Trace gas constraints on vertical transport in models: a case study of Indonesian biomass burning emissions in 2006

Abstract:

We investigate the use of joint Aura TES and MLS CO retrievals in constraining vertical transport in the NASA GISS ModelE2 composition-climate model. We examine September to November 2006 over the tropics. El Nino-induced dry conditions over western Indonesia led to extensive biomass burning and persistent CO greater than 200 ppb in the upper troposphere. This was one of the highest CO episodes over the MLS period since 2004. We show how improvements in the vertical resolution of trace gas retrievals can help to distinguish between errors in parameterized vertical transport and biases in bottom-up emissions estimates.

We simulate the episode using the NASA GISS ModelE2 coupled composition-climate model with different subgrid physics for small ensembles of experiments with perturbed initial conditions. The starting point is the CMIP5 version of the model, in which there was a pronounced vertical CO dipole over the Maritime Continent, but with a CO peak 100 ppb higher than Aura CO in the upper troposphere. With modified cumulus and boundary layer parameterizations, but the same prescribed biomass burning emissions estimates, the upper tropospheric CO bias is significantly reduced. Concurrently, precipitation over the emissions source region is reduced relative to observational estimates, leading to better consistency with the dry conditions under which the burning occurred. We discuss the effects of the physics changes on the roles of convective frequency and depth in reducing the bias.