Estimating Actual Transpiration of Apple Trees Based on Infrared Thermometry

A method was developed based on the radiative properties and energy budget of a single apple leaf to calculate the actual transpiration ($T$) of apple trees. The model uses canopy temperature ($T_c$), air temperature ($T_a$) measured in the orchard, and other meteorological data from a local weather station as inputs. The model was applied to two scenarios, as follows: (1) well-watered, young Fuji apple trees in the 2007 and 2008 growing seasons; and (2) older apple trees, bearing little fruit in the 2013 growing season. Simulated transpiration rates at both scenarios were compared with Penman-Monteith (PM) model predictions corrected by regionally adjusted crop coefficients, i.e., values of $E{T}_c$. In 2007 and 2008, a linear regression analysis of the relationship between daily mean transpiration ($T_{\mathrm{avg}}$) and $E{T}_c$ revealed that they better agreed on warm and dry days (correlation coefficient $R^2=0.57$, $\mathrm{slope}=1.16$, and $\mathrm{intercept}=0.4$) than during cold and humid periods ($R^2=0.48$, $\mathrm{slope}=0.69$, and $\mathrm{intercept}=2.3$). Combining the results of the 2007 and 2008 seasons, $T_{\mathrm{avg}}$ and $E{T}_c$ presented a fairly good agreement, with the relationships $R^2$, slope, and intercept of 0.77, 1.0, and 1.08, respectively. In 2013, the actual water use calculated by a soil water budget approach ($E{T}_{WB}$) was considerably less than $E{T}_c$ while there was no significant difference between the total simulated transpiration ($\mathrm{\Sigma }{T}_{\mathrm{avg}}$) and $E{T}_{WB}$. In 2013, a linear regression analysis of the relationship between midday $T$ ($T_{\mathrm{mid}}$), $T_{\mathrm{avg}}$, and midday stem water potential (${\mathrm{\Psi }}_{\mathrm{stem}}$) showed they were highly correlated ($T_{\mathrm{mid}}$, $R^2=0.85$; $T_{\mathrm{avg}}$, $R^2=0.87$). The experiments presented varied results on the linear relationship between air vapor pressure deficit ($D_a$) and $T_{\mathrm{avg}}$ from year to year. On the other hand, canopy and air temperature difference ($\mathrm{\Delta }{T}_m$) was linearly related to $T_{\mathrm{avg}}$ in all of the seasons. According to the model for the actual transpiration (i.e., the T-model), the apple trees had an intense transpiration in the morning and then there was a decline around solar noon. The transpiration increased again late in the afternoon. Real-time estimations of water use using the T-model can provide a basis for a fully automated system of irrigating apple orchards.