It was anticipated that the substantial databases generated in the field and laboratory studies would then be used to develop and test geochemical and radionuclide transport models.

The uranium ore in the Koongarra No. 1 orebody is elongated over a distance of 450 m and persists to a depth of about 100 m. Secondary uranium mineralisation is present in the weathered schist zone to a depth of 25-30 m and has been mobilised downslope for about 80 m to form a dispersion fan.

Hydrogeological field investigations showed that the site was more complex than initially thought, hence a major site investigation of structures relevant to groundwater flow was also undertaken using geophysical techniques. The "simple" flow model that had been anticipated when initiating the project was seen to be inappropriate.

A geomorphological assessment of the development of the site and variations in the climate during the past million years, with their effect on the site hydrology, was undertaken to determine the time frame in which the uranium dispersion fan had developed, and to determine the timing of the dispersion fan formation.

Field measurement programs to measure chemical parameters of groundwater were carried out at the end of the Northern Territory wet and dry seasons, when water levels were at their highest and lowest levels. Colloids and particles were also collected and measurements were made of isotopes such as deuterium, tritium, 13C and 14C, which enabled groundwater mixing and flow paths to be studied.

An extensive study of the chemistry and mineralogy of the solid phase was also carried out on a large selection of materials sampled during the ARAP. This provided data on the early stages of weathering and the development of isotopic disequilibria under such conditions, the behaviour of rare earth elements, and the associations of uranium with mineral phases. A late finding in the mineralogical studies was the highly significant association of cerium and uranium in the weathered zone, a finding completely unexpected at the beginning of the project, and one which merits further investigation.

A two-phase study of sorption determined experimentally the factors that influence the sorption of U(VI) to well-defined mineral phases and to natural substrates from Koongarra and, on the basis of this experimental data, models of uranyl sorption were developed.

The groundwater and solid phase data provided a geochemical framework of the Koongarra system that was then used to develop and test transport and geochemical models. Initially, aqueous speciation models were used to calculate the composition of representative Koongarra waters. These showed that the present-day groundwaters were strongly undersaturated with respect to uranium-bearing minerals, hence should have been actively dissolving the uranyl phosphate zone. The geochemical modelling study was therefore extended to consider both present-day geochemical processes and those that operated in the geologic past.

A large number of radionuclide transport models were used to interpret the large database of concentrations of uranium series radionuclides in the groundwater and in the solid phases. Although the work focused on the 238U series, and to a lesser extent on the 232Th and 235U series, there were also studies of 23Pu, 129I, 99Tc and 36CI. All of these radionuclides are found in radioactive wastes, therefore the estimates of radionuclide mobility obtained in this study are of relevance to performance assessment modelling.

The modelling approaches had to deal with many uncertainties in representing the Koongarra site and its evolution with time including the hydrogeology, the effect of climatic variations, and the heterogeneity of the weathered zone. Consequently, it was found that a number of models fit the available data, and it was not possible to conclude that any one particular model was correct and that the others were incorrect. Rather, it was possible to identify important processes affecting uranium transport and then apply a broad variety of modelling approaches.

Uranium transport in the secondary dispersion fan was found to have taken place over fairly short distances during a timescale of the order of millions of years. However, it was not possible to exclude the loss of some uranium bearing fluids from the system along flow paths through now eroded rocks. The fact that most of the uranium has moved only of the order of 50 to 100 m over a timescale of the order of millions of years suggests that the uranium is not very mobile, even in a chemically active system with relatively high groundwater flow.

The findings from the technical studies are discussed in the context of assessments of the long-term performance of geological repositories for radioactive wastes, which are being undertaken in many countries. They are also considered in an integrated "Scenario Development" approach, aimed to understand the formation of the ore deposit. Despite their inherent uncertainties, the findings provide a basis for assessing the way in which radionuclides will migrate in environments with a variety of geologic settings and over a range of different geologic timescales.

One of the great challenges of the ARAP, and therefore one of the great benefits because of the successful completion of the project, was that so many contrasting scientific disciplines had to work together to develop an understanding of the system. This was a valuable experience that similar teams will have to develop for the assessment of repositories, which will have to be founded on the characterisation of real sites.

This summary report, which highlights the work and findings of the Alligator Rivers Analogue Project is one of a series of 16 volumes, listed below. Detailed descriptions and results are provided in Volumes 2 to 16. Although an attempt has been made to make the individual reports self-sufficient, in many cases the results and conclusions overlap with other volumes. In these cases, comprehensive cross-referencing should allow the work to be fully accessed. Full acknowledgment to individual contributions is provided in the individual reports, and in Appendix I of this report.