TDR Probes Research Articles

Water relations and photosynthesis as criteria for adequate irrigation management in 'Tahiti' lime trees

Irrigation scheduling based on soil moisture status is one of the most useful methods because of its practicality and low cost. The effects of available soil water depletion on evapotranspiration (ETc), transpiration (E), leaf water potential at predawn (psiP) and midday (psiM), stomatal conductance (gs) and net CO2 assimilation (A) in lime 'Tahiti' trees (Citrus latifolia) were evaluated to improve irrigation schedule and minimize water use without causing water stress. The trees were spaced 7 <FONT FACE=Symbol>´</FONT> 4 m and drip-irrigated by four drippers with the available soil water content (AWC) depleted by suspension of irrigation (40 days). Leaf water potential was measured on a pressure chamber (psiP and psiM) and leaf gas exchange was measured by infrared gas analyzer (E, gs and A). Evapotranspiration was determined with the aid of weighing lysimeter. Water soil content and potential (psiS) were monitored with TDR probes and tensiometers, respectively, installed at 0.3, 0.6 and 0.9 m depths. Meteorological variables were monitored with an automatic weather station in the experimental area. The threshold AWC level for the onset of ETc decline was 43%, and 60% for gs, A, E and Y P. Also, psiP was more sensitive to AWC than psiM, and is therefore a better tool for irrigation. When AWC was around 60%, values of psiP and psis were -0.62 MPa and -48.8 kPa, respectively.

Drainage and nitrate leaching under processing tomato growth with drip irrigation and plastic mulching

Abstract Irrigated agricultural areas are particularly susceptible to groundwater pollution because irrigated crops are abundantly fertilized and occasionally overwatered. Our goal was to evaluate the close link between crop production and groundwater protection. The specific objective of our research was to quantify the effect of three irrigation strategies on yield, drainage and nitrate leaching in processing tomato grown using drip irrigation and plastic mulch. We also assess the effect of high soil mineral N content on nitrate leaching. The study was conducted in a field cropped with tomato ( Lycopersicum esculentum Mill), cultivated under three different irrigation strategies during 2001 and 2002 in the Ebro Valley. The strategies were combinations of two rates and two frequencies of water application. During the establishment period, water was either applied continuously for 8 h, or as 8 applications of 15 min. During the growth period, water was applied either as a one single daily irrigation of 100% ET c , or as 4–8 daily irrigations to 80% ET c . Monitoring transects, composed of 6 TDR probe access tubes inserted into the soil, provided measurements of soil water content in each treatment. Drainage was calculated weekly by applying the water balance to all tubes located in each transect. Nitrate leaching was evaluated as the drainage volume multiplied by the nitrate concentration of the soil solution extracted in 6 porous ceramic cups located in each treatment. At harvest, total and marketable production were recorded for each plot, and the dry weight of the component aboveground plant parts (stems, leaves and fruits) were determined. Crop N uptake for the plant parts was calculated as the product of total biomass and N concentration. Our results showed that excessive irrigation during the crop establishment period may cause large nitrogen losses. High irrigation frequency ensured appropriate soil water content at planting, and reduced both the amount of water applied and lost by drainage. Treatments receiving water at 80% of ET c , recorded almost no drainage during the growth period and no reduction in crop yield. This suggests that further studies are required to adjust crop coefficients to the actual needs of drip-irrigated tomato under plastic mulching conditions. When soil mineral N content before planting is high, nitrate leaching during the growth period can be relevant due to occasional rain events, even if water application is adjusted to crop requirements.

Soil water content inverse profiling from single TDR waveforms

An inverse procedure for the estimation of soil water content profiles along TDR probes is presented. A TDR metallic probe is considered as a transmission line, for which relevant partial derivatives equations apply. The direct problem consists in the integration of transmission line equations, providing V( x, t) along the line. To this aim, the unit length parameters of the transmission line must be known. In particular, unit length capacitance C( x) and transverse conductance G( x) depend on water content distribution along the probe θ( x) through relative permittivity ε r( x) and bulk soil electrical conductivity σ( x), respectively. The inverse procedure consists in finding the water content distribution, and the relevant unit length parameters, giving rise to the best fit between the numerically simulated voltage V ( x ¯ , t ) at the beginning of the line and the experimental voltage trace V exp ( x ¯ , t ) measured by a cable tester. In order to reduce the ill-posedness of the inverse problem, unknown water content profiles are expressed by means of a four parameters functional form. The search for the best fitting parameters vector is carried out with a genetic algorithm. The proposed inverse procedure is successfully applied to the determination of vertical water content profiles along a soil sample in the laboratory by means of a single three rods metallic TDR probe. Water content profiles estimated either in steady flow conditions, or during controlled infiltration-evaporation transients are compared with independent water content measurements carried out by means of horizontal TDR probes at various depths, showing in all cases good agreement.

