Conferences

Ben-Gurion University of the Negev
CWST - Center for Water Science and Technology
Transport and Clean-Up Processes in the Unsaturated (Soil) Zone under Dryland Conditions

Report on a workshop held at the Sde-Boker campus of BGU
June 8-10, 1999
Introduction and Dedication
The unsaturated zone plays an important role in the terrestial water balance; it is a controlling factor of plant growth as well as for the quantity and quality of the recharge waters and of the resultant groundwater. This is particularly true in the arid zone "where 95% of what goes on there hydrologically is governed by processes in the unsaturated zone. It is the dominance of the arid regions by the unsaturated zone that make them special " (J. Philip, 1999).
During the workshop, the hydrologic regime, the geochemical and microbial processes in the unsaturated zone as well as their remedial role in the mitigation of pollution was reviewed. Some case studies from the Negev were presented and gaps in the understanding of the system identified.
The detailed program and summaries of the presentation are attached, followed by comments and evaluation of the invited experts. Of these, the comments of John Philip, characterized as usual by profound wisdom and clear language, will stand out. The tragic death of John immediately after our workshop gives them a special significance.
This report on a subject so dear to John Philip's heart is dedicated in humble homage to our friend and mentor.

Prof. John R. Philip

Scientific Program
Part 1 - "The hydrologic regime"

Hydro-chemical perspectives on groundwater recharge processes in semi-arid and arid regions (C. Kulls)
Water movement in non-swelling and swelling soils (J. Philip)
Matrix flow, by-pass flow and perched ground water (K.P. Seiler)
Extensions of the macroscopic Navier-Stokes equation (S. Sorek)
The saturated-unsaturated interface region (SUIR) of a phreatic aquifer (H. Scher & D. Ronen)

Part 2 - "Infiltration"

Infiltration in 1, 2 and 3 dimensions (J. Philip)
Measurement of hydrologic conductivity in the deep vadose zone by in-situ methods (U. Shani)
Water flow through fractures in unsaturated chalk (O. Dahan)
Dynamic contact angle explanation of flow rate-dependent saturation pressure relationships during transient liquid flow in unsaturated porous media (S. Friedman)

Part 3 - "Solute transport and chemical and biochemical reaction processes in the soil zone"

Modeling solute transport in the unsaturated zone (including multi-species components) (E. Holzbecher)
Using the boundary layer concept for modeling transport in fracture/matrix systems (R. Wallach)
Particle favoured transport in the unsaturated zone (K.P.Seiler)
Transport of bacteria through saturated and unsaturated soils (S. Friedman)
Sorption of organic contaminants in a fractured chalk formation (A.Wefer-Roehl)
The interface between fracture surfaces and aqueous solutions in unsaturated chalk: physical, chemical and colloid generating processes (N. Weisbrod)

Part 4 - "Cleanup and remediation"

Management of remediation of contaminated soils, with reference to organic contaminants (T. Harmon)
The role of bacteria, fungi and plants in in-situ remediation (D. Kaplan)
Transport of organic and inorganic contaminants in desert soil: evaluation of flushing potential from a contaminated soil near Ramat-Hovav Industrial Park (S. Arnon & E. Adar)

Part 5 - Remote sensing

Soil moisture measurements by synthetic aperture radar and scatterometers in the arid and semi-arid environment (D. Blumberg)

Summaries of the presentations
Two contributions to the Encyclopedia of Hydrology and Water Resources, which set the stage, respectively, for part 1 (the hydrologic regime) and part 2 (infiltration) are attached, annotated by John Philip in his own handwriting. The rest of the contributions now follow:

HYDRO-CHEMICAL PERSPECTIVES ON GROUNDWATER RECHARGE PROCESSES IN SEMI-ARID AND ARID REGIONS
C. Kuells & E. Adar
Blaustein Institute for Desert Research, Ben Gurion Univ. of the Negev, Sede Boker, 84990, Israel

Understanding the non-linear response of the unsaturated zone to rainfall events is a key to process-oriented groundwater recharge studies in arid and semi-arid regions. At smaller scales the development of fingering in homogeneous soils and the occurrence of macro-pores or preferred flow-paths in non-homogeneous settings are relevant for groundwater recharge as they create potential bypass connections to the water table. Furthermore, surface and soil properties at infiltration events determine the onset and extent of runoff generation. Once overland flow has developed, the concentration of overland flow becomes the driving factor for the spatial distribution of infiltration and groundwater recharge in the catchment. However, at a basin scale prediction of infiltration, runoff generation and groundwater recharge is difficult, especially in drylands. Therefore, an inverse modelling approach is suggested to locate, identify and quantify recharge processes. This study makes use of stable isotope and extensive hydro-chemical data from the saturated zone to which an inverse mixing-cell model was applied. Examples of the occurrence and distribution of different recharge processes are given for a large scale semi-arid basin in the Kalahari, Namibia.
The inverse hydro-chemical approach indicates that in the Kalahari environment direct rainfall recharge is only of limited importance (< 2% of mean annual rainfall) compared to recharge processes involving runoff generation and concentration. Stable isotope and Chloride data point to the significance of macro-pore flow as a major component of direct recharge. At a regional scale the most efficient processes were found to be fracture surcharge on exposed hard-rock terrain, transmission losses from ephemeral floods and quick percolation of overland flow through karstified calcrete. All these processes exhibited high spatial variability. Since the hydro-chemical approach allows to identity the impact of fingerprints stemming from different recharge sources on the groundwater chemistry, it was possible to delineate preferential recharge zones. Recharge from ephemeral floods was found to be restricted to the confluence zones of major tributaries incised into permeable rock. Recharge through sinkholes only developed along a belt in the mountain front zone. In this area thin Kalahari sand cover and associated plant CO2 production had enhanced the re-dissolution of calcrete. In general, the combination of preferential flow paths (macro-pores, fractures and sinkholes) on the one hand and of runoff concentration along washes and in depressions/pans on the other hand was found to be critical for triggering recharge events in this semi-arid environment.

