Remote sensing of bathing water quality

The European Union (EU) has openly solicited advice on the development of EU bathing water quality policy and made calls for the development of remotely sensed operational real world solutions. This research demonstrates a new approach to estimating water quality using remote sensing and specifically to monitoring bathing water quality by using remote sensing to "flag" failing areas for manual survey. This method meets the environmental demands of the EU, the tourist industry, the water industry and environmental monitoring agencies throughout the world. The results show the genuine potential for a remotely sensed monitoring system that could, with further research, lead to an efficient and effective method of monitoring bathing water quality. These findings are particularly important given the imminent changes in EU Bathing Water policy, an expected increase in monitoring costs (currently estimated by the EU to be 15 million euros for 2001 (EU, 2002)) and the widespread availability of airborne sensors and satellites. Simultaneous water quality and spectral data were collected at Southend-on-Sea pier with a Natural Environment Research Council (NERC) loaned spectroradiometer and water sampling equipment. Simultaneous data enabled the accurate analysis of the relationship between water quality and reflectance, avoiding the normal delays experienced with flown or satellite data. The thesis successfully proposes and investigates a remotely sensed flagging system for bathing water quality monitoring using both statistical and visual analysis to identify optimum wavelengths which identify threshold levels of E.coli, suspended sediments, low pH, nitrates, chlorophyll, faecal coliform and temperature. The findings demonstrate that remote sensing could be used to monitor several of the water quality parameters that are relevant to the EU Bathing Water Directives and, in particular, the monitoring of effluent in bathing waters through the successful identification of high E.coli counts. Through the creation and integration of a localised water quality model, it demonstrates that it is possible to predict when water quality parameters exceed a threshold level through direct remote sensing or through the use of remotely sensed indirect water quality parameters. The success rate of remotely sensed "flagging" of samples above a threshold level was tested and used to yield a "predictor" rating for each parameter. Finally, a spectral physical model was constructed that identifies the parameters, wavelengths and secondary parameters that could be used to flag failing water quality areas. This model could be used to improve monitoring coverage and reduce overall costs. The application of the model, which was based on Case 2 coastal water, to other types of coastal area is suggested as needing further research before it could be widely exploited. Remote sensing information could lead to a greater understanding of the coastal environment and offers potential near real time monitoring, allowing for the first time reactive management of coastal water quality in failing water quality areas. This would provide a solution to many of the issues raised by the EU regarding the current bathing water quality directives and provides the remote sensing community with a practical solution to a real world problem