Biogeochemical and physical processes in the unsaturated soil zone (vadose zone) control the the contaminant fate and transport of diffuse pollution through soils to other environmental compartments.
The complexity of the vadose zone processes arises from several sources. First of all, the soil itself is a complicated disperse system made up of a microscopically heterogeneous mixture of solid, liquid and gaseous phases. The solid phase contains mineral and organic particles of varying sizes, shapes and chemical composition ranging from the molecular sized and colloidal particles to coarse sand and gravel. The organic fraction of the solid phase includes diverse communities of living organisms, plant and animal residues in different stages of decomposition and humification as well as various types of coals and charred organic matter. How soil reacts to long-term changes in the hydrological cycle or by changes in land use is still not understood in its complexity (weathering, dissolution / precipitation of minerals, carbon turnover, release of DOC, wetting and drying properties, permeability…). Studies regarding biological availability and mobility have often resulted in different and apparently contradictory observations and conclusions. However, the underlying physico-chemical processes cannot be in conflict and therefore such contradictions have to be due to limitations of our understanding of the involved factors and processes. The uncertainties due to apparent contradictions may be a major constraint to an accurate management and policy development for dealing with contamination.
Contaminants dissolved in the soil solution migrate by three basic mechanisms: advection, molecular diffusion and mechanical dispersion. From the macroscopic point of view, the process of advection is fully determined by the soil water velocity field. The processes of molecular diffusion and mechanical dispersion are usually described using identical physical means (Fick’s law) and can therefore be conceptually combined into a single process – hydrodynamic dispersion.
Migration, however, is not limited to the dissolved phase. Contaminants may attach to colloidal or suspended particles. Once mobilized, these particles can carry even strongly sorbing organic and inorganic contaminants. While the fundamental physics and chemistry of particle facilitated transport is understood, the relevance for contaminant spreading and transport is still not known. Especially the environmental conditions and porous media properties which favor the formation, mobilization and transport of inorganic and organic colloids are essentially unknown. As the conditions for the retardation and immobilization have to be clarified as well, no valid and reliable assessment of the risk exposed by particle facilitated transport is available.
Contaminants migrating through the soil are reacting with the soil constituents and undergo complex physical, chemical and biological transformations. Most contaminant transport mechanisms in soils are mediated or at least strongly affected by the presence and movement of soil water. Thus, deep knowledge of the physics and chemistry of soil water movement in heterogeneous and dynamic systems is an essential prerequisite for a valid und reliable estimation of the contaminant fate. This, of course, requires our lasting efforts to gain a better understanding of the movement of water in heterogeneous porous media.
Under natural conditions, the macroscopic soil properties often vary considerably in both space and time. Since these properties change from point to point at multiple scales, the heterogeneity and spatial variability needs to be conceptualized as a hierarchical system (see Scales). Another source of complexity is related to the natural soil water and soil air regime. The soil surface is exposed to continually changing radiation fluxes, reflected in evaporation changes and surface temperature gradients. The supply of water to the surface in a form of precipitation is highly irregular in time and space. There are complex interactions between plant roots, rhizosphere micro-organisms, and soil constituents. Vadose zone processes are strongly linked to groundwater recharge quantity and quality.
After entering the soil, contaminants may dissolve/disperse in soil water, attach to mobile and immobile soil particles, partition to soil air, or even form separate non-aqueous phase liquids (NAPLs). Despite the fact that numerous research projects in the past focused on the understanding of the contaminant fate in soil and aquifer systems, we still lack in depth knowledge on how soils process contaminant inputs under natural, i.e., environmental condition in the long run. In particular, long-term effects which result in aging phenomena are known to play a prominent role in the attenuation of contaminants. It is, however, essentially unknown which mechanisms contribute to aging and which properties favor such phenomena.
An essential question of the fate of contaminants in soils is the affect of the structural and functional biodiversity. While numerous studies have proved the effect of microbial populations on the persistence of contaminants at the bench scale (microcosm and biodegradation studies), the interdependence of long-term contaminant inputs in soils with soil flora and fauna and vice versa is essentially unclear. One reason is our principle knowledge gap on the functional biodiversity. Up to now, studies aiming at biodiversity mainly concentrated in the assessment of the structural biodiversity. How soil flora and fauna functioning affects contaminant fate and turnover rates and how contaminants affect the functioning of the subsurface microbial community is a key issue of future soil research.