MATRIX FLOW, BYPASS FLOW AND PERCHED GROUND WATER
K. -P. Seiler
GSF Institute of Hydrology, D-85758 Neuherberg, Germany

In soils and the unsaturated zone below soils, flow paths for seepage have often quite different geometries due to primary and secondary structure and texture properties in sediments and the activities of biota in it. This expresses itself mostly in inhomogeneous flow in the unsaturated zone with two extremes: bypass- or fast flow-and matrix or slow-flow. Many transitions situation occur in between. Studies in cylindrical (suppressed lateral flow) and natural lysimeters (non-suppressed lateral flow) demonstrate the existence of bypass flow in fine and coarse grained sediments as well and bypass flow velocities exceeding decimetres per day as compared to matrix flow of less than a few meters per year. Bypass flow penetrates to depth of several meters in unconsolidated rocks (> 3m) and many ten meters in consolidated fissured rocks, often transforms into interflow and than contributes more than 50% to overland discharge in areas with significant changes in hydraulic conductivities at bedding interfaces paralleling topography of hilly terrains. Bypass flow always interchanges with matrix flow by incorporation due to gradients both in capillary forces and concentrations. Consequently, chemical and isotopic characteristics of bypass flow approaches matrix flow the longer the flow path or the residence time.
Hydraulic functions in unsaturated soils and sediments cover a wide variety of geometries of y/k and y/q relations. Connected to this, the terminology of high and low conducting beds from the saturated zone will often hardly apply to the unsaturated zone, in which clays and sands are often more conductive than gravels under the prevailing water contents (capillary barriers). Since most sediment units consists of sequences of different materials many interfaces between beds may act as the base of a perched aquifer. This is true for consolidated and unconsolidated sediments as well, and increases the variety of flow processes and the disposability of water for plants in unsaturated sediments. The interaction between bypass and matrix flow and the transformation of bypass into interflow in hilly terrains will be used for the first time to approximate quantitatively the bypass flow in vertical profiles (time integral method) using stable environmental isotopes and by sampling interflow during storm events (aerial integration method), respectively, to get lumped and representative information of bypass flow on large scales. In such scale related investigations lysimeter studies of the different type and the sampling of soil moisture and discharges offer the possibility to study processes and representative parameters as well.

EXTENSIONS TO THE MACROSCOPIC NAVIER-STOKES EQUATION
S. Soreka, D. Levi-Hevroni, A. Levy and G. Ben-Dor J. Blaustein Institute for Desert Research, Department of Environmental Hydrology & Microbiology, Sde Boker Campus, 84490, Israel
Department of Mechanical Engineering, Pearlstone Center for Aeronautical Studies, Beer Sheva, 84105, Israel

A mathematical development is presented concerning extensions to the macroscopic momentum balance equation for compressible Newtonian fluids, flowing through saturated porous matrix. This accounts for terms deviating from the dominant average momentum flux terms and Forchheimer terms associated with the relative fluid square velocity and its temporal rate, representing the microscopic momentum exchange between the phases at the solid-fluid interfaces. Non-dimensional investigation provide the conditions when Darcy's linear momentum will become dominant over Forchheimer momentum form or when the momentum balance equation will conform to a nonlinear wave equation. The latter is then solved numerically for a highly deformable elasto-plastic matrix, showing good qualitative agreement with experiments.

THE SATURATED-UNSATURATED INTERFACE REGION (SUIR) OF A PHREATIC AQUIFER
D. Ronena, and H. Scherb
Research Department, Israel Hydrological Service, POB 20365, IL- 61 203 Tel Aviv, Israel;
Department of Environmental Sciences and Energy Research, Weizmann Institute of Science Rehovot, 76100, Israel.

Field measurements of profiles of the water content have been conducted north of the city of Tel Aviv, in the phreatic Coastal Plain aquifer of Israel. The water content as a function of depth was obtained from continuous cores in 13 boreholes that were drilled in relatively homogeneous sandy sediments, within a radius of 5 m, in a region where the water table is at a depth of 7 m. It was found that the distribution of water within the capillary fringe (CF) was compact, meaning that there was an abrupt change in water content with increasing height above the water table. The height of the CF, above the water table was about 1.4 m, and over a horizontal distance of about 4 m, this height varied by up to 33% and 50% before and after the rainy season, respectively. The CF was displaced vertically as the water table rose 35 cm after the winter rains and the distribution of water in the CF remained compact. These findings are in agreement with the key qualitative features of simulation results of capillary dominated multiphase flow in a random porous media.
The saturated-unsaturated interface region (SUIR - extending from the imaginary plane where the sediment is completely saturated and pressure is positive to the top of the CF) has an unusual nature where saturated conditions are found in the CF and unsaturated conditions are detected in shallow regions (up to 1.5 m) below the water table. This 3 -m thick saturated-unsaturated interface region is characterized by a high water content, as compared to the residual water content of the unsaturated zone above, and a gas phase which may occupy up to 40% of the pore volume. The gas phase in the SUIR results from bacterial activity and air entrapped in the porous media during drainage, and imbibition. Dissolved organic carbon (DOC) has been found to be mobile through the unsaturated zone of the aquifer. DOC persistence for more than 15 years under unsaturated conditions, and laboratory BOD (biological oxygen demand) tests of sediments from the unsaturated zone suggest that moisture content may be a major controlling factor for biodegradation. Results of field studies conducted at the SUIR, revealed.

intensive biochemical activity as reflected by the consumption of dissolved O2 with the concomitant oxidation of organic matter, and the production of N2O (up to 400 mg) and CO2 (2%-5%);
the presence of an almost stagnant water layer (q = 0.5 m/y) down to a depth of 60 cm below the water table, and
the presence of microscale isothermal water parcels (characteristic vertical and horizontal length dimensions of the order of less than 1 m) which differ from each other in their chemical composition and density and are characterized by very sharp boundaries between them. At the study site the residence time of recharge water in SUIR was calculated to be > 5 years. During such a long time period chemical and bacteriological processes change the chemical composition of replenishment.

