Published online 24 January 2007
Published in Vadose Zone J 6:105-115 (2007)
DOI: 10.2136/vzj2006.0035
© 2007 Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
ORIGINAL RESEARCH
Preferential Flow Patterns in Paddy Fields Using a Dye Tracer
Till Sander and
Horst H. Gerke*
Institute of Soil Landscape Research, Leibniz-Centre for Agricultural Landscape Research (ZALF), Eberswalder Strasse 84, D-15374 Müncheberg, Germany
* Corresponding author (hgerke{at}zalf.de)
Received 2 March 2006.
 |
ABSTRACT
|
|---|
Reports of chemical leaching from intensively cropped paddy rice (Oryza sativa L.) field sites indicate that soil structure dynamics may possibly affect preferential flow and transport despite repeated "puddling" (i.e., mechanical homogenization of water-saturated soil) and plow pan formation. Our objective was to identify preferential pathways of soils in regularly managed paddy fields, which were drained during temporal fallow. Dye tracer studies were conducted on two experimental 20-yr-old rice fields of about 300 m2 located in the Sunjian watershed in southeast China (Jiangxi province). Brilliant Blue solution was applied by ponding a 50-mm pulse onto four 1-m2 plots with intact and unleveled soil surfaces. Staining patterns were recorded for 11 to 13 vertical and 10 to 15 horizontal soil profiles. Dye tracer penetrated vertically via preferential pathways to depths ranging from 94 (vertical profiles) to 120 cm (horizontal profiles), while most of the soil matrix remained unstained. Stained biopores and cracks were found in the plow pan. Horizontal profiles with stained pathways indicate large spatial heterogeneity of hydraulic conductivity. Dye coverage vs. depth calculated from vertical and horizontal profiles suggests that horizontal spreading of the solution above the plow pan supports access to plow pan macropores. Penetration through the plow pan proceeded with little horizontal spreading. The results obtained here for initially drained paddy soils indicate that preferential flow is a phenomenon that may be relevant for water and nutrient cycling (and losses to groundwater) in paddy fields. If the observed macropores and crack networks persist, preferential flow might even occur under flooding conditions.
Abbreviations: BB, Brilliant Blue dye tracer
|
INTRODUCTION
|
|---|
THE INTENSIFICATION of rice production has largely depended on new varieties and the application of fertilizers and pesticides. Mean N fertilizer application rates were 120 to 140 kg N ha–1 per crop for rice in 1999 (Tong, 2003), which can amount to annual application rates of 500 to 650 kg N ha–1 in intensively cultivated regions (Zhu et al., 2000). The widespread use of agrochemicals carries the risk of unintended losses and contamination of the environment. Awareness of problems associated with nonpoint-source pollution of groundwater and surface water bodies in rice production has recently increased in the People's Republic of China (Zhu et al., 2000).
Topsoil structure of paddy fields is widely assumed to be relatively homogeneous, since the cultivated horizon is repeatedly being homogenized by puddling, while shrinkage cracks are assumed to close due to swelling after flooding the paddy fields. Puddling softens the soil for transplanting of rice seedlings (Kirchhof et al., 2000; Sharma and De Datta, 1986) and creates a plow layer that reduces hydraulic permeability to support ponding of water (Gong, 1986), with the subsoil remaining unsaturated below the plow pan (Wopereis et al., 1994b).
Previous studies of water movement in paddy rice fields were directed toward analyzing the impact of paddy soil management strategies on either water-use efficiency (e.g., Tuong et al., 1994; Cabangon and Tuong, 2000; Kirchhof et al., 2000) or groundwater recharge (Chen and Liu, 2002). Most simulation and infiltration studies of paddy fields assumed homogeneous hydraulic properties (e.g., Tuong et al., 1994; Chen et al., 2002; Liu et al., 2003). In these studies, the plow pan was assumed to be the major factor controlling water flow because of the relatively low hydraulic permeability. Since the hydraulic resistance of the cultivated horizon was found to be comparably small (e.g., Wopereis et al., 1994b), some modeling results assumed flow resistance in the topsoil to be negligible in puddled fields (e.g., Chen et al., 2002). If cracks and biopores perforate the plow pan, however, vertical flow will largely be controlled by cracks and macropores, and preferential or bypass flow (e.g., Germann and Beven, 1985) may be possible.
Preferential flow due to crack formation in paddy soils of wetland rice fields was previously observed in only a few studies (e.g., Wopereis et al., 1994a; Liu et al., 2003). Wopereis et al. (1994a) studied bypass flow in drained, cracked paddy soils using a morphological staining technique. After applying white dye and drawing dominant pathways of three horizontal profiles at 0.05-, 0.25-, and 0.5-m depths, Wopereis et al. (1994a) observed horizontal cracking at the interface between puddled and unpuddled layers and detected vertical cracks down to 0.65 m. The drying process of paddy fields has been studied and simulated successfully using hydraulic and shrinkage parameters (e.g., Ringrose-Voase and Sanidad, 1996; Kirby and Ringrose-Voase, 2000). Although crack formation in general has been the focus of many numerical studies (e.g., Chertkov and Ravina, 1998; Colina and Roux, 2000; Yoshida and Adachi, 2004), quantitative studies of the macropore system of paddy soils in rice fields are limited and mostly do not provide the qualitative and quantitative knowledge required to fully understand flow processes in paddy soils.
