Using palaeoecological records to disentangle the effects of multiple stressors on floodplain wetlands

Abstract

Ecosystems worldwide are subject to the deleterious effects of multiple anthropogenic stressors. Understanding and mitigating the effects of these stressors is difficult both because stressors are confounded in space and have the potential to act both synergistically and antagonistically. Palaeoecological approaches applied to systems where stressors may be confounded in space but not in time offer a way to explore the effects of multiple stressors. This multiproxy study of sediment records from four floodplain lakes (billabongs) on a dryland river floodplain subject to grazing, commencing in the late 1800s, and irrigated cropping, commencing in the late 1900s, tests this approach. The results suggest that the effects of both grazing and irrigation on floodplain ecosystems can be detected in the pollen and diatoms preserved in sedimentary records of billabongs. For the pollen, these changes are inconsistent, but appear to reflect local shifts in dominance among major tree taxa and among key understorey plant families. For the diatoms, the changes were also not consistent across sites, but can be generalised as reductions in epiphytic diatoms and increases in planktonic and facultative planktonic taxa that likely reflect increased fluxes of sediments and nutrients and reduced flood frequency. Overall, the effects of grazing appear to have been greater than irrigated cropping. The results also show that the relative effects of grazing and irrigated cropping on floodplain and riparian vegetation and on diatom communities vary between billabongs, with some evidence that at least some of this variation relates to the level of hydrological connection to the mainstream. Finally, the study suggests that for the most part, grazing and irrigated cropping act antagonistically in the way they impact these floodplain ecosystems, a pattern that likely reflects a release from grazing pressure associated with the increase in irrigated cropping. Future applications of the approach should increase spatial and temporal replication and develop more sophisticated frameworks that account for temporal variation in driver intensity and proxy indicators of the specific stressors that influence ecosystem structure and function.