DYNAMIC CONTACT ANGLE EXPLANATION OF FLOW RATE-DEPENDENT SATURATION-PRESSURE RELATIONSHIPS DURING TRANSIENT LIQUID FLOW IN UNSATURATED POROUS MEDIA
S.P. Friedman
Institute of Soil, Water and Environmental Sciences, Agricultural Research Organization,
The Volcani Center, Bet Dagan 50250, Israel

The common assumption when modeling transient liquid flow in an unsaturated porous medium is that the capillary pressure-saturation degree relationship is independent of the macroscopic liquid flux, This assumption is not always applicable, and one reason for this is the dependence of the solid-liquid-gas contact angle at the moving liquid-gaseous interface on the flow velocities, as found in systems such as long cylindrical capillaries. In the present theoretical study a conjecture is made that at a prescribed capillary pressure the criterion for the liquid phase to invade an empty pore is defined by the Young Laplace equation, but with the expected dynamic contact angle used instead of the static one. An iterative procedure, based on a simplified description of the pore system, enables a quantitative estimation of the extent of the liquid flux dependence of the capillary pressure-saturation degree relationship. For a given capillary pressure the degree of liquid saturation decreases with increasing liquid flow velocity, for wetting processes, and vice versa for drainage. This effect of the liquid flux is more pronounced as the width of pore-size distribution increases.

MODELING SOLUTE TRANSPORT IN THE UNSATURATED ZONE (INCLUDING MULTI-SPECIES COMPONENTS)
E. Holzbecher
Institute of Freshwater Ecology, Department of Eco-Hydrology
Rudower Chaussee 30, D-12489 Berlin (Adlershof), Germany

Transport of substances through the unsaturated zone is an important topic. Pollutants, deposited intentionally or unintentionally at the surface or in the ground near to the surface may migrate towards the groundwater table. Other substances, like saline components or nutrients, enter the unsaturated zone as a result of human activities. For example agriculture has a tremendous influence on the conditions in the soil, particularly in dry regions where irrigation is practiced. In order to obtain a scientific basis for decisions to be taken, it is necessary to have a fundamental understanding of the relevant processes. The presentation focuses on the underlying conceptual model which is not only fundamental for numerical modeling, but for theoretical derivations and experimental design also. Deterministic modeling is based on the transport equation. This is a mass balance equation derived from mass conservation principle. If there are several components several conservation equations have to be formulated. The following processes are considered in the equation:
- advection
- diffusion
- dispersion
- sorption
- degradation and decay
- reaction
The transport equation is thus derived generally and can be applied in every hydrological compartment. Usually it is applied in single phase environments, as in streams. There are some modifications required when it is transferred to a two phase environment as in saturated porous media, in which the porous medium represents the solid phase and the water filled pore space the fluid phase. The presentation shows that further generalizations are necessary in the unsaturated zone.
In the unsaturated zone three phases need to be taken into account: solid (soil/rock), liquid (water) and gaseous (air) phase. The interaction of processes across the interfaces of these phases complicates the description of transport. Just to mention is the difficulty to characterize the flow field in which substances migrate. As other lectures have tackled that point extensively, the focus of the presentation is on genuine transport parameters, like diffusivity, tortuosity, dispersivity and retardation. In the final part it is demonstrated how transport and geochemistry can be coupled.
Fick's Law is one of the fundamental relationships determining transport in the environment. The classical form has to be extended in several ways. One adjustment has to take into account that diffusion occurs in the pore space only. As a second adjustment tortuosity has to be introduced. For unsaturated porous media further adjustments are necessary to take into account that the pore space is partially saturated only. Laboratory experiments show that tortuosity is strongly saturation dependent.
Dispersion is a velocity dependent mixing process which is of great relevance in porous media. Again experiments were carried out in order to study the dependence on saturation. In column experiments at Technical University Berlin with chloride has been observed an increase of dispersion length by one order of magnitude when saturation decreased from 100% to 30%. A detailed description and discussion of these experiments is given in the presentation. Sorption denotes the interaction of solid and fluid phases in porous media. Mass balance equations can be formulated for all phases involved. It is common to simplify the set of differential equations using the assumption that interphase exchange at the surface of the porous medium is a fast process. Under that condition an equilibrium is established and the general mathematical formulation can be simplified to a transport equation with an additional retardation factor. In the presentation it is shown that the retardation factor does not depend on saturation, when the usual assumption is made that volume and surface area change equally with saturation.
The complex interaction of processes is demonstrated for a system which is mainly governed by decay and linear sorption. It can be shown that the secular equilibrium for mother- and daughter-nuclides from the sing1e phase situation are not valid anymore and has to be replaced by an extended formula which takes into account: nuclide specific retardation.
Finally a multi-species system is treated as an example of how a complex system of differential equations can be simplified extensively. The species and their reactions form the major part of a aqueous system in carbonate porous medium. The initial system of eight transport equations can be simplified by using four equilibrium reactions and two constant activity species (calcite, water). Thus the entire system can be described by two differential equations.

USING THE BOUNDARY LAYER CONCEPT FOR MODELING CHEMICAL TRANSPORT IN A FRACTURE-MATRIX SYSTEM
R. Wallach
Department of Soil and Water Sciences Faculty of Agricultural, Food and Environmental Sciences
The Hebrew University of Jerusalem, Israel

Many experimental observations have shown that preferential flow and transport is taking place in fractures, fissures, and cracks that are commonly found under most field conditions. In most cases, the volume of water within the preferential paths is much lower than the volume of the stagnant fluid, but a notable part of the displacing solution may be moving within the well-defined preferential paths ahead of the main flow. The common approach to model the transport of fractured rocks is the dual porosity. This approach is based on the assumption that naturally fractured reservoirs behave as two porous structures, fractures and surrounding matrix, rather than one. The dual porosity models have also been used to describe flow and transport in structured porous media where the intra-aggregates porosity and the dead-end pores in the structured-soils are considered as the low-conductivity matrix while the inter-aggregate porosity are the paths where preferential flow takes place. The advection within the regions with the relatively low hydraulic conductivity is usually assumed zero, therefore, these domains act as sink/source components and the dual-porosity, dual-velocity model becomes a mobile-immobile model.
Several rate-limited mechanisms are involved in the solute-transfer process between the mobile and immobile solutes:

advective-dispersive transport from bulk solution to the boundary layer at the interface between the mobile and immobile porosities;
external mass transfer in the boundary layer (film diffusion);
pore diffusion within the immobile region.

