Global to Regional Perspectives on Intensification of the Hydrologic Cycle: Implications for Extreme Events
T.G. Huntington, U.S. Geological Survey
Climate warming is expected to intensify or accelerate the global hydrologic cycle, resulting in increases in rates of evaporation, evapotranspiration (ET), and precipitation and an increase in the concentration of atmospheric water vapor. The strength of the hydrologic response, or sensitivity of the response for a given amount of warming, is a critical outstanding question in hydroclimatology. An assessment of the published record on observations of trends in various components of the hydrologic cycle and associated variables provides insight into this question. The weight of evidence from global and regional trends in evaporation, ET, and atmospheric water-vapor concentration supports an ongoing intensification of the hydrologic cycle. Global trends in precipitation, runoff, and soil moisture are more uncertain, in part because of high spatial and temporal variability and lack of consistent, high-quality, long-term records. Changes in regional ocean salinity indicate possible increasing evaporation at low latitudes and increasing freshwater inputs (precipitation, runoff, and melting ice) at high latitudes. Ongoing lengthening of the growing season may contribute to increasing ET rates. The evidence for an increase in the frequency, intensity, or duration of extreme weather events like hurricanes and floods is mixed; consequently, regional to global trends remain uncertain.
Understanding Changes in Precipitation and Runoff with a Changing Climate
Kevin E. Trenberth, National Center for Atmospheric Research
The global hydrological cycle and its changes over time are examined in light of observations and current understanding. A particular focus is on how precipitation changes as the climate changes and changes in extremes, including risk of flooding and drought. Net changes in surface evaporation are fairly modest, and a much larger percentage change occurs in the water-holding capacity as atmospheric temperatures increase (7% per C). In this talk we will examine the consequences of this, especially noting the differences over ocean, where water supply is unlimited, and over land. A description will also be given of the understanding of other large-scale changes in patterns and amount of precipitation, soil moisture, and drought. It is important to understand not only changes in mean precipitation, but also the intensity, frequency, duration, and type, and this also applies to the storms that bring precipitation. Understanding these profound consequences of climate change is especially important for water managers.
Is Precipitation Becoming More Intense?
Pavel Groisman, National Oceanic and Atmospheric Administration
An overview of 12-year-long National Climatic Data Center (NCDC) studies of changes of intense precipitation during the period of instrumental observations will be presented with a focus on North America. NCDC has created a database of daily and hourly time series of high scientific quality for use in assessment of changes in precipitation characteristics over the regions where we have sufficient amount of information to answer the question outlined in the talk’s title.
Prior to 2005, NCDC constructed various time series of precipitation characteristics and analyzed their trends. Now (in addition to routine updates of these time series), we have
• analyzed the factors that control intense precipitation (e.g., CAPE and land-falling tropical cyclones trajectories),
• assessed the rainfall distribution characteristics (e.g., hourly rainfall rates), their changes, and their relationships with global and regional surface air temperatures, and
• investigated changes in “direct impact” characteristics of precipitation spectra such as prolonged no-rain periods, fire weather indices, and maximum rainfall intensity.
Our past and ongoing studies (as well as findings by other foreign researchers) embolden our opinion that in the past several decades over most of the extratropics precipitation became more intense. However, the changes in intense precipitation also occur with changes in several other precipitation characteristics and they too deserve our thorough attention.
A Process-Based “Bottom-Up” Approach for Addressing Changing Flood-Climate Relationships
Katie Hirschboeck, University of Arizona
In response to the unprecedented persistence of extreme drought conditions in the western United States, some western water managers have moved beyond conventional approaches to plan for future extreme low flow conditions in innovative ways involving paleo-records, scenarios, and climate projection modeling. In contrast, flood hazard managers are far more constrained in developing ways to incorporate climate change information operationally, in part because of existing flood policy, but also because of the short-term, localized, and weather-based nature of the flooding process itself. What is needed is information that is presented in an operationally useful format for flood managers and that describes how changes in the large-scale climatic drivers of hydrometeorological extremes will affect flooding variability in specific watersheds. This presentation outlines a framework for linking global climatic change to the gauged time series of peak flows in individual watersheds. Using a process-sensitive “bottom up” approach, each individual peak in a gauged record is associated with its flood-producing storm type and circulation pattern. This approach highlights the underlying physical reasons for
The Ghost of Flooding Past, Present, and Future
Harry F. Lins, U.S. Geological Survey
An element of human-enhanced greenhouse theory is that the hydrological cycle will accelerate. This has led to the hypothesis that extreme events, such as floods and droughts, will increase in frequency and/or severity. Published studies indicate that precipitation has increased over the past century, and this increase has been characterized as occurring in “extreme” and “intense” precipitation. However, empirical studies from North America and Europe find no evidence of an increase in flood frequency or magnitude during the 20th century, although increases in low to moderate streamflows have been widely reported. What, then, are the likely effects of an accelerated hydrological cycle on streamflow in general, and on floods in particular? This question is considered using data and the published literature with respect to two issues: What is known about the sensitivity of various return-period floods and annual precipitation? What is the likely impact of a given percentage change in precipitation on a flow quantile (e.g., Q100 versus Qmean)? Results indicate that the precipitation sensitivity of mean streamflow is much greater than that of peak streamflow, and that precipitation sensitivity decreases as flood return period increases. This suggests that human-induced greenhouse warming may be more likely to produce noticeable and significant changes in the mean state of hydrological regimes than in hydrological extremes.
