VZJ sign up for etocs
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online 8 March 2006
Published in Vadose Zone J 5:405-411 (2006)
DOI: 10.2136/vzj2005.0031
© 2006 Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (3)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Täumer, K.
Right arrow Articles by Wessolek, G.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Täumer, K.
Right arrow Articles by Wessolek, G.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Täumer, K.
Right arrow Articles by Wessolek, G.
Related Collections
Right arrow Vadose Zone Processes and Chemical Transport
Right arrow Preferential Flow

SPECIAL SECTION: FROM FIELD- TO LANDSCAPE-SCALE VADOSE ZONE PROCESSES

Seasonal Dynamics of Preferential Flow in a Water Repellent Soil

K. Täumer, H. Stoffregen* and G. Wessolek

Department of Site Evaluation and Soil Protection, Berlin Technical University, Salzufer 11-12, 10587 Berlin, Germany
* Corresponding author (heiner.stoffregen{at}tu-berlin.de)

Received 2 March 2005.



   ABSTRACT
TOP
 EXECUTIVE SUMMARY
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
The temporal dynamic of water repellency in soils has a strong influence on water flow and the appearance of preferential flow paths at potentially water repellent sites. To quantify this effect, field investigations were conducted at a sandy site near Berlin, Germany. A large number of soil samples were collected at 32 different times during a 3-yr period. Additionally, a time domain reflectometry (TDR) array with 63 probes measured water contents hourly along a transect of 130 by 60 cm. On the basis of these sampling campaigns the area share of water repellent soil regions was measured. Water content changes were observed with the TDRs at high spatial and temporal resolution after several rainfall events. Heterogeneities in water content changes were analyzed. To quantify the heterogeneity (i.e., the degree of preferential flow) we propose the use of the effective cross section, ECS, for water flow. This parameter was calculated by fitting the ß function to the cumulative values of the water content change over a horizontal cross section at a depth of 25 cm. Sampling and TDR measurements showed similar seasonal dynamics of preferential flow, with the highest occurrence in summer and early autumn and a maximum accessible soil volume in the spring. Preferential flow in the winter month was enhanced not only by water repellency, but also by freezing and melting. We also tried to calculate the ECS from climatic data. It was possible to calculate the ECS using a linear relationship with the initial soil moisture at the 10-cm depth, the duration and amount of precipitation, and the potential evapotranspiration rate.

Abbreviations: ECS, effective cross section • TDR, time domain reflectometry • WDPT, water drop penetration time


   INTRODUCTION
TOP
 EXECUTIVE SUMMARY
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
WATER REPELLENCY is a widespread phenomenon that can significantly affect water flow in soils. Water repellency occurs on most continents and in a variety of land uses and climatic conditions. Repellency can occur naturally, as reported by Dekker et al. (1999), Harper and Gilkes (1994), and York and Canaway (2000), or can be affected by fire (Doerr et al., 1996). Repellency may significantly impact surface runoff and can lead to reduced wetting rates of dry soils. Water repellent soils tend to restrict water flow, creating fingerlike wetting fronts, which leads to a reduction in plant-available water and limits solute transport to relatively narrowed flow channels. Pollutants and nutrients consequently have shorter residence times in the affected soil layers and move much faster through the vadose zone. The effects can be observed as irregular moisture patterns or tracer distribution in the soil (Ritsema and Dekker, 1998; Arbel et al., 2005), patchy growth of plants, or erosion in sloping regions (Witter et al., 1991). Water repellency is not a permanent property. Soils become water repellent when they desiccate during dry periods in the summer, with water repellency often diminishing or vanishing when the soil becomes wet in the autumn or during winter (Dekker and Ritsema, 1994).