If adsorption to the solid phase takes place, the appropriate equations should be added explicitly. Any or all of these three components may be rate-limiting steps and their relative contribution to the non-ideal breakthrough curves (BTCs) is not fully determined. Two different concepts have been used to describe the solute exchange between the fractures (mobile) and the surrounding matrix (immobile). The first is an instantaneous equilibration between fracture and matrix concentrations at their interface. The second is a rate limited exchange that is proportional to the difference between the laterally averaged concentrations in both fracture and matrix. The first approach is widely used by hydrogeologists and the second by soil physicists who deal with aggregated soils where the mobile solution is flowing within the macropores; the boundary layer (film) is the water surrounding the aggregates, and the immobile region is the aggregate solution. The last concept is known as the mobile-immobile model (MIM). The difference between the two concepts is embedded in the relative time scales of the affecting mechanisms. The exact nature of the immobile pore system could vary significantly for various systems. Stagnant water could be found in aggregates, surrounding solid phase particles, or in "dead-end" pores or a matrix surrounding cracks or fractures. Different diffusion length scales are involved in these different cases, where the length scale of the intra-aggregate diffusion is a few millimeters or less, while the length scale of diffusion into a matrix between well-defined fractures could be ten centimeters and more. The current study focuses on the dynamics of solute exchange between a single crack surrounded by a matrix with diffusion length scales of a few centimeters and more in the immobile solute. The effect of the detailed lateral chemical distribution within the matrix on the breakthrough curve at the crack outlet is analyzed as well.
The system setup is made of a single discrete fracture imbedded in an infinite porous rock matrix. The contaminant transport takes place in the fracture while the solution in the matrix is assumed stagnant. A boundary layer is formed along the interface between the crack and the matrix through which the solute concentration varies from its value in the flowing solution to its value in the matrix at the interface. This boundary layer is assumed to have a constant thickness for the steady flow conditions. The chemical transfer between the flowing and the stagnant solutes is a rate limited first-order process controlled by diffusion through the stagnant thin fluid film. Both mobile and stagnant fluid concentrations vary with time and depth. The flow and dissolved chemical concentrations normal to the flow direction are essentially uniform in the crack except for the boundary layer. The concentration in the matrix is not assumed to be spatially uniform, as is usually assumed in the mobile-immobile model. The lateral flux of dissolved chemicals in the matrix is by diffusion due to local concentration gradients and calculated by the diffusion equation. Due to distinctive time scales, differences between concentration variation in the vertical direction, by convection in the crack and by diffusion in the matrix, the chemical vertical fluxes in the matrix are neglected. The effect of both molecular diffusion and velocity-dependent dispersion on the concentration distribution within the crack is assumed to be minor. Thus, the dissolved chemical transport in the crack is assumed to be a sharp front convective process and is modeled by the kinematic-wave approach. An approximate analytical solution has been developed by applying the Laplace transform with respect to time. The model output has been successfully compared with measured data for 30-cm-long granitic drill cores (20-cm diameter) with a natural fissure that runs parallel to its axis and with 55-cm long undisturbed columns from pedal soil horizons with well-developed vertically oriented pores. The physical meaning of the fitted parameters is discussed.
The relative role of the two rate-limited processes, namely film transfer and diffusion in the stagnant matrix solute on the overall chemical exchange and BTC shape is analyzed by the non-dimensional version of the mass balance equations. The displacement duration has been divided into two stages. Soon after its initiation, the chemical transfer through the stagnant film controls the chemical exchange between the crack and matrix. The duration of this stage depends on different properties of the system. Subsequently, the matrix-diffusion controls the chemical exchange between the two domains and the model can then be simplified by replacing the rate-limited transfer by a local equilibrium. However, for cases in which the current study is dealing, the preferential flow is very fast and the rate-limited transfer through the stagnant film dominates the BTC shape. Using the local equilibrium at the crack-matrix interface underestimates the crack concentration during short and intermediate times after first breakthrough. The deviation between the two concepts depends on the system dimensions and flow rate in the crack.

PARTICLE FAVOURED TRANSPORT IN THE UNSATURATED ZONE
K. P. Seiler
GSF-Institute of Hydrology D-85758 Neuherberg, Germany

Reactive processes in subsurface waters often occur with slow kinetics. This sorption process, however, may become ineffective with respect to any export from the subsurface when quick flow and particle favoured transport intervene.
Observations in the unsaturated zone prove that clays, sands and the oxihydrates of iron, manganese, aluminium as well as DOC often act in subsurface waters as a carrier for nutrients and pollutants. As an example, DOC occurs accumulated in the effective root zone and drops below the it rather rapidly to concentrations close to those in groundwater (<2mg/L). This drop is attributed to sorption and microbial disintegration as well and appears less pronounced when organics occur in sediments. -During storm events the export of DOC increases to concentrations as high as in the active root zone, which obviously releases DOC; in agricultural areas this export is mostly lower than in forest areas, because disintegration, leaching and crops uptake of DOC are more pronounced in non-forest areas. Groundwater experiments in Quaternary gravels demonstrate that DOC of low molecular weight behaves quite conservatively and has often low transport capacity; contrary DOC of high molecular weight experiences generally a velocity dependent retardation. These retardation is lower in fine than in coarse-grained sediments and depends also on the age of DOC. High molecular weight DOC, however, experiences to have a high transport capacity.
The simultaneous observation of applied DOC and heavy metals (elementary), which have been applied earlier to the DOC experiment, indicates that sorpt heavy metals get exported through migrating DOC.
By convention, DOC is differentiated into fractions like humic and fulvic acids representing old and young, low and high mulecular weight fractions or hardly and easily disintegratable fractions, respectively. Fulvic acids contribute more to particle favoured transport and coprecipitation of heavy metals than humic acids; fulvic acids undergo also stronger mechanical filtering, because of its greater molecule size.

