Surface Winds and Dust Biases in Climate Models. Issue 2 (27th January 2018)

• Record Type: Journal Article
• Authors: Evan, A. T.
• Is Part Of: Geophysical research letters. Volume 45:Issue 2(2018)
• Page Start: 1079
• Page End: 1085
• Publication Date: 2018-01-27
• Subjects: dust -- climate -- models; Geophysics -- Periodicals; Planets -- Periodicals; Lunar geology -- Periodicals; 550
• DOI: 10.1002/2017GL076353
• Languages: English
• ISSNs: 0094-8276

Abstract

Abstract: An analysis of North African dust from models participating in the Fifth Climate Models Intercomparison Project (CMIP5) suggested that, when forced by observed sea surface temperatures, these models were unable to reproduce any aspects of the observed year‐to‐year variability in dust from North Africa. Consequently, there would be little reason to have confidence in the models' projections of changes in dust over the 21st century. However, no subsequent study has elucidated the root causes of the disagreement between CMIP5 and observed dust. Here I develop an idealized model of dust emission and then use this model to show that, over North Africa, such biases in CMIP5 models are due to errors in the surface wind fields and not due to the representation of dust emission processes. These results also suggest that because the surface wind field over North Africa is highly spatially autocorrelated, intermodel differences in the spatial structure of dust emission have little effect on the relative change in year‐to‐year dust emission over the continent. I use these results to show that similar biases in North African dust from the NASA Modern Era Retrospective analysis for Research and Applications (MERRA) version 2 surface wind field biases but that these wind biases were not present in the first version of MERRA. Plain Language Summary: Why do climate models seem to have such a difficult time simulating past changes in dust storms from the Sahara? This paper shows that the reason stems not from how they treat the physics of dust emission but rather is due to poor representation of the meteorology at the surface of the Sahara desert. Key Points: Climate models exhibit large errors in dust emitted from North Africa These dust errors are due to biases in modeled surface winds Similar errors in MERRA‐2 dust stem from biases in MERRA‐2 surface winds