For more information about this poster please contact Gerard Devine School of Earth and Environment Environment University of Leeds Leeds LS2 9JT Email gerardenvleedsacuk Tel(044) (0113) 3431598

The influence of subgrid variability on chemical transport in a deep convective environment

Gerard Devine1 Doug Parker1 Ken Carslaw1 Jon Petch21School of Earth and Environment University of Leeds

2UK Meteorological Office

Deep convective clouds play an important role in the transport of chemical species from the planetary boundary layer (PBL) to the upper troposphere and lower stratosphere In modelling such transport global chemical transport models resolve scales larger than that of a typical convective cloud event However the subgrid-scale dynamical features of convective cloud systems may be important in controlling the abundance of chemical species in the PBL and ultimately how much is transported vertically In this study we use a 2-D cloud-resolving model (CRM) to examine the influence of such subgrid-scale features on the concentration and vertical transport of dimethyl sulphide (DMS) in a deep convective environment

Sensitivity Experiments

EXPERIMENT WIND FLUX DMS

CRMRES Resolved Resolved Resolved

AVWIND AveragedResolved from

averaged windsResolved

AVFLUX Resolved Averaged Resolved

AVDMS Resolved Resolved Averaged

Convective Simulation Model forced using observational data from TOGA-COARE

Chemistry Setup

Smooth

surface

Rough surfac

e

Breaking Waves

We simulate dimethyl sulphide (DMS) an important pre-cursor gas to sulphur dioxide (SO2) and sulphate aerosol

CRMRES ndash 10 m wind DMS flux and DMS concentration all fully resolved

Model Setup

UK Met Office Large Eddy Model (LEM)

Domain width of 256 km and domain height of 20 km Horizontal resolution - 1 km Vertical resolution at 20 m near the surface rising to 500 m at domain top

(1) Deriving fluxes of DMS from the ocean surface using a spatially averaged surface wind representative of a global model reduces the domain-mean DMS concentration by approximately 50 Emission of DMS from the sea surface is greater in the CRM because it resolves the localized high wind speeds embedded in the dynamical structures associated with the convective cloud systems

6-day simulation characterized by periods of active convection

The figure opposite shows x-t plots of (a) surface precipitation rate above 1 mmhr and (b) 10 m horizontal wind speed (ms)

Tropical OceanGlobal Atmosphere ndash Coupled Ocean Atmosphere Response Experiment

Sea-air flux of DMS based on parametrization by Liss and Merlivat [1986]

Parametrization consists of three wind lsquoregimesrsquo resulting in a piecewise linear relationship (see figure opposite)

Sink term representing oxidation by the hydroxyl radical is also included

AVWIND ndash 10 m wind vector averaged across the domain at each timestep DMS flux calculated from resultant average wind

AVFLUX ndash 10 m wind resolved but an averaged flux field applied across the domain at each timestep

AVDMS ndash 10 m wind field and DMS flux resolved but the resultant DMS concentration in the boundary layer horizontally averaged at each timestep

Results Summary

Boundary layer

Mid troposphe

re

Time evolution of average DMS (ppt) in each of three atmospheric regions

CRMRESAVWINDAVFLUXAVDMS

Mid and upper troposphere ndash

concentration significantly underpredicted in a model

with averaged wind and averaged DMS fields

Mid and upper troposphere ndash

concentration significantly underpredicted in a model

with averaged wind and averaged DMS fields

We have highlighted two issues

(2) The spatial pattern of DMS concentration in the boundary layer is positively correlated with the pattern of convective updraughts Using a mean DMS concentration field reduces vertical transport to the upper troposphere by more than 50 The explanation is that secondary convection occurs preferentially on the edges of spreading cold pools where DMS concentrations are higher than the domain mean

Explanation

On several occasions the mean wind is in a less efficient lsquoregimersquo than the resolved winds associated with

convective activity

Upper troposphe

re

AVWIND

Mean Wind Maximum and minimum

resolved winds

Breaking wave regime

Rough surface regime

Smooth surface regime

AVDMS

Dotted line represents total

domain DMS surface flux for

CRMRES Dashed line

represents total domain DMS

surface flux for AVWIND

Comparing AVDMS and CRMRES shows strong updraughts are correlated with areas with higher than the horizontally averaged concentration of DMS

Discussion

Through the convective cold pool convective clouds create an environment in which vertical transport of DMS is enhanced

Low DMS in cold pool

Initiation of secondary

convective cells at leading edge

of spreading cold pool

High DMS outside of cold

pool

Ascending air

Timescale for recovery of DMS by enhanced fluxes within the

lsquogustyrsquo conditions of the cold pool is longer than that for secondary

initiation

Underestimates of the transport due to neglecting these effects may have significant consequences for predictions of the chemical state of the upper troposphere and lower stratosphere

Theta line depicting

convective cold pool