Possible Regional Probability Distribution Type of Canadian Annual Streamflow by L-moments

For effective planning, design, and management of water resources engineering, the probability distribution of annual streamflow is necessary. The method of L-moments is applied to identify the probability distribution type of annual streamflow in different climatic regions of Canada. In the Pacific...

Full description

Bibliographic Details
Main Authors: Sheng Yue, Chun Wang
Format: Article in Journal/Newspaper
Language:unknown
Subjects:
Online Access:http://hdl.handle.net/10.1023/B:WARM.0000049145.37577.87
Description
Summary:For effective planning, design, and management of water resources engineering, the probability distribution of annual streamflow is necessary. The method of L-moments is applied to identify the probability distribution type of annual streamflow in different climatic regions of Canada. In the Pacific and southern British Columbia mountains (regions 1 and 2), the generalized extreme value (GEV) distribution fits the observations best with the 3-parameter lognormal (LN3) and log Pearson type III (LP3) as potential candidates. In Yukon and northern British Columbia (region 3), the LN3 distribution corresponds to observations best with the LP3 and P3 as potential candidates. In the northwestern forest (region 5), the LP3 distribution matches observations best with the P3 and GEV as potential candidates. In Arctic Tundra (region 10), the 3-parameter Weibull (W3) is the best one with the LN3 and P3 as potential candidates. The P3 distribution provides a best-fit to observations in the Prairies (region 4), northeastern forest (region 6), great Lakes and St. Lawrence (region 7), Atlantic (region 8), and Mackenzie (region 9) with the LN3, LP3, and GEV as potential candidates. Copyright Kluwer Academic Publishers 2004 annual streamflow, discordancy, L-moments, regional frequency analysis, regional hydrology, statistical hydrology