Soil Science Society of America Journal 64:1893-1904 (2000)
© 2000 Soil Science Society of America
DIVISION S-10-WETLAND SOILS
Morphological Changes in Soils Produced When Hydrology Is Altered by Ditching
W.A. Hayes, Jr. and
M.J. Vepraskas
North Carolina State Univ., Dep. of Soil Science, Box 7619, Raleigh, NC 27695-7619 USA
michael_vepraskas{at}ncsu.edu
 |
ABSTRACT
|
|---|
A soil's hydrology (seasonal saturation occurrence) must be estimated in the field to delineate jurisdictional wetlands and to evaluate soil suitability for on-site waste disposal. It is difficult to predict soil hydrology on lands that contain ditches, because the areal extent of hydrologic alteration by an individual ditch is generally unknown. This study evaluated whether morphological changes occurred in soils after a drainage ditch had been installed. Four transects of plots were established parallel to a ditch with plots at distances of 7, 30, 60, and 80 m from the ditch. Each transect contained plots in the following soils: Aquic Paleudults, Aeric Paleaquults, and Typic Paleaquults. Soils within 30 m of the ditch had a significantly (0.10 level) greater volume of Fe masses at depths of 40 to 100 cm than soils further from the ditch. Duration of saturation did not vary significantly with distance from the ditch, but within 30 m of the ditch water tables fluctuated more frequently than those in soils further away. Concentrations of Fe(II) in groundwater at a depth of 60 cm were higher at 7 m from the ditch than at 60 m, but redox potentials at a depth of 60 cm were <500 mv for shorter periods of time at 7 m than at greater distances from the ditch. We hypothesized that groundwater flowing into the soils within 30 m of the ditch introduced Fe(II) into the Bt horizons. The Fe(II) oxidized and formed Fe masses as the water table fell. Our results indicate that soil colors can change within 30 yr as a result of ditching. We suggest that the major area of soil influenced by the ditch can be identified by where the Fe masses in the argillic horizons increase as one approaches the ditch.
Abbreviations: ANOVA, analysis of variance procedures • Eh, redox potential • PVC, polyvinyl chloride
|
INTRODUCTION
|
|---|
PERIODICALLY SATURATED SOILS frequently develop color patterns consisting of low chroma or gray colors in their E, B, and C horizons. The saturation does not change the soil color directly, it simply prevents atmospheric O2 from entering the soil. Soil colors change after a soil becomes saturated when microbial respiration of organic tissues consumes available O2 causing the soil to become anaerobic. Under anaerobic conditions, microbially mediated reduction of ferric iron [Fe(III)] results in an increase of ferrous iron [Fe(II)] in the soil solution (McBride, 1994). Iron reduction eliminates the red and yellow colors produced by Fe(III) compounds, and allows the low (
2) chroma or gray soil colors of silicate minerals to predominate in seasonally saturated soils (Vepraskas, 1994). Because low chroma colors, or redox depletions, frequently form in saturated soil, these color patterns are commonly used to predict the depth of seasonal saturation (Evans and Franzmeier, 1986; Franzmeier et al., 1983; Pickering and Veneman, 1984). In some soils, chromas 3 have also been related to seasonal saturation (Vepraskas and Wilding, 1983; Evans and Franzmeier, 1986).
Estimating the depth to seasonal saturation with low chroma colors is a reliable prediction at sites where the natural soil's hydrology has not been altered. Hydrology can be altered by ditching the land or installing drainage tiles to lower the water table. When hydrologic alterations have occurred we must know whether the soil's low chroma colors reflect the modified water table position, or soil conditions prior to drainage if we are to make reliable predictions of the depth to seasonal saturation. When the low chroma colors reflect conditions prior to drainage, they are considered relics of the former moisture regime and cannot be used to predict current levels of the seasonal high water table (James and Fenton, 1993). Unfortunately, identifying relict low chroma colors in altered soils is difficult unless the extent of hydrological alteration on saturation frequency and duration is known (Vepraskas, 1994).
In the USA, approximately 40 million ha of farmland had been drained as of 1980 (Gosselink and Maltby, 1990). Furthermore, over 50% of the wetland soils in the USA (soils with aquic conditions) have been altered by agricultural drainage or development (Dahl, 1990). These data suggest that for soils with low chroma colors within 0.5 m of the surface, the depth considered for aquic conditions, at least one half of the area these soil's cover may be influenced by a drainage ditch or other form of hydrological alteration.