Estabilidade temporal de medidas do teor e do potencial mátrico da água no solo em uma transeção

O elevado nível de variabilidade espacial de observações experimentais de campo da condutividade hidráulica do solo dificulta o planejamento de amostragem e a utilização e extrapolação de valores médios, em estudos de campo de dinâmica da água no solo. Nesse sentido, o objetivo deste trabalho foi verificar a estabilidade temporal de medidas do teor e do potencial mátrico da água no solo e sua aplicação como ferramenta para racionalizar a estratégia de amostragem para avaliação da condutividade hidráulica do solo em condições de campo. O experimento de drenagem foi desenvolvido num Latossolo Vermelho-Amarelo em Piracicaba (SP), numa parcela experimental com 50 pontos de observação distanciados de 1 m entre si, nos quais foram medidos o teor de água (com TDR) e o potencial mátrico (com tensiômetros) durante três semanas de redistribuição da água no solo. Os resultados mostraram que o método da estabilidade temporal permite identificar, com exatidão, os locais mais adequados para as amostragens, possibilitando a redução do número de amostras necessárias e do custo de execução do esforço amostral para o planejamento de sistemas de uso e manejo de água na agricultura.

(372) Water Requirements of High bush Blueberry Cultivars Planted at Normal and High Density Spacings

Many growers in the Pacific Northwest are planting blueberry fields at higher densities to improve yields and increase the number of berries harvested per acre in the first few years after planting. The objective of this study was to determine the effect of high-density planting on blueberry water requirements. Although close spacing reduces individual plant size, we expected that plants spaced closer together would require more irrigation per unit land area than those spaced further apart due to increased canopy coverage within rows. The study utilized a 5-year-old planting of highbush blueberry, consisting of three cultivars, `Duke', `Bluecrop', and `Elliott', planted at 0.5- and 1.2-m in-row spacings. Plant water use was calculated from changes in soil water content measured using TDR probes for shallow depths and a neutron probe and access tubes for deeper depths. Stem water potentials were also measured periodically using a pressure chamber to determine how well irrigation was meeting crop water demands throughout the season. Surprisingly, plants spaced 0.5 m apart required only slightly more water than those spaced 1.2 m apart. They did, however, require more frequent irrigations due to their smaller root systems, especially during fruit filling. Water use by each cultivar increased during fruit filling and then rapidly decreased after harvest. `Duke' required the most water among cultivars, using 5–10 mm/day from mid-May to mid-August, while `Elliott' required the least, using 3–5 mm/day.

Measuring and Modeling of Direct Ground Wave Depth Penetration Under Transient Soil Moisture Conditions

This study represents the first attempt to investigate the C;PR direct ground wave sampling depth by comparing GPR estimated soil moisture contents with data from horizontally and vertically installed TDR probes at different depths. The GPR direct ground wave method (200 MHz centre frequency) was used to estimate the temporal soil moisture dependence during uniform irrigation and drainage. Uniform irrigation and drainage experiments were conducted in an experimental pit (2.5 x 1.01 x 0.8 m) filled with repacked sandy loam soil. The GPR moisture contents measurements were more consistent with the moisture contents from vertically installed TDR than horizontally installed TDR. An analytical solution for onedimensional drainage of water was used to estimate the change in GPR ground wave sampling depth during drainage. The analytical solution was first fit to vertical TDR data to obtain an estimate of the soil hydraulic parameters and the GPR sampling depth was then estimated by fitting the drainage solution to the measured GPR data. The GPR direct ground wave sampling depth using the analytical solution during drainage varied from - 20 cin at high moisture content to -50 cm at the lowest moisture content

Using a rainfall simulator and a physically based hydrological model to investigate runoff processes in a hillslope

