An infiltration model based on flow variability in macropores: development, sensitivity analysis and applications

Simulating infiltration in soils containing macropores still provides unsatisfactory results, as existing models seem not to capture all relevant processes. Recent studies of macropore flow initiation in natural soils containing earthworm channels revealed a distinct flow rate variability in the macropores depending on the initiation process. When macropore flow was initiated at the soil surface, most of the macropores received very little water while a few macropores received a large proportion of the total inflow. In contrast, when macropore flow was initiated from a saturated or nearly saturated soil layer, macropore flow rate variation was much lower. The objective of this study was to develop, evaluate, and test a model, which combines macropore flow variability with several established approaches to model dual permeability soils. We then evaluate the INfiltration–INitiation–INteraction Model (IN³M) to explore the influence of macropore flow variability on infiltration behavior by performing a sensitivity analysis and applying IN³M to sprinkling and dye tracer experiments at three field sites with different macropore and soil matrix properties. The sensitivity analysis showed that the flow variability in macropores reduces interaction between the macropores and the surrounding soil matrix and thus increases bypass flow, especially for surface initiation of macropore flow and at higher rainfall intensities. The model application shows reasonable agreement between IN³M simulations and field data in terms of water balance, water content change, and dye patterns. The influence of macropore flow variability on the hydrological response of the soil was considerable and especially pronounced for soils where initiation occurs at the soil surface. In future, the model could be applied to explore other types of preferential flow and hence to get a generally better understanding of macropore flow.     

Effects of antecedent moisture and macroporosity on infiltration and water flow in frozen soil

Infiltration into frozen soil plays an important role in soil freeze–thaw and snowmelt-driven hydrological processes. To better understand the complex thermal energy and water transport mechanisms involved, the influence of antecedent moisture content and macroporosity on infiltration into frozen soil was investigated. Ponded infiltration experiments on frozen macroporous and non-macroporous soil columns revealed that dry macroporous soil produced infiltration rates reaching 10³ to 10⁴ mm day⁻¹, two to three orders of magnitude larger than dry non-macroporous soil. Results suggest that rapid infiltration and drainage were a result of preferential flow through initially air-filled macropores. Using recorded flow rates and measured macropore characteristics, calculations indicated that a combination of both saturated flow and unsaturated film flow likely occurred within macropores. Under wet conditions, regardless of the presence of macropores, infiltration was restricted by the slow thawing rate of pore ice, producing infiltration rates of 2.8 to 5.0 mm day⁻¹. Reduced preferential flow under wet conditions was attributed to a combination of soil swelling, due to smectite-rich clay (that reduced macropore volume), and pore ice blockage within macropores. In comparison, dry soil column experiments demonstrated that macropores provided conduits for water and thermal energy to bypass the frozen matrix during infiltration, reducing thaw rates compared with non-macroporous soils. Overall, results showed the dominant control of antecedent moisture content on the initiation, timing, and magnitude of infiltration and flow in frozen macroporous soils, as well as the important role of macropore connectivity. The study provides an important data set that can aid the development of hydrological models that consider the interacting effects of soil freeze–thaw and preferential flow on snowmelt partitioning in cold regions.     

Estimating macropore and matrix flow using the hydrograph separation procedure in an experimental forest plot

This study was conducted to estimate macropore space, macropore flow and matrix flow in an experimental forest plot in the Ouachita Mountains of Arkansas. Lateral soil water fluxes and soil capillary potentials were observed in the isolated plot during applied rainfall experiments. Rainfalls were applied 17 times during the period 17 July to 10 October 1991. The subsurface hydrograph separation technique was used to estimate macropore space, macropore flux and matrix flux. The boundary between macropore and matrix flow was statistically determined by covariance analysis. The maximum estimated lateral macropore space was approximately 0.006 (cm³ cm^?3). The maximum estimated lateral macropore and matrix flow were 0.042 and 0.00066 cm s^?1, respectively. This report also emphasizes the need for further research on the hydrograph separation procedure for estimating macropores and macropore flow.     

