Development of a catchment risk assessment procedure for the protection of potable water supply intakes

Water supply companies operating within the European Union (EU) have a statutory duty to provide water that is wholesome to drink. To ensure continued compliance with EU water quality regulations together with reductions in operating costs it is necessary to optimise the use of raw water supplies. In order to do this successfully, a knowledge of the potential risks to water quality within a supply catchment is required. This thesis describes the development of a novel methodology to assess pollution risks in public water supply catchments. The Risk Assessment of Supply Catchments And Land (RASCAL) methodology consists of four interrelated stages to identify, estimate, evaluate and manage risks to water quality. The methodology is generic in nature thus allowing the transfer of the approaches developed to other catchment areas.

The potential of geographical information systems (GIS) as a tool assessing pollution risks in supply catchments is identified from an extensive review of water quality hazards and the use of GIS in the UK water industry. This potential is explored further through the example application of the RASCAL methodology to the Upper Wharfe catchment area.

The thesis describes an important development in assessing catchment risks, the construction of catchment-scale hazard maps using logical structures (Source Trees) and GIS data models. In this way, the probable extent of point and diffuse sources of pollution are identified for the entire catchment area of a surface water intake. The use of generically available digital data allows the production of hazard maps for ten water quality issues potentially affecting the Upper Wharfe source. These digital hazard maps are also used in the development of a raw water sampling programme. The results from this programme confirm the ability of the digital maps to identify the source of water quality hazards using generic data.

Quantitative approaches relying on the acquisition of large amounts of historical or background data are found to be inherently inappropriate for a generic methodology. The availability of water quality data is identified as a fundamental barrier to the successful preparation of such quantitative frequency estimates. This research combines quantitative and qualitative approaches to provide a valid methodology of risk assessment for generic application.

A fundamental characteristic of the methodology developed is this mixture of quantitative, semi- quantitative and qualitative techniques encompassed within it. This represents a significant contrast to the traditional hard-engineering approaches to risk assessment applied in past studies where the emphasis has been on the quantitative estimation of risk. The use of digital data and GIS allow the most suitable technique to be identified and displayed.