Ensuring model assumptions support effective...
URL: https://www.mdba.gov.au/sites/default/files/publications/ensuring-model-assumptions-support-effective-decision-making.pdf
Authors: Georgia K. Dwyer, Galen Holt, Rebecca E Lester
Date of publication: September 2025
Ecological models help predict how nature will respond to changes in climate or water use, but their results can vary a lot depending on how they are built. Even small decisions — like how information is grouped or which time periods are included — can lead to big differences in outcomes .
In managing river systems like the Murray–Darling Basin, thousands of flow targets are used to guide water allocation, but the way these targets are modelled and combined can greatly influence the predicted outcomes.
This research explores how sensitive environmental models are to different assumptions, especially in how water needs for ecosystems are represented. By identifying which choices have the biggest impact on predictions, and which approaches are more cautious, the study aims to support more informed, transparent, and effective environmental decision-making.
Key findings / recommendations:
-
Simply changing some of the assumptions in the model could lead to very different results: with differences in the predicted outcomes as large as from 30% to 80% of water targets being met, even under the same climate and management conditions.
-
Simple or “default” modelling choices can give more optimistic results that may overlook real environmental risks.
-
Differences are especially noticeable during dry periods or in complex ecosystems, where it is crucial to get decisions right.
-
To improve reliability, experts suggest keeping models as simple as possible while still reflecting key biological processes, comparing multiple models side-by-side instead of blending them together, and working closely with scientists and environmental managers.
-
The best way to know whether a model’s predictions are trustworthy is by testing them against real-world data.
You can copy and paste the embed code into a CMS or blog software that supports raw HTML