Radon global simulations with the multiscale chemistry and transport model MOCAGE
Peuch, V. -H.
- Publisher: Tellus B
(issn: 1600-0889, eissn: 0280-6509)
We present an evaluation of the representation of subgrid scale transport in the new multiscale global chemistry andtransport model MOCAGE. The approach is an off-line computation of vertical mass fluxes due to convective andturbulent processes, using only large-scale variables archived in meteorological analyses. Radon is a naturally emittedgas with a radioactive half-life of 3.8 days and is a useful tracer of tropospheric transport processes. A 1-yr (1999)simulation of atmospheric radon concentration has been performed, using 6-hourly meteorological analyses for theforcings. Two different mass flux convection schemes have been tested: a simplified version of the Tiedtke (1989)scheme and Kain–Fritsch–Bechtold (Bechtold et al., 2001). We compare model outputs with observations at differenttime and space scales, showing good overall results. A new interpretation is given to the more contrasted results obtainedin Antarctica, as for other models. The state-of-the-art representation of synoptic scale activity around Antarctica ismarkedly worse than in other parts of the world, both due to oversimplifications of the seasonal evolution of the extent ofsea ice, and to the scarcity of observations. Twelve-hourly simulated concentrations are evaluated at two sites for 1999.At Amsterdam Island results are satisfactory: correlation between observed and modelled concentrations is of the orderof 0.5. The model reproduces well “radonic storm” events. At the coastal site of Mace Head in Ireland, simulations areavailable at two different horizontal resolutions. The correlation between observations and the model is of the order of0.7. This result is mainly determined by the synoptic scale context, even though local-scale circulations such as breezesinterfere on occasions. Finally, it appears that the off-line approach in MOCAGE for subgrid transport is a practical onefor chemistry and transport multiscale modelling.DOI: 10.1111/j.1600-0889.2004.00112.x