Due to the complexity of the processes determining the transport of contaminants in soils, it seems to be practically impossible to formulate a completely unified quantitative description of the phenomena, which would adequately explain processes at all involved time and space scales. Instead, specific methodologies are used to solve predefined classes of problems. On the other hand unifying views are necessary when complex large scale contamination problems are to be solved.
Large gaps exist in our knowledge and understanding of soil-plant-atmosphere interactions at the ecosystem level such as soil-vegetation-atmosphere interactions or hierarchical and dynamic root systems. Consequently, there is an urgent need for much closer cooperation between the researchers representing different fields of science to come up with new interdisciplinary solutions, so that more complex real world problems can be addressed. Following are selected examples of the 'interdisciplinary' research gaps which may have serious impact on transport of contaminants in soils:
· Basic soil hydraulic properties (such as retention curves and hydraulic conductivities) are routinely measured for soil layers identified by regular soil surveying. However, very little attention is usually paid to the properties of the first several centimeters of soil near the soil surface. In soils fully covered with vegetation, this thin layer is very different from the average topsoil material. Often the standard measurement techniques are useless because of the unstable nature of the soil organic matter. New/existing methods for assessing the near surface hydraulic properties need to be developed/improved and brought to common use.
· The origin and development of a root system in soils is a complex and yet scarcely understood subsurface phenomena. We understand that soil structure and pore system morphology together with other properties determine direction of root growth and root system development. How these physical and biological processes and properties work together resulting in complex structured hierarchical and dynamic root systems, and how such systems interact and communicate with the biotic and abiotic environment, is of fundamental importance for the understanding of contaminant fate in the highly active and dynamic rhizosphere.
· Ecologically, the effect of soil water flow and retention is important for the availability of water for crops and vegetation. This availability is known to be spatiotemporal variable, and this renders root water extraction spatiotemporal variable. For schematised situations, these dependencies have been addressed, but nevertheless, to relate the type of vegetation quantitatively to the soil water regime with a mechanistic basis is now an emerging field of research.
Structured soils contain a highly-permeable macropore or fracture pore system through which water and solutes can move at considerably higher velocities than in the porous matrix. Consequently, local (sub-macroscopic) nonequilibrium conditions in the transient pressure head and solute concentration may develop. Such preferential flow phenomena severely limit the prediction of water and solute movement. Preferential flow related to soil structure has been widely reported in soils containing wormholes, root channels, and inter-aggregate fissures.
Additional types of preferential flow have been linked to textural differences rather than structural effects. Two types of preferential flow phenomena, that belong to this category, are fingering and funneled flow. The evolution of finger-type preferential flow paths is associated with gravity-driven flow instability. Fingering occurs in water repellent soils, when water percolates from a fine-textured into a coarse-textured layer, or when the air pressure increases ahead of infiltration front. To what depth preferential flow takes place is largely unknown and depends much on subsurface heterogeneity.
Most field soils exhibit different types of spatial heterogeneity, such as soil spatial variability and soil structure, which often coexist. When quantitative description of soil water flow is based on the traditional continuum approach, the notion of spatial variability relates to spatial distribution of macroscopic model parameters, such as hydraulic conductivity. However, in soils with strongly developed structure, microscale effects sometimes become so dominant that they affect macroscopic flow and transport processes. Both spatial variability in soil hydraulic properties and structure-induced heterogeneity can contribute to the initiation of preferential pathways. The challenge is to adequately account for both types of spatial heterogeneity.
Small-scale heterogeneities related to soil structure can be modeled either by using a discrete fracture network model or a multi-continuum approach. Within the discrete fracture network concept, a map of the structural geometry must be known, while with the multi-continuum approach two or more continua, representing matrix and fracture systems, share the same space domain. Crucial components of these types of models are transfer terms governing the exchange of water and/or solutes between the fracture and matrix pore systems. Empirical and semi-empirical expressions exist that are applicable to transient unsaturated flow, however more research is needed to establish more adequate and computationally feasible relationships and to develop experimental methodologies needed to determine the additional constitutive parameters for dual-porosity/permeability models.
