HSPF modelling of the Msunduzi river catchment

Abstract

Hydrological modelling, simulating surface runoff in a catchment, was a relatively new approach in South Africa in the mid seventies. The advent of hydrological modelling on a practical level in South Africa lagged that of other countries by several years at that stage. This delay could be ascribed to the fact that the awareness concerning surface water, draining concepts and hydrological modelling was only then becoming more prevalent in South Africa. Today, following international trends, there is a growing number of hydrological modelling systems for integrated water resource management on a catchment basis in South Africa. Such systems are likely to be installed for operational use in ongoing learning, research, strategic planning and consensus building amongst role players in the Catchment Management Agencies (CMAs). These installed systems are poised to fundamentally change the way modelling is approached in South Africa. Hydrological models are the logical and irreversible response to the enormous forces, which have led to the revision of the National Water Act (NWA), 1998 (Act 36 of 1998) and the water resource management paradigms. The primary objective of this modelling process is the application of the streamflow generation component of the Hydrological Simulation Program - Fortran (HSPF) model in order to provide hydrological information essential to those responsible for planning, development and management of the Msunduzi River Catchment. The focus of this modelling process is on the development and implementation of all input data and the testing of HSPF's continuous modelling system to correctly represent the hydrological components of the hydrological cycle. The Msunduzi River Catchment forms an integral part of the southern portion of the Mgeni River Catchment. By 1985, the Mgeni River Catchment was already supplying water to 3.6 million people, industry and agriculture, which contribute 20% of South Africa's Gross National Product. The margin of error in the simulated annual water balance varied between two and 11% during calibration, while it varied between five and 16% during the period of verification. There is a slight bias in the seasonal calibration for under-estimating the streamflow in the upper sub-catchments during the wet period and over-estimating the baseflow during autumn and follow-on dry periods. In the lower sub-catchments, the wet periods were over-simulated and the baseflow was correctly simulated during the dry period. The response time and interflow recession rates of the hydrographs were accurately simulated, except for some cases during autumn where the interflow recession rate was incorrectly simulated. The over-simulation of the various single storm events (higher streamflow peaks) was the most consistent error that occurred during the period of calibration. This can probably be ascribed to the poor representation and areal distribution of precipitation data, which did not account accurately for the spatial variation in precipitation- and storm distributions. Hydrological simulations are a prerequisite for further expansion of the model for water quality simulations. Therefore an accurate hydrological simulation forms an essential first step in developing a full HSPF application for a catchment. The fact that the Msunduzi River Catchment constitutes a substantial portion of the Mgeni River Catchment emphasises the importance of this project's results to help manage the hydrology and water quality of the larger Mgeni River Catchment of approximately 4000 km2 . HSPF can also be used to develop model parameters to test scenarios of future development within the catchment such as land-use changes, which may affect runoff and streamflow. HSPF, together with other installed modelling systems, will enable CMAs to address the contradictory calls for affordable modelling at the same time as recognising greatly increased complexity, scalability, accessibility and integration. The successful implementation of the HSPF model will address the timeliness issues of today's fast moving world in which contentious water resource issues are becoming much more commonplace.