When evaluating land on-site for septic system installation, or when determining whether a site has jurisdictional wetlands, it is necessary to know how far from the ditch the water tables are being affected. Hydrologic models can be used to compute the dynamic position of a water table following ditching or other forms of land drainage. Hydrological models such as DRAINMOD have been developed to determine optimum drain spacing to achieve a predetermined water table level (Skaggs, 1978). Such models require input of a soil's pore size distribution, saturated hydraulic conductivity, hourly rainfall, and other data that can be expensive to acquire. Simpler models such as the ellipse equation can be used to estimate water table positions between drains (USDA-NRCS, 1997). This equation also requires input of average soil hydraulic conductivity, drainage rate, and vertical distance after drawdown of the water table, among other parameters which cannot be easily obtained. The ellipse equation was developed for sites having a network of parallel drains, and it is inappropriate for use in areas with ditches on perimeters of fields (R. W. Skaggs, 1998, personal communication).
Soil morphology might be used to evaluate a ditch's effectiveness because morphology has been found to change when hydrology is altered. Daniels and Gamble (1967) showed that soils along stream-dissected edges of the Middle Coastal Plain had redder colored Bt horizons than soils further from the edge. This "red-edge effect" was believed to have occurred after the Coastal Plain had been incised and drained by streams. Prior to stream incision, restricted surface and subsurface drainage caused saturated and Fe-reduced conditions to persist for long periods. Following stream incision and headward erosion, improved surface and subsurface drainage resulted in a lowering of watertables and the oxidation of Fe, producing red and yellow soil colors in areas with improved drainage that are adjacent to streams. On the other hand, soils further from the edge remained saturated and reduced for long periods and slowly lost Fe(II) as water moved to the stream. Consequently, the differences in Bt horizon color were produced by changes in Fe oxide contents as distance from the edge increased. The time required for these changes to occur was not known, but Daniels et al. (1975) speculated that tens of thousands of years were likely needed for the dissection alone.
James and Fenton (1993) documented changes in drained and undrained members of a catena of Mollisols (Udolls and Aquolls). The drainage system had been installed approximately 90 yr prior to the time of the study. All soils apparently contained little Fe, and the redoximorphic features present consisted largely of redox depletions. James and Fenton (1993) found that while the drainage system lowered the water table to depths exceeding 40 cm there were no apparent changes in soil color.
More rapid changes in soil color were documented in Aquolls and Udalfs along a created marsh by Vepraskas et al. (1999). They monitored changes in soil color along four landscape positions: in the marsh, at the marsh edge, in a transition to an upland, and in an upland position. They found that at the edge of the marsh where soils were saturated for approximately 6 mo/yr, the soil chroma decreased by two units over a 5-yr period. Changes in chroma of 1 unit occurred during a single year for soils in the transition zone that were saturated for a shorter time than soils at the edge. The color changes monitored in the marsh, edge, and transition were great enough to create hydric soil field indicators over the 5-yr period (USDA-NRCS, 1998). This study showed that detectable changes in soil color can occur in periods of less than 5 yr when Fe(II) is present in the reduced soil solution at concentrations between 5 and 10 mg/L.
The objectives of this study were to: (i) determine if soil morphology had changed in response to ditching at a site in the Lower Coastal Plain of North Carolina; (ii) relate any observed changes to current hydrology and oxidation-reduction dynamics; and (iii) assess the effects of ditching on soil classification and hydric soil classification.
|
Materials and methods
|
|---|
The study site was located in Pitt County, NC, approximately 5.1 km southwest of Greenville at N 35° 34' 10'', and W 77° 26' 26''. It is on a broad interstream divide of the Lower Coastal Plain, approximately 7.6 m above sea level (Daniels et al., 1984). Soils were mapped as part of this study and found to be members of the following series: Goldsboro (Aquic Paleudults), Lynchburg (Aeric Paleaquults), and Rains (Typic Paleaquults). All series were members of the fine-loamy, siliceous, subactive or semiactive, thermic families. The vegetation covering the site was forest consisting primarily of white oaks (Quercus alba L), red maples (Acer rubrum L), and loblolly pines (Pinus taeda L) up to approximately 50 yr old.
A ditch bounded a portion of the site as shown in Fig. 1
. The ditch was 1.25 m wide and 0.7 m deep and was dug in two sections. The section running predominantly north to south is at least 50 yr old according to the property owners but detailed historical records were not kept. The section running east to west was approximately 30 yr old according to evidence from aerial photographs.
