Search

Irrigated agriculture is threatened by soil salinity in numerous arid and semi-arid areas of the world, chiefly caused by the use of highly salinity irrigation water, compounded by excessive evapotranspiration. Given this threat, efficient field assessment methods are needed to monitor the dynamics of soil salinity in salt-affected irrigated lands and evaluate the performance of management strategies. In this study, we report on the results of an irrigation experiment with the main objective of evaluating time-lapse inversion of electromagnetic induction (EMI) data and hydrological modelling in field assessment of soil salinity dynamics. Four experimental plots were established and irrigated 12 times during a 2-month period, with water at four different salinity levels (1, 4, 8 and 12 dS m−1) using a drip irrigation system. Time-lapse apparent electrical conductivity (σa) data were collected four times during the experiment period using the CMD Mini-Explorer. Prior to inversion of time-lapse σa data, a numerical experiment was performed by 2D simulations of the water and solute infiltration and redistribution process in synthetic transects, generated by using the statistical distribution of the hydraulic properties in the study area. These simulations gave known spatio-temporal distribution of water contents and solute concentrations and thus of bulk electrical conductivity (σb), which in turn were used to obtain known structures of apparent electrical conductivity, σa. These synthetic distributions were used for a preliminary understanding of how the physical context may influence the EMI-based σa readings carried out in the monitored transects as well as being used to optimize the smoothing parameter to be used in the inversion of σa readings. With this prior information at hand, we inverted the time-lapse field σa data and interpreted the results in terms of concentration distributions over time. The proposed approach, using preliminary hydrological simulations to understand the potential role of the variability of the physical system to be monitored by EMI, may actually allow for a better choice of the inversion parameters and interpretation of EMI readings, thus increasing the potentiality of using the electromagnetic induction technique for rapid and non-invasive investigation of spatio-temporal variability in soil salinity over large areas.

Management of saline water irrigation at the field scale requires electrical conductivity to be monitored regularly in the generally shallow soil layer explored by plant roots. Non-invasive electromagnetic induction (EMI) sensors might be valuable for evaluating large-scale soil salinity. However, obtaining information on soil surface salinity from depth-integrated EMI measurements requires either an inversion of the electromagnetic signal or an empirical calibration. We opted for an empirical calibration of an EM38 sensor, which requires a reference dataset of local bulk electrical conductivity, σb, for comparison with EMI readings for estimating regression coefficients. We used time-domain reflectometry (TDR) to replace direct sampling for local σbmeasurements. With empirical approaches, the different soil volumes involved with the EMI and TDR sensors become problematic. We resolved this issue by analysing large EMI and TDR datasets recorded at several times along three transects irrigated with water at 1, 3 and 6 dS m−1degrees of salinity. A Fourier filtering technique was applied to remove the high frequency part (at small spatial scales) of the variation in the original data, which was the main source of dissimilarity between the two datasets. Therefore, calibration focused on the lower frequency information only; that is, information at a spatial scale larger than the observation volume of the sensors. We show that information from the TDR observations derives from a combination of local and larger scale heterogeneities, and how it should be managed for calibration of the EMI sensor. Analysis enabled us to identify characteristics of the calibration data that should be included to improve prediction. Highlights: EMI and TDR sensors have different observation volumes and are not immediately comparable. With Fourier filtering the different observation volumes were accounted for in calibration of the EMI data. We improved the empirical calibration between EMI and TDR datasets. We identified criteria to select data series for EMI sensor calibration.

The knowledge of the longitudinal dispersivity parameter is necessary to simulate solute transport in porous media. The direct measurement of this parameter is expensive and time consuming. Indirect methods based on the inverse modelling procedure provide an easier and reliable alternative, if measurements of solute concentration and soil water content are available. In this study, Hydrus 1D model was coupled with apparent electrical conductivity measured by means of the simple and rapid electromagnetic induction (EMI) technique in order to determine the longitudinal dispersion parameter in a sandy soil. By means of forward Hydrus 1D simulations the temporal distribution of soil water content in 1 m depth soil profile was obtained. The soil water content and the inverted soil electrical conductivity were implemented in the Rhoades linear equation in order to determine the distribution of the soil solution electrical conductivity. Thereby, the longitudinal distribution parameter was finally determined by carrying out Hydrus 1D inversions. The estimated dispersion parameter was 10.23 cm. The proposed procedure allows to obtain reasonable values of longitudinal dispersion with a good degree of approximation.

