Influence of inundation characteristics on the distribution of dryland floodplain vegetation communities

•Water management affects vegetation distribution by altering the natural flow regime.•A well-calibrated and spatially explicit fine-scale inundation model was used.•Electivity and GAMs were applied to quantify the inundation-vegetation associations.•E. camaldulensis variation was well explained by inundation metrics.•Woodlands appear to be more drought tolerant than previously reported findings. Semi-arid floodplain vegetation is an essential component of floodplain and terrestrial ecosystems due to the wide range of services provided for waterbirds, woodland birds, amphibians and mammals, together with their contribution to natural carbon sequestration. Since overbank flooding has been considered the driving factor of vegetation distribution on floodplains, sustainable water resource management requires a better understanding of the influence of inundation on vegetation distribution. We examined the relationship between the distribution of four flood-dependant vegetation communities and associated long term inundation metrics on the upper Darling River floodplain using electivity analysis and a generalised additive model (GAM) by generating a total of 10,478 individual inundation maps with a high spatial resolution over a period of 29 years. Our results show that the four dominant vegetation communities are situated differently in relation to inundation attributes. The inundation metrics better explained the variation in the distribution of River Red Gum than those of Black box, Coolabah and Lignum communities. We found that the floodplain forests and woodlands in the upper Darling can survive longer periods of drought and shorter inundation duration than previously reported. The results provide practical information for ecosystem management by offering a means of predicting changes in vegetation distribution in relation to alteration in flow regime resulting from water planning arrangements or climate change.