The process of water repellency has been analyzed since the 1960s. Its spatial variability and the creation of preferential flow paths have recently received much attention. For example, Wang et al. (2000) conducted infiltration experiments on water repellent and wettable soils in the laboratory. They used the saturated hydraulic conductivity and an air entry value of the matric potential as conditions for unstable infiltration fronts. Clothier et al. (2000) studied infiltration experiments under unsaturated conditions in the laboratory and in the field. They observed an increase in the infiltration rate after 100 min, after which the water repellency broke down. Selker et al. (1992) derived the size of flow fingers from the unsaturated hydraulic conductivity, the air entry value, and the flow rate. Wang et al. (1998) reported finger diameters between 2 and 23 cm. An approach for calculating the finger size is included in the one-dimensional model SWAP (van Dam et al., 1996). Most experiments involved infiltration into dry soils. Glass et al. (1989) and Liu et al. (1993) found that the distributions of water contents and flow paths were strongly influenced by earlier infiltrations with fingered flow. Previous finger paths were preserved because of hysteresis in the soil moisture retention.

Dekker and Ritsema (1994) established the concept of a transition zone, or a critical soil moisture zone, to distinguish between water repellent and wettable conditions. The soil becomes water repellent when the water content decreases below a certain critical value, and it becomes wettable again when its water content exceeds another critical water content. Water contents for these two thresholds have been found to vary widely for different soils (Doerr and Thomas, 2003; Dekker et al., 2001; Ziogas et al., 2003). Täumer et al. (2005) described the influence of soil organic matter content on the critical water content.

Doerr et al. (1996) determined the spatial variability of water repellency in forestry soils in Portugal. Ritsema and Dekker (1998) analyzed the three-dimensional distributions of water repellency, bromide, and pH values, while Dekker and Ritsema (2000) described the water repellency of a clay soil. Using several sampling campaigns, they showed the highly spatial variability and temporal nature of water repellency.

Most field investigations of water repellency have been conducted on disturbed or undisturbed soil samples. The disadvantage of this method is that it destroys the soil. Only few investigations have used nondestructive methods. Ritsema et al. (1998) used an array of TDR probes to demonstrate the occurrence of fingering after three rainfall events; however, they did not determine seasonal changes in preferential flow.

Without information about the seasonal dynamics of preferential flow, water and solute transport in a water repellent soil cannot be fully described. Therefore, the objective of this study was to quantify the seasonal dynamics of preferential flow under water repellent field conditions. The concept of the effective cross section is introduced as a new parameter to describe the area share of the preferential flow paths and the water repellent soil volume. Quantifying seasonal variations in this parameter allows one to include the effects of water repellency in the seasonal simulations of water and solute movement.


   MATERIALS AND METHODS
TOP
 EXECUTIVE SUMMARY
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Study Site
The study site is located at the northern city limit of Berlin, Germany (49°29'31'' N, 11°02'59'' E). Untreated sewage wastewater was applied to 13000 ha of farm land around Berlin beginning in 1890. In 1985, the wastewater application ended, the soil surface was leveled, and an effort was made to reforest the fields. Most of the trees died, mainly due to water shortage during summer, nutrient deficiency, and heavy metal contamination (Schlenther et al., 1996). The average air temperature of the site is 7.5°C, and the annual precipitation is 580 mm (340 mm from April to September, 240 mm from October to March).

The area presently is dominated by dry grassland, mainly couch grass [Elytrigia repens (L.) Desv. ex Nevski]. The soil, a hortic Anthrosol, consists of 40 to 60 cm of organic topsoil over medium-sized sand. In most areas very dense rooting can be found in the upper 10 to 15 cm of the topsoil. The organic matter content of the topsoil horizon ranges mostly between 40 and 60 g kg–1, but with several spots showing up to 300 g kg–1. Below a depth of 40 cm the organic matter content is approximately 10 g kg–1, while further decreasing to <5 g kg–1 below 60 cm. Except for its organic matter content the sand is homogenous up to a depth of 5 m. No visible layering or small-scale texture changes could be observed. Therefore, particle size distributions were obtained only at selected locations. The clay content of the noncalcareous fluvial sand is <10 g kg–1. Due to the reforestation effort, the surface is slightly undulated (wavelength 3.3 m; amplitude 10–15 cm). The bulk density up to a depth of 20 cm is about 0.9 to 1 g cm–3 in the upper parts of the undulated surface (Fig. 1 , distance 40–120 cm). In the lower parts (Fig. 1, distance 120–180 cm) the bulk density of the topsoil is 1.2 g cm–3. Rapid turnover processes of organic matter occur, and increased leaching of heavy metals and nitrate in the area were observed by Hoffmann (2002).


