Abstract of PhD Thesis

A flushing study analysis of selected Irish waterbodies

Thomas Dabrowski (2005) - National University of Ireland, Galway

This study begins with an extensive literature review of existing flushing analyses methods that have emerged over the last fifty years following the research conducted by Ketchum (1951). Various flushing characteristics are defined and their meaning discussed. The methodologies for calculating flushing characteristics include those based on the water volume and salt budgets (including box models), conservative dye distribution as well as methods based on sophisticated theories, such as the return flow factor model and the jet-sink-like circulation model. Numerical models are used to derive flushing characteristics. Passive tracer transport simulations were carried out to achieve the goal. Flushing studies were conducted for the Irish Sea and nine dissimilar Irish estuaries. The two-dimensional (DIVAST) and three-dimensional (ECOMSED) models were applied to the Irish Sea. Firstly, tidal models were developed and calibrated against available data on tidal currents and tidal amplitudes. The calibrated ECOMSED model was further developed to include meteorological forcing in order to account for wind-driven and thermohaline circulations. The fully baroclinic and barotropic model was capable of reflecting the temperature field properly, providing appropriate choices were made with regards to the model parametrisation. Modelling effort was focused on the western Irish Sea, where organised cyclonic baroclinic flows were reported. The importance of various forcing functions upon flushing was investigated. The net flow is highly variable throughout the year due to the joint action of wind and density gradients; it is northward if averaged over the year and equals c.2.50 km/d. Baroclinc circulation has slackening effects on flows. Fully forced ECOMSED model predicted average residence times of the entire Irish Sea region at the level of 386 and 444 days in the cases of the passive tracer transport simulations commencing in June and December, respectively. Spatial distributions of residence times were examined to reveal ‘flushing paths’ and to investigate their percentage distribution within the area. The St. George’s Channel is flushed in the first place; a backwater is formed to the southeast of the Isle of Man. Transit times from the Sellafield’s nuclear plant outfall site to various locations have been proved to vary seasonally. Retention in the western Irish Sea is twice longer due to developing thermal stratification and associated cyclonic circulation. A particle tracking model may well be utilised for the purpose of flushing studies, since delivered estimates are in close agreement with those based on conservative dye transport simulations. The thesis then describes the application of the two-dimensional DIVAST model to nine Irish estuaries. Hydrodynamic models as well as the results from the passive tracer transport simulations are presented and discussed. Several flushing characteristics were calculated, namely the average residence times, average exchange per tidal cycle coefficients, average transit times and pulse residence times. The influence of various wind conditions and river discharges upon flushing was examined and relationships developed. In the cases of quickly flushed reservoirs neap, mean and spring tidal ranges were considered separately. Spatial distributions of residence times were also examined. Subsequently, the estuaries were classified in terms of the tidal range, average residence time, primary factor governing flushing and in terms of flushing efficiency. The latter two are new ways of classification proposed in this research. Individual categories are clearly defined. It also appears that the vulnerability of flushing of a given estuary to wind can be assessed on the basis of its surface area, length and width. The classification schemes have been proved helpful in developing a simplified formula relating the average residence time to basic physical estuarine characteristics. Such relationship was developed for primarily tidally flushed estuaries, giving more accurate estimates than other methods not based on numerical model simulations. The last part of the thesis deals with the utilisation of the information on flushing properties of an estuary for the prediction and control of the eutrophication process. Simulations based on DIVAST and carried out on a set of rectangular harbours show that the greater the residence time the higher the predicted basin-averaged chlorophyll_a concentration under the same nutrient loading. The critical value of the average residence time, below which the increase in basin-averaged chlorophyll_a concentration was not observed in the simulations, equalled 3.1 days. In limnology, the maximum permissible phosphorus loading to avoid eutrophic conditions was related to lake’s residence time, Vollenweider (1976). In this research the critical phosphorus loading was successfully related to the estuarine average residence time and exchange per tidal cycle coefficient in the case of diffused sources and to the pulse residence time in the case of point sources, such as the river inflow. Therefore, the knowledge about flushing properties of the waterbodies has been proved to be helpful in understanding its ecology.