Tracing solute infiltration using a combined method of dye tracer test and electrical resistivity tomography in an undisturbed forest soil profile

An accurate prediction of solute infiltration in a soil profile is important in the area of environmental science, groundwater and civil engineering. We examined the infiltration pattern and monitored the infiltration process using a combined method of dye tracer test and electrical resistivity tomography (ERT) in an undisturbed field soil (1 m × 1 m). A homogeneous matrix flow was observed in the surface soil (A horizon), but a preferential flow along macropores and residual rock structure was the dominant infiltration pattern in the subsurface soil. Saturated interflow along the slopping boundaries of A and C1 horizons and of an upper sandy layer and a lower thin clay layer in the C horizon was also observed. The result of ERT showed that matrix flow started first in A horizon and then the infiltration was followed by the preferential flows along the sloping interfaces and macropores. The ERT did not show as much detail as the dye‐stained image for the preferential flow. However, the area with the higher staining density where preferential flow was dominant showed a relatively lower electrical resistivity. The result of this study indicates that ERT can be applied for the monitoring of solute transportation in the vadose zone. Copyright © 2010 John Wiley & Sons, Ltd.     

Differences in preferential flow with antecedent moisture conditions and soil texture: Implications for subsurface P transport

Preferential flowpaths transport phosphorus (P) to agricultural tile drains. However, if and to what extent this may vary with soil texture, moisture conditions, and P placement is poorly understood. This study investigated (a) interactions between soil texture, antecedent moisture conditions, and the relative contributions of matrix and preferential flow and (b) associated P distributions through the soil profile when fertilizers were applied to the surface or subsurface. Brilliant blue dye was used to stain subsurface flowpaths in clay and silt loam plots during simulated rainfall events under wet and dry conditions. Fertilizer P was applied to the surface or via subsurface placement to plots of different soil texture and moisture condition. Photographs of dye stains were analysed to classify the flow patterns as matrix dominated or macropore dominated, and soils within plots were analysed for their water‐extractable P (WEP) content. Preferential flow occurred under all soil texture and moisture conditions. Dye penetrated deeper into clay soils via macropores and had lower interaction with the soil matrix, compared with silt loam soil. Moisture conditions influenced preferential flowpaths in clay, with dry clay having deeper infiltration (92 ± 7.6 cm) and less dye–matrix interaction than wet clay (77 ± 4.7 cm). Depth of staining did not differ between wet (56 ± 7.2 cm) and dry (50 ± 6.6 cm) silt loam, nor did dominant flowpaths. WEP distribution in the top 10 cm of the soil profile differed with fertilizer placement, but no differences in soil WEP were observed at depth. These results demonstrate that large rainfall events following drought conditions in clay soil may be prone to rapid P transport to tile drains due to increased preferential flow, whereas flow in silt loams is less affected by antecedent moisture. Subsurface placement of fertilizer may minimize the risk of subsurface P transport, particularily in clay.     

Macroporosity and infiltration in blanket peat: the implications of tension disc infiltrometer measurements

Little is known about the processes of infiltration and water movement in the upper layers of blanket peat. A tension infiltrometer was used to measure hydraulic conductivity in a blanket peat in the North Pennines, England. Measurements were taken from the surface down to 20 cm in depth for peat under four different vegetation covers. It was found that macropore flow is a significant pathway for water in the upper layers of this soil type. It was also found that peat depth and surface vegetation cover were associated with macroporosity and saturated hydraulic conductivity. The proportion of macropore flow was found to be greater at 5 cm depth than at 0, 10 and 20 cm depth. Peat beneath a Sphagnum cover tends to be more permeable and a greater proportion of macropore flow can occur beneath this vegetation type. Functional macroporosity and matrix flow in the near‐surface layers of bare peat appear to have been affected by weathering processes. Comparision of results with rainfall records demonstrates that infiltration‐excess overland flow is unlikely to be a common runoff‐generating mechanism on blanket peat; rather, a saturation‐excess mechanism combined with percolation‐excess above much less permeable layers dominates the runoff response. Copyright © 2001 John Wiley & Sons, Ltd.     

