Modelling day-time concentrations of biogenic volatile organic compounds in a boreal forest canopy

Abstract. Three different models for day-time atmospheric methanol, acetaldehyde, acetone, isoprene and monoterpene concentrations were developed using measurements above a boreal forest stand in Southern Finland in 2006–2007 and tested against an independent dataset from the same forest measured in summer 2008. The models were based on the exponential relationship between air temperature and the concentration of biogenic volatile organic compounds (BVOC). Our first model for BVOC concentrations was a simple exponential function of air temperature (T-model). The T-model could explain 27–66% of the variation of all the compounds, but it failed to catch the extremely high concentration peaks observed in summer. To improve the temperature model we developed two other models. The second model, a Temperature-State of Development- model (T-S model), included two explaining variables: air temperature and the seasonal photosynthetic efficiency. This model performed slightly better compared to the T-model for both datasets and increased the fraction of variation explained to 29–69%, but it still could not explain the high concentration peaks. To explain those we modified the T-S model to include environmental triggers that could increase the concentrations momentarily. The triggers that improved the model most were high photosynthetically active photon flux density (PPDF) compared to the seasonally available radiation and high ozone concentration. The Trigger model described the peak concentrations somewhat better than T or T-S model, thus the level of explanation was improved and was 30–71%. This study shows the importance to include seasonal variations in photosynthetic efficiency when modeling BVOC concentrations and presents the idea of a trigger model for explaining high peak concentrations of BVOCs. Our study suggests that when developing a trigger type modelfurther the model and the triggers should be more compounds-specific.