastructure with design lifetimes on the order of 50 or more years operate on a timescale over which there is less certainty about various climate change scenarios, modeling results, and thus hydrologic stocks and fluxes. In such cases, measures of robustness of alternative future scenarios together with economic criteria are useful because design decisions, once implemented, are not easily changed over time (NRC, 2010a), which might necessitate larger safety margins. The City of New York has developed one of the most comprehensive approaches
to climate change adaptation to date (NRC, 2010a). Workshop participants noted that infrastructure can be constructed in adaptable modular units, with shorter expected longevities (or design revisit times) on the order of 10-20 years, which in practical terms reflects the non-stationarity concept. In general, insights drawn from historical and paleohydrologic records, plus practical assumptions, can provide a basis for establishing margins-of-error on designing infrastructure or repositioning existing assets (e.g., to higher ground). Such strategies have been used to avoid future urban flood damage in the Mississippi River basin (Interagency Floodplain Management Review Committee, 1994; NRC, 2009b; USACE, 2011).
Basic monitoring of key elements of the hydrologic cycle is essential to support analysis of hydrologic extremes with any confidence.
Addressing basic questions on the hydrology of extremes will require continuing commitments to monitoring networks and routine observations, through climate, weather, and hydrologic monitoring networks and their integration (see also, NRC, 2010b). Although the United States has an enviable record of hydrologic measurement, its hydrologic networks have become increasingly fragmented (NRC, 2009a). Absent firm commitments to retain observational networks by federal, state, and municipal agencies, estimation of the risk of hydrologic extremes will be compromised, as will the ability to prepare, adapt, and mitigate the impacts of these extremes as climate conditions change. As one example, the USGS stream gaging network has monitored flow in the nation’s rivers since 1889 (http://water.usgs.gov/nsip/history1.html), yet it and other USGS water monitoring programs are under continuing pressure to provide data with decreasing resources (NRC, 2009a).
Decisions about the design of hydrologic monitoring networks are increasingly confounded by hydrologic non-stationarity, not only because of a changing climate but also because of anthropogenic land cover change and water management effects. Procedures developed in the 1960s and 1970s to determine when stations can be discontinued are based on stationary statistical assumptions that are no longer defensible. These protocols therefore lead to management decisions regarding monitoring stations that are diametrically opposed to those that are applicable to a non-stationary world. The next generation of observational networks may include paleoclimatic and paleohydrologic data sets as well as newer technologies, such as precipitation radars, to augment basic data from stream gage networks.
The COHS-hosted workshop raised many questions and challenges in terms of characterizing hydrologic extremes, translating scientific knowledge to the policy and management communities, and identifying a productive future role for hydrologic sciences. Workshop participants confirmed the research findings that show that the water cycle is changing and indeed accelerating, but noted the many unknowns that remain with respect to the drivers of this change, the system response, and the implications for society. The issues discussed at the workshop are fundamental to the nation’s environmental security as it embarks on a major
climate adaptation strategy. The workshop was admittedly a modest step forward in uniting the perspectives of a broad research community in climate and hydrology as well as planners and engineers. Creating a more productive dialogue is essential in order to safely and efficiently deploy our 21st century strategic water investments.