Comparing the effects of different filtering formulas on base flow separation based on daily flow data (Case study: West Rivers of Urmia Lake)

The purpose of this study was to compare the different baseflow extraction methods based on the five hydrometric stations data during 2001-2011 located in the west of Lake Urmia. For this purpose, Fixed Interval Method, Sliding Interval Method, Local Minimum, Lyne and Hollick algorithm with four numerical filters and Eckhardt algorithm with two numerical filters were used. Then, baseflow index values were analyzed by the absolute mean error, flow correlation and root mean squares. The results revealed that the average annual index for all stations in the whole period fluctuated between 74.5 and 89.8. So, this value is indicating the high contribution of subsurface flow in supplying of the river flow in the study region. As a result, the method of Lyne and Hollick algorithm with a numerical filter of 0.9 and, besides it, Eckhart with the filter of 0.9 is the most suitable methods and can be useful in the absence of tracing with chemicals to extract the base flow in the study area. After choosing the best method, the base flow at the Band, Babaroud and Tapik stations were calculated using longterm flow data and dam and agricultural consumptions impact on baseflow also discussed. The annual trend of Band station's baseflow was then evaluated using Mann-Kendall test. The results of the trend analysis indicate that the baseflow of this station has a severe downward trend.