A large rainfall simulator (LRS, area 970 m 2 ) is being used to measure flow, sediment and contaminant runoff from a pasture hillslope in the North Island of New Zealand. Results from a winter experiment corresponding to a 1 in 8 year storm were used to calibrate a complex, physically based hydrological model (SHETRAN, Ewen et al., 2002). The ability of SHETRAN to reproduce firstly the observed runoff hydrograph and secondly the soil moisture response, was assessed. Surface runoff was widely distributed over the simulator plot but was not uniform because of uneven topography and possibly because of spatial variations in soil properties. Most model parameters were determined from field measurements, but two (vertical hydraulic conductivity and surface friction) were ‘calibrated’ by matching the observed and predicted runoff hydrograph. The calibrated model had a coefficient of determination (R 2 ) of 98%, and reproduced accurately the shape of the hydrograph and the total volume of runoff. The model predicted soil moisture that matched TDR probe measurements within the calibration uncertainties, at one location, and underestimated the reduction in soil moisture following the end of rain, possibly indicating that the calibrated value of vertical hydraulic conductivity was too low. This study assumed no-flux boundary conditions under and around the LRS, and while this approximation may be acceptable when modelling a single storm, vertical and downslope drainage will need to be quantified in future longer-term applications. From both field data and model simulations, it was concluded that surface runoff occurred principally as a result of the infiltration excess (IE) mechanism. Saturation excess (SE) and variable source area (VSA) were rejected as the principal generation mechanisms. Copyright  2005 John Wiley & Sons, Ltd.

Étude de la migration du soluté d’un lixiviat dans un sol non saturé par la méthode TDR

The protection of the environment requires building waste containment systems made up of liners of compacted soil. In this paper, infiltration tests are carried out on soil columns equipped with TDR probes to investigate how water and pollutants migrate in the liners. The TDR method is a technique to measure simultaneously the volumetric water content and the electrical conductivity lor the impedance) of the soil in a continuous manner. Using six soils and three different leachates, new relationships between the volumetric water content (θ) of the soil and its dielectric constant (e) is established. These relationships appear to be more accurate than the one of Topp et al. (1980). They are used to determine water content profiles in a soil column. The relationship between the leachate solute concentration and the electrical conductivity allows to compute the solute concentration profiles in the soil column and the solute breakthrough curve. These curves are used to determine solute convection rate and solute apparent dispersion coefficient. Infiltration test results show that the soil permeability and the diffusivity increase when the leachate is infiltrated. The infiltration rate and the solute convection rate are higher than the leachate outflow rate and the three rates yield to a same value which depends on the soil permeability. The Green and Ampt (1911) model applied to the infiltration shows that the movement of the wet front is faster than the movement of the pollution front. The apparent dispersion coefficient and the solute convection rate are computed by applying the statistic distribution law to the moisture and the solute concentration profiles. It is observed that the apparent dispersion coefficient increases during the saturation of the sample. The electrical conductivity and the solute concentration of the percolated leachate are scaled with the initial values. The evolution curves of these scaled values with elapsed time are used to compute the apparent dispersion coefficient. There is little difference between the apparent dispersion values obtained with the scaled electrical conductivity evolution curve and the scaled solute concentration evolution curve. The use of the TDR measurements allows comparison between solute and water migration parameters and the obtained results can be applied to in situ permeability measurements.

Effects of tillage systems in dryland farming on near-surface water content during the late winter period

Tillage systems modify, at least temporarily, some of the physical properties of soil, such as soil porosity. Tillage also has an indirect effect on soil water content throughout the growth cycle, particularly in areas with a Mediterranean climate. This paper presents the results of monitoring the water content in the topsoil (0–0.20 m) of three adjacent plots during February to May cycles starting in 1994–1995 and ending in 1998–1999. Each of the plots had a surface area of 2700 m 2, an 8% slope and Calcic Cambisol soil. Starting in 1994, three different tillage systems were applied: conventional tillage, which is typical of the area (CT); minimum tillage (MT); and no-tillage (NT). Two vertical 200 mm TDR probes were permanently installed in each plot and measurements were taken every week. The results show that, under an NT system, the soil had significantly higher water content than the other two soil plots. However, this increased quantity of water did not denote increased crop production; on the contrary, these preliminary data point to a decrease in crop production.

TDR detection of nonaqueous phase liquids in sandy soils using the eigendecomposition method

In this study, the presence of nonaqueous phase liquids (NAPL) in sandy soils are detected using a TDR probe system and the eigendecomposition method of analysis. As a demonstration, five NAPLs with different physicochemical properties (acetone, benzene, heptane, trichloroethylene, and xylene; Table 1) were used. Samples were prepared in such a way that the soil pore fluid has different contents of deionized water and NAPLs. For each experiment, a pulse signal with known characteristics was used and reflected signals were captured by an oscilloscope and analyzed using the eigendecomposition method. Autoregressive modeling and singular value decomposition were used to calculate the eigenvalues. The most significant eigenvalues were identified based on their power spectrum. The relative eigenvalue of the first mode (Eow), which is a measure of the power carried by the signal, was calculated and correlated to NAPL type and content, and octanol water partition coefficient (log Kow). The results indicated that for the same NAPL content, as log Kow increases, Eow decreases due to increase of hydrophobicity. For the same log Kow, as the organic content in soil pore fluid increases, Eow increases due to decrease of dielectric properties of the pore fluids.

