In this study, we analyse the spatio-temporal variability of evapotranspiration (ET) fluxes within a small maize cropped field. ET fluxes were determined with a soil water balance approach on 28 different locations during two long dry periods of 24 days during the summer of 1999. Results of this study show firstly that the estimated ET fluxes are in the vast majority normally distributed. Furthermore, our results show that there is no apparent spatial dependence between the ET fluxes at least at the considered scale. Secondly, the study show that the spatial variability of the ET fluxes is quite large with coefficients of variation ranging between 15 and 46 %. It is worth to note that this high variability tends to decrease inversely to the length of the considered period. The high spatial variability suggests that an intensive sampling is necessary to estimate the mean value with a good precision. Indeed, our results show that the estimation of the mean ET flux with an accuracy of 10% and 20% for a probability level of 95 % requires respectively between 10 and 90 samples, and between 4 and 22 samples. This study shows further that the spatially variable ET fluxes are well correlated with the Leaf Area Index and the dry matter yield (r=0.65) measured on each location. This suggests that these variables could be used as auxiliary variables for the upscaling of locally measured ET fluxes. In conclusion, this study shows that the evapotranspiration process is effectively spatially variable even within a small apparently homogeneous cropped field. Such observed variability should be taken into account into distributed hydrological models.
Hupet, F., & Vanclooster, M. (2003). Spatio-temporal variability of evapotranspiration fluxes within a small agricultural field. Geophysical Research Abstracts, p. EAE03-A-5323. https://hdl.handle.net/2078.5/154289