Comparison of the annual minimum flow and the deficit below threshold approaches: case study for the province of New Brunswick, Canada

Abstract

A low-flow frequency analysis using the annual minimum flow (AMF) and the deficit below threshold (DBT) approaches was carried out for 30 hydrometric stations across the province of New Brunswick. The AMF method considers only the annual minimum events, and the DBT method considers all low flows below a certain threshold (or truncation level). In the present study, the DBT method characterizes low flow in terms of deficit intensity, which is becoming increasingly important in both water and aquatic resources management. The annual minimum series were fitted using the three-parameter Weibull distribution (AMF–WEI3), and the intensity data series were fitted using the two-parameter Weibull distribution (DBT–WEI2) and the generalized Pareto distribution (DBT–GP). All parameter estimates were obtained using the maximum likelihood method. The AMF–WEI3 and DBT–GP approaches provided a good fit to at-site data in terms of annual minimum flow and deficit intensity, respectively. However, the present study showed that the DBT–WEI2 approach underestimated the right tail of low-flow intensity. The Anderson–Darling statistic was also calculated for the data series and can provide insight into which distribution may adequately model the low-flow characteristics. A regionalization study was also performed using the AMF–WEI3 and DBT–GP methods.