Modelling relative richness of flooding-response groups to predict hydrology-driven change in wetland plant communities

• Wetlands are highly threatened by climate change, but impacts are hard to predict. • To quantify hydrological risk, wetland plants can be grouped by flooding response. • Modelled thresholds between groups indicate points of ecological transition. • These indicators show good ability to predict drying-related change in composition. • Maximum depths >8 cm below ground predict transition from wetland to upland plants. Wetlands are sensitive to drying climates but projecting future impacts is challenging. Grouping plant species by flooding tolerance simplifies hydrological response modelling, with changes in group representation along the hydrological gradient defining both distinct vegetation zones and transition thresholds between them. We test this approach, using the relative number of species adapted to terrestrial, amphibious, and submerged conditions (hydrological group richness, HGR) as an indicator of plant composition and observed water depth in the growing season as our hydrological predictor. Using published data to model HGR ( n = 813 observations across 12 wetland complexes in temperate Australia), we inferred depth thresholds from transitions between terrestrial, amphibious and submerged groups and validated these thresholds using independent data collected from the same region over the preceding 5 years ( n = 198, 23 wetland complexes). The model explained 0.72 of variation in HGR suggesting a strong predictive relationship with depth. Thresholds distinguishing submerged-terrestrial (median ± [95 % credible intervals] = 13 [7, 22] cm) and submerged-amphibious group transitions (62 [50, 72] cm) were modelled directly, but amphibious-terrestrial transitions were predicted (via extrapolation) to occur below ground (−8 [−18, −2] cm). Predictive performance between obligate wetland (amphibious and submerged) species was limited, suggesting factors other than hydrology were involved. However, predictions against terrestrial species were reasonable (Cohen’s kappa > 0.58), suggesting the approach can identify hydrological tipping points associated with wetland drying. Modelling critical hydrology-driven transitions between wetland and terrestrial plants will require data on shallow sub-surface saturation (e.g., within ∼30 cm) rather than inundation depth.