Figure 1
View larger version (24K):
[in this window]
[in a new window]
  		Fig. 1. Cross section of the soil profile and positions of the TDR probes.

 
Soil Sampling
From April 2001 to April 2004 about 3000 soil samples were taken at 32 different times, in most cases as disturbed samples. The sampling positions were distributed across a plot of 25 by 150 m. The TDR transect was also part of the plot. Soil samples were taken eight times with high spatial resolution along a transect or in a grid (10 by 10 or 5 by 5 cm). The samples allowed excellent description of the spatial variability of several soil properties. Special attention was given to the area share, the degree of water repellency, and the water content at a depth of about 20 to 30 cm. The actual water repellency was measured on field moist samples using the water drop penetration time (WDPT) test as described by Krammes and DeBano (1965). At the other 24 sampling times, the area share of the water repellent profile was estimated from the visible pattern of lighter and darker color, associated with dry and wet parts of the soil (Fig. 2 ). The area share was estimated directly from observations of the profile or from photographs of the trench. The sampling data, especially the water content and WDPT, were used to quantify the spatial variability. The spatial analyses of the water content from the first sampling campaigns helped to determine the optimal setup for the TDR array.


Figure 2
View larger version (154K):
[in this window]
[in a new window]
  		Fig. 2. Patterns of the soil moisture showing the relatively dry water repellent areas (light color) and moist, wettable areas (darker color).

 
TDR Measurements
To monitor short term moisture changes in the topsoil, a TDR transect using D-LOG/mats (EASY TEST Ltd., Lublin, Poland) device was installed permanently. The device operates using needle pulse type generated electromagnetic waves with 300 ps rise time and electromagnetic radiation with a frequency range from 30 MHz to 1.6 GHz (Malicki and Skierucha, 1989). The D-LOG is designed for periodic recording of instantaneous soil moisture profiles at selected time intervals. To switch between the probes the D-LOG unit contains a built-in first-level microwave switch and eight second-level switches. Each second-level switch is connected to eight TDR probes.

The TDR probes consist of a thin-wall PVC body that is 2 cm in diameter and 15 cm long. The probe has two parallel 10-cm-long waveguides. The waveguides are stainless-steel rods (diameter 2 mm) that are 16 mm apart. The region of influence of the sensor is approximately a cylinder with a length of 12 cm and a diameter of 5 cm. Outside of this region the electromagnetic wave of the TDR probe has little or no influence on the velocity.

To obtain the correct probe positions, the soil surface and horizon boundaries were copied onto plastic foil, which was fixed onto the profile wall. Since the surface undulated, a reference depth of zero was made to the smoothed soil surface. After removing the foil with the topography of the profile, the 63 TDR probes were installed horizontally into the profile wall through pilot bore holes of 15 cm. The TDR transect covered the complete topsoil for a representative area, from the top of the elevation to the middle of the lower profile range. In view of results of the spatial analyses, spacing of the probes was 10 cm in the humus-rich topsoil. Smaller spacings could result in interactions between the single TDR probes and lead to a more disturbance of the surrounding soil. Larger spacing, on the other hand, would not reflect the heterogeneities in water content changes. After the installation the trench was refilled carefully with the original soil

Each probe represented a 10-cm compartment, except for the top probes (at the 10-cm depth). A depth range between 0 and 15 cm was established for these probes. Thirteen probes were installed at each depth to analyze the heterogeneities. Figure 1 shows the positions of the TDR probes in the profile. Measurement of the entire profile took about 2.5 min. The measurements were usually taken at regular intervals of 1 h. Water contents were calculated by using the calibration function of Roth et al. (1992).