Dye tracers are frequently used to study soil structure and preferential flow phenomena. Nevertheless, a number of experimental problems have been reported that may be relevant to the study of paddy soils; for instance, alterations of boundary conditions by leveling (Kulli et al., 2003; Yasuda et al., 2001) or homogenizing the soil surface (Flury et al., 1994; Ghodrati and Jury, 1990) may lead to changes in flow patterns (Schwartz et al., 1999). In preferential flow studies, relatively high cumulative infiltration depths (Nobles et al., 2004, about 360 mm; Ghodrati and Jury, 1990, 100 mm) or some smaller depths (Forrer et al, 2000, 50–200 mm; Flury et al., 1994, 40 mm; Schwartz et al. 1999, 40 mm) were used to facilitate macropore flow.
The aim of this study was to describe preferential pathways for water and chemical fluxes in paddy rice fields on the pedon scale using a dye tracer. Photos were taken from vertical and horizontal soil profiles to combine the advantages of analyzing stained pathways in both spatial directions (Ghodrati and Jury, 1990). Image analysis of photos was used since this technique resulted in higher spatial resolution and better objectivity than manual drawings (e.g., Ewing and Horton, 1999; Hangen et al., 2004).
|
MATERIALS AND METHODS
|
|---|
The study was carried out in the typical subtropics of the Peoples' Republic of China in a small watershed named Sunjian Catchment. The experimental site is located approximately 5 km northwest of the Ecological Experimental Research Station of Red Soil in Yingtan, Jiangxi Province, which belongs to the Chinese Academy of Sciences in Nanjing (28.5 N/116.91 E). Soils developed from quaternary clay. Soils of the two selected experimental paddy fields of this study were both classified as Stagnic Cambisols according to the FAO classification scheme (FAO, 1998). The texture of the topsoil down to 42 cm is clay loam (Table 1). The compacted plow pan starts at about 14-cm depth. The location of the plow pan was determined from visual inspection of color changes and horizontal cracks, separating the plow pan from the puddled topsoil. Bulk density gradually decreased from the plow pan downward to about 30-cm soil depth. This apparent decrease in soil bulk density was later verified from bulk density measurements obtained from x-ray computed tomography scanning of intact soil columns (data not shown). Relatively high soil bulk densities were measured in the plow pan from sampling of three 300-cm3 (6-cm height and 8-cm diameter) soil cores: 1.46 g cm–3 with a standard deviation of 0.017 g cm–3 for the 15- to 21-cm soil depth in Field 1 and 1.47 g cm–3 with a standard deviation of 0.008 g cm–3 for the 16- to 22-cm soil depth in Field 2. Data on soil bulk density for the top- and subsoil horizons were not available at the time of the study.
The compacted plow pan and an adjacent cemented Bw2 horizon (Table 2) showed prismatic and blocky structures, which included strongly embedded granular aggregates and crumbs (Ball and Douglas, 2003). The very top of the plow pan exhibited a platy horizontal structure, whereas the subsoil was softer, with a granular structure. An additional strongly structured deep plinthite horizon was found in Paddy Field 2 below the 80-cm soil depth.
For a period of at least 20 yr, the annual cropping system was early rice–early rice–fallow, whereas Chinese pear (Pyrus spp.) trees were grown before that time. On Field 2, peanut (Arachis hypogaea L.) was the first crop in 2004; rice was cultivated during the second part of the season (Table 3). The soil was manually puddled twice a year by a rotary implement attached to a water buffalo (Bubalus bubalus bubalis). The groundwater level was at the 110-cm depth on 5 Nov. 2004 in Field 1. The 120-cm-deep profiles in Field 2, however, did not show a groundwater table during the tracer experiments; later, on 4 Dec. 2005 after 3 d with 41.3 mm of precipitation, a water table was measured at the 83-cm depth at Field 2 (Zhang Bin, Red Soil Station, personal communication, 2004).
View this table:
[in this window]
[in a new window]
|
Table 3. Cultivation cycle during 1 yr (2003–2004) in relation to experimental procedures for Paddy Fields 1 and 2.
|
|
The paddy fields were usually kept flooded during the growing season, but the lack of water occasionally forced the farmers to interrupt ponding between the middle of June and middle of August (Table 3). During periodic fallow conditions, flooding of the bare and cracked soil occurred frequently, resulting either from natural rainfall or flood irrigation of surrounding paddy fields. Additionally, the farmer irrigated to allow germination of legume seeds after the growing season; this crop, however, rarely covered the soil surface in November. Experimental Fields 1 and 2 were drained on 28 Oct. and 17 Nov. 2004, respectively. Four dye tracer experiments were conducted during the winter fallow period. At Field 1, dye tracer application was on 3 (Plot 1a) and 5 Nov. (Plot 1b) 2004. At Field 2, the experiments were on 17 (Plot 2a) and 18 Nov. (Plot 2b) 2004.