MACROPORE FLOW AND MASS WASTING OF GULLIES IN THE LOESS PLATEAU, CHINA

Due to the extensive gullying from historically excessive erosion in the loess plateau of China, much of this region is being converted to native grass and shrub vegetation. Tunnel scour and mass wasting are important gully erosion processes resulting from preferential flow through macropores （ pores 〉 1 mm diameter）. The objective of this study is to assess the changes with time in macropore flow characteristics of soils on the Loess Plateau following conversion to grass vegetation and the associated degree of mass wasting of gully faces. Ridge areas that had been revegetated for 1 year, 6 years, and 〉 15 years following tilling, and for 6 years following contour-ditching and the adjacent gully faces were characterized for their macropore and soil matrix properties on a 50 cm by 50 cm area. The total number of macropores increased from 11.6/m^2 to 39.6/m2 from 1 to 6 years and to 51.6/m2 after 15 years of revegetation following tillage. The macroporosity increased from 0.0008 m^3/m^3 to 0.0018 m^3/m^3 from 1 to 6 years of revegetation following tillage but the lowest macroporosity （0.0005 m3/m3） was 6 years of revegetation following contour-ditching. The contourditched area had the lowest infiltration rate （95 m/d） through the soil matrix （areas without macropores） with the tilled areas having similar infiltration rates regardless of the number of years of revegetation （averaged 146 m/d）. Due to tunnel scour erosion of macropores during infiltration into the area revegetated for 1 year, pore diameters enlarged by more than 200% resulting in this condition having the highest individual macropore infiltration rates （7967 m/d）. Macropores in all other areas were stable with no tunnel scour erosion of macropores. The total capacity for infiltration through macropores increased significantly with time following revegetation. The number of macropores on the gully faces was triple （92.8/m2） and the macroporosity quadruple （0.004 m3/m3） that of the ridge surfaces. The upper gully faces exhibited 1.1 slumps m^-1 for a total soil loss of 48622 kg per ha.     

Soil macroporosity and infiltration characteristics of a forest podzol

Spatial distribution of soil macroporosity was determined for a forest podzol from tension infiltrometer measurements at the soil surface. Surface‐derived macroporosity values were compared with point infiltration characteristics obtained from soil water content and soil water chemistry measurements during an experimental irrigation, and with parameters of a kinematic wave model applied to soil water content data. Macroporosity estimated by the tension infiltrometer ranged from 0·00087 to 0·0219% of soil volume, and infiltration at these two sites was dominated by propagation of a well‐defined wetting front through the soil profile and bypass flow via soil macropores, respectively. Infiltration at sites with intermediate macroporosities reflected a combination of these two processes, although results were inconclusive at one site owing to lateral flow at the base of the soil profile. There was no agreement between macroporosities estimated by the tension infiltrometer and the kinematic wave model. The maximum soil conductance parameter within the profile at a site, however, was related directly to the surface‐derived macroporosity. The partial agreement between surface‐derived macroporosity estimates and point infiltration characteristics shown here supports the use of tension infiltrometry as a rapid, non‐destructive method of assessing spatial variations in the relative contribution of macropore flow to the infiltration process. Copyright © 2000 John Wiley & Sons, Ltd.     

A model to couple overland flow and infiltration into macroporous vadose zone

Most vegetated land surfaces contain macropores that may have a significant effect on the rate of infiltration of water under ponded conditions on the ground surface. Owing to the small-scale variations of the land topography (microtopography), only portions of the land area may get ponded during the process of overland flow. As the macropores transmit water at much higher rates than the primary soil matrix, higher macropore activation in ponded areas produces larger effective infiltration rates into the soil. Therefore, overland flow and infiltration into the macroporous vadose zone are interrelated. Representing the microtopographic variation of the land surface by a simple sine wave function, a method was developed to relate the ponding area to the average ponding depth which was determined by overland flow. A numerical model coupling overland flow and infiltration into the macroporous vadose zone was developed. Overland flow was simulated using the St. Venant equations with the inertia terms neglected. A single macropore model was used to simulate the infiltration into the macroporous vadose zone. The interaction between overland flow and the infiltration into the macroporous vadose zone was analyzed for a hypothetical watershed. The sensitivity analysis revealed that the interaction of macropore flow and overland flow is significant. For the conditions tested, the macropore flow and the overland flow were found to be more sensitive to the macroporosity and less sensitive to the microtopographic surface variation.     