Specifying Preferential Flow from TDR Measurements
For water repellent soils, the area share of the flow paths is an important value reflecting the preferential flow paths for water and solutes. The flow paths can be identified by determining profile positions where the TDR readings show a high change in water content during single rainfall events. An index for the degree of preferential flow is necessary to compare different flow events. The index should be unambiguous for comparing and contrasting different distributions of changes in the soil water. We used an approach based on the Beta function. The standard Beta function (p) was defined by Bronstein and Semendajajew (1987) as

Formula 1[1]
where {Gamma} is the Gamma function (or Euler's integral of the second kind) and {alpha} and {zeta} are free parameters. The Beta function is defined over the interval [0,1].

Calculation of the Effective Cross Section
We point out that the ECS concept is only useful for finger flow in the soil matrix, not for rapid flow through soil macropores. In the latter case, the TDR probes would not be able to measure the water content changes. At our site we found that the flow paths had typical dimensions in the range of decimeters. The TDR readings were examined for single rainfall events. A typical rainfall event and the corresponding response in water content changes for the whole soil profile are shown in Fig. 3 . The water content change in the soil was delayed after the rainfall. A small difference existed between the amount of rainfall measured at this site and the amount of water recovery in the profile. Differences could be caused by interception and by higher water contents in the top 10 cm, above the first TDR probes.


Figure 3
View larger version (13K):
[in this window]
[in a new window]
  		Fig. 3. Measured cumulative rainfall and water content changes for the entire profile as derived from TDR readings. The time of the maximum water content tmaxWC is 16 h.

 
Only single rainfall events were analyzed to keep the share of stationary flow (flow without changes in water content) as low as possible. A single rainfall event is defined here as a rainfall event with a dry period of at least 3 d before the event.

The time (tmaxWC) of the maximum water content in the profile was determined for each rainfall event. For the example displayed in Fig. 3, tmaxWC was established as 16 h. The gain in water content ({theta}) from the beginning of the rainfall (t0) until tmaxWC was calculated for each TDR probe at position x and depth z. The ratio fx,z between the water content change at position x and the total change in water content in that layer was calculated for each probe using

Formula 2[2]

Afterward the values of fx,z for that layer were ranked in descending order. Each probe represented a section equivalent to 1/13th of the transect. Figure 4 shows a plot of the cumulative ratio fx,z against the cumulative cross-sectional area. The Beta function was fitted to the data by optimizing the parameter {alpha} and {zeta}. The function expresses the share of the water content changes as a function of the area share. We define now the ECS as that fraction of the total area that realizes 90% of water content change as measured by the TDR probes at a certain depth. We used this value to quantify the heterogeneity in water flow in the soil.


Figure 4
View larger version (21K):
[in this window]
[in a new window]
  		Fig. 4. Cumulative water content changes for (a) homogeneous flow and flow affected by relatively (b) low and (c) high degrees of heterogeneity. The effective cross section, ECS is 0.9 for Case a, 0.7 for Case b, and 0.2 for Case c.

 
Case a in Fig. 4 shows the hypothetical situation of having a uniform distribution with every compartment contributing the same amount to the water content change (piston flow). Ninety percent of the water content change takes place on 90% of the cross-sectional area. Cases b and c show nonuniform distributions on the basis of measured values. The graphs in those cases are curved. Every measurement represents an area of 10 by 10 cm. The first compartments in Fig. 4 have a higher share of the water content change than later ones, which leads to a deflection of the curve. For case b 90% of the water content change occurred in approximately 70% of the cross-sectional area. The shape of the curve reflects the degree of heterogeneity. The more the data points and the fitted ß distribution depart from the 1:1 line, the higher is the extent of preferential flow. A very high degree of preferential flow can be seen for Case c, where the two compartments with the highest change realize 85% of the total water content change in that layer. Only 20% of the cross-sectional area is responsible for 90% of water content change. The concept of ECS can be used to quantify the area share that actually takes part in the water flow and appears most useful for making comparisons between different rainfall events.


