The response of coastal ecosystems to anthropogenic nutrient enrichment is challenging to observe due to multiple nutrient sources, rapid transformations and the limitations of nutrient sampling methods. Early detection of broadscale impacts are particularly difficult to distinguish from the natural temporal evolution and variability present in coastal waters. We use the CSIRO Environmental Modelling Suite to construct a digital twin simulation of water quality in southeast Tasmania, to quantifying the impact of expanding salmon aquaculture. This region includes multiple estuaries, complex coastal morphology and is influenced by seasonally modulated ocean boundary currents. The numerical model was statistically validated against observations from satellites, moorings, landers, gliders and discrete samples over a hindcast period of 5 years to confirm that it was ‘fit for purpose’. Simulations characterise the variability in regional circulation, nutrient supply, water quality and sediment dynamics. Multiple digital twin scenario simulations (with more or less salmon aquaculture) predict changes in water and sediment quality due to contrasting nutrient load. Near real time simulations have the capability to distinguish anthropogenic nutrient impacts from natural variability to understand the drivers of variability in monthly monitoring data. This research is informing environmental regulators who are making decisions on future salmon aquaculture.