Detection of trends in low streamflows

A study of trends and variability of low flow characteristics was conducted for the Reference Hydrometric Basin Network (RHBN, established in the mid 1990s, for detection, monitoring, and assessment of climate change in Canada). RHBN stations are characterized by either pristine or stable hydrological conditions, with 20 or more years of good quality record. The 7-, 14-, 21-, and 30-day low-flow indices of RHBN hydrometric stations from the beginning of their records to 2003 were extracted using a moving average procedure, and examined to detect trends in Canadian lowstream flows for different record lengths. Also, the time of occurrence of 7-day low flow index was extracted and tested to detect changes in the timing of summer and winter 7-day low-flows for different record lengths. A modified Mann-Kendall (MK) nonparametric trend test was applied to the time series at a 0.05 significance level. The variance of the S statistic was modified if the absolute value of serial correlation was significant at a 0.1 significance level.;Statistical analysis indicated that regardless of record length, the percentages of significant trends for different low flow indices are in a good agreement. The number of significant trends in the quantity of low flows amounts to about 33% of the total tested stations. In 65% of stations with significant trends, low-flows experienced downward trends. On the other hand, low flows experienced a significant increase in 11% of tested stations. Numerical analysis indicated that northern Canada (stations located above latitude 60°N) has undergone a significant increasing trend in low-flows. A significant downward trend dominated the Atlantic Provinces and southern British Columbia. No evidence of significant trends in the Prairies and eastern Ontario was found.;It was also found that there is a positive relationship between record length and the percentage of detected significant trends in low flows, i.e. the percentage of significant downward trends increased as study period increased whereas the percentage of significant upward trends decreased as study period increased.;Numerical analysis showed that the timing of summer 7-day low flow in 15% of stations experienced significant trends where in 52% of tested stations it shifted toward earlier dates. As record length increased the number of detected significant upward trends decreased. It was found that in the east half of the country (Atlantic Provinces and southern Ontario) summer 7-day low-flow shifted to arrive earlier whereas in the west and northwest (British Columbia, Yukon Territories, and North West Territory) it shifted to arrive later. There were no significant trends in timing of summer low flow in central Canada for longer record length. However, regardless of record length, timing of summer low flows showed a shift toward earlier dates in the higher latitudes while it showed a shift toward later dates in lower latitudes of central Canada.;It was found that although the majority of the low flow time series were dominated by slight positive autocorrelations, the timing of winter low-flows was dominated by negative autocorrelation. It was also observed that if the autocorrelation coefficient was insignificant, its impact on acceptance or rejection of the null hypothesis was almost insignificant as well.;It was observed that the timing of winter 7-day low flow in 20% of stations experienced significant trends. This percentage of significant trends was greater compared to the summer period. Furthermore, in 74% of the detected significant trends, the date of occurrence of winter 7-day low-flow shifted toward earlier dates. The results of this study indicate that trends in timing of both summer and winter 7-day low-flow experienced a shift toward earlier dates in eastern Canada whereas in western Canada winter trends were downward while summer trends were upward. There was no evidence of significant trends in Central Canada.