View larger version (19K):
[in this window]
[in a new window]
|
Fig. 1 Schematic map of the study site. One ditch occurred along a portion of the perimeter of the site. The four transects were at differing distances from the ditch. Each transect consisted of at least three soil plots, with a plot in the Goldsboro, Lynchburg, and Rains series. Soil boundaries approximate topographic changes, with the Goldsboro plots lying at an elevation approximately 1 m higher than the Rains plots
|
|
To evaluate the effect of the ditch on soil properties, four transects of plots were laid out at increasing distances from the ditch. A randomized complete block design was used with three replications. Soils of each soil series were considered a separate replicate within a transect, and transects were considered to be treatments (i.e., distance from ditch). As shown in Fig. 1, four treatments were used with transects placed at 7, 30, 60, and 80 m from the ditch. Because the effect of the ditch on soil hydrology was not known prior to these experiments, transect distances were selected arbitrarily to include areas that likely were and were not affected by the ditch. Each transect contained three plots, with a plot in each of the Goldsboro, Lynchburg, and Rains soils. Each plot measured approximately 3 m by 3 m and was instrumented to monitor water table levels, redox potential, soil temperature, and to extract water samples for chemical analysis.
Water table levels were measured daily in each plot with an electronic well (Remote Data Systems, Inc., Wilmington, NC). Wells were installed by digging a 10-cm-diam. hole to a depth of 2.25 m, inserting the well, and backfilling the hole with sand to the level of the slits in the well pipe. The top of the sand was then covered with a 3-cm thick layer of dry bentonite pellets which were then covered with soil. A conical mound of soil was placed around the portion of the well that was above ground to direct rainfall falling on the well away from the well and onto undisturbed soil.
Rainfall was measured in an open field adjacent to the site using a recording rain gauge. Soil temperature was measured weekly using thermocouples installed at depths of 15, 30, and 60 cm. The thermocouples were constructed following the procedure of Culik et al. (1982). The three thermocouples for each plot were placed in 1-in- diam PVC pipe. The thermocouple tips extended through holes in the pipe so they were in contact with soil at depths of 15, 30, and 60 cm.
Redox potentials were monitored weekly at depths of 15, 30, and 60 cm in each plot. Five Pt electrodes were installed at each depth. Electrodes were constructed by fusing 20-gauge Pt wire to metal brass brazing rods (3/32 inch diam.) which were then insulated with heat-shrink tubing and epoxy. The Pt electrodes were checked for accuracy using the redox buffer solution of Light (1972). Electrodes agreeing to within +/- 10 mv in the buffer solution were then tested in distilled water, and those agreeing to within +/- 20 mv were selected for use. Electrodes were installed in the field in 2.5-cm diam. holes backfilled with a mud slurry made from the extracted soil. Redox potential was measured using an Accumet pH/mv meter (Fisher Scientific, Inc., Pittsburgh, PA) and a Ag/AgCl reference electrode filled with a saturated solution of KCl. Field electrode readings were converted to redox potential or Eh by adding a correction factor of 200 to 210 mv to each reading depending on soil temperature.
Water samples were collected from piezometers installed at depths of 30 and 60 cm. Piezometers consisted of a 1-cm-long section of well screen (2-in diam.) glued to PVC pipe (2-in diam.). The piezometers were installed like the electronic wells. Water samples were collected weekly during the "wet" season. Piezometers were first purged of water by hand pumping, and were allowed to refill for approximately 1 h. They were then pumped again to collect two water samples. One sample was placed in a 50-mL glass vial and transported to the laboratory on ice. A second sample was collected and several drops of a 0.2% solution of 
, '-dipyridyl dye were added to it to visually detect the presence of Fe(II) (Childs, 1981). When a positive reaction was noted, it was recorded as being either weak, moderate, or strong depending upon the intensity of the color. For samples brought to the laboratory, pH was measured electrometrically within 4 h of sample collection. Samples were then analyzed for total dissolved organic C, Cl, and Fe(II) by the procedures of US EPA (1979). Because of funding constraints, these analyses could not be performed on all samples. However, we did correlate the observed positive reactions to dye obtained in the field to measured quantities of Fe(II) determined on samples brought to the laboratory. This allowed us to extend our Fe(II) observations in the water samples to periods beyond which measurements were made.
