loading page

Improving the parameterization of dust emission threshold in the Community Earth System Model (CESM)
  • +6
  • Danny Min Leung,
  • Jasper Kok,
  • Longlei Li,
  • Natalie Mahowald,
  • Laurent MENUT,
  • Catherine Prigent,
  • Martina Klose,
  • Carlos Pérez García-Pando,
  • David Lawrence
Danny Min Leung
University of California, Los Angeles

Corresponding Author:dannymleung@ucla.edu

Author Profile
Jasper Kok
University of California, Los Angeles
Author Profile
Longlei Li
Cornell University
Author Profile
Natalie Mahowald
Cornell University
Author Profile
Laurent MENUT
Laboratoire de Météorologie Dynamique
Author Profile
Catherine Prigent
Observatoire de Paris-Meudon
Author Profile
Martina Klose
Barcelona Supercomputing Center
Author Profile
Carlos Pérez García-Pando
ICREA, Catalan Institution for Research and Advanced Studies
Author Profile
David Lawrence
National Center for Atmospheric Research
Author Profile

Abstract

A key challenge in accurate simulations of desert dust emission is the parameterization of the threshold wind speed above which dust emission occurs. However, the existing parameterizations yield a unrealistically low dust emission threshold in some climate models such as the Community Earth System Model (CESM), leading to higher simulated dust source activation frequencies than observed and requiring global tuning constants to scale down dust emissions. Here we develop a more realistic parameterization for the dust emission threshold in CESM. In particular, we account for the dissipation of surface wind momentum by surface roughness elements such as vegetation, rocks, and pebbles, which reduce the wind momentum exerted on the bare soil surface. We achieve this by implementing a dynamic wind drag partition model by considering the roughness of the time-varying vegetation as quantified by the leaf area index (LAI), as well as the time-invariant rocks and pebbles using satellite-derived aeolian roughness length. Furthermore, we account for the effect of soil size on dust emission threshold by replacing the currently used globally constant soil median diameter with a spatially varying soil texture map. Results show that with the new parameterization dust emissions decrease by 20–80% over source regions such as Africa, Middle East, and Asia, thereby reducing the need for the global tuning constant. Simulated dust emissions match better in both spatiotemporal variability and emission frequency when compared against satellite observed dust activation frequency data. Our results suggest that including more physical dust emission parameterizations into climate models can lessen bias and improve simulation results, possibly eliminate the use of empirical source functions, and reduce the need for tuning constants. This development could improve assessments of dust impacts on the Earth system.