Effects of parameter indeterminacy in pelagic biogeochemical modules of Earth System Models on projections into a warming future: The scale of the problem. Issue 7 (21st July 2017)

• Record Type: Journal Article
• Authors: Löptien, U.; Dietze, H.
• Is Part Of: Global biogeochemical cycles. Volume 31:Issue 7(2017:Jul.)
• Page Start: 1155
• Page End: 1172
• Publication Date: 2017-07-21
• Subjects: biogeochemical ocean modeling -- model verification and validation -- parameter estimation -- future projections -- uncertainty -- Earth System Model; Biogeochemical cycles -- Periodicals; Electronic journals; 577.1405
• DOI: 10.1002/2017GB005690
• Languages: English
• ISSNs: 0886-6236

Abstract

Abstract : Numerical Earth System Models are generic tools used to extrapolate present climate conditions into a warming future and to explore geoengineering options. Most of the current‐generation models feature a simple pelagic biogeochemical model component that is embedded into a three‐dimensional ocean general circulation model. The dynamics of these biogeochemical model components is essentially controlled by so‐called model parameters most of which are poorly known. Here we explore the feasibility to estimate these parameters in a full‐fledged three‐dimensional Earth System Model by minimizing the misfit to noisy observations. The focus is on parameter identifiability. Based on earlier studies, we illustrate problems in determining a unique estimate of those parameters that prescribe the limiting effect of nutrient‐ and light‐depleted conditions on carbon assimilation by autotrophic phytoplankton. Our results showcase that for typical models and evaluation metrics no meaningful "best" unique parameter set exists. We find very different parameter sets which are, on the one hand, equally consistent with our (synthetic) historical observations while, on the other hand, they propose strikingly differing projections into a warming climate. Key Points: Revisiting optimization techniques of poorly constrained biogeochemical model parameters in a three‐dimensional Earth System Model For typical models, noisy observations and evaluation metrics no unique "best" parameter set may exist Different parameter sets can be equally consistent with historical observations, while they feature strikingly differing future projections