TM3 results compared with SONEX observations Ernst Meijer, Michiel van Weele, and Bram Bregman
Focus: role NMHCs on HO x chemistry Analysis of model results: Two simulations: with and without NMHC chemistry Output of mixing ratios and chemical rates that contribute to the OH concentration Each model time step linear interpolation of relevant model parameters to the position of the observation (1-minute averaged) Observed chemical rates were constructed from TM3 rate constants (JPL) and observed concentrations, temperature, water vapour and pressure. Only samples between hPa were considered to restrict the analysis to the UTLS region
NMHC-induced changes in HO x (July) HO x with NMHCs HO x HO x (%) HO x without NMHCs
Comparison of tracer concentrations: O 3
Comparison of tracer concentrations: CO
Comparison of tracer concentrations: OH and HO 2
Average modelled and observed OH rates
Improvement mainly due to removal of observations of very high NO x that were strongly underestimated by TM3 is approx. P(O 3 )! 70% of the samples remain Criterion: NO x /NO y > 0.5 NO y > 500 pptv For background UTLS chemistry
For air exposed to convection and lightning
Summary NMHCs play an important role in UTLS chemistry. Model predictions: [OH] up to -10% in July, UTLS, NH [HOx] up to +20% in July, UTLS, NH There is a good correlation between modelled and observed HO x, but the model gives too low mean values. The TM3 model is capable of simulating background UTLS chemistry with a good agreement for the rate of NO+HO 2. For air exposed to recent convection and/or lightning the agreement is much worse. This is mainly due to inefficient convective input of air with high NO x concentrations from the BL.