Experimental and computational techniques currently used for determination of soil hydraulic and transport properties are often inadequate for solution of many practical field scale problems of water flow and contaminant transport. Regarding flow, significant advances have been made conceptually, numerically (including software tool development), and by analytical studies. Nevertheless, for a particular type of flow problems, i.e., the multiphase flow of water, gas, and NAPLs, models have been developed only during the last decade and the scenarios that have been studied are still relatively scarce and simplified. Consequently, well based predictions on the basis of a well developed understanding and theory are still left for the future. For instance, the rate of lateral spreading of floating LNAPL lenses at the groundwater table have been mainly considered for rather homogeneous situations (both numerically and analytically). Also the spreading of NAPLs above coarse/fine (and vice versa) layer interfaces has been studied, but only for relatively simple boundary conditions (homogeneous layers). Hence, one of the main problems, the volume of soil and aquifer contaminated by erratic flow of DNAPLs, in realistically layered nonhomogeneous porous medium, cannot be assessed. For similar reasons, the dissolution of free liquid into the aqueous phase, as well as the volatilisation of important NAPLs to the soil atmosphere, and the concentration levels and time periods involved, are still practical questions to which current science cannot provide a reasonably accurate answer.
Geochemical and biological processes are predominant factors of the fate and transport of contaminants in soils and the unsaturated zone. Often these processes are studied separately. Detailed modeling approaches have been developed to couple the description of water flow and geochemical interactions. In parallel, attempts have been performed to couple the description of microbiological activities with geochemical interactions. They are functionally strongly related, where small scale heterogeneity serves as an important factor to provide a niche for surviving organisms, a starting point for colonising other parts of the soil, and where an adequate linkage of biotic and abiotic factors is necessary for the assessment of the response of the ecosystem to contamination, the ability to buffer and transform pollutants, and of the integrity of the soil ecosystem/food chain.
In addition to the understanding and modelling of biogeochemical interactions, it is necessary to predict the impact of these coupled processes at transient flow conditions in the unsaturated zone. The complexity of these coupling phenomena is one of the major barrier for practical application of the present knowledge in each discipline (e.g. soil physics, geochemistry, plant physiology and microbiology). To overcome this gap, more interdisciplinary modelling approaches must be proposed, following a step by step experimental identification and validation of the model modules. The development, test, and validation of such coupled models for simulating the fate and transport of contaminants in soil and the unsaturated zone is an important task in the environmental sciences.
Another, yet still neglected aspect of the bigeochemical interactions in soil is the mutual interdependence of long-term contaminant input and the presence and the structural and functional biodiversity of soils and aquifers. While the principal role of the subsurface community for the contaminant fate is obvious (degradation, mineralisation, metabolisation, humification, immobilisation and formation of bound residues), we essentially lack the knowledge with respect to the effect of contaminants on the microbial community and its activity. Thus, neither the formation of metabolites nor the export of contaminants, metabolites and degradation production nor the development of contaminant turnover rates can be estimated à priori.
Our current understanding of natural attenuation is based on assumptions that the present biological activity as well as the environmental boundary conditions stay constant over time. With this, however, it is intrinsically presumed that the active microbial community at a site und thus the functional diversity is not affected by the contaminants. This assumption, besides most other assumptions on the biodiversity in soils, is completely unproven. The most prominent reason for this fact, of course, is our principle knowledge gap both with respect to the functional biodiversity and the structural biodiversity in soils. The few studies on soil biodiversity, conducted up to now, aimed at the assessment of the structural biodiversity, while the aspect of functional biodiversity is a key issue within the biodiversity research in general.
Heavy metals and other toxic elements are subject to a complex geochemical speciation in the unsaturated zone. Thermodynamic data of the main part of inorganic water/mineral reactions are with respect to pure minerals. The conceptual approach of interface interactions between solid compounds (including complex minerals and natural organic matter) is well developed. It includes the description of ionic exchanges on charged interfaces, surface complexation, and coprecipitation in specific mineral solid solutions. But the validation of the different concepts and the quantification of related parameters is still uncertain and subject to scientific discussion. The state-of-the art for the interactions between solutes and organic ligands or organo-mineral colloids is approximately at the same level. As a consequence of this fragmented knowledge, detailed geochemical codes are available for scientists using precise thermodynamical data of equilibrium systems, but the uncertainty concerning the reaction kinetics is an important obstacle for understanding and hence modeling of real world fate and transport of toxic elements even in quite simple and well-controlled lab-conditions. For some heavy metals or other toxic elements, the speciation can be extended to organo-mineral geochemistry. In this case, some chemical species are known and their behavior is more complex due to potential volatilisation and biotransformation (for example Hg, Sn, As).
