The Role of Asymmetric Uncertainty Visualizations in Decision-Making by Environmental Decision Support System Users when Evaluating Conservation Plans for Peak Flood Reduction

Watershed conservation planning can be influenced by physical uncertainties and uncertainties in human behavior that influence stakeholder preferences and adoption of recommended conservation practice plans. Representation and communication of uncertainty are essential to facilitate effective and timely decisions about conservation practices needed to resolve pending watershed issues and climate change adaptation. However, individuals may misread, or disregard visualizations intended to communicate uncertainty in information. Our objective was to study how uncertainty representation of peak flow reductions (a measure of flood mitigation benefits) can affect conservation plan evaluations in a decision support system. We also investigated whether test-users may be used to evaluate decision support system visualization approaches (e.g., when end-user testing is not feasible), and whether users' statistical backgrounds can influence decisions they make after viewing uncertain information. A new Asymmetric Uncertainty Metric (AUM) was developed to manipulate levels of bias in uncertainty. In two online sessions, one with and one without visualized uncertainty, 77 participants (53 test users and 24 end users) evaluated conservation plans. Linear regression models were used to assess the association between AUM and user evaluations of the plans. Results show that users' evaluations made based on visualized uncertainty differed from those made without (p < .001), indicating that the visualization of uncertainty can influence participants' evaluations of conservation plans. Visualized uncertainty influenced end-users more than test-users (p < .001), and participants who had a statistical background were also more sensitive to visualized uncertainty than those who did not (p < .001).