flood variations in specific watersheds, defines how mixed flood distributions and outlier events may be linked to climate shifts, and challenges the underlying “iid” assumption that flood peaks are independently and identically distributed. Linking extreme flood events to meteorological causes driven by shifting circulation features can provide water managers with critical climate-based interpretive information for how flood probability distributions are likely to respond within individual watersheds under future climate change scenarios.

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



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement