How uncertainties are tackled in multi-disciplinary science? A review of integrated assessments under global change

Abstract

Integrated assessment (IA) modelling can be an effective tool to gain insight into the dynamics of coupled earth system (land use, climate etc.) and socio-economic components. Quantifying and communicating uncertainties is a challenge of any scientific assessment, but is here magnified by the complex and boundary-crossing nature of IA models. Understanding the dynamics of coupled earth and socio-economic systems require data and methods from multiple disciplines, each with its own perspective on epistemological uncertainties (parametric and structural uncertainties), and its own protocols for assessing uncertainty. During the Paris Agreement, the lack of uncertainty analyses (UA) in IAs was risen (Rogelj et al. 2017) and calls for close collaboration of scientists coming from different fields. In this study, we review how uncertainties are tackled in a range of science disciplines that are related to global change including climate, hydrology, energy and land use, and which contribute to IA modelling. We conducted a meta-analysis to identify the contributing disciplines, and review which type of uncertainties are assessed. We then describe sources of uncertainty (e.g. parameter values, model structure), and present opportunities for improved assessment and communication of uncertainties in IA modelling. We show in our meta-analysis that parametric uncertainty is the uncertainty analysis that has been applied the most, while structural uncertainty is less commonly applied, with the exception of the energy scientific discipline. We finish our study with key recommendations to improve uncertainty analysis such as including risk analysis. By embracing uncertainties, resilient and effective solutions for climate change mitigation and adaptation could be better communicated, identified and implemented.