Organic pollutants are also subject to complex interactions in soils. The main difficulty is the potential biotransformation of these compounds and the limitation of the bioavailability of these compounds due to physico-chemical interactions with organic matter and other reactive soil components. The consequence of biological transformations is the occurrence of many metabolites before the hypothetic complete mineralisation of contaminants. These metabolites will be characterized by different toxic properties, different fate and transport patterns, and by potential interactions with the original contaminants. This “speciation” of organic contaminants is related to biochemistry and microbiology. The more classical concept for describing organic interactions in soils is based on linear partitioning between organic matter and the aqueous solution, and linear kinetic biodegradation. In many real field situations, this concept is not sufficient. Linear partitioning is only justified if unsoluble and soluble organic matter remains constant quantitatively and qualitatively during the fate and transport of the organic pollutant. Linear kinetic biodegradation is only justified for an apparent one step biodegradation, independent to other biological processes. The presence of solutes modifying the solvent properties of the aqueous phase (e.g. more or less polar solutes), the modification of the physical state of the unsoluble organic matter (e.g. solid-liquid transitions depending on temperature), and the complexity of micro-organism community metabolisms induce generally non-linear behaviors which must be taken into account to model systems.
Mobile sorbents such as dissolved, colloidal phase and suspended inorganic and organic particles affect flow of water and transport of solutes in soils. Major processes between the solution and the solid phase, such as sorption, partitioning, speciation and ion-exchange are influenced by the interactions with mobile sorbents. Their presence affects the solubility of solutes due to complexation, solubilisation, carrier association and the solvophobic effect. In recent years, research has focused on processes leading to mobility enhancement of organic and inorganic pollutants. The issue was to understand to what extent mobile sorbents may facilitate contaminant transport in porous media with respect to risk assessment, soil and groundwater reclamation and clean-up. Major compounds which have been shown to increase the solubility and thus the mobility of nutrients and contaminants are surfactants, co-solvents, (hydr)oxides, clay and other minerals, humic substances and humin-coated inorganic colloids. In contrast to aquatic environments, the major amount of dissolved and colloidal-sized constituents in soils is either biotic or organic in nature or composed of inorganic matter with organic coatings. Due to their chemical properties, i.e., high concentration of functional groups and the chemical diversity, these substances interact with the immobile solid phase as well as with other dissolved and colloidal phase components in the liquid phase. The retention of mobile sorbents may cause reduced overall contaminant mobility due to immobilization. The underlying process has been described as co-sorption. Sorption of mobile organic substances can as well lead to an increase of the organic carbon content of the bulk soil, thus increasing the number of potential contaminant binding sites, a process, described as cumulative sorption.
While the fundamental physics and chemistry of colloid and carrier affected transport is understood, its relevance for contaminant spreading is essentially not known. In particular, understanding of the environmental conditions and porous media properties which favor the formation, mobilization and transport on one hand and the retardation and immobilization on the other hand is needed. Moreover, the effect of mobile sorbents on the mobility of contaminants under changing environmental settings has to be addressed in future research.
During the last 15 years huge differences in the sorption capacity of natural organic matter have been found for predominantly organic pollutants. Whereas on one hand differences in chemical composition and structure of the organic matter are discussed as being responsible for these differences, more and more evidence appears showing the soils and sediments contain a complex mixture of particulate organic matter. Within this mixture carbonaceous particles predominate sorption of many hydrophobic organic compounds especially at environmentally low concentrations. Such carbonaceous particles comprise charcoal, bituminous coal fragment, all sorts of soot (all belonging to the ill defined group of black carbon). Methods to quantify the fraction of this highly reactive constituents in soils and sediments are lacking. Methods proposed so far are operationally defined which likely produces artefacts.
In addition, a fundamental understanding of the sorption mechanisms and kinetics in these carbonaceous phases is needed. This will help to explain the often observed but not understood sorption/desorption hysteresis in soils which is crucial to assess bioavailablity of sorbed pollutants.
A complete lack exists concerning the long-term impact of weathering on these carbonaceous phases. This applies for carbon turnover rates (new data indicate that these type of carbon does not really participate in the carbon turnover in soils and sediment and is therefore termed dead carbon). Question concerning the secondary release of pollutants associated with these carbonaceous phase upon weathering is completely open.
Authors: Tomas Vogel, Peter Grathwohl, Sjoerd van der Zee, Kai-Uwe Totsche; Michel Jauzein
SOWA – Integrated Soil and water Protection: Risks from diffuse pollution