At the time electronic wells were installed, soil samples from the 10-cm-diam holes were collected in 15-cm depth increments to a depth of 2.25 m and bagged. Samples were air-dried and ground to pass a 2-mm-mesh sieve. Free iron oxides were determined on each sample by a dithionite extraction (Coffin, 1963). Total C was determined on the samples from the upper 60 cm of each plot by a CHN Elemental Analyzer (Perkin-Elmer Corp., Norwalk, CT) following the procedures of Nelson and Sommers (1982). Particle size distribution was determined on all samples by the hydrometer method of Gee and Bauder (1985).
Soil profiles from all plots were described from pits (1.0 m wide, 1.5 long, and 1.5 m deep) which were dug within 3 m of each plot. Detailed profile descriptions were completed to a depth of approximately 1 m. Redoximorphic feature percentages were estimated by eye by each author separately using proportion charts, and the results from both authors were averaged for each depth in each profile. The descriptions from pits and auger samples were compared to assess variability in horizon color and redoximorphic feature abundance around the plot.
Treatment effects (i.e., distance from ditch) on soil morphology, hydrology, and chemistry were tested for by an analysis of variance (ANOVA) procedure for a randomized complete block design (SAS, 1985). Three blocks were used with each block being soils of different soil series. Values for a selected variable were averaged across series defined soils for a given depth within each transect. These means were compared by the ANOVA procedures. When significant differences were detected, Tukey's w procedure was used to identify where differences occurred (SAS, 1985).
|
Results
|
|---|
Soil Morphology
Profile descriptions for each soil of transect 3 are given in Table 1
to illustrate the pertinent soil characteristics found in the landscape. Soils from other transects had similar types of features for a comparable landscape position but percentages of individual features differed among transects. Each soil contained redox concentrations and redox depletions in at least a portion of the Bt or Btg horizons. The hydric soil field indicators shown in Table 1 were identified from profile descriptions (USDA-NRCS, 1998). The Umbric Surface field indicator (F13) was found in the Rains soils within each transect, and these were the only hydric soils identified. Soil taxonomic classification and hydric soil status shown in Table 1 did not change across transects.
Results of the ANOVA procedure for comparing the percentages of redox concentrations, redox depletions, and concentrations of colors with chroma of 3 are shown in Table 2
. Only results for the Bt horizons (40 to 100 cm) are presented because no significant (P < 0.10) differences were found for any feature between depths of 0 and 40 cm. As shown in Table 2, no significant difference was found for any feature at the four distances from the ditch shown for the Goldsboro soils. For the Rains and Lynchburg soils the percentages of redox concentrations were significantly higher at distances of 7 and 30 m from the ditch than at the two greater distances. In the Rains soils, the redox concentrations nearly doubled within 30 m of the ditch compared with further away. The effect that the redox concentrations had on Bt horizon color in the Rains plots is illustrated in Fig. 2
for soils at 7 and 60 m from the ditch. The yellowing of the Bt horizons is readily apparent and can be easily detected in the field.
View this table:
[in this window]
[in a new window]
|
Table 2 Comparison of redoximorphic feature abundance as a function of distance from the ditch. All data are for depths between 40 to 100 cm (Bt horizons) because that is where most features occurred
|
|
View larger version (111K):
[in this window]
[in a new window]
|
Fig. 2 Profile photographs showing the development of Fe masses in the Rains plots at 7 m (A) and 60 m (B) from the ditch. Approximately twice as many Fe masses (40- to 100-cm depths) were found at 7 m than at 60 m from the ditch
|
|
In the Lynchburg soils, the percentage of redox depletions also appeared to be greater within 30 m of the ditch as compared with further away (Table 2), but no significant (0.10 level) differences were found. Our statistical design required soils at all positions be analyzed as a group for a given transect. Because no differences in redox depletions were found in the other two soils, significant effects in the plots of Lynchburg soils were not identified. Percentages of colors with chromas 3 and values 
4 also did not significantly change with distance from the ditch.
The increase in redox concentrations found within 30 m of the ditch for the Rains and Lynchburg soils was related to a general increase in the amount of dithionite-extractable Fe as shown in Fig. 3
. Data were expressed as the Fe:clay ratio to remove the effect of clay distribution on changes in Fe concentration with depth. The increase in Fe per unit of clay with depth shows the effects of Fe movement caused by reduction rather than by clay illuviation. At a distance of 7 m from the ditch in the Rains soils (Fig. 3A), the Fe:clay ratios of 200 to 300 between the depths of 40 to 100 cm corresponded with the zone where an increase in redox concentrations was observed. The Goldsboro and Lynchburg soils at 7 m from the ditch had Fe:clay ratio changes with depth that suggested an accumulation Fe between the depths of 100 and 160 cm. At a distance of 60 m from the ditch (Fig. 3B) the Fe:clay ratios were generally <100 for the same depths in the Rains with one small increase at 69 cm. In the Goldsboro and Lynchburg soils, the Fe:clay ratios tended to be more uniform with depth than those found at a distance of 7 m from the ditch.
