Usually, Ozone can be as a tracers of the background pollutants from natural sources. The inclusion of Ozone is an indirect estimation of the nitrogen dioxide from combustion processes. The period of simulation must be strictly related to the inclusion of the seasonally effects on yours model.

AOT40 is usually calculated over the growing season of a crop. Thus, you need at least the data for that particular period. We normally calculate AOT40 as 3 month average (running average if you have data for longer period) for rice and compare with the WHO guideline, for example, to assess potential effect on crop yield.

I can suggest to include, together with Ozone, also the NOx to evaluate this effect. You can use a multi regression model to elaborate the data.