Trends in snowmelt-related streamflow timing in the conterminous United States

JOURNAL OF HYDROLOGY

卷:547

Article

Dudley, R. W.¹  Hodgkins, G. A.¹  McHale, M. R.²  Kolian, M. J.³  Renard, B.⁴ 

[1] US Geol Survey, New England Water Sci Ctr, 196 Whitten Rd, Augusta, ME 04330 USA

[2] US Geol Survey, New York Water Sci Ctr, 425 Jordan Rd, Troy, NY 12180 USA

[3] US EPA, Off Atmospher Programs, 1200 Penn Ave NW, Washington, DC 20460 USA

[4] Irstea, Hydrol Hydraul, Lyon, France

关键词: Snowmelt;    Streamflow;    Trends;    Climate;    Hydrology;   

DOI  :  10.1016/j.jhydrol.2017.01.051

来源: Elsevier

【 摘 要 】

Changes in snowmelt-related streamflow timing have implications for water availability and use as well as ecologically relevant shifts in streamflow. Historical trends in snowmelt-related streamflow timing (winter-spring center volume date, WSCVD) were computed for minimally disturbed river basins in the conterminous United States. WSCVD was computed by summing daily streamflow for a seasonal window then calculating the day that half of the seasonal volume had flowed past the gage. We used basins where at least 30 percent of annual precipitation was received as snow, and streamflow data were restricted to regionally based winter-spring periods to focus the analyses on snowmelt-related stream flow. Trends over time in WSCVD at gages in the eastern U.S. were relatively homogenous in magnitude and direction and statistically significant; median WSCVD was earlier by 8.2 days (1.1 days/decade) and 8.6 days (1.6 days/decade) for 1940-2014 and 1960-2014 periods respectively. Fewer trends in the West were significant though most trends indicated earlier WSCVD over time. Trends at low-to-mid elevation (<1600 m) basins in the West, predominantly located in the Northwest, had median earlier WSCVD by 6.8 days (1940-2014, 0.9 days/decade) and 3.4 days (1960-2014, 0.6 days/decade). Streamflow timing at high-elevation (>= 1600 m) basins in the West had median earlier WSCVD by 4.0 days (1940-2014, 0.5 days/decade) and 5.2 days (1960-2014, 0.9 days/decade). Trends toward earlier WSCVD in the Northwest were not statistically significant, differing from previous studies that observed many large and (or) significant trends in this region. Much of this difference is likely due to the sensitivity of trend tests to the time period being tested, as well as differences in the streamflow timing metrics used among the studies. Mean February-May air temperature was significantly correlated with WSCVD at 100 percent of the study gages (field significant, p < 0.0001), demonstrating the sensitivity of WSCVD to air temperature across snowmelt dominated basins in the U.S. WSCVD in high elevation basins in the West, however, was related to both air temperature and precipitation yielding earlier snowmelt-related streamflow timing under warmer and drier conditions. Published by Elsevier B.V.

【 授权许可】

Free