TRANSPORT OF BACTERIA THROUGH SATURATED AND UNSATURATED SOILS
S.P. Friedman
Institute of Soil, Water and Environmental Sciences, Agricultural Research Organization,
The Volcani Center, Bet Dagan 50250, Israel

This review discusses the phenomenon of bacterial transport through soil, and concentrates on the mechanisms involved in the movement and attachment of bacteria which are transported by water flowing in soil and similar granular media The macroscopic approach to describing the transport of bacteria in a porous medium is reviewed first, followed by the microscopic approach for describing the transport and attachment mechanisms for colloidal particles. By means of the microscopic approach it is possible to estimate the parameters needed for describing and predicting the transport according to the macroscopic approach. Next, experimental studies of transport of bacteria in soil columns under controlled conditions in the laboratory are reviewed. By understanding the mechanisms acting and by using the experiment a 2l results, it is possible to evaluate the effects of physical, chemical and biological factors on the mobility of the bacteria. These factors can be grouped into four categories: the characteristics of the bacterium, of the soil, of the soil solution and of the flow regime. In general, the bacterial movement is enhanced in coarse-textured soils, when the water flux is high, at high temperatures, with high pH and low ionic strength of the soil solution, when the adsorbed cations are of low valence, and in the presence of dissolved organic matter. Fe-hydroxide coatings of the sand grains and calcite minerals hinder the movement of the bacteria, while cracks, through which most of the water flow takes place under saturation conditions, enhance the mobility. The metabolic motility of the bacteria significantly improves their mobility in soil. Irrigated soils are usually unsaturated with water, which means that some soil pores are filled with water, while others are filled with air. The important problem of the effect of water content on the bacterial mobility still remains unsolved, and further research is needed to solve it. Transport of bacteria through soil has agricultural as well as environmental aspects. Knowledge of the effects of the various factors can help to determine irrigation practices involving effluent water or recharge with treated water for enriching groundwater, and in assessing the risk of pathogenic bacteria reaching the groundwater. In addition, it will serve to improve the predictability of processes in which soil bacteria are involved, mainly those related to the decomposition of organic matter and the plant nutrient cycles.

SORPTION OF ORGANIC CONTAMINANTS IN A FRACTURED CHALK FORMATION
A. Wefer-Roehl,
Blaustein Institute for Desert Research, Ben Gurion Univ. of the Negev, Sede Boker, 84990, Israel

To understand and predict transport of organic pollutants in a contaminated fractured chalk formation, the sorption attenuation capacity of the chalk, fracture filling material, and mineralized fracture wall coating was evaluated in batch tests using four aromatic compounds: ametryn, phenanthrene, m-xylene, and 2,4,6 tribromophenol, and the aliphatic molecule 1,2-dichloroethane. Sorption of these substances is very high and except for 1 ,2-dichloroethane non-linear on unoxidized grey chalk. The non-linear isotherms give comparable capacities (average 6.6 mmol/kg) for all four aromatic compounds, adsorption sites are thus considered to be the same. Sorption capacity of grey chalk is 100-1000 fold greater than the sorption capacity of oxidised white chalk; this difference is explained by a function of the nature of organic carbon (reduced versus oxidised). Grey chalk sorption capacity normalised to organic matter content gives a value comparable to those reported for activated carbon. Sorption on the fracture wall coating is nearly non-existent, as such, the distribution of coating along the walls of fractures, and the diffusion through the coating, will play important roles in organic compounds attenuation. Breakthrough curves obtained from laboratory column studies on chalk cores with a natural pronounced fracture show a clear retardation of 2,4,6 tribromophenol towards a nonsorbing tracer. Also a clear difference in retardation between white and grey chalk is demonstrated. The breakthrough curves show peaks and discrete plateaus, which indicates preferential flow paths in the fracture. Fluctuation of the flow rates despite a constant hydraulic head can be related to the dynamic channeling effects prevail during flow along a fracture within soft chalk matrix. This is supported by a decrease of flow rate with time, which can be explained by the detachment and transport of particles from the fracture wall and as a consequence, decreasing in the effective hydraulic aperture. The results of the column experiments were used in a flow and transport model with a numerical solution. The model was calibrated in the sorption stage using also diffusion coefficients, obtained from a diffusion experiment. The fracture aperture, the longitudinal dispersivity and the exchange coefficient were obtained from this model. The same parameters were later used in the desorption stage for validation.

THE INTERFACE BETWEEN FRACTURE SURFACES AND AQUEOUS SOLUTIONS IN UNSATURATED CHALK: PHYSICAL, CHEMICAL AND COLLOID GENERATING PROCESSES
N. Weisbrod
Department of Soil and Water Sciences Faculty of Agricultural, Food and Environmental Quality Sciences
The Hebrew University of Jerusalem, P.O.B.12 Rehovot 76100, Israel

