Doctor of Philosophy
School of Civil, Mining and Environmental engineering - Faculty of Engineering
Bodhinayake, Nanayakkara Dayananda, Influence of hydrological, geomorphological and climatological characteristics of natural catchments on lag parameters, PhD thesis, School of Civil, Mining and Environmental engineering, University of Wollongong, 2004. http://ro.uow.edu.au/theses/385
Catchment lag time is considered as a key factor in flood hydrograph modelling and design. The extensive literature investigation of this study revealed that most of the lag time equations that have been developed include various hydrological, geomorphological and climatological characteristics of the catchment. However, different studies use different combinations of these variables, and therefore, the appropriate context of the relation is not known with confidence. The intention of this research is to determine to what extent the above mentioned catchment characteristics influence the lag parameter, which is directly related to the catchment’s lag time. In order to assess the influence of catchment characteristics on the lag parameter, reliable and valid rainfall and flow data must be analysed. Therefore, at the outset of this research, the reliability and validity of rainfall data of seventeen rural catchments in Queensland, Australia, were examined. These catchments belong to five river basins and they are, Mary, Haughton, Herbert, Don and Johnstone. A total of 254 storm events on these catchments were analysed. To compute the lag parameters of the catchments, the computer based Watershed Bounded Network Model (WBNM) was selected due to its in-built non-linearity property as well as other capabilities. These include the ability to model spatially varying rainfall, the simplicity of data files and the requirement of a minimum amount of data. The constantslope method was adopted to separate the base flow from the recorded total hydrograph in order to derive the ordinates of the surface runoff hydrograph, which is one of the essential components for the input file of WBNM. The time variation of the rainfall was examined by means of mass curves of rainfall and the spatial variability of the rainfall was studied with the help of isohyetal plots. Thereafter the rainfall and flow data, as well as the physical features of the catchments, were incorporated into WBNM to generate hydrographs for all 254 storm events. The lag parameter was altered until WBNM generated a hydrograph that closely resembled the recorded surface runoff hydrograph. This process was repeated for each storm event to obtain its lag parameter value. From this method, lag parameter values were derived for all 254 storm events on the seventeen catchments. The next stage of the analysis involved testing to determine whether the lag parameter is related to a range of hydrological, geomorphological and climatological variables. To carry out the analysis the necessary hydrological characteristics were extracted from the storm data. Other useful geomorphological and climatological characteristics were obtained from AUSLIG maps and the Bureau of Meteorology. If the lag relations built into WBNM are sufficient to account for those variables, then no significant relation between the lag parameter and those variables should exist when the lag parameter is plotted against each variable. The lag parameter (C) versus a range of hydrological, geomorphological and climatological characteristics of all seventeen catchments were plotted to examine their correlation. Two tailed statistical t-tests were carried out for each plot to find out whether the gradients of best-fit straight lines of those plots are significantly different from zero at 5% level of significance. The results of this research have shown that there are no strong relationships between the lag parameter (C) and the range of catchment characteristics selected for this study. Therefore, the lag parameter can be considered as an independent factor applying to a wide range of catchments. While this research was carried out for the WBNM model, its essential findings, that the non-linearity power is near to 0.23, and that the dominant variable influencing catchment lag time is the catchment area, also apply to other flood hydrograph models.