The saturation condition of the soil during the tracer application was characterized using field tensiometric data. Three rows of tensiometers (Tensimeter, UGT Ltd., Müncheberg, Germany) were vertically installed in Field 1 through the 40-cm soil depth on 29 Oct. 2004, and pressure heads were recorded daily during the drying process. The pressure head distributions (Fig. 1
) in Field 1 near the plots of the dye tracer experiments show relatively uniform drying conditions from the topsoil during the monitoring period. The pressure head data (Fig. 1) indicate that soil moisture in and below the plow pan was relatively constant during the experiments. While there was no significant rain the week before the experiments in Field 1, there was a 67.5-mm rainfall between 5 and 16 November (Zhang Bin, Red Soil Station, personal communication, 2004) prior to the start of the Field 2 experiment. Drying conditions prevailed, however, between 16 (0.5-mm rain) and 18 November. Since the soil dried rapidly, differences in soil moisture conditions in and below the plow pan between Fields 1 and 2 were not significant.
View larger version (22K):
[in this window]
[in a new window]
|
Fig. 1. Soil water pressure head as a function of time (mean values and standard deviation of three replicates) measured using insertion tensiometers at six depths. Black arrows mark the dye tracer application times. Precipitation (P) is in millimeters per day.
|
|
The dye tracer Brilliant Blue (BB) (Vitasyn Blue, AE85, Hoechst AG, Frankfurt am Main, Germany) was applied in solution (5 kg m–3 water) as a 50-mm ponding pulse across an area of 1 m2. After cutting a slit in the soil with a knife down to the plow pan, a steel frame of 1- by 1- by 0.4-m depth was pushed through the slit, using a 5-kg hammer, until the blade reached the plow pan. The frame was horizontally leveled. The soil outside the steel frame was manually compacted using the hammer at Plots 1b, 2a, and 2b to avoid possible lateral leakage of applied tracer solution, which occurred at Plot 1a (data are excluded).
The soil surface inside the frame was kept intact. Surface (micro)topography was measured on a 10-cm grid using a ruler in reference to a scaled wooden bar on top of the leveled frame. In another plot in Field 2, the topography of the plow pan was measured by the same technique after removal of the cultivated topsoil. Crack network and rice stubbles at the soil surface were photographed. The soil surface inside the frame was covered with a plastic sheet before 50 L of dye tracer solution was applied. After removing the plastic sheet, the entire 1-m2 plot was flooded within about 2 or 3 s. Another plastic cover was subsequently placed over the plots to prevent evaporation.
One day later, the plastic cover and steel frame were removed and an access soil pit was excavated along one side for profile analysis. Between 11 and 13 vertical profiles of 1-m depth were excavated at horizontal intervals of 5 cm or less until the center of the 1-m2 area was reached (Fig. 2
). For the second half of the pit, 10 to 15 horizontal profiles were created at vertical intervals of up to 5 cm for the topsoil. Excavation of horizontal profiles continued below the 20-cm depth, in intervals of about 10 cm downward to a depth of 1 m.
View larger version (54K):
[in this window]
[in a new window]
|
Fig. 2. Schematics of creation of profiles: vertical profiles were created toward the center of the plot first, horizontal profiles followed afterward. Examples show vertical face at the center, (left) horizontal face at 20-cm depth for Plot 2b, and (right) all horizontal profiles created for Plot 2a.
|
|
Altogether, 49 horizontal and 51 vertical profiles were exposed. Smeared surfaces and crumbs were carefully removed from the profiles using a knife. Photos were taken with a digital camera (Canon Powershot Pro 70, 1536 by 1024 pixels) using a wooden frame as a spatial reference. The frame was 1 by 1 m and had a 50-cm-square grid. A sheet was used to avoid strong contrasts by unequal light exposure of the photographs, as caused by exposure to direct sunlight.
The images were orthogonalized with the georeferencing software WGEO (WASY GmbH, 2006) by transforming four corners on the photographed reference frame from internal raster coordinates to real coordinates. The images were prepared and analyzed using image processing tools of MATLAB 6.5 (MathWorks, 2006). The background was corrected to eliminate variable illumination, the contrast was linearly stretched, and the images were "thresholded" to create binary images with stained and unstained areas. By using manually selected training sites in comparison with the original image, a single value per image was identified as a threshold and saved to make the procedure reproducible. Images of vertical profiles were produced by assembling four smaller images with higher quality. Pixels that were mistakenly identified as stained were removed. One image pixel corresponded to 1 mm2 of soil. The outer 20 mm and upper 30 mm of the profiles were excluded during analysis in order to avoid inclusion of edge effects such as staining along the frame and disturbances caused by the removal of the frame. Dye coverage was calculated from the percentage of stained pixels vs. depth according to Flury et al. (1994), but taking into account all of the 11 to 13 vertical profiles by calculating averaged values. The same procedure was applied analogously for analyzing horizontal profiles. Here, relative dye coverage was calculated for each horizontal cross-section, yielding single values of relative dye coverage for each depth of a horizontal profile.
