Modelling river systems under change can create deep uncertainty in representing dynamic water demands, climate variability and climate change. This is particularly challenging in mine rehabilitation, where large local hydrologic changes can propagate downstream and interact with vulnerable water-dependent environmental values. Yet water balance modelling for mine closure and Environmental Impact Assessments typically applies few fixed demand assumptions and climate scenarios, using uniform scaling of historical flows. This can obscure how ecologically relevant flow components change over time and limit assessment of whether environmental values remain within acceptable thresholds under plausible futures. Nor does it allow study of multiple dimensions of change, such as reduction in rainfall and shift in seasonality.
This paper presents a bottom-up stochastic modelling framework developed to support threshold-based assessment of environmental values under deep uncertainty. The framework couples dynamic water demands that change over time and respond to climate with stochastic climate sequences to represent natural variability. Climate change is represented as a range of plausible rainfall and temperature end states applied progressively through time, rather than as static adjustments. Large ensembles of simulations are run stochastically using a Source river system model to quantify the distribution of outcomes and the likelihood of environmental performance thresholds being exceeded.
Using the Yallourn mine rehabilitation project in Victoria’s Latrobe Valley, we show how the framework improves attribution of risk to climate versus operational decisions, supports comparison of alternative water management strategies, and provides a transparent, updateable basis for stress-testing environmental flow tolerance under uncertain futures.