Abstract
The AssESs project (Assessing the extent and significance of uncertainty in offshore wind assessments), funded by the Offshore Renewables Joint Industry Programme (ORJIP) for Offshore Wind, aims to improve the treatment of uncertainty within the assessment process for ornithological impacts, to reduce risks and delays to the consenting of offshore wind developments.
A key motivation for the project is an urgent need to quantify current levels of uncertainty across the assessment process, and sensitivities of estimated impacts to different sources of uncertainty. This is delivered through a review of existing approaches to the treatment of uncertainty within assessments, and of the evidence basis that informs these approaches, which is then used to structure a quantitative evaluation of the sensitivity of key metrics of impact to uncertainty in parameter values and model assumptions. The second key motivation for the project is a need to improve, via stakeholder engagement, the ways in which information on uncertainty is translated into decision-making within the context of a precautionary approach.
Within this work package (WP2) we assess the extent to which key impact assessment outputs, such as metrics associated with Population Viability Analysis (PVA), are sensitive to uncertainties in inputs, and to assumptions around model structure. Uncertainty ranges and distributions are derived from the WP1 review (AssESs – Summary report of uncertainty and approaches to evaluating uncertainty review (WP1)) wherever feasible, supplemented by information obtained through stakeholder engagement. We also assess sensitivity to the ways in which uncertainty is treated within assessment tools, considering different scenarios for the use of uncertainty that reflect current practice and feasible alternatives that do not consider uncertainty in each step of the assessment process in isolation but instead allow for uncertainty to be propagated between stages. These comparisons helped to inform stakeholder discussions in WP3 (AssESs – Summary report of stakeholder engagement (WP3)) around the way that uncertainty is interpreted in relation to precaution. The sensitivity analysis (SA) also informs WP4 recommendations (AssESs – Recommendations and roadmap (WP4)) around future data collection and research, and treatment and interpretation of uncertainty in relation to precaution within the assessment process.
Uncertainty ranges and distributions considered within the sensitivity analysis, and scenarios that capture current practice around the use of these ranges within assessments, have been agreed following consultation with stakeholders (Defining scenarios with stakeholders)) to ensure relevance of the results. Using existing tools, and approaches to link these tools developed within the Scottish Marine Energy Research (ScotMER) Cumulative Effects Framework (CEF), we have evaluated the sensitivity of key metrics of impact to uncertainties in inputs across multiple assessment tools (Uncertainty in relation to inputs)). This task has also evaluated the sensitivity of outputs to alternative scenarios around the treatment of uncertainty within the process, maximising the ability of the SA to structure stakeholder discussions around decision-making in WP3.
Structural uncertainties arise from the fact that current assessment approaches and tools make simplifying assumptions, and omit biological features. Systems-based Sensitivity Analysis) investigates whether a holistic modelling of the uncertainties in the assessment process can lead to a robust representation of the sensitivities in the system. We therefore aim to examine whether the types of system interdependencies in the seabird/renewables interaction, that are currently not considered within assessments, lead to significant divergence of risk estimates, and over/under-precautionary predictions of population viability. This forward-looking investigation of uncertainty, although purely exploratory at the present time, will help us determine which biological features currently omitted from assessments need to be introduced, and which of the existing features may be simplified without loss in accuracy or precision.