   RESULTS AND DISCUSSION
TOP
 EXECUTIVE SUMMARY
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
Soil Sampling—Spatial Distribution of the Soil Water Content
Figure 5 shows the distribution of the water content and actual water repellency (WDPT) of a transect sampled 22 Jan. 2002, during a period of mild weather following a strong frost. The soil layers at this position were found to be similar to the layers of the TDR transect (Fig. 1). The soil moisture distribution in the topsoil was clearly very heterogeneous. The higher water contents coincided with the wettable areas, while the dryer ranges were related to the water repellent areas. No differences in bulk density or particle size distribution were found for these adjacent wet and dry regions. We also excluded nonuniform water uptake by plant roots as a cause of the observed differences in soil moisture. This is because sampling was performed in the middle of winter, outside the vegetation period (April–October); also, no differences in root penetration were found between wet and dry areas. We did find a strong relation between the actual water repellency (which triggers the preferential flow process on the site) and the water content and the organic matter content (Täumer et al., 2005).


Figure 5
View larger version (37K):
[in this window]
[in a new window]
  		Fig. 5. Spatial distribution of (a) the gravimetric water content and (b) the water drop penetration time (WDPT) for a vertical profile. Samples were taken every 6 cm in the horizontal direction (51 samples per layer) and with a vertical resolution of 10 cm.

 
We analyzed the data geostatistically to quantify the range of spatial correlation (Webster and Oliver, 2001). Experimental semivariances were calculated in a horizontal direction, separately for the topsoil and subsoil samples. Figure 6 shows the semivariance in water content for samples taken at depths of 30 and 90 cm. The variance for the topsoil shows a steep increase in the first 20 cm. A local minimum is visible every 40 cm, indicating a recurring property at regular spacings, that is, the average distance between the flow channels or the dry spots. This spacing was not found for the subsoil, where no preferential flow paths were found. The water contents in the topsoil range between 0.039 and 0.242 g g–1, and in the subsoil between 0.026 and 0.116 g g–1. Therefore, the value of the variance for the subsoil was three to four times smaller than for the topsoil. A spherical model (blue solid lines in Fig. 6) was applied to the data:

Formula 3[3]
where A is the sill (%), h the distance (cm), and b the range (cm). The range b was 26 cm at the 30-cm depth. For the subsoil (90-cm depth) the range was significantly larger (76 cm). The analysis of other soil samplings at this site showed similar results.


Figure 6
View larger version (14K):
[in this window]
[in a new window]
  		Fig. 6. Semivariogram of the measured gravimetric water content for samples taken at the 30- and 90-cm depths. The solid lines are the fits of a spherical model.

 
Results of the spatial analysis of the TDR transect data suggested the need for a different experimental setup for the topsoil and the subsoil. To observe the rapid spatial changes in water content, the spacing between the TDR probes should be smaller than the range. With the selected grid of 10 cm, neighboring TDR probes were situated at distances of one half- to one-third of the range.

TDR Data
The moisture contents were measured from April 2002 until April 2004, excluding some periods where the measurements failed due to problems with the TDR device or power supply. Figure 7 shows data for two typical time frames from five neighboring TDR probes, each at the 10-cm depth. The measurements show the drying of the topsoil in spring 2002 (Fig. 7a) and the rewetting after summer (Fig. 7b). Different responses to rainfall events are apparent during the two periods. While some of the TDR probes indicated large changes in water content, others did not react at all. Rainfall events caused an abrupt increase of the water content at some positions of the profile. After these abrupt increases, the water content decreased quickly, within the first hours after the rainfall event, due to rapid drainage into deeper soil layers.


Figure 7
View larger version (32K):
[in this window]
[in a new window]
  		Fig. 7. Water content measurements during spring (top) and autumn (bottom) 2002 for five neighboring TDR probes at the 10-cm depth. The probes in the flow paths show very abrupt reactions, while probes in the water repellent regions show little or no reactions.

 
In the spring we observed a decrease in water content from the end of April to the beginning of June. On average, the water content of the top soil layer decreased from 0.18 to 0.06 m3 m–3 during that time. During rewetting in the fall, especially at the end of September, large differences occurred in the TDR readings among the different probes. At the end of October more probes reacted than at the beginning of September. Water content at several positions remained constant after the rainfall events, thereby indicating the water repellent areas of the soil.

