Factors affecting diatom distribution in floodplain lakes of the southeast Murray Basin,Australia and implications for palaeolimnological studies

Diatom assemblages of surface sediments in 46 billabongs from four river floodplains in the southeast Murray-Darling Basin, Australia were sampled to investigate drivers of species distribution. The principal purpose of the study was to derive information to aid interpretation of diatom-based palaeoecological studies of these systems and of floodplain lakes more generally. Patterns in billabong diatom assemblages in relation to river reach, hydrology and farming intensity on surrounding land were examined, as were correlations with water quality variables. Seasonal variation in billabong water quality was high relative to spatial variation, and spatial patterns in billabong water quality were weak. In contrast, strong patterns were evident in diatom assemblages. Three main patterns were observed: (1) a distinction between billabongs dominated by planktonic diatoms from those dominated by benthic and attached forms; (2) differences in diatom assemblages in billabongs on different river reaches; and (3) differences in assemblages in billabongs with different hydrology. Of all water quality variables tested, total phosphorus (TP), total nitrogen (TN) and pH exerted the strongest independent influence on diatom distribution; however, only TP remained an important variable when species variation due to river reach, hydrology, and aquatic plant cover was taken into account. The weak influence of water quality on diatom distribution is interpreted as reflecting the dichotomy between plankton and non-plankton-dominated billabongs, the influence of hydrology and biogeography, the lack of strong spatial water quality gradients and the high degree of temporal variability in water quality. The findings show that diatom records from billabong sediments can provide evidence of long-term changes in the abundance of aquatic macrophytes and hydrology. They also suggest that merging calibration data sets across regions for the purpose of improving diatom transfer functions for water quality reconstruction is of limited value for floodplain lakes, and that performance is more likely to be gained by boosting site numbers within regions.