Soil hydrological response under simulated rainfall in the Dehesa land system (Extremadura,SW Spain) under drought conditions

Soil hydrology was investigated in the Guadelperalón experimental watershed in order to determine the influence of land use and vegetation cover on runoff and infiltration within the Dehesa land system. Five soil–vegetation units were selected: (1) tree cover, (2) sheep trials, (3) shrub cover, (4) hillslope grass and (5) bottom grass. The results of the simulated rainfall experiments performed at an intensity of 56·6 mm h⁻¹ during one hour on plots of 0·25 m², and the water drop penetration time test indicate the importance of water repellency in the Dehesa land system under drought conditions. Low infiltration rates (c. 9–44 mm h⁻¹) were found everywhere except at shrub sites and in areas with low grazing pressure. Soil water repellency greatly reduced infiltration, especially beneath Quercus ilex canopies, where fast ponding and greater runoff rates were observed. The low vegetation cover as a consequence of a prolonged drought and grazing pressure, in conjunction with the soil water repellency, induces high runoff rates (15–70 per cent). In spite of this, macropore fluxes were found in different locations, beneath trees, on shrub-covered surfaces, as well as at sites with a dominance of herbaceous cover. Discontinuity of the runoff fluxes due to variations in hydrophobicity causes preferential flows and as a consequence deeper infiltration, especially where macropores are developed. © 1998 John Wiley & Sons, Ltd.     

Macropores and preferential flow—a love‐hate relationship

Preferential flow is of high relevance for runoff generation, transport of chemicals and nutrients, and the transit time distribution of water in the soil or watershed. However, preferential flow effects are generally ignored in lumped hydrological models. And even most physically‐based models ignore macropores and preferential flow features at the soil and hillslope scale. Keith Beven was never satisfied with this situation and he tried again and again to convince the scientific community to focus their research on the complex topic of macropore and preferential flow. Although he recognized how difficult it is to correctly include preferential flow in hydrological models, he made substantial progress defining and describing macropore flow and showing its relevance, developing models to simulate preferential flow, and in particular, the interaction between macropores and the soil matrix. In this short commentary, I reflect on these achievements and outline a vision for research in preferential flow experiments and modeling.     

Topographic controls upon soil macropore flow

Macropores are important components of soil hydrology. The spatial distribution of macropore flow as a proportion of saturated hydraulic conductivity was tested on six humid–temperate slopes using transects of tension infiltrometer measurements. Automated water table and overland flow monitoring allowed the timing of, and differentiation between, saturation‐excess overland flow and infiltration‐excess overland flow occurrence on the slopes to be determined and related to tension‐infiltrometer measurements. Two slopes were covered with blanket peat, two with stagnohumic gleys and two with brown earth soils. None of the slopes had been disturbed by agricultural activity within the last 20 years. This controlled the potential for tillage impacts on macropores. The proportion of near‐surface macropore flow to saturated hydraulic conductivity was found to vary according to slope position. The spatial patterns were not the same for all hillslopes. On the four non‐peat slopes there was a relationship between locations of overland flow occurrence and reduced macroporosity. This relationship did not exist for the peat slopes investigated because they experienced overland flow across their whole slope surfaces. Nevertheless, they still had a distinctive spatial pattern of macropore flow according to slope position. For the other soils tested, parts of slopes that were susceptible to saturation‐excess overland flow (e.g. hilltoes or flat hilltops) tended to have least macropore flow. To a lesser extent, for the parts of slopes susceptible to infiltration‐excess overland flow, the proportion of macropore flow as a component of infiltration was also smaller compared with the rest of the slope. The roles of macropore creation and macropore infilling by sheet wash are discussed, and it is noted that the combination of these may result in distinctive topographically controlled spatial patterns of macropore flow. Copyright © 2008 John Wiley & Sons, Ltd.     

Effect of macropores on soil freezing and thawing with infiltration

An understanding of heat transport and water flow in unsaturated soils experiencing freezing and thawing is important when considering hydrological and thermal processes in cold regions. Macropores, such as cracks, roots, and animal holes, provide efficient conduits for enhanced infiltration, resulting in a unique distribution of water content. However, the effects of macropores on soil freezing and thawing with infiltration have not been well studied. A one‐directional soil‐column freezing and thawing experiment was conducted using unsaturated sandy and silt loams with different sizes and numbers of macropores. During freezing, macropores were found to retard the formation of the frozen layer, depending on their size and number. During thawing, water flowed through macropores in the frozen layer and reached the underlying unfrozen soil. However, infiltrated water sometimes refroze in a macropore. The ice started to form at near inner wall of the macropore, grew to the centre, and blocked flow through the macropore. The blockage ice in the macropore could not melt until the frozen layer disappeared. Improving a soil freezing model to consider these macropore effects is required. Copyright © 2016 John Wiley & Sons, Ltd.     

