Dynamics of microbial pollution in aquatic systems

Abstract

Microbial pollution of surface waters and coastal zones is one of the foremost challenges facing the water industry and regulatory authorities. Yet despite the concern and increasing pressures on water resources in both developed and developing countries, understanding of microbial pollutants in the aquatic environment is fairly scattered. There is a need for an improved ability to quantify the processes that control the fate and distribution of enteric organisms to support decision - making and risk management activities. The aim of this thesis has been to advance the understanding of the dynamics of microbial pollution in aquatic systems through review, experimentation and numerical modelling. Initially, a new module for simulating the protozoan pathogen, Cryptosporidium, was developed and implemented within a three - dimensional ( 3D ) coupled hydrodynamic - water quality model ( ELCOMCAEDYM ). The coupled 3D model was validated against a comprehensive dataset collected in Myponga Reservoir ( South Australia ), and without calibration, performed to a high degree of accuracy. The investigation then sought to examine the experimental dataset in more detail and found a significant difference between protozoan pathogens and the bacterial and viral indicators. To examine the role of bacterial association with particles in more detail, a second experimental campaign was carried out in Sugarloaf Reservoir ( Victoria ). This campaign was used to gain insights into the association of coliform bacteria with suspended sediment and to quantify their sedimentation dynamics based on in situ measurements. Using an inverse technique, particle profile data was used to create a simple Lagrangian model that was applied to back - calculate the sedimentation rates of the coliform bacteria and the fraction that were attached to the particles. The results indicated that 80 - 100 % were associated with a small - sized clay fraction. This result was in contrast with the Cryptosporidium dynamics in Myponga Reservoir, where it was concluded that oocysts did not settle with the inorganic particles. These findings indicated the current models for simulating the array of organisms of interest to regulatory authorities are inadequate to resolve the level of detail necessary for useful predictions and risk management. Large differences between the protozoa, bacteria and phages were being observed due to different particle association rates and sedimentation dynamics, order of magnitude differences in natural mortality rates, and different sensitivity to sunlight bandwidths. The original model implemented within CAEDYM was therefore rewritten to be more complete and generic for all microbial pollutants and different types of aquatic systems. The model was built using a generic set of parameterizations that describe the dynamics of most protozoan, bacterial and viral organisms of interest. The parameterizations dynamically account for sensitivities to environmental conditions, including temperature, salinity, pH, dissolved oxygen, sunlight, nutrients and turbidity, on the growth and mortality of enteric organisms. The new model significantly advances previous studies in several areas. First, inclusion of the growth term allows for simulation of organisms in warm, nutrient rich environments, where typical die - off models tend to over - predict loss rates. Second, the natural mortality term has been extended to independently account for the effects of salinity and pH, in addition to temperature. The salinity - mediated mortality has also been adapted to account for the nutrient status of the medium to simulate the importance of nutrient starvation on the ability of an organism to survive under osmotic stress. Third, a new model for sunlight - mediated mortality is presented that differentially accounts for mortality induced through exposure to visible, UV - A and UV - B bandwidths. The new expression has capacity to simulate the photo - oxidative and photo - biological mechanisms of inactivation through included sensitivities to dissolved oxygen and pH. Fourth, the model allows for organisms to be split between free and attached pools, and sedimented organisms may become resuspended in response to high shear stress events at the water - sediment interface caused by high velocities or wind - wave action. Fifth, the enteric organism module has been implemented within the bio - geochemical model CAEDYM, thereby giving it access to dynamically calculated concentrations of dissolved oxygen, organic carbon, and suspended solids, in addition to pH, shear stress and light climate information. Without adjustment of the literature derived parameter values, the new model was validated against a range of microbial data from three reservoirs that differed in their climatic zone, trophic status and operation. The simulations in conjunction with the experimental data highlighted the large spatial and temporal variability in processes that control the fate and distribution of enteric organisms. Additionally, large differences between species originate from variable rates of growth, mortality and sedimentation and it is emphasized that the use of surrogates for quantifying risk is problematic. The model can be used to help design targeted monitoring programs, examine differences between species and the appropriateness of surrogate indicators, and to support management and real - time decision - making. Areas where insufficient data and understanding exist are also discussed.