We also observed some diurnal variations in water content, especially during the dry periods in May. Such diurnal changes in water content were caused by evapotranspiration during the day and water redistribution during the night. Temperature effects may also have been present (Stoffregen, 1998; Parlange et al., 1998).

Figure 8 shows moisture distribution along the transect at the start of rainfall event on 22 Sept. 2002 and 9, 24, and 44 h afterward. The evolution of flow paths and the influence of the undulated soil surface are clearly noticeable on the plot. The upper right side of the profile stayed much drier than the lower left side. Several water repellent areas without much or any changes in the water content were present. Probes that showed strong reactions to rainfall events (e.g., 140 cm, 10-cm depth) at the beginning of rewetting also showed strong changes with subsequent events. Apparently once a flow path is established, this path will be followed during following rainfall events. Regions between the flow paths showed no or only slight changes in water content (e.g., 130 cm, 10-cm depth). Some of these dry regions were very persistent. It took several rainfall events to overcome water repellency and to have most of the profile become wet again. In time the number and size of the flow regions also increased.


Figure 8
View larger version (18K):
[in this window]
[in a new window]
  		Fig. 8. Plots showing changing moisture distribution in the topsoil (a) initially and (b) 9 h, (c) 24 h, and (d) 44 h after a rainfall event on 22 Sept. 2002.

 
Heterogeneity
The wettable areas estimated from the sampling data and the ECS from the TDR data clearly showed a seasonal trend. They reached minimum values in summer and maximums at the beginning of spring (Fig. 9 ). Note that the sampling data (yellow and orange squares) were collected under varying conditions, mostly a few days after a rainfall event. The ECS values were always calculated from the TDR measurements (blue triangles) directly after rainfall events. Hence, the TDR data never showed a totally water repellent soil, while for the sampling data these conditions did sometimes occur during summer.


Figure 9
View larger version (16K):
[in this window]
[in a new window]
  		Fig. 9. Seasonal changes in the wettable area as estimated by soil sampling (squares) and in the effective cross section as derived from the TDR data (triangles). The gray dashed lines indicate the range of values during the year, excluding frost periods.

 
From February to the beginning of April nearly all of the profile was wettable; the area share of the water repellent spots was mostly smaller than 10%. At the beginning of the growing season in April and May, the water content in the topsoil decreased. The evapotranspiration rate at that time was higher than the precipitation rate. During prolonged dry periods the entire topsoil dried out and became water repellent. Rainfall events caused incomplete wetting, with the ECS being reduced to 20 to 40% of the total cross-sectional area.

With decreasing water demand by the plants and lower temperatures in the fall the soil water balance slowly became positive. This caused a rewetting of the soil and hence an increase in the ECS. Especially at the beginning of May, the TDR data and sampling results showed an extremely high variability. The variability was caused by the variation in the climatic conditions. The year 2002 (820 mm annual precipitation) was unusually wet, while 2003 (410 mm) was very dry, especially during the first part from February to June.

Values of the TDR measurements in January were influenced considerably by the occurrence of frost. Sealing of the soil surface due to frozen parts caused runoff and preferential flow when the snow melted. Frozen water has a much lower dielectric constant than liquid water. Therefore, the time when the soil thaws can be detected by the TDR measurements. Differences in the melting time between neighboring profile positions at the same depth were observed for several hours up to 2 d. In the soil sampling, on the other hand, this effect cannot be detected by measuring the water repellent area share.

Predictions of Preferential Flow
We next examined the dependence of the ECS area on climatic parameters and water content. Data from periods with frost in the soil profile were excluded from the analysis. An important factor for water repellent soils is the initial moisture in the topsoil. A drier soil generally has a higher probability of becoming water repellent and, therefore, poses a higher risk of preferential flow. Several climatic conditions seem to favor the creation of preferential flow. For example, short intensive rain tends to trigger preferential flow, while long, steady rains tend to moisten the soil more uniformly. Figure 10 displays the relation between the ECS and the average initial water content at a depth of 10 cm. A general dependence can be seen, although the regression coefficient r2 is only 0.35. Notice that the results differ as a function of the average intensity of the rainfall event. For lower (<0.35 mm h–1) and medium intensities we found a higher correlation, while no correlation was found for higher intensities (>0.9 mm h–1).


