An integrated surface water, groundwater and wetland plant model of drought response and recovery for environmental water management

Modelling and associated reporting of ecosystem drought recovery across the Murray Darling Basin (MDB) is critical for managing inter-and intra-annual delivery of environmental flows. Such modelling and reporting can enable the provision of timely information during phases of critical decision making, especially when deciding which assets (locations) most need environmental water. To enable integrated (conjunctive) management of surface water and groundwater resources for environmental outcomes, such timely reporting of wetland response to drought cannot always rely on the availability of complex groundwater and surface water models because such models rely on detailed data, which is not always available. We develop an autoregressive model of adjusted river flow versus plant biomass for two key sites within the Great Cumbung Swamp (GCS, Lachlan Valley, MDB). The model operates over drought and flood conditions, and incorporates previous flows (and hence antecedent conditions). The incorporation of both previous and current flows is used to represent change in wetland water volume (∼ local flood depth), which in turn is a key factor in influencing plant survival and growth. Cumbungi (Typha species; hereafter Typha) and Common Reed (Phragmites australis; hereafter Phragmites) are represented using a surrogate measure for wetland plant biomass. These wetland plant taxa represent local plant communities (dominating groundcover scores) at environmental flow monitoring sites. Phragmites is known to better utilise local, shallow groundwater and therefore we predicted that Phragmites would be more resilient and resistant to drought. The dependence of plants on both surface flows and groundwater is also explored, by examining changes in biomass during both moderate and dry periods, and the potential to access shallow groundwater (plant data was not available for a wet period). The models indicate the extent to which groundwater versus surface flows influence the observed localized, species-specific responses. We also explore local depth-conductivity relationships to estimate the extent of groundwater loss or evaporative concentration, with the hypothesis that conductivity does not change. Previous models of groundwater loss for the GCS as-a-whole indicate that the GCS is a losing system, and hence does not concentrate solutes through evapotranspiration. We found, as expected, that Phragmites abundance takes longer to fall during a drought and recovers more quickly; indicating an overall greater resistance and resilience to drought effects. However, during wetter periods this relationship breaks down, and presumably other factors (e.g., temperature) are the main determinants of plant growth. Also, and surprisingly, in spite of previous groundwater models and the probable high dependency of Phragmites on shallow groundwater during drought, there appears to be very little groundwater loss at monitoring sites. This, and previous studies on the generally impermeable clays within the GCS, indicates that the monitoring sites are operating as perched wetland systems. Phragmites is accessing groundwater that is not infiltrating locally within the study sites, or diffusely across the entire GCS, but instead is accessing groundwater recharged within numerous other locations across the GCS. We conclude that such models help inform management actions so that they can focus on species-specific water requirements, particularly under drought conditions. If such studies were widespread, then comparisons of these water availability-plant response dynamics among MDB wetlands could be used to draw further conclusions about the relative resilience of plant species and communities to drought