Dual-porosity and kinematic wave approaches to assess the degree of preferential flow in an unsaturated soil

Abstract

The purpose of this study was to assess the degree of preferential flow in an unsaturated soil column using two different models: the dual-porosity model, MACRO, and the kinematic wave approach (KWA) based on boundary-layer flow theory. The soil column experiments consisted of six infiltrations with intensities varying from 15 to 101 mm h^?1. Bromide solution was also infiltrated at an intensity of 79 mm h^?1 and a concentration of 80 mg l^?1. Both MACRO and the KWA indicated the absence of pure preferential flow. The KWA indicated intermediate flow with dispersion of the wetting front with depth, whereas MACRO indicated flow dominated by the diffusion of capillary potential. These results shed light on the transition between flows dominated by momentum dissipation and by diffusion of capillary potential. The absence of pure macropore flow in the structured sandy soil is mainly due to efficient lateral mass exchange in this material.     

Interactions between perched and saprolite aquifers on a small,salt-affected and deeply weathered hillslope

A small hillslope was chosen to investigate the role of throughflow as a mechanism responsible for the movement of soil water and solutes towards a saline seep and as a source of recharge to a permanent, regional aquifer at depth. The hydraulic properties, chemical characteristics and physical responses of both systems were studied on a deeply weathered, salt-affected hillslope. Additional data were also obtained from other sites in south-western Australia. Regional groundwater flow occurred in a variably textured, deeply weathered material in which the hydraulic conductivity varied from < 0·001 to 0·14m day^?1. Perched groundwater flow (throughflow) occurred in the higher permeability (? 1·5 m day^?1), near-surface soil materials. Throughflow occurred throughout winter, contributing approximately 530 m³ of fresh (? 160 mg l^?1 Cl) water to a saline seep. By contrast, the deep aquifer discharged approximately 1100 m³ of waters with salt concentrations of 2000–6000 mg l^?1 Cl. Recharge and discharge rates to and from the deep aquifer, were estimated to be of the order of 5–20 mm a^?1 and 50–300 mm a^?1 respectively. Saturated conditions existed throughout winter within the seep and the immediately adjacent non-saline area, with up to 60 per cent of the hillslope soils becoming saturated after major rainfall events ( > 20 mm day^?1). In the mid-slopes, in particular along a central depression, saturation of the shallow soils caused macropore channel recharge to take place through the clay-textured subsoils. Water-level responses suggest that approximately 25–30 per cent of annual recharge occurred from one storm studied in September 1984. Recharge through macropore channels is a significant mechanism in the concave slope areas on the hillslope. Throughflow was found to be a major source of water, but not salt, contributing to the saline seep. In general, the contribution of throughflow was found to decrease further inland at other sites studied. However, at inland sites where perennial, perched aquifers have developed in deep sands, saline areas have been caused by throughflow and not by deep aquifer discharge.     

Parametric study of a physically-based,plot-scale hillslope hydrological model through virtual experiments

Abstract

A physically-based hillslope hydrological model with shallow overland flow and rapid subsurface stormflow components was developed and calibrated using field experiments conducted on a preferential path nested hillslope in northeast India. Virtual experiments were carried out to perform sensitivity analysis of the model using the automated parameter estimation (PEST) algorithm. Different physical parameters of the model were varied to study the resulting effects on overland flow and subsurface stormflow responses from the theoretical hillslopes. It was observed that topographical shapes had significant effects on overland flow hydrographs. The slope profiles, surface storage, relief, rainfall intensity and infiltration rates primarily controlled the overland flow response of the hillslopes. Prompt subsurface stormflow responses were mainly dominated by lateral preferential flow, as soil matrix flow rates were very slow. Rainfall intensity and soil macropore structures were the most influential parameters on subsurface stormflow. The number of connected soil macropores was a more sensitive parameter than the size of macropores. In hillslopes with highly active vertical and lateral preferential pathways, saturation excess overland flow was not evident. However, saturation excess overland flow was generated if the lateral macropores were disconnected. Under such conditions, rainfall intensity, duration and preferential flow rate governed the process of saturation excess overland flow generation from hillslopes.