Planning for Non-Stationary Extreme Events: Statistical Approaches
Richard M. Vogel, Tufts University
It is no longer possible to consider streamflow and other hydrologic processes as a stationary process. Nearly all of the methods developed for the planning, management, and operation of water resource systems assume stationarity of hydrologic processes. Non-stationarity can result from a myriad of human influences ranging from agricultural and urban land use modifications, to climate change and water infrastructure. Most previous work in trend detection associated with extreme events has focused on the influence of climate change, alone. This study takes a different approach by exploring flood and low flow trends in watersheds that are subject to a very broad range of anthropogenic influences. We define a decadal flood magnification factor as the ratio of the T-year flood in a decade to the T-year flood today. Using
historical flood data across the entire United States we obtain typical flood magnification factors in excess of 2-5 for many U.S. regions, particularly those regions with higher population densities.
A simple statistical model is developed that can both mimic observed flood trends as well as the frequency of floods in a non-stationary world. This model is used to explore a range of flood planning issues in a non-stationary world. Importantly, non-stationarity in both extreme high and low flows is shown to result from a variety of processes including land use, climate, and water use, with likely interactions among those processes making it very difficult to attribute trends to a particular cause. Multivariate regression models are shown to provide a useful tool for developing the type of conditional forecasts of the moments of extreme events necessary for planning in a non-stationary world.
Planning in a non-stationary and uncertain world is not a new challenge for engineers, because the classic “capacity expansion problem” and other planning problems have always involved both non-stationarity and uncertainty. What is new is the increased variety of sources of uncertainty and non-stationarity that are now inherent in nearly all water resource planning problems, making it essential to incorporate non-stationary planning models of the type discussed here.
Planning for Non-Stationarity and Floods: A Management Perspective
Gerald E. Galloway, University of Maryland
Recent decades have seen a growing increase in flood damages across the nation. A resultant focus on reducing these flood damages has brought long-neglected attention to the systematic assessment and improvement of the quality of existing flood damage reduction structures and pleas for “protection” for areas not now ringed by levees, floodwalls, or other such structural measures. The specter of climate change has led many agencies, both in the United States and abroad, to closely examine how they would deal with more frequent and more severe floods and consider how they might adapt to these future conditions. Flood risk management has replaced flood damage reduction in the lexicon of federal engineers, and considerable effort is now focused on both how they might best manage flood risk and how they might communicate the level of future risk to the public. Given the uncertainties surrounding the calculation of recurrence intervals, how do managers and engineers decide how high their levees should be and how structural measures fit with non-structural actions such as zoning, floodproofing, evacuation, etc.? In 2008, a committee chartered by the Netherlands government recommended to the Parliament that standards for coastal and riverine defense (recurrence intervals) be raised by a factor of 10 to deal with the myriad flood and storm uncertainties faced by that nation. What guidance can be given today to U.S. planners to deal with an uncertain future? They must do something now, but what should this something be?
Mechanisms for Global Warming Impacts on the Large-Scale Atmospheric Branch of the Hydrological Cycle
Richard Seager, Columbia University
It is a robust prediction of state-of-the-art climate models that greenhouse gas-induced global warming will cause the wet regions of the planet (in the deep tropics and the mid to high latitudes) to get wetter while the subtropical dry zones get drier. It is also projected that the subtropical dry zones will expand poleward. Here we analyze the 13 models that made available all the required data to determine the mechanisms responsible for these changes in the hydrological cycle. The mechanisms are divided into first, thermodynamic ones that only rely on a change in specific humidity, second, dynamic ones that only rely on changes in the mean circulation and, third, changes in transient eddy moisture fluxes. Much of the basic pattern of change in precipitation—evaporation (P-E) is accounted for thermodynamically as humidity rises in a warmer atmosphere and intensifies existing patterns of moisture transport. However, changes in circulation are required to explain many changes of P-E in the tropics and, especially, to explain the poleward expansion of the subtropical dry zones. Increases in poleward transient eddy heat moisture fluxes also assist in drying the subtropics and moistening the higher latitudes. Causes of the increased transient eddy fluxes are shown to be complex.
