An assessment of precipitation adjustment and feedback computation methods. Issue 19 (14th October 2016)

• Record Type: Journal Article
• Authors: Richardson, T. B.; Samset, B. H.; Andrews, T.; Myhre, G.; Forster, P. M.
• Is Part Of: Journal of geophysical research. Volume 121:Issue 19(2016)
• Page Start: 11, 608
• Page End: 11, 619
• Publication Date: 2016-10-14
• Subjects: precipitation -- rapid adjustments -- feedbacks -- radiative forcing -- PDRMIP -- climate models; Atmospheric physics -- Periodicals; Geophysics -- Periodicals; 551.5
• DOI: 10.1002/2016JD025625
• Languages: English
• ISSNs: 2169-897X

Abstract

Abstract: The precipitation adjustment and feedback framework is a useful tool for understanding global and regional precipitation changes. However, there is no definitive method for making the decomposition. In this study we highlight important differences which arise in results due to methodological choices. The responses to five different forcing agents (CO2, CH4, SO4, black carbon, and solar insolation) are analyzed using global climate model simulations. Three decomposition methods are compared: using fixed sea surface temperature experiments (fSST), regressing transient climate change after an abrupt forcing (regression), and separating based on timescale using the first year of coupled simulations (YR1). The YR1 method is found to incorporate significant SST‐driven feedbacks into the adjustment and is therefore not suitable for making the decomposition. Globally, the regression and fSST methods produce generally consistent results; however, the regression values are dependent on the number of years analyzed and have considerably larger uncertainties. Regionally, there are substantial differences between methods. The pattern of change calculated using regression reverses sign in many regions as the number of years analyzed increases. This makes it difficult to establish what effects are included in the decomposition. The fSST method provides a more clear‐cut separation in terms of what physical drivers are included in each component. The fSST results are less affected by methodological choices and exhibit much less variability. We find that the precipitation adjustment is weakly affected by the choice of SST climatology. Key Points: Important physical and quantitative differences in precipitation adjustment and feedback results arise due to computation method Fixed SST method provides a consistent and clear mechanistic decomposition Fixed SST method is less dependent on methodological choices and exhibits less variability