Assessing certainty and uncertainty in riparian habitat suitability models by identifying parameters with extreme outputs

The aim of this paper is to introduce a computationally efficient uncertainty assessment approach using an index-based habitat suitability model. The approach focuses on uncertainty in ecological knowledge regarding parameters of index curves and weights. A case study determines which of two 15-year periods has more suitable surface water and groundwater regimes for riparian vegetation. The uncertainty assessment consists of defining constraints on index curves and weights. Linear programming is used to identify parameters that yield two extreme outputs: maximising and minimising differences between the two periods. Because they are extremes, if both outputs agree on which period is better (e.g. maximum and minimum differences are both positive), then all other models will also agree. Identifying models with extreme outputs prompts learning about the boundaries of our knowledge and identifies patterns about what is considered certain. It helps build an understanding of what is already known despite the high uncertainty.