Figure 10
View larger version (14K):
[in this window]
[in a new window]
  		Fig. 10. Effective cross section and initial soil moisture at the 10-cm depth for relatively low (squares), medium (circles), and high (triangle) average rainfall intensities.

 
The ECS calculated from the TDR data was found to be a function of several parameters. Using results of 20 rainfall events during 1 yr, we examined several empirical correlations to predict the ECS. The factors we used in the prediction were the initial water content at the 10-cm depth; the amount, intensity, and duration of rainfall; the potential evaporation rate; and the climatic water balance (precipitation minus potential evapotranspiration) for different periods before the rainfall events. Different functional dependences (e.g., linear, quadratic, power function, hyperbolic) were tested. The sum of square roots between calculated and measured ECS values were for this purpose optimized by using a numerical solver.

The initial soil moisture at a depth of 10 cm at the beginning of the rainfall event was found to have the greatest influence on the occurrence of preferential flow. Good results were achieved using the linear relationship:

Formula 4[4]
where ECS is the effective cross section (m2 m–2), {theta}10 is the average initial water content at 10 cm (m3 m–3), I is the average intensity of a particular rainfall event (mm h–1), Ep is the potential evaporation (mm) for a 24-d period, and R is the rainfall amount (mm). Figure 11 shows a comparison between the ECS values calculated from TDR measurements and the predicted values from Eq. [4]. The use of quadratic equations gave only slightly better results, while hyperbolic or power functions produced worse results.


Figure 11
View larger version (14K):
[in this window]
[in a new window]
  		Fig. 11. Values of the effective cross section, ECS, calculated from TDR measurements, and predicted values with Eq. [4].

 

   CONCLUSIONS
TOP
 EXECUTIVE SUMMARY
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
Our field data showed the considerable seasonal changes in the extent of water repellent areas and the area share of preferential flow paths. The ECS for the water flow was calculated from the TDR transect with high spatial and temporal resolution. ECS provides a useful quantitative measure of the extent of preferential flow during the seasons. This parameter can also be used for comparing numerical models and measurements for heterogeneous water flow. It was possible to calculate ECS using the average water content and climatic parameter. The average water content was found to have the strongest influence on the ECS. While the concept of ECS itself could be transferred to other sites, the parameters of Eq. [4] are limited to the examined site. The equation may be used in numerical models to predict the ECS for different climatic conditions. Our results are based on data from 1 yr. Measurements during subsequent years may be needed. Use of the ECS as such, and in numerical models, may lead to a better estimation of the amount of plant-available water and contaminant transport at particular sites than is currently possible.


   ACKNOWLEDGMENTS
 
We thank the German Research Foundation (DFG) for the financial support of the research group INTERURBAN (www.interurban.de), in which the work was done.


   REFERENCES
TOP
 EXECUTIVE SUMMARY
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 




This article has been cited by other articles:


Home page
 Vadose Zone JHome page
W.-L. Liang, K. Kosugi, and T. Mizuyama
Heterogeneous Soil Water Dynamics around a Tree Growing on a Steep Hillslope
Vadose Zone J., November 20, 2007; 6(4): 879 - 889.
[Abstract] [Full Text] [PDF]

Home page
 Vadose Zone JHome page
D. L. Corwin, J. Hopmans, and G. H. de Rooij
From Field- to Landscape-Scale Vadose Zone Processes: Scale Issues, Modeling, and Monitoring
Vadose Zone J., March 8, 2006; 5(1): 129 - 139.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (3)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Täumer, K.
Right arrow Articles by Wessolek, G.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Täumer, K.
Right arrow Articles by Wessolek, G.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Täumer, K.
Right arrow Articles by Wessolek, G.
Related Collections
Right arrow Vadose Zone Processes and Chemical Transport
Right arrow Preferential Flow

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Agronomy Journal Crop Science
Journal of Natural Resources
and Life Sciences Education		 Soil Science Society of America Journal
Journal of Plant Registrations Journal of
Environmental Quality		 The Plant Genome