In April 2021, I spent a weekend touring wineries and talking ecology with . . . an engineer. His models were noisy, deterministic and calculated solutions; mine were stochastic and calculated relative probabilities, and for two days we sipped fine wines and struggled towards mutual understanding. As the last topaque was drained, we decided to gather our colleagues and build a model together. I will present this lovechild; an integrated systems model of 70 km in the Lower Murray River between locks 2 and 3. Therein, a modelled river flows alongside modelled redgum (Eucalyptus camaldulensis) and lignum (Duma florulenta) wetlands. Modelled southern bell frogs (Ranoidea raniformis) breed and hop around the modelled landscape, wary of the modelled fish and crustacean predators that build up in long-lived waterbodies.
Currently in this reach, environmental water is spread thinly. Management plans tend to favour redgum and lignum deliveries, while efforts to support the Vulnerable southern bell frog are constrained by competition for this limited resource. We used a genetic algorithm to move selectively through the possible solutions, in search of optimal strategies to prioritise environmental water deliveries. Across 430,000 stochastic evaluations we refined a strategy that improved the condition of redgum and lignum and doubled the number of modelled southern bell frog populations using the same resources that currently prioritise the two plants.
This shared exercise turned into a 2-way learning opportunity with implications for conservation, optimisation, post-hoc analyses, decision support tools and your next winery tour.