Over the past two decades, it has been recognized that in rock formations a large fraction of water and contaminants is transported through fractures rather than through the entire rock matrix. Therefore, considerable effort has been extended toward understanding the characteristics of flow in fractures. In this study, the various phenomena occurring at the interface between the chalk fracture surface and the flowing fluid were explored under laboratory and field conditions. In a preliminary set of experiments, fracture surface erosion was studied under stagnant conditions and a new method of accurately assessing the magnitude of erosion was developed. An improved setup was later developed to study surface erosion in a system that simulates intermittent flow along unsaturated fractures under field conditions. In this newly developed "flow-through" system, the impact of the interaction between the surface coating and rainwater/industrial wastewater percolation on the fracture surface was studied. The influence of wetting and drying cycles on salt deposition and its dissolution during successive infiltration events was also studied. The results of this research suggest that erosion of chalk fracture surfaces can occur under conditions of periodic infiltration events. Properties such as surface roughness, flow channels and coating may rapidly change within hours of flow, resulting in temporal variations of the aperture geometry. Erosion was caused mainly by particle detachment from the fracture surface. These particles may clog the fracture aperture or enhance colloidal-facilitated transport of contaminants. The results presented in this study suggest that models describing periodic flow through fractures in the vadose zone should take into account the variability of the aforementioned properties. The exact impact of these variabilities depends on the matrix and coating properties, the duration of the flow periods and the chemical composition of the infiltrating solution. Wetting and drying cycles of the fracture surface resulted in salt accumulation on it due to water evaporation at the surface followed by capillary transport of water and salts from the bulk matrix toward the fracture surface. These salts were subsequently dissolved by the water infiltrating along the fracture surface (during the wetting stage) The aforementioned processes were quantitatively studied in the laboratory and qualitatively confirmed through field experiments.

MANAGEMENT OF REMEDIATION OF CONTAMINATED SOILS, WITH REFERENCE TO ORGANIC CONTAMINANTS
T. C. Harmon
Civil and Environmental Engineering, UCLA, (tch@seas.ucla.edu 310-206-3735)

This presentation begins with an overview of soil remediation technologies currently seeing appreciable use in the United States. Advantages and disadvantages of each are discussed with regard to appropriate soil and chemical types, technical feasibility and project cost and duration. Then, the presentation presents the details of an on-going UCLA project aimed at testing the feasibility of natural gas-fueled combustion process to remediate heavily contaminated urban soils containing a toxic residue called lampblack. The primary contaminants of interest in lampblack are polycyclic aromatic hydrocarbons (PAHs). The in situ approach combines low temperature thermal desorption (300-400 0C) with soil vapor with the objective of accelerating PAH removal without combusting them. Natural gas combustion was to fuel the process in order to produce steam and lower energy costs.
Batch and bench-scale prototype experiments are presented for removal from a silty sand and urban fill mixture containing lampblack and 11 measurable PAHs. First, a one-dimensional glass soil column was packed with lampblack-laden soil. A steel tube was placed longitudinally through the core of this column as a gas-burning system. Temperatures ranging from roughly 700 0C (at the core) to 250 ~C (at the outer walls) were achieved and maintained for 10 hours. In a larger, three-dimensional bench-scale test, an air-extraction system was introduced to enhance the PAH removal process. This box system was packed with a silty sand and lampblack matrix. The total PAH (TPAH) soil concentration was reduced from over 4000 ppm (by wt.) to less than 100 ppm during a 35 day residence time at 350 0C. The observed contaminant removal rate is modeled reasonably well as a multicomponent vaporization process, assuming an ideal, nonaqueous PAH-surrogate hydrocarbon mixture. However, the proposed model failed to reproduce the slow removal observed in the final stages of the experiment. The observed residue may persist at this temperature as a low vapor pressure tar mixture and/or as a slowly desorbing fraction. Given the presence of this fraction, the overall feasibility of the proposed in situ thermal desorption technology will depend on soil heterogeneity, cost/benefit analysis and risk assessment. Finally, early results from a pilot-scale study employing the process are presented. Preliminary indications are that the process can provide a significant savings by avoiding excavation costs provided that efficient heat distribution and moisture control can be achieved.

TRANSPORT OF ORGANIC AND INORGANIC CONTAMINANTS IN DESERT SOIL:-EVALUATION OF CONTAMINANTS FLUSHING POTENTIAL FROM A CONTAMINATED SOIL NEAR RAMAT-HOVAV INDUSTRIAL PARK
S. Arnona, E. Adara,b, Z. Ronenb, A. Yakirevichb
Department of Geological and Environmental Sciences Ben Gurion University of the Negev, 84105 Beer Sheva, Israel
The Jacob Blaustein Institute For Desert Research, 84990 Sede Boker, Israel

During the years 1987-1991 a total area of 26 acres was sprayed with wastewater from Rarmat - Hovav industrial park. The spraying was an additional method for diminishing the volume of wastewater. The wrong assumption that the contamination will be adsorbed within the topsoil caused a severe soil and groundwater contamination that finally led to stop this method of discharge. The cluster of contaminants in the topsoil is considered as a potential source for a continuous groundwater contamination.
The main objective of this research was to evaluate the possibility of reducing groundwater and downstream contamination by flushing out the contaminants from the soil. Dozens of organic and inorganic contaminants were identified mainly in the topsoil. Four major contaminants, TBP (Tribromophenol), TBBPA (Tetrabromobisphenol A), Cl- and Br were chosen as tracers for the research experiments. An undisturbed contaminated soil taken from the sprayed area was packed in columns and flushed by water. According to the breakthrough curves the average retardation factors for TBP and TBBPA were 2 and 3 respectively. The residual concentrations in the flushed soil were mainly those of TBBPA because it's low solubility. In soils that are contaminated for a long period of time, a biphasic desorption (fast and slow) is a common phenomenon. Fast desorption and transport of TBBPA in the soil occurs mainly in areas located close to the main flow channels. Slow desorption and/or diffusion rates compared to the advective flow rate may account for the slow release of TBBPA in the second phase. This slow release of TBBPA can last months or even years. Hence, the amount of TBBPA that was not flushed and removed during the fast phase was found in the soil as the residual concentration.
Because the spraying area is located mostly on a slope and the soil contains large amount of sodium that upon wetting causes dispersion of soil particles, even small amounts of rainfall infiltrates and seals the topsoil. The rest of the rainfall turns into runoff. The sealing of the soil and the strong evaporation results in almost no deep percolation toward groundwater and therefore most of the contaminants are concentrated in the upper part of the soil. Despite of the above-mentioned, contaminant migration toward groundwater can occur through preferential flow paths in the soil and through fractures in the chalk formation. Also, contaminants might be transported in the runoff as solutes or sorbed on soil particles. The contamination in the runoff drains to the nearby stream and into the alluvial aquifer, or into the Eocene aquitard via fractures in the underlying Chalk.