|
RESULTS
|
|---|
The dye tracer Brilliant Blue exhibited good visible contrast against the investigated reddish (Munsell hue 5) soil (see Fig. 3
). The maximum depths of dye penetration were 74.7, 94, and 77.5 cm for the vertical and >90-, >120-, and >80-cm depth for the horizontal profiles of Plots 1b, 2a, and 2b, respectively. Most of the matrix was bypassed and staining occurred only along preferential pathways (Fig. 3). A horizontal network of cracks connected to surface cracks was visible at the 5-cm depth in the cultivated horizon. Many continuous cracks and biopores extended through the plow pan and could be detected to about the 30-cm soil depth (Fig. 4
). Crack networks (Fig. 5
) above (5-cm depth) and below the plow pan (30 cm) were dissimilar and seemed not to be spatially correlated or connected. Thus, the deeper cracks in and below the pan were probably not formed within the actual drying period. Horizontal cracks were observed above the plow pan during excavation of the profiles. Superimposed vertical profiles of relative dye coverage (Fig. 6
) show distinct local maxima of about 61% at the 7.6-cm depth (Plot 1b), 72% at the 9.8-cm depth (Plot 2a), and 76% at the 11.8-cm depth (Plot 2b). Horizontal profiles show a dye coverage maximum of 60% at the 12-cm soil depth (Plot 1b), 70% at the 10-cm depth (Plot 2a), and 61% at the 15-cm soil depth (Plot 2b). All maxima occurred near the plow pan for both vertical and horizontal profiles. Second peaks of 21% dye coverage occurred at 38.3 cm (Plot 1b) and 37% at the 41.1-cm soil depth (Plot 2b) for vertical profiles, and 27% at 45 cm (Plot 1b) and 30% at the 40-cm depth (Plot 2b) for horizontal profiles (Fig. 6). The mean dye coverage along the soil depth calculated from the vertical profiles (Fig. 6, right, solid lines) is in relatively good agreement with the mean values that were obtained from the horizontal profiles (Fig. 6, right, crosses).
View larger version (108K):
[in this window]
[in a new window]
|
Fig. 3. Examples of horizontal profiles: stained surface cracks at 4- and 5-cm depths, more horizontally extended dye pattern at 10-, 12-, and 15-cm depths, stained cracks and biopores at 17- to 30-cm depths, and more horizontally extended dye pattern below 40-cm depth.
|
|
View larger version (79K):
[in this window]
[in a new window]
|
Fig. 4. Example of staining patterns: upper left corner (0.5 by 0.5 m, 0–50-cm depth) of a vertical profile at 10-cm horizontal distance from the center of Plot 2b. Arrows indicate the plow pan at about 15-cm depth and a second soil structural change at about 40-cm depth.
|
|
View larger version (28K):
[in this window]
[in a new window]
|
Fig. 5. Binary images of horizontal staining pattern for (top) Plot 2a and (bottom) Plot 2b (left) in the cracked cultivated topsoil, (center) in the compacted plow pan with cracks and biopores, and (right) in the uncompacted subsoil with radially spread dye around macropores.
|
|
View larger version (44K):
[in this window]
[in a new window]
|
Fig. 6. Spatial distribution of (left) superimposed two-dimensional vertical cross-sections and (right) horizontally integrated one-dimensional vertical values of relative dye coverage vs. depth for (top) Plot 1b, (center) Plot 2a, and (bottom) Plot 2b. Solid lines indicate values obtained from vertical profiles, crosses indicate values obtained from horizontal profiles. The grey-scale bar (Plot 2a, center) for the dye coverage ranges between values of 0 (white) and 1 (black), indicating the relative numbers of stained pixels from all vertical two-dimensional cross-sections. Arrows indicate peaks of dye coverage.
|
|
The topography of the soil surface, the location of crack networks, and rice plants residues after harvest are presented in Fig. 7
. The largest variations in topography within plots were 6.2 (Plot 1b), 6.3 (Plot 2a), and 6.1 cm (Plot 2b). Half of Plot 1b (only one half was photographed) included eight, whereas the two other plots included 16 rice stubble residues, with six (Plot 1b) and 11 (Plots 2a and 2b) stubbles intersected by cracks. The topography of the surface of the plow pan was measured using an additional 1-m2 plot in Field 2, showing maximum elevation differences of the plow pan surface of >8 cm (Fig. 8
) and a different microtopography as compared to the soil surface.
