(Erosion Prediction in Ungauged Basins: Integrating Methods and Techniques (Proceedings of Symposium HS01 held during IUGG2003 at Sapporo, July 2003), IAHS Publication No. 279, 2003, pp. 33–42).
Lake Samsonvale (North Pine Dam, NPD) is a major reservoir, of approximately 22 km2 with a capacity of 215,000 Ml and drainage area of 347 km2, which supplies drinking water to Brisbane, the capital city of Queensland in eastern Australia. Historically, there are large and persistent blooms of potentially toxic cyanobacteria in NPD, and it was hypothesized that their occurrence and biomass were related to input of sediment-bound phosphorus from the drainage basin. The NPD drainage basin is ungauged and few estimates of (sub) drainage basin sediment loads exist. This paper develops a Bayesian approach utilizing endmember models to estimate, based on major and trace element geochemical signatures, the proportion of various drainage basin sediment sources in sediment samples taken from NPD. This approach not only allows for the incorporation of prior knowledge about the geochemical composition of the sources (or endmembers), but also allows for correlation between spatially continguous samples and the prediction of the sediment composition at unsampled locations.
Algal bloom, Australia, Bayesian mixing model, Geochemistry, Monte Carlo Markov chain, Principal component analysis, Reservoir, Trace elements
