Water resource management traditionally depends on use of highly complex hydrological models designed originally to manage water for abstraction but increasingly relied on to determine ecological impacts and test ecological rehabilitation opportunities. These models are rarely independently tested. We compared a relatively simple statistical model, integrated flow and flood modelling (IFFM), with a complex hydrological model, the integrated quality and quantity model (IQQM), on the highly regulated Macquarie River of the Murray-Darling Basin, southeastern Australia. We compared annual flows (1891-2007) at three gauges to actual data and modelled output: before dams and diversions (unregulated) and after river regulation (regulated), using the goodness of fit (Nash-Sutcliffe coefficient of efficiency) and nonparametric tests. IQQM underestimated impacts of river regulation respectively on median and average flows at the Macquarie Marshes (Oxley gauge) by about 10% and 16%, compared to IFFM. IFFM model output more closely matched actual unregulated and regulated flows than IQQM which tended to underestimate unregulated flows and overestimate regulated flows at the Ramsar-listed wetland. Output was reasonably similar for the two models at the other two flow gauges. Relatively simple statistical models could more reliably estimate ecological impact at floodplains of large river systems, as well as an independent assessment tool compared to complex hydrological models. Finally, such statistical models may be valuable for predicting ecological responses to environmental flows, given their simplicity and relative ease to run.