Dimensões de bulbo molhado no solo, a partir de fonte pontual de água, para manejo de água na cultura da pupunha (Bactris gasipaes, Kunth)

Aiming to describe the soil body, wetted by water applied on soil surface from a punctual source (drip emitter), this work was conducted at experimental irrigation center of Universidade Estadual de Maringá, at Maringá, Paraná, Brasil. A PVC Box with 1,0m diameter and 0,7m deep was filled by a sand soil, usually founded on a Northeast Region of Parana State. To identify the wetted front position, some TDR probe were placed horizontally each 0,05m deep, from soil surface to 0,60m. Some TDR probes were installed vertically from soil surface, each 0,08m from punctual source of water. Water was applied by 1,0; 1,8 and 3,4 litters per hour outflow emitter, and TDR readings were done at a 10 minutes interval, after beginning of water application. Results show that good evaluations of wetted front position can be obtained by this methodology use. For the studded conditions, the results showed that drip irrigation time, using this emitter, must be lower than 120 minutes, for the Pupunha palm tree, cultivated at this Parana State Region

Distribuição da água no solo para o dimensionamento da irrigação por gotejamento

O objetivo do dimensionamento de um sistema de irrigação por gotejamento é escolher adequadamente o "layout" e os componentes do sistema para promover a distribuição precisa da água (e fertilizante) para todas as plantas, ao longo do campo. Com vistas a este propósito, avaliaram-se as contribuições capazes de fornecer elementos básicos para o dimensionamento da irrigação por gotejamento, utilizando-se sondas de TDR. Assim, o estudo experimental foi dividido em duas partes; na primeira, tem-se o efeito de irrigações sucessivas nas dimensões dos raios horizontal e vertical do bulbo molhado, analisado em condições de laboratório e, na segunda, o processo de distribuição da água na formação do bulbo molhado em campo. Com os resultados apresentados nesta pesquisa pôde-se concluir que: a umidade inicial do solo, o volume aplicado, a vazão do gotejador, o disco saturado e a condutividade hidráulica, são elementos importantes e devem ser conhecidos para o dimensionamento e o manejo adequado da irrigação.

Interactions of maize and Italian ryegrass in a living mulch system: (2) Nitrogen and water dynamics

Water and nitrogen availability may limit the growth of the main crop competing with a cover crop in a living mulch system. Some aspects of the dynamics of water (soil water content and deep percolation) and nitrogen (concentration in soil solution and leachate) were studied in maize (Zea mays L.) sown into a bare soil (BS, conventional cropping) or into a living Italian ryegrass (Lolium multiflorum Lam.) mulch (LM) during three years. Eight lysimeters (1.0 by 1.0 m square surface area and 1.1 m deep) with ceramic suction cups, TDR probes and a drainage pipe were used each. In LM a 0.3 m wide strip was kept free of grass around the maize row. The living mulch reduced the soil water content between 0.3 and 0.9 m soil depth, which remained lower even after intense rainfall. Deep percolation over the entire maize crop season was at least 40% lower in the LM compared to the BS treatment. In LM the nitrate concentrations in the soil solution and in the leachate (usually <10 mg L−1) were very low. In BS the nitrate concentration in the leachate reached as much as 70 mg L−1. Losses of N in LM did not reach 1% of the values observed in BS. Reduced water and N availability in LM contribute to explain the decrease in growth and yield of the maize plants, and are in good agreement with the dense root system developed in this cropping system as compared to BS. The challenge for the development of living mulch systems is to improve the uptake of water and nitrogen by the roots of the main crop in a competitive environment without affecting the capacity of the cover crop to prevent N losses by leaching.