View larger version (61K):
[in this window]
[in a new window]
|
Fig. 7. Horizontal views of the plots showing the soil surface topography (lowest locations are blue and highest locations are red-brown). Location of cracks (black lines) and rice stubbles (black solid circles) were obtained from photos. For Plot 1b, cracks and stubbles are not shown for half the plot since photos from the soil surface above the vertical profiles were not available.
|
|
View larger version (31K):
[in this window]
[in a new window]
|
Fig. 8. Topography of the surface of the plow pan, determined at a separate 1-m2 plot in Field 2. The lowest spot was used as the reference height.
|
|
|
DISCUSSION
|
|---|
In the puddled horizon, shrinkage-induced cracks are formed after drainage of the paddy field. Through these cracks, dye tracer and dissolved chemicals can move rapidly downward to the surface of the plow pan, where a first zone of spreading of dye was observed. The location of staining in the cultivated horizon (Fig. 3 and 5) is not governed by surface topography, but only by the location of surface cracks. Surface cracks, in turn, may be governed by the location of rice plants and topography (Fig. 7). Above the surface of the relatively low-permeability plow pan, dye tracer movement was restricted, causing the applied solution to spread laterally along horizontal cracks, as observed during excavation of the soil profiles. Intact macropores were present in the plow pan, and were not completely destroyed by puddling. These biopores probably originated from decayed roots and earthworms found during excavation. Staining patterns (Fig. 3 and 4) show that both cracks and biopores serve as preferential pathways through the plow horizon. While the hydraulic properties of cracks may change due to swelling of the surrounding soil matrix, the properties of vertical biopores are not affected since they are generally regarded as mechanically more stable due to their cylindrical shape and composition (e.g., Hartge and Sommer, 1980; Blackwell, 1990; Horn, 1994). Furthermore, biopores, unlike cracks, may also form under wet conditions.
A second peak of dye coverage was observed below the compacted zone at about the 40-cm soil depth in Plot 1b and 2b (Fig. 6), indicating a second zone of horizontal spreading of BB (i.e., a "dispersion zone", according to Flühler et al., 1996). A similar phenomenon of "pulse splitting" near the plow pan was observed by other studies (e.g., Kasteel et al., 2002), probably caused by the rapid supply of dye through the macropores with a higher vertical unsaturated hydraulic conductivity than the underlying soil matrix. The horizontal spreading could also be caused by the higher horizontal permeability of the uncompacted and less structured soil (i.e., fewer macropores, no prismatic aggregates), enhancing radial spreading of dye around macropores (Fig. 5, right). In contrast, vertical flow and transport in and below the compacted horizons might be directed by the macropore network since water and solute mass transfer from the macropores into the surrounding soil matrix is restricted either by the low permeability of coated macropore walls or a relatively denser surrounding soil matrix. The differences in the distance of dye spreading from macropores or cracks into the soil matrix demonstrate different properties that control the rates of dye tracer exchange between pore domains. Figure 9
shows that preferential pathways with relatively lower and higher matrix exchange of dye tracer may coexist in the same horizontal profile; however, no information is available on the local direction of dye movement at the time of these observations. Additional temporal information (i.e., by local pressure head distribution) would be necessary to better understand nonequilibrium conditions between flow domains.
View larger version (101K):
[in this window]
[in a new window]
|
Fig. 9. Example of (right) stained structures with little lateral dispersion close to (left) stains with strongly laterally dispersed dye at one horizontal profile of Plot 2a at 40-cm depth.
|
|
Although Plots 1b and 2b both show a second maximum in dye patterns, these maxima look dissimilar (Fig. 6). The distinct peak in Plot 2b on the left side of the profile indicates the presence of a restricting soil layer at the 40-cm soil depth that originates from the absence of macropores rather than from a lower permeability soil matrix. This pattern correlates with a change in soil structure and the location of the boundary between the B2 and C1 horizons (Table 2). The slightly inclined structure of the C1 horizon could have been caused by the removal of soil material some 40 yr earlier. The pattern of dye coverage in Plot 1b (Fig. 6, top) reflects the same bypass effect through the B2 horizon as in Plot 2b, but without such a significant zone of flow resistance on top of the C1 horizon.
A comparison of crack networks in the cultivated horizon (e.g., at 5-cm depth) with stained cracks within and below the plow pan (e.g., at about 30-cm depth) reveals that the crack network at depth looks dissimilar (Fig. 5) in shape with respect to the location of surface cracks. Thus, cracks below the plow pan can hardly be seasonal cracks, and may have been developed during earlier drying periods and were preserved. The relatively high degree of dye coverage directly over the plow pan (first peak, Fig. 6) indicates lateral spreading, which allows dye tracer solution to enter those subsoil macropores that are not directly in contact with or located directly below seasonal topsoil cracks. Visual inspection showed that the surface of the plow pan was nearly completely stained. Since the plow pan surface is rather uneven (Fig. 8), however, the peaks in dye coverage calculated for fixed depth intervals are unable to reflect vertical continuity of staining in greater detail.
Puddling may indirectly affect the pore continuity in the unpuddled subsoil by local compression or incorporation of dispersed particles. Nevertheless, the existence of stained macropores below the plow pan proves that puddling does not completely destroy the connectivity of these pathways through the plow pan or that connectivity may possibly become reestablished during the cropping season.
