| 英文摘要: | Heavy precipitation has increased worldwide, but the effect of this on flood magnitude has been difficult to pinpoint. An alternative approach to analysing records shows that, in the central United States, floods have become more frequent but not larger. There is a widely held perception that floods are increasing around the planet. Precipitation data show significant increases in the frequency and magnitude of heavy precipitation events in many areas1, and economic damage due to floods is on the rise2. Yet most analyses of flood trends do not conclusively show that the size of floods is increasing3. This apparent contradiction has been widely noted4,5, and the Fifth Assessment Report of the IPCC concluded that “... there continues to be a lack of evidence and thus low confidence regarding the sign of trend in the magnitude and/or frequency of floods on a global scale”6. Writing in Nature Climate Change, Mallakpour and Villarini7 describe an approach to detecting flood trends that may help resolve this puzzle, and use it to demonstrate widespread increases in the frequency — but not the size — of floods across parts of the central United States. Floods have the potential to cause economic damages and loss of life: the Great Flood of 1993, for example, caused $US15 billion in damages in the Mississippi River Basin8. But floods also replenish reservoirs and enhance ecosystem health and are major drivers of sediment and nutrient transport. Quantifying the changing magnitude and frequency of floods is crucial to optimizing our response to them, including mapping flood hazard zones, setting flood insurance rates, designing bridges and other infrastructure such as water and wastewater treatment plants, and protecting and restoring ecosystems. Warmer air can hold more moisture, so atmospheric warming is anticipated to increase heavy precipitation events and affect flood regimes9. To understand and quantify changes in flooding, hydrologists use historical observations from long-term stream gauges, which continuously measure river flow. Most analyses of flood records focus on what hydrologists term the annual peak discharge record. For example, in a 50-year record of streamflow observations, the annual peak discharge record is made up of the highest instantaneous value of discharge in each year. In a drought year, this annual peak may be so small that no one would consider calling it a flood, but it is included in the study of flood frequency and used in the evaluation of flood trends. Mallakpour and Villarini use a different approach, termed a peaks-over-threshold (POT) record, to develop a flood record from which to evaluate trends. The resulting time series registers all events in which discharge exceeded a selected high value (see Fig. 1 as an example), thus focusing only on flows that could actually be considered floods. A POT record permits analysis of changes in the frequency of flood events, which can't be done with the annual peak discharge record.
|
