Comparison of Effective Radiative Forcing Calculations Using Multiple Methods, Drivers, and Models

T. Tang, D. Shindell, G. Faluvegi, G. Myhre, D. Olivié, A. Voulgarakis, M. Kasoar, T. Andrews, O. Boucher, P. M. Forster, Hodnebrog, T. Iversen, A. Kirkevåg, J. F. Lamarque, T. Richardson, B. H. Samset, C. W. Stjern, T. Takemura, C. Smith

Research output: Contribution to journalArticlepeer-review

17 Citations (Scopus)


We compare six methods of estimating effective radiative forcing (ERF) using a set of atmosphere-ocean general circulation models. This is the first multiforcing agent, multimodel evaluation of ERF values calculated using different methods. We demonstrate that previously reported apparent consistency between the ERF values derived from fixed sea surface temperature simulations and linear regression holds for most climate forcings, excluding black carbon (BC). When land adjustment is accounted for, however, the fixed sea surface temperature ERF values are generally 10–30% larger than ERFs derived using linear regression across all forcing agents, with a much larger (~70–100%) discrepancy for BC. Except for BC, this difference can be largely reduced by either using radiative kernel techniques or by exponential regression. Responses of clouds and their effects on shortwave radiation show the strongest variability in all experiments, limiting the application of regression-based ERF in small forcing simulations.

Original languageEnglish
Pages (from-to)4382-4394
Number of pages13
JournalJournal of Geophysical Research: Atmospheres
Issue number8
Publication statusPublished - Jan 1 2019

All Science Journal Classification (ASJC) codes

  • Geophysics
  • Forestry
  • Oceanography
  • Aquatic Science
  • Ecology
  • Water Science and Technology
  • Soil Science
  • Geochemistry and Petrology
  • Earth-Surface Processes
  • Atmospheric Science
  • Space and Planetary Science
  • Earth and Planetary Sciences (miscellaneous)
  • Palaeontology


Dive into the research topics of 'Comparison of Effective Radiative Forcing Calculations Using Multiple Methods, Drivers, and Models'. Together they form a unique fingerprint.

Cite this