SOIL MOISTURE MEASUREMENTS BY SYNTHETIC APERTURE RADAR AND SCATTEROMETERS IN THE ARID AND SEMI-ARID ENVIRONMENT
Dan G. Blumberg
Department of Geography and Environmental Development, Ben-Gurion University of the Negev, Be'er Sheva, Israel.

Water is a scarce commodity in desert regions, yet, most remote sensing experiments of soil moisture have been conducted in temperate areas. The ability to detect and reliably assess soil moisture in the arid environment has important implications both at the micro scale such as the individual farm and at the macro scale for climate modeling.
A series of experiments utilizing an airborne scatterometer operating at P-band were conducted to assess the ability to measure soil moisture over farmlands in the Negev Desert by remote sensing. The soil moisture is determined by measuring the backscattered radar signal at the scatterometer.
Results show that even in the arid environment of the Negev scatterometer data correlate well with field measured soil moisture. The values measured were between 2 and 7%. Some extreme values were measured at the Ramat Negev experimental farm. A later visit to the farm confirmed a faucet leak at the location of the extensive values (>20%.) We could also detect leaks from a sewer pond adjacent to Mashabei Sadeh with this system.
A second series of experiments were conducted with spaceborne synthetic aperture radar using the ERS SAR system. This system is single wavelength operating at C-band HH polarization. We used two images to achieve two incidence angles for the same site. Results show that 43% of the variance over individual fields based on the soil moisture measurements.
The future work will concentrate on measurements of surface roughness and soil moisture to improve the explained variance and the model.

Comments on the workshop
The workshop was attended by a number of leading experts and we are fortunate that they were willing to put their impression before us bluntly and constructively. As aptly put by John Philips: "the true fruits of workshops are: the lubrication of person to person contacts which are the heart of real research collaboration. They jolt participants with conflicting views and viewpoints and they remind everyone of the big world outside their own knowledge and experience".
First here is an edited selection of critique of the workshop, as expressed by these experts, and their recommendations for possible future work;
J. Philip: The practical goal of unsaturated flow research is to establish an economical, usable, macroscopic, Darcy-scale phenomenological approach to describing and understanding the various processes. There was mental confusion between the scales of discourse; with concepts and quantities on the microscopic (Navier-Stokes) scale muddled up with those on the macroscopic (Darcy) scale. Of course microscopic studies can provide helpful critique of macroscopic formulations, but we need to understand that is what we're doing.
Too many people may be impressed with impenetrable mathematical adventures telling us nothing new nor correct about the real physics. There are two broad classes of scientific mind, which we might call physical and biological. The physical mind seeks for the useful simplification, the economical generalization. It attempts to home in on and cling to the relevant; and not to be distracted by the superficial and irrelevant. On the other hand, the biological mind is alert to the exception. Sometimes this is extremely valuable and existing theory must be challenged. But the biological mind can get excited over what are, in reality, red herrings. I confess to having a physical mind; and so I lamented what I saw as an excess pursuit of red herrings at the workshop.
One excellent aspect of the biological mind is that it is happy to go out in the field and gather data, often painstakingly and at great expense. This is scientific virtue in a world where modellers, all readily, turn their backs on the real world and generate space-age fairy tales. But there is a drawback to this 'biological' field naturalism. In small-catchment studies, for example, one can only speculate on the significance of concentration data at the outlet, when we don't have data on the 3-dimensional above - and below - surface water flows of the catchment. Field naturalists can sit back and suppose the research is done when it only has begun.

E. Holzbecher: Human activities shift further into the dry regions, which have been mostly unrecognized by men in the past. Israel is an excellent example, where this shift can be observed. But the same thing happens all over the world, where increasing population, particularly in fertile regions and under good climate, increases the pressure to expand in regions with less favorable conditions. It is crucial to have a scientific basis for all activities connected with the unsaturated zone in arid semi-arid regions. The unsaturated zone is very important, because activities like water consumption, waste storage and agriculture - and others - all have an influence on the unsaturated zone. Thus the subject of the workshop was a very good choice.
Many presentations showed that processes in the unsaturated zone are very complex. That is not new and amazing in itself. From most compartments of the real world the unsaturated zone is nearest to the ground surface where most complex processes and different activities are going on. From our studies in Germany I conclude that the vadose zone is complex everywhere, not only under dryland conditions.
Water flow itself is influenced in a complex manner by several processes - this was the message in several presentations of the workshop. The theory of flow in the unsaturated zone states complex nonlinear interactions (Philip), infiltration is a special topic of flow (Philip), phenomena like by-pass flow complicate the situation (Seiler), the capillary fringe needs extra considerations (Ronen). The problem to find appropriate measurements for the flow estimations was mentioned (Dahan, Friedman). A topic aside from all that is the flow in fractured porous media which was touched in several presentations.
Finally it can be summarized that some scientific and application fields were well represented, some not well in relation to their importance. Hydraulics and biogeochemistry (incl. bioremediation) had a lot of contributions. Transport in the unsaturated zone had an appropriate representation. Less weight was given to hydrology and geo-chemistry. The complex aspects of hydrology was treated by Kulls, talking about field studies in the Kalhari desert, and from the point of measurements by Blumberg, talking about remote sensing to determine soil moisture. A conceptual model of geo-chemistry was given by Holzbecher - aside from that chemistry was hardly recognized.
As for application, technology, cleanup and remediation were in the focus of the workshop. Several presentations were related to pollution problems in a nearby chemical production plant. Other activities are equally interesting, but were somewhat underrepresented. Most of all agriculture, which has a lot of influence on soil (flow is changed from irrigation, substances are brought into the subsurface) and is vice-versa influenced by soil conditions.
Altogether the workshop was very useful. It lies in the nature of the subject that there seemed to be more questions than answers.