Characterizing unsaturated solute transport by simultaneous use of dye and bromide

Variability in solute transport properties in the unsaturated zone is of growing concern due to environmental hazards. By combining dye and salt tracers, a detailed description of the solute transport can be made in both time and space. In this paper, qualitative and quantitative effects of preferential flow were investigated in a sandy loam. The experimental site was equipped with 64 TDR probes. The plot was irrigated with pure water for 12 days to achieve steady state conditions. Then the plot was irrigated with a solution containing a dye tracer and KBr over three days. During the experiment, bromide concentration and soil moisture were measured hourly by the TDR probes. On the day following the last day of irrigation, a trench was dug and 14 vertical slices were excavated and photographed. A Sigma Probe (SP) was used to investigate the bromide concentration in an individual finger. The SP is a new dielectric measuring technique that gives readings of soil solution electrical conductivity. The dye photographs provided visualization of finger flow. Horizontal flow was evident in the E horizon at 0.3–0.4 m depth. Also the bromide concentration was elevated at these depths. The occurrences of high bromide as measured by the SP concentration corresponded fairly well to the dye patterns. The visible lower limit of dye stained soil corresponded to a bromide concentration of 0.10. A comparison between the bromide and dye penetration showed that the retardation factor of the dye as compared to bromide was 1.5. This is within the range reported by other authors and shows that the SP can be a useful tool to investigate the retardation of dyes as compared to bromide.

A Variable-Volume TDR Probe for Measuring Water Content in Large Soil Volumes

A Variable-Volume TDR Probe for Measuring Water Content in Large Soil Volumes

Comparison of CERES, WOFOST and SWAP models in simulating soil water content during growing season under different soil conditions

A lysimeter experiment conducted on three soil types in a main agricultural production region of Austria in Marchfeld (latitude 48°12′N, longitude 16°34′E and altitude 150 m above sea level), was used to test the performance of the three widely used crop models, CERES, SWAP and WOFOST. The soils included chernozem, sandy chernozem and fluvisol with a 2.0 m profile depth. Daily measurements of the soil water content were taken using TDR probes (one per 0.3 m of depth) in six replicates for each soil type. The analysis was carried out for winter wheat and spring barley grown on the site during seasons 2000 and 2001 and included a detailed comparison of the simulated and measured soil water contents as well as an analysis of seasonal soil water balances, root front velocities and an evaluation of the modeled crop yields. CERES and SWAP, in contrast to WOFOST, simulated the grain yield of barley and wheat well. All three models simulated soil water content in the profile with similar results. The root mean square error (RMSE) range of soil water content was 0.71–4.67% for barley and 2.32–6.77% for wheat, depending on the model and soil type. None of the models simulated total soil water content in the profile significantly better, but there was a general tendency for the models to overestimate soil water depletion. Both CERES and SWAP mimicked the soil water content dynamics well in the top 0.3 m of the soil. The study shows that the multiple layer approach models (SWAP or CERES) including more sophisticated estimation methods for root growth and soil water extraction should be preferred in comparable environments. Further adjustments of evapotranspiration subroutines to the local conditions should be considered prior to the model use for drought impact assessment, yield forecasting or climate change impact studies.

Evaluation of TDR Use to Monitor Water Content in Stem of Lemon Trees and Soil and Their Response to Water Stress

The purpose of the study was to compare the response of TDR‐determined stem (θstem) and soil (θsoil) water content to different irrigation managements. θstem (L L−1) was measured with three‐rods TDR probes (70 mm) installed vertically or horizontally into predrilled holes in the trunk of 5‐yr‐old lemon [Citrus limon (L.) Burman f.] trees in a semiarid region (Israel). Four irrigation treatments were established to deliver 100% (“full”), half of this amount on one side of the tree (“50%”), same amount as in 50% but applied to alternate sides (3‐wk intervals) of the tree (“split”), and 0% (“dry”) of the orchard's normal irrigation volume (typically 150–200 m3 wk−1 ha−1). Treated sewage water (σw = 0.9 dS m−1) was used to irrigate the trees for 75 d (end of June and mid September). Changes in θsoil, and θstem were monitored at weekly intervals. Leaf water potential and temperature measurements were used to verify the achieved water stress levels. θstem of the full treatment fluctuated by about 0.02 to 0.03 L L−1 above and below the “prestress” reference level (beginning of the season). In contrast, θstem of the 50% treatment declined (by about 0.07 L L−1) steadily over the season. For the split and dry treatments, θstem decreased by about 0.12 L L−1 relative to the reference level. θsoil and θstem values suggest that some surplus irrigation was applied to the full and 50% treatments. It was found that water stress was reflected in TDR‐measured θstem changes but that these changes were too small for routine irrigation control.