View larger version (18K):
[in this window]
[in a new window]
|
Fig. 3 Iron to clay ratios with depth for the Goldsboro, Lynchburg, and Rains soils at 7 m from the ditch (A) and 60 m from the ditch (B). Prior to ditching the soils may have had a depth function like that shown in B. Following ditching, the increase in Fe found near the ditch apparently caused an increase in Fe:clay, particularly at depths below 50 cm
|
|
Saturation
Rainfall data collected for the study period are shown in Table 3 along with soil temperature. The normal range in rainfall values show that the periods measured in 1996 and 1998 fell within the normal range, while rainfall in 1997 was below normal. Rainfall in 1998 during the months of January through April, along with July and August were all above normal and this indicated that 1998 was a relatively wet year. Temperature data showed that the soils remained above 5° C throughout the study period. This indicated the biologically-mediated reactions responsible for Fe reduction could be expected to occur throughout the year (Megonigal et al., 1996).
Cumulative periods of saturation for the two wet seasons evaluated are shown in Table 4
. There were apparent differences in cumulative durations of saturation among soils for a given depth, but only one significant difference was found. The Rains soils at a depth of 15 cm and at a distance of 7 m from the ditch were saturated for significantly shorter periods than were soils at distances of 30, 60, or 80 m from the ditch. The ditch had very little effect on cumulative saturation duration at depths between 40 and 100 cm which is the zone where the volume of Fe masses increased toward the ditch.
The ditch did have a major influence on the frequency of water table fluctuation as shown in Fig. 4
. In the Rains soils at a distance of 7 m from the ditch the water table fluctuated often between the surface and a depth of 50 cm between Day 335 in 1996 and Day 120 in 1997, and a similar effect was noted in 1998. Each time the water table fell at the 7-m distance probably allowed oxygen to penetrate into the Btg horizon as macropores (e.g. root channels) drained. At a distance of 60 m from the ditch the saturation was virtually continuous below 25 cm for a similar period. In fact, the effect of the ditch on water table fluctuation disappeared at distances 30 m from the ditch.
View larger version (18K):
[in this window]
[in a new window]
|
Fig. 4 Water table fluctuations in Rains plots at 7 m (A) and 60 m (B) from the ditch. The water table in A showed greater fluctuation for the periods indicated, and allowed some oxygen to penetrate to depths of 50 cm
|
|
Iron Reduction
The occurrence of Fe reduction was evaluated by means of measurements of Fe(II) concentrations in water samples and also by Eh measurements. Changes in Fe(II) concentrations over time for a depth of 60 cm are shown in Fig. 5
for the Rains soils. In 1997, the amounts of Fe(II) in the soil water were generally higher at a distance of 7 m than those found at 60 m from the ditch even though the soil at 60 m remained saturated throughout this period. In 1998, dissolved Fe(II) levels were about one-half those found in 1997. The dissolved Fe(II) levels at a distance of 7 m from the ditch were comparable to those found at a distance of 60 m, and may have been higher just before the soil drained.
View larger version (17K):
[in this window]
[in a new window]
|
Fig. 5 Concentrations of Fe(II) in groundwater samples extracted at a depth of 60 cm in Rains soils at 7 and 60 m from the ditch in 1997 (A) and 1998 (B). The soils at 7 m from the ditch developed higher Fe(II) levels over time, which were maintained in 1997 even after periods of drainage and resaturation. Concentrations estimated visually by dye reaction in the field are shown as circled symbols
|
|
Iron(II) was also detected in water samples from the Lynchburg soils as shown in Fig. 6
. The intermittent saturation of the soil at the 7-m distance from the ditch limited us to collecting two samples, and therefore data for the plots at 30 m from the ditch are shown for comparison. In 1997, relatively high levels of Fe(II) were found at a distance of 60 m from the ditch which were approximately twice as high as found at a distance of 30 m. Soils at both transects drained between Days 100 and 120 in 1997, but Fe(II) was again detected in solution following resaturation, with the higher concentrations found at 60 m than closer to the ditch. A similar trend was found in 1998. These results suggest that a portion of the Fe(II) entering the Rains soils near the ditch may originate in the Lynchburg soils at a distance of 60 m and more from the ditch. Some Fe(II) probably originates in the Goldsboro soils at similar distances from the ditch as well, but sampling wells were too shallow in these soils to have water in them.
