The epistemological status of general circulation models

Article

Abstract

Forecasts of both likely anthropogenic effects on climate and consequent effects on nature and society are based on large, complex software tools called general circulation models (GCMs). Forecasts generated by GCMs have been used extensively in policy decisions related to climate change. However, the relation between underlying physical theories and results produced by GCMs is unclear. In the case of GCMs, many discretizations and approximations are made, and simulating Earth system processes is far from simple and currently leads to some results with unknown energy balance implications. Statistical testing of GCM forecasts for degree of agreement with data would facilitate assessment of fitness for use. If model results need to be put on an anomaly basis due to model bias, then both visual and quantitative measures of model fit depend strongly on the reference period used for normalization, making testing problematic. Epistemology is here applied to problems of statistical inference during testing, the relationship between the underlying physics and the models, the epistemic meaning of ensemble statistics, problems of spatial and temporal scale, the existence or not of an unforced null for climate fluctuations, the meaning of existing uncertainty estimates, and other issues. Rigorous reasoning entails carefully quantifying levels of uncertainty.

Keywords

Climate change General circulation models Model testing Epistemology 

Supplementary material

382_2017_3717_MOESM1_ESM.docx (948 kb)
Supplementary material 1 (DOCX 947 KB)

Copyright information

© Springer-Verlag Berlin Heidelberg 2017

Authors and Affiliations

  1. 1.National Council for Air and Stream Improvement, Inc. (NCASI)NapervilleUSA

Personalised recommendations