Recent research shows how soil information can be generated quickly and cheaply, by using a state-of-the-art electromagnetic (EM) instrument and the inversion method to generate high-resolution EM conductivity images (EMCI). EM instruments have the advantages of being faster, less expensive, and also to collect data easier than the common Electrical Resistivity Tomography (ERT) method, which makes EM adequate to cover a larger area in regional studies. However, the capability of the technique is limited in mapping soil dynamic parameters (i.e. water content, salinity) due to the lack of a time-lapse inversion algorithm. While several time-lapse inversion algorithms have been developed for modeling of time-lapse ERT data in order to reduce the inversion artefact, no attempt has been made to implement a time-lapse inversion algorithm to invert the EM data. Developing a time-lapse inversion algorithm for EM data is, therefore, a determined step forward for application of EM instruments in characterization of the soil's dynamic parameters where a multiple datasets is required. We developed a time-lapse inversion algorithm where a spatiotemporal objective function is minimized and we evaluated the algorithm by performing several synthetic tests. The developed program will be applied to the field data collected in the scope of the SALTFREE project to image spatiotemporal variability of soil salinity. The project consortium is formed by five partners from four countries around the Mediterranean - Egypt, Italy, Portugal, and Tunisia in the scope of ARIMNET2 program and aims to develop a framework for the evaluation of the salinization risk in the management scale.

This paper dealt with the calibration of an EMI sensor for monitoring the time dynamics of root zone salinity under irrigation with saline water. Calibration was based on an empirical multiple regression approach largely adopted in the past and still applied in practice for its relative simplicity. Compared to the more complex inversion approaches, it requires an independent dataset of local σb measured within discrete depth intervals, to be compared to horizontal and vertical electrical conductivity (ECaH and ECaV) readings for estimating the parameters of the empirical regression equations. In this paper, we used time domain reflectometry (TDR) readings to replace direct sampling for these local σb measurements. When using this approach, there is the important issue of taking into account the effect of the different sensor observation volumes, making the readings not immediately comparable for empirical calibration. Accordingly, a classical Fourier’s filtering technique was applied to remove the high frequency part (at small spatial scale) of the original data variability, which, due to the different observation volume, was the main source of dissimilarity between the two datasets. Thus, calibration focused only on the lower frequency information, that is, the information at a spatial scale larger than the observation volume of the sensors. By this analysis, we showed and quantified the degree to which the information of the set of TDR readings came from a combination of local and larger scale heterogeneities and how they have to be manipulated for use in EMI electromagnetic induction sensor calibration.

The use of treated wastewater (TWW) as an alternative resource to freshwater (FW) for irrigation purposes is becoming increasingly important, especially in semiarid and arid regions. However, achieving success in crop production largely depends on the adoption of appropriate on-farm management strategies aimed at optimizing crop yields, maintaining soil productivity and safeguarding the environment. For this purpose, predictive models are of particular interest. A safe irrigation management (SIM) model developed and tested in this research was used to schedule irrigation under controlled management tailored to the use of i) TWW and ii) FW and to reproduce farmers' strategies. These management strategies aim to improve actual irrigation practices, accounting for water quality, soil characteristics and crop yield. The results of the application of SIM on a citrus farm in Souss-Massa, Morocco, show that the management strategy adopted by farmers withdraws considerable amounts of water and results in substantial drainage volumes compared to those in the SIM strategy. In the specific case of TWW, the strategy simulated by the SIM model resulted in a decrease in yield of approximately 4%, compared to the 23% decrease derived from the farmers' traditional strategy. Moreover, SIM allowed for great savings in terms of fertilizing elements and for the reduction in the movement of water and salts beyond the root zone, usually considered the main source of groundwater contamination. These results confirm the appropriateness of using prediction models and the accuracy of the SIM model in adapting irrigation strategies to TWW, which will be an integral part of the strategies that encourage their use in irrigated agriculture. This article is protected by copyright. All rights reserved.