Preservation of the surface topography (no surface leveling) and a ponding depth of only 50 mm were chosen to obtain more realistic experimental boundary conditions. The selected infiltration experiment should compare well to a field situation of initial ponding after a fallow (drying) period. The tracer application, however, by pulsing was faster than typical flooding practices, as time periods until complete ponding by rain or flood irrigation typically range between 0.5 h and several days. So we cannot exclude that preferential flow was slightly exaggerated in the experiments. Probably, slower soil wetting conditions allowed for gradual swelling of the soil matrix and crack closure, thus reducing preferential flow effects.
For this study, we used Brilliant Blue as a dye tracer to visualize preferential flow paths. Although others (e.g., Flury and Flühler, 1995; Kasteel et al., 2002; Ketelsen and Meyer-Windel, 1999) have shown that BB can be adsorbed to the soil, it is considered a good compromise between visibility, toxicity, and mobility (Flury and Flühler, 1995). The evaluation of the dye-stained areas combining horizontal and vertical profiles is innovative, because data of relatively high spatial resolution in vertical and horizontal directions help reducing the risk of bias in the analysis, and provide a crosscheck for the accuracy of the calculated dye-stained areas.
The phenomenon of lateral spreading of dye above a plow pan or its decrease below an Ap horizon has been recognized before (e.g., Forrer et al., 1999; Schwartz et al., 1999; Nobles et al., 2004). In other studies, the soil surface was homogenized to some depth, thus not allowing any comparison (Flury et al., 1994; Yasuda et al., 2001), or no plow pan was detected (Kulli Honauer, 2002; Zehe and Flühler, 2001). The effect of spreading above the plow pan seems to be less pronounced in studies conducted on nonpaddy fields than the fields investigated in this study, or was even not recognized (Ghodrati and Jury, 1990). The effect of pressure buildup on the pan, which increases the susceptibility to preferential flow, may be typical for paddy soils where soil structure differs significantly between puddled topsoil and pan.
The location of preferential flow paths reflects a particular situation for periodic fallow soil conditions, thus representing the effects of seasonal crack systems of the cultivated horizon. We focused on this situation since we assumed that most preferential flow, if any, would occur during the period of initial flooding of a rice paddy, in which the soil has been subjected to drying during an unflooded period after rice harvest. Our study is not conclusive for preferential flow patterns under flooded conditions. In the presence of relatively stable subsoil biopores and crack systems, however, we may hypothesize that preferential flow patterns during ponded conditions will depend on pore continuity through the plow pan to the underlying macropore system. Investigations of flow during continuous ponded conditions was not the focus of this study, but will be addressed separately.
The observed preferential flow paths are representative for a transport regime during periodic fallowing. Irrigation or heavy rainfall may lead to rapid movement of water and chemicals toward greater soil depths under these conditions. This study reveals the vulnerability to contamination of groundwater bodies under such land use. Leaching of chemicals from paddy fields could occur as distinct peaks, especially during rewetting after periodic fallow conditions. Chemical leaching, however, is also partly controlled by the initial chemical distribution in the soil. Depending on water and soil management practices, chemicals are usually applied to either the soil or the water surface (i.e., in the case of ponding) as with irrigation water. Chemical residuals in the topsoil may accumulate along crack surfaces due to evaporation-induced transport toward cracks. For a better understanding of leaching, initial chemical distributions as well as chemical concentrations in the groundwater must be determined. The results of our experiment can only be interpreted in terms of leaching for those nonreactive chemicals that are already in the irrigation water or those chemicals that may readily enter the soil solution during flooding after their desorption from exposed soil surfaces. During ponded conditions in the puddled topsoil, chemical mobilization toward the flowing solution phase may be large for chemicals that are within an intensively mixed soil paste.
|
CONCLUSIONS
|
|---|
This study used ponded dye tracer application to analyze flow paths in paddy field soils. The findings revealed the presence of an extensive (new) crack network in the topsoil, some cracks in the plow pan, and a less dense (older) crack network and (macro) biopores in the subsoil. These crack networks and macropores serve as preferential flow paths, despite prevalent flooding at the field site causing swelling conditions. The plow pan did not prevent preferential flow, but enhanced horizontally redistributed flow prior to downward channeling at localized spots. The study thus demonstrates that crack networks below the plow pan in paddy fields may persist during ponding to form preferential flow paths. Hence, preferential flow is a phenomenon that may clearly be relevant for water and nutrient cycling (and losses to groundwater) in paddy rice fields. Puddling does not necessarily prevent preferential flow in paddy fields and does not completely destroy the hydraulic continuity of macropores in the plow pan.
Preferential leaching of chemicals depends on the application rate and initial presence of chemicals in the soil matrix. Chemicals that are applied to the irrigation water will be readily available for leaching if preferential flow occurs under flooding conditions.
The experimental setup with undisturbed surface topography and maximum ponding depth of 5 cm visualizes macropores only if they serve as flow pathways during typical boundary conditions. Brilliant Blue was found to be useful to demonstrate the occurrence and importance of macropores during infiltration.The evaluation of the dye-stained areas by combining both horizontal and vertical profiles reduced the risk of bias in the analysis and improved the accuracy of calculated dye-stained areas.