Much of the thermodynamic-induced change in P-E can itself be accounted for simply by atmospheric warming under fixed relative humidity. The mechanisms for projected drying of southwestern North America will be analyzed. This region will dry no matter what, but it is also shown that the character of the tropical Pacific atmosphere-ocean response to increasing greenhouse gases will determine the relative magnitude of the drying. Recent climate change is reviewed for evidence of these changes already occurring, but it is concluded that recent trends have been dominated by large-amplitude natural decadal atmosphere-ocean variability. Near-term hydroclimate prediction therefore must account for both anthropogenic change and the evolution of natural modes of variability.
Connecting Global-Scale Variability to Regional Drought: Mechanisms and Modeling Challenges
Siegfried Schubert, NASA’s Goddard Space Flight Center
Recent research has linked long-term drought (or more specifically extended periods of reduced precipitation) to a number of factors including slowly varying sea surface temperatures (SSTs), the influences of the land surface (e.g., atmosphere/soil moisture feedbacks, aerosols, and vegetation changes), as well as the chance occurrence of extended runs of dry years that can occur even in the absence of any year-to-year memory in the climate system. The possibility of predicting long-term drought rests largely on the strength of the SST linkages to the land component of the hydrological cycle, and of course on our ability to predict the relevant SST changes. The U.S. CLIVAR (Climate Variability and Predictability) working group on drought recently initiated a series of global climate model simulations forced with idealized SST anomaly
patterns, designed to address a number of uncertainties regarding the impact of SST forcing and the role of land-atmosphere feedbacks on regional drought. This talk reviews some of those and related results, with a focus on the U.S. Great Plains, although the basic mechanisms appear to be relevant to drought in many other regions of the world. Issues to be addressed include the seasonality of the global SST response, the impact of soil moisture feedbacks, the potential predictability associated with SST changes, as well as model deficiencies currently limiting our ability to simulate and predict long-term drought.
Do We Need to Put Aquifers into Atmospheric Simulation Models? Evidence for Large Water Table Fluctuations and Groundwater Supported ET under Conditions of Pleistocene and Holocene Climate Change
Mark Person, New Mexico Tech
Aquifer-atmosphere interactions can be important in landscapes where the water table is shallow (<2m) and the watershed topography is gentle. Regional climate models that include aquifer hydrodynamics indicate that between 5 to 20% of evapotranspiration is drawn from the aquifer. The groundwater-supported fraction of evapotranspiration is higher under drought conditions, when evapotranspiration exceeds precipitation. The response time of an aquifer to drought conditions can be long—on the order of 200-500 years—indicating that feedbacks between these two water reservoirs act on disparate timescales. Analysis of Holocene and late Pleistocene paleowater table records suggests that water table fluctuations can be as great as 50 m during drought conditions. With recent advances in the computational power of massively parallel supercomputers, it may soon become possible to incorporate physically based representations of aquifer hydrodynamics into GCM land surface parameterization schemes. This may help to improve our predictions of the long-term consequences of droughts on water resources and climate dynamics.
Breaking the Hydro-Illogical Cycle: The Status of Drought Risk Management in the United States
Mike Hayes, University of Nebraska-Lincoln
This presentation will focus on drought risk management within the United States given the context of climate variability, climate change, and extremes. As the last presentation in the workshop, an attempt will be made to connect comments and issues addressed within previous presentations and breakout groups. A focus will be placed on drought monitoring, impact assessment, mitigation, and planning efforts taking place now across the country, and on suggesting where current efforts need more concentration. The National Integrated Drought Information System (NIDIS) will also be highlighted. Drought fits well into the enhanced efforts by the climate community to create and provide “services” and decision support tools. Each service and tool being designed for drought helps define the “big picture” of drought for policymakers and others who need that scale of information. But they also work to localize drought, putting valuable information in the hands of agricultural producers and community, tribal, and
other grassroots decision-makers—exactly what is needed to boost drought risk management through the rest of this century.