Editor D. Koutsoyiannis; Associate editor C. Perrin     

Synergistic effects of water repellency and macropore flow on the hydraulic conductivity of a burned forest soil,south‐east Australia

Research shows that water repellency is a key hydraulic property that results in reduced infiltration rates in burned soils. However, more work is required in order to link the hydrological behaviour of water repellent soils to observed runoff responses at the plot and hillslope scale. This study used 5 M ethanol and water in disc infiltrometers to quantify the role of macropore flow and water repellency on spatial and temporal infiltration patterns in a burned soil at plot (<10 m²) scale in a wet eucalypt forest in south‐east Australia. In the first summer and winter after wildfire, an average of 70% and 60%, respectively, of the plot area was water repellent and did not contribute to infiltration. Macropores (r > 0·5 mm), comprising just 5·5% of the soil volume, contributed to 70% and 95%, respectively, of the field‐saturated and ponded hydraulic conductivity (K_p). Because flow occurred almost entirely via macropores in non‐repellent areas, this meant that less than 2·5% of the soil surface effectively contributed to infiltration. The hydraulic conductivity increased by a factor of up to 2·5 as the hydraulic head increased from 0 to 5 mm. Due to the synergistic effect of macropore flow and water repellency, the coefficient of variation (CV) in K_p was three times higher in the water‐repellent soil (CV = 175%) than under the simulated non‐repellent conditions (CV = 66%). The high spatial variability in K_p would act to reduce the effective infiltration rate during runoff generation at plot scale. Ponding, which tend to increase with increasing scale, activates flow through macropores and would raise the effective infiltration rates at larger scales. Field experiments designed to provide representative measurements of infiltration after fire in these systems must therefore consider both the inherent variability in hydraulic conductivity and the variability in infiltration caused by interactions between surface runoff and hydraulic conductivity. Copyright © 2010 John Wiley & Sons, Ltd.     

Double funnelling in a mature coastal British Columbia forest: spatial patterns of stemflow after infiltration

Many studies have focused on the amount of stemflow in different forests and for different rainfall events, but few studies have focused on how stemflow intensity varies during events or the infiltration of stemflow into the soil. Stemflow may lead to higher water delivery rates at the base of the tree compared with throughfall over the same area and fast and deeper infiltration of this water along roots and other preferential flow pathways. In this study, stemflow amounts and intensities were measured and blue dye experiments were conducted in a mature coniferous forest in coastal British Columbia to examine double funnelling of stemflow. Stemflow accounted for only 1% of precipitation and increased linearly with event total precipitation. Funnelling ratios ranged from less than 1 to almost 20; smaller trees had larger funnelling ratios. Stemflow intensity generally was highest for periods with high‐intensity rainfall later in the event. The maximum stemflow intensities were higher than the maximum precipitation intensities. Dye tracer experiments showed that stemflow infiltrated primarily along roots and was found more frequently at depth than near the soil surface. Lateral flow of stemflow was observed above a dense clay layer for both the throughfall and stemflow experiments. Stemflow appeared to infiltrate deeper (122 cm) than throughfall (85 cm), but this difference was in part a result of site‐specific differences in maximum soil depth. However, the observed high stemflow intensities combined with preferential flow of stemflow may lead to enhanced subsurface stormflow. This suggests that even though stemflow is only a very minor component of the water balance, it may still significantly affect soil moisture, recharge, and runoff generation. Copyright © 2016 John Wiley & Sons, Ltd.     

Solute transport and water content measurements in clay soils using time domain reflectometry

Abstract

Clayey and saline soils have been shown to be problematic for time domain reflectometry (TDR) measurements. This study presents some of these problems and discusses solutions to them. Thirteen solute transport experiments were carried out in three undisturbed soil columns of swelling clay soil from Tunisia, labelled S1, S2, and S3 respectively. The columns were collected at three different physiographical regions within a catchment. Water fluxes ranged from 1.2 to 7.2 cm day^?1. The large solute transport heterogeneity and large tailing indicated that preferential flow was most pronounced in S1. The preferential flow took place in voids between structural elements and in wormholes. In S3, preferential flow was also evident, but not to the same extent as in S1. In S2, the solute transport was more uniform with little preferential flow. The heterogeneity of the solute transport increased with the water flux in S1 and to a smaller extent in S3, whereas it remained constant in S2. In a previous dye experiment in the field, preferential flow in cracks was observed at those sites where S1 and S3 were collected. In the column experiments, preferential flow in these cracks was less due to the higher initial water content compared to the dye experiments, indicating that the desiccation cracks were closed by the swelling clay.