K.-P. Seiler: The recent workshop on Transport and clean up processes in the unsaturated zone of consolidated and unconsolidated rocks under dryland conditions covered the fields of

soil physics,
tracer techniques and numerical modeling in seepage studies,
the physical and chemical behaviour of pollutants in both types of rocks
and microbia aspects of remediation technologies;

it did not cover the wide field of rain agriculture, irrigation practice and botany and just touched the role of remote sensing techniques in drylands.
Drylands are known to represent very sensitive ecosystems in an environment of huge variations of climatic conditions. They contribute

to plant growth mainly by long term water retention and capillary water movement,
to quite distinctive ground water recharge and
to resilience mechanisms with respect to man induced changes, due to their microbiologic inventory.

Quantities and limits of these properties, however, are not well understood or even unknown, although much research has been initiated in these and related fields. Especially, scale related research to cover simultaneously process mechanisms on the small scale and the highly aggregated scale and a focus on the interaction of the governing factors on different scales are missing.
In the field of microbiology, the laboratory scale dominates both world wide and in Sede Boqer and it should be supplemented by more field research; knowledge of natural biodiversity in drylands and its variability as a bench mark for future activities in drylands is further needed. The use of modern molecular biologic tools, to count and characterize individually and through its activity microbial communities, should be strongly promoted. Also studies of botanical adaptation and sequences in untouched and cultivated drylands are needed.
In the field of ground water recharge, more emphasis is needed on input/output relations both by rain and condensation processes and outflow and evaporative processes. Such studies should include local as well as regional efforts including tracer techniques on the small scale and environmental isotope and geochemical tools on the catchment scale.
The role of the unsaturated zone as a reservoir and as throughflow compartment of atmospheric water plays an important role in developing strategies for a sustainable use of dessert ecosystems by man. It determines especially soil salinisation processes.
For all these biotic and abiotic activities a common crystallization point should be created as e.g. a research catchment to study desert hydrologic ecosystems of Mediterranean or Afro-Arabic type, in which all the research activities on virgin conditions and on the consequences of man induced changes of this environment should get concentrated. The local and individual activities of other areas from the past and in the future should get aggregated to this crystallization point aiming to regionalise results, to develop long-term strategies of conservation of selfpurification processes and to develop early warning tools to prevent unpredicted changes in time.
Such an activity has to include three components:

a long-term monitoring program,
process research and
the regionalisation of results.

These components imply to engage a permanent staff to guarantee monitoring and continuity of ongoing studies and a fluctuating staff that might consist of master and PhD students, sabbatical scientists and postdocs as well. if possible the state research institutions should get involved in this university led study as well. Since it will probably not be possible to install such a crystallization area at once it is suggested to select a respective area and to aggregate stepwise research programs to this area. To do so this selection should be prepared interinstitutionally.
The area between Sede Boqer and Beer Sheva seems to be predestinated for such a crystallization catchment because it belongs to an arid close to a semiarid area and thus could cover most of the dryland conditions, features and gestations in the Mediterranean area.
To optimize this research all data and results should be sampled on a data bank and be evaluated by modeling techniques to

formulate working hypothesis,
validate and calibrate models,
prove the consistency of data,
develop future management strategies including socio-economic aspects and develop tools of early warning systems.

The author is aware of many projects planned in the future by Israelis which have common backgrounds but not yet any local or interinstitutional coordination. It is suggested to try to coordinate these planning's in brain storming actions before applying for it on the international or national floor. As an example there exist efforts in the use of waste waters for irrigation. This has different aspects as e.g. hydrogeological, water quality and socio-economics. There is, however, no concerted action in formulating better efficiencies in the interaction of these activities.
Special tasks for further and additional research are seen in the following fields:

Transfer of unproductive evaporation losses into transpiration losses in areas with deep ground water tables, Reduction of unproductive evaporation losses in areas with water tables close to the surface,
Study of the migration of micro-organisms and colloids in unsaturated compacted and uncompacted dryland sediments, Study of incubation times of microbiologic disintegration and adaptation processes in dryland areas in response to changes in land use, Integrative studies of mechanisms and quantities of ground water recharge and discharge in time and space,
Start of remote sensing, environmental isotope, microbiologic and hydrochemical monitoring.

Synopsis and Reflections
J.R. Gat
The simple translation of processes, models and methods in the study of the unsaturated zone from the more humid environment to arid conditions is usually inadequate, as it does not account for the non-steady inputs, the patchy and variable nature of the hydrologic regime, the wetting and drying cycles and the interactions of soil minerals, plants and microbial activity with the moisture flux, which differ substantially from those of the non-arid setting. Phenomena such as swelling, cracking, variable plant cover with its associated evapo-transpiration demands, aeolian inputs, soil erosion and sediment transport, all have an influence both on the fluid transport and the water quality. Of special concern is the non-reversible nature of changes, such as result in desertification, salinisation and deterioration of the system.
While the hydraulic aspects have attracted some attention, the geochemical and microbiological implication apparently are not well known. As a result it is difficult to predict the remedial role of the soil zone in the case of pollution, and its contribution to the resilience of the arid system.
The suggested steps to be taken, as a first measure, is the documentation of the changes of as many parameters as possible, i.e. temperature, soil moisture, surface runoff and percolation (both water salinity and isotopic composition), the ecological structure, chemical load and microbial activity, on representative landscape units. Some of these should be native and others subject to changes in land use, irrigation, etc. Long term monitoring over seasonal as well as annual cycles (especially drought conditions) is essential. Ways will have to be found to characterize the soil system for its different function, namely as a conduit, reservoir and modifying medium for both water and its chemical load.
The long-term monitoring can then be accompanied by tracer applications in order to study the dynamics of the system, hysteresis effects and irreversible processes using either in-situ measurements, controlled lysimeters or smaller scale test facilities.