Time Stable Representative Position determination as affected by the considered part of an irrigation cycle

•Redistribution during and after irrigation rises uncertainty in defining temporal stability.•Data obtained during and shortly after irrigation should not be used to detect temporal stability.•The measurements period used to evaluate temporal stability affects optimum sensor positioning.•Evapotranspiration prediction depends on the positioning of sensors for data collection. A recurrent issue in irrigation management refers to the number and optimal positioning of water content sensors in the root zone of a crop. In the field, the definition of the number and location of sensor installation is still arbitrary. How many sensors and where to install them has been studied using the Time Stable Representative Positions (TSRP) concept introduced by Soulis and Elmaloglou (2016). However, the effect of the period of the soil water content data used for the determination of the TSRP is not usually considered. The main objectives of this work are to analyze how the soil water content data collection period within an irrigation cycle affects the determined optimal sensor positioning based on TSRP; and, to evaluate if the sensors positioning affects the prediction of crop evapotranspiration (ETc) by numerical modeling. Over 60,000 soil water content data were obtained during 36 days from 12 TDR probes installed at different monitoring positions in the root zone of an irrigated banana plant. The optimal position of sensors was determined based on the concept of TSRP, considering soil water content data obtained during the entire and during the second part of the irrigation cycles. The SWAP hydrological model was also used to investigate the effect of sensors positioning in prediction of ETc by numerical modeling. It was verified that the use of soil water content data obtained at the beginning of an irrigation cycle - at times when infiltration occurs and there is a high intensity of redistribution of water in the soil - increases the uncertainties regarding the estimation of temporal stability for purposes of irrigation management. The result of the determination of the optimal positioning for sensors installation varies according to the part of the irrigation cycle from which water content measurements are considered. This variation affects the prediction of ETc by numerical modeling. As values of soil water content obtained at the beginning of the irrigation cycle increase the uncertainty of the statistical indicators and are not of practical interest for irrigatio