The results suggest that hydraulic conditions in paddy fields may be dynamically changing. Simulation models do not correctly describe flow processes, as they often assume homogeneous soil properties, thereby underestimating the infiltration depth and velocity of water and solutes through the plow pan. The occurrence of both cracks and biopores with different susceptibilities to swelling has to be considered for understanding the influence of moisture conditions on flow due to swelling. With respect to a quantitative description of flow and transport processes in paddy fields, results suggest that separate flow domains must be considered.
|
ACKNOWLEDGMENTS
|
|---|
Fieldwork was carried out in collaboration with the Ecological Experimental Station of Red Soil and the Institute of Soil Science, Chinese Academy of Sciences, Nanjing. We thank our Chinese collaborators, Prof. Dr. Zhang Bin, Li Jiangtao, Bi Lidong, and Yue Xing, as well as our colleagues Robert Bartsch and Norbert Wypler from the ZALF in Müncheberg. This project was financially supported by the Deutsche Forschungsgemeinschaft, Bonn (GE 990/5-1). Valuable comments of J. Maximilian Köhne are gratefully acknowledged.
|
REFERENCES
|
|---|
- Ball, B.C., and J.T. Douglas. 2003. A simple procedure for assessing soil structural, rooting and surface conditions. Soil Use Manage. 19:50–56.[CrossRef]
- Blackwell, P.S. 1990. Responses of biopore channels from roots to compression by vertical stresses. Soil Sci. Soc. Am. J. 54:1088–1091.[Abstract/Free Full Text]
- Cabangon, R.J., and T.P. Tuong. 2000. Management of cracked soils for water saving during land preparation for rice cultivation. Soil Tillage Res. 56:105–116.[CrossRef]
- Chen, S.-K., and C.W. Liu. 2002. Analysis of water movement in paddy rice fields: I. Experimental studies. J. Hydrol. 260:206–215.[CrossRef]
- Chen, S.-K., C.W. Liu, and H.-C. Huang. 2002. Analysis of water movement in paddy rice fields: II. Simulation studies. J. Hydrol. 268:259–271.[CrossRef]
- Chertkov, V.Y., and I. Ravina. 1998. Modeling the crack network of swelling clay soils. Soil Sci. Soc. Am. J. 62:1162–1171.[Abstract/Free Full Text]
- Colina, H., and S. Roux. 2000. Experimental model of cracking induced by drying shrinkage. Eur. Phys. J. E1:189–194.
- Ewing, R.P., and R. Horton. 1999. Discriminating dyes in soil color with image analysis. Soil Sci. Soc. Am. J. 63:18–24.[Abstract/Free Full Text]
- FAO. 1998. World reference base for soil resources. World Soil Resour. Rep. 84. FAO, Rome.
- Flühler, H., W. Durner, and M. Flury. 1996. Lateral solute mixing processes—A key for understanding field-scale transport of water and solutes. Geoderma 70:165–183.[CrossRef][ISI]
- Flury, M., and H. Flühler. 1995. Tracer characteristics of Brilliant Blue FCF. Soil Sci. Soc. Am. J. 59:22–27.[ISI]
- Flury, M., H. Flühler, W.A. Jury, and J. Leuenberger. 1994. Susceptibility of soils to preferential flow of water: A field study. Water Resour. Res. 30:1945–1954.
- Forrer, I., R. Kasteel, M. Flury, and H. Flühler. 1999. Longitudinal and lateral dispersion in an unsaturated field soil. Water Resour. Res. 35:3049–3060.[CrossRef]
- Forrer, I., A. Papritz, R. Kasteel, H. Flühler, and D. Luca. 2000. Quantifying dye tracers in soil profiles by image processing. Eur. J. Soil Sci. 51:313–322.
- Germann, P., and K. Beven. 1985. Kinematic wave approximation to infiltration into soils with sorbing macropores. Water Resour. Res. 21:990–996.[CrossRef]
- Ghodrati, M., and W.A. Jury. 1990. A field study using dyes to characterize preferential flow of water. Soil Sci. Soc. Am. J. 54:1558–1563.[ISI]
- Gong, Z. 1986. Origin, evolution, and classification of paddy soils in China. Adv. Soil Sci. 5:179–200.
- Hangen, E., H.H. Gerke, W. Schaaf, and R.F. Hüttl. 2004. Flow path visualization in a lignitic mine soil using iodine-starch staining. Geoderma 120:121–135.[CrossRef][ISI]
- Hartge, K.H., and C. Sommer. 1980. The effect of geometric pattern of soil structure on compressibility. Soil Sci. 130:180–185.
- Horn, R., H. Taubner, M. Wuttke, and T. Baumgart. 1994. Soil physical properties related to soil structure. Soil Tillage Res. 30:187–216.
- Kasteel, R., H. Vogel, and K. Roth. 2002. Effect of non-linear adsorption on the transport behaviour of Brilliant Blue in a field soil. Eur. J. Soil Sci. 53:231–240.
