A source-orientated approach for estimating daytime concentrations of biogenic volatile organic compounds in an upper layer of a boreal forest canopy

Biologically justified statistical models for daytime atmospheric concentrations of methanol, acetaldehyde, acetone, isoprene and monoterpene were tested using measurements at a boreal forest stand in southern Finland in 2006―2007 and in summer 2008. The canopy-scale concentrations of all compounds except monotepene were closely correlated with shoot-scale concentrations indicating a strong link to biological emission source. All the models were based on the exponential relationship between air temperature and atmospheric concentration of biogenic volatile organic compounds (BVOCs). The first model — an exponential function of air temperature (T model) — could explain 27%―64% of the variation in BVOC daytime concentrations in the test data. The second model — a Temperature-State of Development model (T-S model) having two explaining variables (air temperature and seasonal photosynthetic efficiency) — was derived from an empirical adjustment of seasonality. This model slightly increased the fraction of explained variation but it still could not explain the high concentration peaks, which accounted for most of the unexplained variation. To better analyse these peaks we tested the Trigger model including two potential environmental triggers, a PAR index (high photosynthetically active photon flux density (PAR) and high ozone concentration, that could increase the concentrations momentarily. However, the Trigger model described the peak concentrations only somewhat better than the T or T-S model. It seems that it is very difficult to explain more than 32%―67% of variation in BVOC concentrations by a straightforward source-oriented modelling without deep understanding of biological and physical processes. In order to improve the models profound studies on specific stress factors and events inducing BVOC emissions are needed.