Advection -dispersion model for nutrient dynamics in River Swale
In LOIS very little research has been directed to d[...]
Preview referat: Advection -dispersion model for nutrient dynamics in River Swale
In LOIS very little research has been directed to detailed study of short river stretches. Rather than for the detailed modelling studies of short river stretches (e.g. plume dispersion studies, pollutant transport after accidental release), QUESTOR is suitable for the modelling of large rivers at daily time resolution. Swale is also the subject of several studies focused on theestimation of nutrient dynamics, but they are not focused on the prediction of pollutant transport along the river. In the present research the analytical solution of the fundamental advection-dispersion equation (ADE) for mass transport in rivers was used to develop a detailed mathematical model for nutrient transport in River Swale (ADModel).
These kind of advection-dispersion analytical models have as main advantages the quality of results, the short computation time they need, lower computation resources compared to numerical model and user-friendly post processing of results. Possible disadvantages of such analytical model are related to: (1) the representation of hydraulic non-uniformity of the river stretch; and (2) the inclusion of influences (tributaries, pollution sources, abstractions).
To attend these issues the river stretch is divided in reaches, each one having a constant average value of hydraulic parameters and including influences at the beginning of reach.This approach complicates the applicability of the model to other rivers, and has been replaced in the present research by a more useful method explained elsewhere (Ani et al., 2009b). ADModel offers a different and more detailed perspective of studying pollutant transport compared to many existing studies and models (e.g. QUESTOR).
Existing approaches are typically based on a broad characterisation of chemical status and how it varies within large river basins. These existing models (1) represent the river as a perfect mixed tank or as a succession of perfect mixed reaches; (2) assume constant average parameters of the river/reaches; (3) make predictions at large time steps (daily); and (4) locate pollution sources and abstractions at reaches boundaries; while ADModel (1) represents the river as a continuous computational domain; (2) with variable parameters along it; (3) predicts concentration at smaller time steps (hourly); and (4) locate pollution sources and abstractions at the real place along the river.
This enables ADModel to be useful to predict the propagation of the four investigated nutrients at any place along the river stretch (under normal and accidental discharge) while existing models predict pollutant dynamics just at reach boundaries. Study and explanation of nutrient dynamics in the river stretch in terms of the temporal variability of the nutrient transformation rates is also possible.
The model development, calibration and validation rely on experimental data obtained during ten monitoring campaigns carried out with the river in low and medium flow, but also under storm conditions. Monitoring of concentration, water flow and water depth, was done at up to four sites (further referred to as M1 to M4) along the river stretch. These variables are used in the model as time series. Measurements of river channel parameters (water depth, channel width, river bed slope) are also available. Detailed information on: campaigns, monitoring sites, monitored parameters; are published elsewhere.
« mai multe referate din Engleza