Toward Optimal Irrigation Management at the Plot Level: Evaluation of Commercial Water Potential Sensors

Proper irrigation practice consists of applying the optimum amount of water to the soil at the right time. The porous characteristics of the soil determine the capacity of the soil to absorb, infiltrate, and store water. In irrigation, it is not sufficient to only determine the water content of the soil; it is also necessary to determine the availability of water for plants: water potential. In this paper, a comprehensive laboratory evaluation-accuracy and variability-of the world's leading commercial water potential sensors is carried out. No such comprehensive and exhaustive comparative evaluation of these devices has been carried out to date. Ten pairs of representative commercial sensors from four different families were selected according to their principle of operation (tensiometers, capacitive sensors, heat dissipation sensors, and resistance blocks). The accuracy of the readings (0 kPa-200 kPa) was determined in two soils of contrasting textures. The variability in the recordings-repeatability and reproducibility-was carried out in a homogeneous and inert material (sand) in the same suction range. The response in terms of accuracy and value dispersion of the different sensor families was different according to the suction range considered. In the suction range of agronomic interest (0-100 kPa), the heat dissipation sensor and the capacitive sensors were the most accurate. In both families, registrations could be extended up to 150-200 kPa. The scatter in the readings across the different sensors was due to approximately 80% of the repeatability or intrinsic variability in the sensor unit and 20% of the reproducibility. Some sensors would significantly improve their performance with ad hoc calibrations.

Challenges and progresses in the detailed estimation of sediment export in agricultural watersheds in Navarra (Spain) after two decades of experience

Soil erosion is a very serious environmental problem worldwide, with agriculture considered the main source of sediment in inland waters. In order to determine the extent and importance of soil erosion in the Spanish region of Navarra, in 1995 the Government of Navarra established the Network of Experimental Agricultural Watersheds (NEAWGN), which consists of five small watersheds representative of local conditions. In each watershed, key hydrometeorological variables, including turbidity, were recorded every 10 min, and daily samples were taken to determine suspended sediment concentration. In 2006, the frequency of suspended sediment sampling was increased during hydrologically relevant events. The main objective of this study is to explore the possibility of obtaining long and accurate time series of suspended sediment concentration in the NEAWGN. To this end, simple linear regressions between sediment concentration and turbidity are proposed. In addition, supervised learning models incorporating a larger number of predictive variables are used for the same purpose. A series of indicators are proposed to objectively characterize the intensity and timing of sampling. It was not possible to obtain a satisfactory model for estimating the concentration of suspended sediment. This would be mainly due to the large temporal variability found of the physical and mineralogical characteristics of the sediment, which would be affecting the turbidity value, independently of the sediment concentration, per se. This fact would be particularly important in small river watersheds such as those of this study, and especially if their physical conditions are spatially and temporally radically disturbed by agricultural tillage and by a constant modification of the vegetation cover, as is the case in cereal basins. Our findings suggest that better results could be obtained by including in the analysis variables such as soil texture and exported sediment texture, rainfall erosivity, and the state of vegetation cover and riparian vegetation.