Method for automatic determination of soybean actual evapotranspiration under open top chambers (OTC) subjected to effects of water stress and air ozone concentration

The present study describes an operational method, based on the Katerji et al. (Eur J Agron 33:218-230, 2010) model, for determining the daily evapotranspiration (ET) for soybean inside open top chambers (OTCs). It includes two functions, calculated day par day, making it possible to separately take into account the effects of concentrations of air ozone and plant water stress. This last function was calibrated in function of the daily values of actual water reserve in the soil. The input variables of the method are (a) the diurnal values of global radiation and temperature, usually measured routinely in a standard weather station; (b) the daily values of the AOT40 index accumulated (accumulated ozone over a threshold of 40 ppb during daylight hours, when global radiation exceeds 50 Wm(-2)) determined inside the OTC; and (c) the actual water reserve in the soil, at the beginning of the trial. The ensemble of these input variables can be automatable; thus, the proposed method could be applied in routine. The ability of the method to take into account contrasting conditions of ozone air concentration and water stress was evaluated over three successive years, for 513 days, in ten crop growth cycles, excluding the days employed to calibrate the method. Tests were carried out in several chambers for each year and take into account the intra- and inter-year variability of ET measured inside the OTCs. On the daily scale, the slope of the linear regression between the ET measured by the soil water balance and that calculated by the proposed method, under different water conditions, are 0.98 and 1.05 for the filtered and unfiltered (or enriched) OTCs with root mean square error (RMSE) equal to 0.77 and 1.07 mm, respectively. On the seasonal scale, the mean difference between measured and calculated ET is equal to +5% and +11% for the filtered and unfiltered OTCs, respectively. The ability of the proposed method to estimate the daily and seasonal ET inside the OTCs is therefore satisfactory following inter- and intra-annual tests. Finally, suggestions about the applications of the proposed method for other species, different from soybean, were also discussed.