DO3SE Model Analysis

a. DO3SE model:
Basically, DO3SE model is derived from the Jarvis-type multiplicative algorithm:

gsto = gmax X min (fphen, fO3) X flight X max (fmin, (fVPD X ftemp) (1)

where gsto is the actual stomatal conductance for O3 (mmol O3 m-2 unit projected leaf area (PLA) s-1), gmax is the species-specific maximum gsto (mmol O3 m-2 PLA s-1), and fmin is the relative minimum gsto (fraction of gmax). The parameters including ftemp, fphen, flight, fO3 and fVPD are the response functions expressed in relative terms. These factors represent for the influences of phenology, O3 (in terms of PODo, accumulated hourly stomatal O3 flux above a flux threshold of 0 nmol O3 m-2 PLA s-1) and the environmental variables including air temperature, light (photosynthetic
2. flight
The response function flight of gsto to light is given in Equation 5 flight = 1 － exp((-Light_a) × PPFD) , (5) where PPFD is the photosynthetic photon flux density (μmol m-2 s-1) and Light_a is the parameter.
3. ftemp
The response function ftemp of gsto to air temperature is given after LRTAP (2010) as follows: , for Tmin < T < Tmax (6) ftemp = fmin, for T ≥ Tmax or T ≤ Tmin, (7) where T is the air temperature (°C), Tmin and Tmax are the minimum and maximum temperatures, respectively, when the temperature is most limiting gsto, Topt is the optimum temperature, and bt is defined 4. fVPD
The response function fVPD of gsto to humidity is given in Equation 8 , (8) where VPD is the atmospheric vapor pressure deficit (kPa), and VPDmin, VPDmax, and fmin are the parameters.
5. fO3
The response function fO3 of gsto to the ozone-induced senescence is given in Equation 9. ,
Monthly average of modelled ozone values were consistently higher than the observation at all stations. Model results were somehow seriously overestimated in Nha Be station but relatively better in other stations. However, in term of seasonal variation of ozone that is important for assessment on crops, model seemed to be able to capture the pattern. Higher concentrations were seen in the dry months as compared to the wet months due to higher solar intensity and temperature in addition to the local build-up of precursor emissions. However, the discrepancy can be also affected by the coarse resolution of the model output as compared to the point-based observation station. In the remote site model result seemed to be better than the complex mixed site of emission sources such as in the urban