Summary We used daily rainfall and streamflow time series from two large catchments in the seasonal tropics of Queensland, Australia to investigate the hydrological impacts of woodland clearing. The Comet catchment (16,440 km 2) had 45% of the native woodland cleared during the mid-1960s. In the Upper Burdekin catchment (17,299 km 2) clearing decreased native woodland extent from 83% to 58% between 1998 and 2009. An earlier modelling study concluded that clearing in the Comet catchment increased annual streamflow by more than 40%. Here, several published inference methods to separate land use effects from climate variability were applied. Trend analysis of daily rainfall and streamflow data showed that interannual changes in mean streamflow in the Comet catchment were mostly due to changes in rainfall. In particular, a series of La Niña events after clearing led to an unusual lack of dry periods and an apparently associated temporary increase in runoff coefficient. The overriding importance of climate variability was further confirmed using a conceptual framework that was used to interpret changes in the long-term coupled water–energy budget. Even so, there was some evidence for a slight increase in streamflow for the first few years after clearing. Fitting a Budyko-type model for two climatically similar pre- and post-clearing periods (1920–1953 and 1979–2007) did not suggest a considerable change in the catchment water balance after clearing. Analysis of daily streamflow metrics did reveal some changes however, with enhanced peak flows and reduced low flows. In the Upper Burdekin catchment, trend analysis revealed a change in baseflow dynamics after clearing, while event storm flow for large rainfall events increased. In summary, woodland clearing in northern Queensland appears to have had a smaller impact on mean and interannual streamflow than might be expected from studies at sites and in small experimental catchments, but changes in daily streamflow patterns do suggest a modest change in catchment dynamics.