Sensitivity of optimal baseflow filter parameter to catchment soil characteristics

Abstract

Baseflow estimation is an important problem in both low flow and flood hydrology and is commonly carried out using recursive filters. However, this approach is based solely on the observed streamflow hydrograph without consideration of physical catchment characteristics, such as soil properties, topography, vegetation and climate, etc. This paper investigates the performance of the commonly used Lyne and Hollick filter in baseflow separation for catchments with sandy soils with different properties. This is done by comparing the baseflow hydrographs obtained using the Lyne and Hollick filter with simulated baseflow from a 3D hypothetical catchment using the fully-integrated HydroGeoSphere model. In addition, the sensitivity of filter performance to different values of the Lyne and Hollick filter parameter is investigated. The results obtained indicate that the performance of the Lyne and Hollick filter varies significantly with soil characteristics and can be improved considerably by adjusting the Lyne and Hollick filter parameter based on soil characteristics.