Preferential flow and solute transport in a large lysimeter, under controlled boundary conditions.

https://doi.org/10.1016/S0022-1694(98)00262-5Get rights and content

Abstract

Laboratory studies of solute transport in soils (soil columns) are not totally representative of field conditions (spatial variability, soil structure etc.). Field studies hardly allow quantification of fluxes and mechanisms. In this article, and intermediate approach is suggested, using a lysimeter (1.7 m3) of an almost undisturbed soil, with controlled boundary conditions, the aim being to be able to quantify fluxes and mechanisms at a scale closer to field conditions, thus yielding results that better depict reality. Two experiments, with constant water fluxes of 1.05 and 1.48 mm h−1 were conducted. Solutes were introduced as concentration pulses. Species 2H2O, C1 and Br were used as tracers, and K+, NH4+, NO3, atrazine as interactive and/or reactive solutes. Elution curves were analyzed by the method of moments. Results show that about 20% of the water are immobile. As a consequence of anion exclusion, anion tracers appear at the outlet with an advance of about 10% in time as compared to isotopic tracers. The added NH4+is mostly nitrified, and K+ undergoes cation exchange with Ca2+ and Mg2+. Under our experimental conditions, leaching of atrazine is significant with low degradation. A third experiment was conducted, in which the flow was interrupted while the solute peak was within the lysimeter, and 400 soil samples were extracted from the lysimeter. Soil–water content distributions exhibit coefficients of variation within layers between 5% and 27%. Concentration distributions exhibit coefficients of variation within layers between 22% and 59%. There is no correlation between concentration and water content. The observed spatial variability suggests the occurrence of preferential flow. Concentrations in suction cups were 55%–136% of those measured in corresponding soil samples.

Introduction

Leaching of surface-applied fertilizers and pesticides beyond the root zone is of interest to agronomists, as they reduce the efficiency of the applied agrochemicals. The quality of the underlying groundwater is also of concern. Many studies of the transport of solutes through repacked laboratory soil columns were performed (Brigham, 1974; Gaudet et al., 1977; Kookana et al., 1993; Krzyszowska et al., 1994). However, these studies do not take into account the spatial variability of the soil properties, because of soil repacking and small column sizes. As a matter of fact, processes such as preferential flow might have a great influence on solute transport. However, experiments were run on large intact soil monoliths (Saffigna et al., 1977; Bergström, 1990; ECPA, 1974, but mostly submitted to natural (i.e. transient) boundary conditions for water fluxes, and to complex boundary conditions for solutes, e.g. solute formulation method (Ghodrati and Jury, 1992; Wietersen et al., 1995). However, it is difficult to use these experiments in predictive models of solute transport.

A few authors adopted an intermediate approach (Bowman et al., 1994; Phillips et al., 1995; Poletika et al., 1995). Using undisturbed soil blocks, larger than usual laboratory columns (lengths and widths between 30 and 80 cm), submitted to controlled boundary conditions for water (constant flux) and solutes (pulses or on-tine application). These studies, however were carried out with high water fluxes (5–20 mm.h−1) and without interactive solutes (except for Poletika et al., 1995). According to the soil types, these lysimeter sizes might be too low to depict preferential flow correctly (Netto et al., 1998). The aim of this article is therefore to use a larger undisturbed lysimeter. Lower water fluxes (closer to natural conditions) use simultaneously different types of tracers and reactive solutes. The scale of the lysimeter used here is the same as for a classical monitoring site in the field, equipped with a neutron probe access tube, tensiometers and soil solution samplers (suction cups) at different depths (Kengni et al., 1994).

Section snippets

The soil

The study was conducted on the Experimental Farm of the Lycee Agricole, La Cote Saint Andre, located 40 km northwest of Grenoble, France, The site is a typical glacial terrace, with a soil approximately 1 m thick overlying a layer of gravel and stones of high permeability of 10–20 m thickness. There is a water-table aquifer, which varies in depth between 9 and 15 m from the soil surface. It has a high sensitivity to nitrate pollution. Large number of wells reach a nitrate concentration close to or

Neutron probes

Volumetric water content was measured at depths between 10 and 120 cm by neutron probes, using 10 cm increments. Calibration was carried out according to Kengni (1993). Mean volumetric water contents were θ=0.253 for #1B, and θ=0.242 for #2.

Soil samples

Volumetric water contents form soil samples (#2), averaged in each layer, are compared to neutron probe water contents in Fig. 3. Water content values averaged on the 6 central sections (greyed in Fig. 2) are also shown. Because they are expected to be closer

Conclusions

The lysimeter used under controlled boundary conditions allows accurate quantification of transport processes for tracers and interactive/reactive solutes, through elution curve analysis and soil sampling. Such an accuracy is seldom reached at scales larger than laboratory soil columns. As a result of compaction problems near the edges, comparison with soil solution samplers (suction cups) was not good.

The moment method gave a preliminary estimation of the hydrodynamic and (bio) geochemical

Acknowledgements

The authors wish to thank the Lycée Agricole and the experimental farm of La Côte Saint André for technical support, the Service Central d’Ánalyses for chemical analyses. The Physiologie Cellulaire et Végétale laboratory of the Université Joseph Fourier, Grenoble, for immunoassay analyses of atrazine, the Section d’Application des Traceurs (CEA/DAMRI/SAR/SAT, Grenoble) for technical support, Michel Sardin for fruitful discussions and David E. Elrick for careful reading of the manuscript.

References (27)

  • C.B. England

    Comments on A technique using porous cups for water sampling at any depth in the unsaturated zone by Warren W. Wood

    Water Resour.Res.

    (1974)
  • J.B. Fleming et al.

    Bromide transport detection in tilled and nontilled soil solution samplers vs. soil cores

    Soil Sci. Soc. Am. J.

    (1995)
  • J.P. Gaudet et al.

    Solute transfer with exchange between mobile and stagnant water, through unsaturated sand

    Soil Sci. Soc Am. J

    (1977)
  • Cited by (79)

    • Retention and transport of mecoprop on acid sandy-loam soils

      2018, Ecotoxicology and Environmental Safety
      Citation Excerpt :

      Curves were further studied using temporal moment analysis as per Valocchi (1985), Stagnitti et al. (2000) and Kamra et al. (2001), following the descriptions given in full in Paradelo et al., (2016, 2017). The following parameters, used for the description of the breakthrough curves, were calculated: the first normalized moment (τ), that indicates the mean concentration breakthrough time; a relative retardation factor with respect to the inert tracer (R); the standard deviation (σ) that is a measure of the typical spread of the curve in relation to the mean breakthrough time; the dispersivity (λ), indicating dispersion of a given element within the column (Stagnitti et al., 2000; Schoen et al., 1999), and the non-dimensional skewness parameter, S, estimated from the third central moment (µ3) that characterises the asymmetry of the breakthrough curve. The quantification of mecoprop in all the samples was carried out on a HPLC liquid chromatograph (Dionex Corporation, Sunnyvale, USA) equipped with a P680 quaternary pump, an ASI-100 autosampler, a TCC-100 thermostatized column compartment and a UVD170U detector.

    View all citing articles on Scopus
    View full text