WIT Press

Validation Of A Decision Support Model To Predict Concentration After A Chemical Spill Into A River


Free (open access)

Paper DOI









342 kb


D. Jabbour, A. Dandrieux-Bony & G. Dusserre


The transport of dangerous materials (road, fluvial, pipeline…) increases the risk of accidental releases into streams. The consequences of these accidents can be very serious, both for the stream’s environment and for people’s lives if the pollution reaches water treatment plants, irrigation water pumping… To limit the consequences of these accidents, it is necessary to use a decision support model that predicts pollutant concentrations in a stream. This work focuses on laboratory model pollution simulations in order to validate the dispersion model that gives a concentration of chemical miscible product versus time or distance from the downstream releasing point. The channel model initial dimensions (34 cm of width, 14 m of length and 16 cm of depth) can be varied with experimental conditions and scaling requirements. The current velocity can vary up to 20 cm/s. The flow characteristics are turbulent and fluvial. These characteristics can be modified to extend the representativeness of the river model to different types of rivers, channels or streams. A specific sampling system is developed to take 144 samples across the section. The mixing length and the dispersion coefficient are evaluated by using a Uranine tracer and samples are analyzed by sphectophotometry UV-Vis. Finally, the dispersion model is validated and modified for one dynamic configuration of the channel model. The predictions are excellent with a correlation coefficient between observed and predicted data equal to unity. Keywords: decision support model, chemical releases, dispersion model, channel model and validation.


decision support model, chemical releases, dispersion model, channel model and validation.