An easily installable groundwater lysimeter to determine waterbalance components and hydraulic properties of peat soils

Abstract. A simple method for the installation of groundwater lysimeters in peat soils was developed which reduces both time and financial effort significantly. The method was applied on several sites in the Rhinluch, a fen peat land 60 km northwest of Berlin, Germany. Over a two-year period, upward capillary flow and evapotranspiration rates under grassland with different groundwater levels were measured. The installation of tensiometers and TDR probes additionally allowed the in situ determination of the soil hydraulic properties (water retention and unsaturated hydraulic conductivity). The results of the measurements of the unsaturated hydraulic conductivity demonstrate that more than one single method has to be applied if the whole range of the conductivity function from saturation to highly unsaturated is to be covered. Measuring the unsaturated conductivity can be done only in the lab for an adequately wide range of soil moisture conditions. Keywords: peat soils, soil hydraulic properties, evapotranspiration, capillary flow, root distribution, unsaturated zone

Ground water pollution by roof runoff infiltration evidenced with multi-tracer experiments

The infiltration of urban roof runoff into well permeable subsurface material may have adverse effects on the ground water quality and endanger drinking water resources. Precipitation water from three different roofs of an industrial complex was channelled to a pit and infiltrated into a perialpine glaciofluvial gravel-and-sand aquifer. A shaft was constructed at the bottom of the pit and equipped with an array of TDR probes, lysimeters and suction cups that allowed measuring and sampling soil water at different depths. A fast infiltration flow was observed during natural rainfall events and during artificial infiltration experiments. For a better understanding of the behaviour of contaminants, experiments were conducted with cocktails of compounds of different reactivity (ammonium, strontium, atratone) and of non-reactive tracers (uranine, bromide, naphthionate), which represent different classes of pollutants. The experiment identified cation exchange reactions influencing the composition of the infiltrating water. These processes occurred under preferential flow conditions in macropores of the material. Measuring concentration changes under the controlled inflow of tracer experiments, the pollution potential was found to be high. Non-reactive tracers exhibited fast breakthrough and little sorption.

Variations of self-potential and unsaturated water flow with time in sandy loam and clay loam soils

Accurate assessment of soil–water fluxes is essential in soil physics due to its direct implications in environmental, agronomical or hydrological applications. Field estimations of soil–water fluxes by ‘classical’ hydraulic methods are often difficult to obtain. Moreover, water fluxes are highly variable in space and time. The obtainment of a reasonable estimate for this variable would require numerous measurement sites. However, such a requirement is rarely met. Thony et al. [CR Acad. Sci. Paris, Earth Planetary Sci. 325 (1997) 317] presented the experimental evidence of a linear relationship between the self-potential (SP) and the unsaturated soil–water flux. Therefore, this relationship would allow the indirect assessment of the water flux using electrical measurements. Such an approach would appear much more flexible and easier to perform than the current hydraulic measurements. The aim of this study is to experimentally investigate the existence and robustness of the flux–SP relationship for different soil types and pedoclimatic conditions.The soil–water fluxes and the SP were monitored in a long-term experiment involving two types of soils, contrasting in hydraulic and electric properties. The soils were placed in lysimeters which were instrumented with tensiometers and TDR probes for monitoring hydraulic heads and moisture content, respectively. Unpolarizable SP electrodes, temperature sensors and suction cups (for collecting pore water) were also installed in the lysimeters. The SP and the fluxes were measured or calculated in the 30–40cm depth section.Results show that the variations of the SP with time were clearly linked to both rainfall events and evaporation. However, in the long-term, the linear relationship between the unsaturated water flux and the SP evolves from strongly correlated to almost not correlated. The slope (sensitivity) of the flux–SP relationship varies with the soil type, decreasing with more electrically conductive soil. Taking into account a varying soil–electrode contact greatly improves the flux–SP relationship at the scale of the rainfall event, particularly when considering infiltration and drainage phases separately. Nevertheless, at the scale of a year, with alternated rainfalls and evaporation phases, the robustness of the relationship decreases (i.e. the coefficients of the relationship vary between events). This variability could be related to time variations in electrical conductivity, not so much to that of the soil–water, but rather to that of the water from the salted soil mud added to the SP electrodes at the time of installation.This study points out methodological problems associated with the measurement of SP in shallow unsaturated soils over the long-term and the need for designing specific electrodes for this purpose. However, in deep soils beneath the root zone, environmental conditions generally vary slowly and lightly in comparison to surface horizons. In this case and with our present set of SP measurement devices, the flux–SP relationship could be more stable than in the surface soil horizons and useful for examining aquifer recharge, capillary rises or contaminant transfer.