- Ketelsen, H., and S. Meyer-Windel. 1999. Adsorption of Brilliant Blue FCF by soils. Geoderma 90:131–145.[CrossRef][ISI]
- Kirby, J.M., and A.J. Ringrose-Voase. 2000. Drying some Philippine and Indonesian puddled rice soils following surface drainage: Numerical analysis using a swelling soil flow model. Soil Tillage Res. 57:13–30.[CrossRef]
- Kirchhof, G., S. Pryono, W.H. Utomo, T. Adisarwanto, E.V. Dacany, and H.B. So. 2000. The effect of soil puddling on the soil physical properties and the growth of rice and post-rice crops. Soil Tillage Res. 56:37–50.[CrossRef]
- Kulli, B., M. Gysi, and H. Flühler. 2003. Visualizing soil compaction based on flow pattern analysis. Soil Tillage Res. 70:29–40.
- Kulli Honauer, B. 2002. Analysis of flow pattern: the influence of soil compaction and soil structure on the infiltration pathways of dye tracer solutions and the quantitative evaluation of flow patterns. Ph.D. diss. ETH Zürich, Switzerland (ETHZ no. 14658).
- Liu, C.-W., S.-W. Cheng, W.-S. Yu, and S.-K. Chen. 2003. Water infiltration rate in cracked paddy soil. Geoderma 117:169–181.[CrossRef][ISI]
- MathWorks. 2006. MATLAB—The language of technical computing. Available at www.mathworks.com/products/matlab/ (verified 22 Oct. 2006). MathWorks, Natick, MA.
- Nobles, M.M., L.P. Wilding, and K.J. McInnes. 2004. Pathways of dye tracer movement through structured soils on a macroscopic scale. Soil Sci. 169:229–242.[CrossRef]
- Ringrose-Voase, A.J., and W.B. Sanidad. 1996. A method for measuring the development of surface cracks in soils: Application to crack development after lowland rice. Geoderma 71:245–261.[CrossRef][ISI]
- Schwartz, R.C., K.J. McInnes, A.S.R. Juo, and C.E. Cervantes. 1999. The vertical distribution of a dye tracer in a layered soil. Soil Sci. 164:561–573.
- Sharma, P.K., and S.K. De Datta. 1986. Physical properties and processes of puddled rice soils. Adv. Soil Sci. 5:139–178.
- Tong, Y. 2003. Nitrogen loss assessment and environmental consequences in the loess soil of China. Ph.D. diss. Dep. of Forest Ecol., Swedish Univ. of Agric. Sci., Umea.
- Tuong, T.P., M.C.S. Wopereis, J.A. Marquez, and M.J. Kropff. 1994. Mechanism and control of percolation losses in irrigated puddled rice fields. Soil Sci. Soc. Am. J. 58:1794–1803.[Abstract/Free Full Text]
- WASY Gmbh. 2006. Willkommen zu WGEO. Version 4.0. Available at www.wasy.de/deutsch/produkte/wgeo/hilfe/html/welcome_module.htm (verified 22 Oct. 2006). WASY Gmbh, Berlin.
- Wopereis, M.C.S., J. Bouma, M.J. Kropff, and W.B. Sanidad. 1994a. Reducing bypass flow through a dry, cracked and previously puddled rice soil. Soil Tillage Res. 29:1–11.
- Wopereis, M.C.S., B.A.M. Bouman, M.J. Kropff, H.F.M. ten Berge, and A.R. Maligaya. 1994b. Water use efficiency of flooded rice fields: I. Validation of the soil-water balance model SAWAH. Agric. Water Manage. 26:277–289.
- Yasuda, H., R. Bendtsson, H. Persson, A. Bahri, and K. Takuma. 2001. Characterizing preferential transport during flood irrigation of a heavy clay soil using the dye Vitasyn Blau. Geoderma 100:49–66.[CrossRef][ISI]
- Yoshida, S., and K. Adachi. 2004. Numerical analysis of crack generation in saturated deformable soil under row-planted vegetation. Geoderma 120:63–74.[CrossRef][ISI]
- Zehe, E., and H. Flühler. 2001. Preferential transport of isoproturon at a plot scale and field scale tile-drained site. J. Hydrol. 247:100–115.[CrossRef]
- Zhu, J.G., Y. Han, G. Liu, Y.L. Zhang, and X.H. Shao. 2000. Nitrogen in percolation water in paddy fields with a rice/wheat rotation. Nutr. Cycling Agroecosyst. 57:75–82.[CrossRef]
This article has been cited by other articles:
|
|
|
|
 
M. Janssen and B. Lennartz
Characterization of Preferential Flow Pathways through Paddy Bunds with Dye Tracer Tests
Soil Sci. Soc. Am. J.,
October 30, 2008;
72(6):
1756 - 1766.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
T. Sander and H. H. Gerke
Noncontact Shrinkage Curve Determination for Soil Clods and Aggregates by Three-Dimensional Optical Scanning
Soil Sci. Soc. Am. J.,
August 9, 2007;
71(5):
1448 - 1454.
[Abstract]
[Full Text]
[PDF]
|
|
|
TOP
ABSTRACT
INTRODUCTION