View larger version (15K):
[in this window]
[in a new window]
|
Fig. 6 Concentrations of Fe(II) in groundwater samples extracted at a depth of 60 cm from the Lynchburg soils at 30 and 60 m from the ditch in 1997 (A) and 1998 (B). The soils at 60 m from the ditch developed higher Fe(II) levels over time, which were maintained in 1997 even after periods of drainage and resaturation. Concentrations estimated visually in the field are shown as circled symbols. Water samples collected at 7 m from the ditch contained only trace levels of Fe(II)
|
|
The number of days that Eh values were <500 mv in soils as a function of depth and distance from the ditch are shown in Table 5
for a period lasting approximately one year. The value of 500 mv was used for these comparisons because when Eh values were less than 500 mv dissolved Fe(II) could be detected in extracted water samples. The Eh values were never below 500 mv in the Goldsboro soils, and were <500 mv in the Lynchburg soils only at a depth of 60 cm within a distance of 30 m from the ditch. Rains soils had Eh values <500 mv at each depth examined at all distances from the ditch (Table 5). Although the plots located at 7 m from the ditch had the higher Fe(II) levels, the Eh values were <500 mv for the shortest duration of time. This suggested that the Fe(II) found at 7 m from the ditch was coming from a source further from the ditch.
View this table:
[in this window]
[in a new window]
|
Table 5 Comparison of cumulative days when mean redox potential values were <500 mv and Fe reduction was occurring for all three soils at different distances from the ditch for the period between 1 Dec. 1997 through 31 Dec. 1998. Redox potentials in Goldsboro plots were >500 mv for entire period
|
|
Changes in Eh following saturation are compared for the Rains soil at a depth of 60 cm for distances of 7 and 60 m from the ditch in Fig. 7
. Both horizons became saturated on the same date in 1997. The Eh in the horizon at a distance of 60 m from the ditch began to decrease almost immediately following saturation. On the other hand, the Eh in the horizon at a distance of 7 m from the ditch increased slightly following saturation, then remained between 600 and 700 mv for 3 wk during a period when 130 mm of rain fell. The Eh at this depth also increased slightly on Day 46 (15 February) after 14 mm of rain fell, and again on Day 125 (5 May) after 28 mm of rain fell. The influx of aerated rainfall presumably retarded the decrease in Eh in this horizon following saturation. As can be seen in Fig. 7, the Btg horizon at 7 m from the ditch developed Eh values <500 mv 20 d later than the horizon at 60 m from the ditch. In addition, the horizon at 7 m from the ditch drained earlier than the horizon at 60 m, and its Eh remained above 500 mv for at least 43 d while its counterpart at 60 m from the ditch maintained an Eh <300 mv.
View larger version (25K):
[in this window]
[in a new window]
|
Fig. 7 Redox potential measurements for Rains soils at 7 m and 60 m from the ditch at a depth of 60 cm. Data are means of five replicate measurements. Saturation occurred on the same day in both plots, but a redox potential of 500 mv was reached 20 d later in the plot 7 m from the ditch compared with the plot at 60 m from the ditch. This is explained by influxes of aerated water occurring frequently at 7 m from the ditch. The "spikes" in redox potential found at 7 m coincide with large rainfall events and show that some aerated rainfall apparently reached the 60 cm depth. An earlier fall of the water table also caused redox potentials at 7 m from the ditch to rise earlier than at 60 m. A redox potential of 500 mv was selected for reference because at this point and lower Fe(II) was detected in water samples. The range in Eh values for both distances from the ditch was approximately 150 mv, but increased to 350 mv when the mean Eh values decreased after saturation. Soil pH's ranged from 4.5 to 5.2 during the study
|
|
The slower development of reducing conditions (Eh values <500 mv) in the Rains plots at a distance of 7 m from the ditch was not related to a limited amount of organic C either in the soil or in the soil solution. As shown in Table 6
, soil horizons at distances of both 7 and 60 m from the ditch had similar levels of organic C in the soil and in solution. Thus, we speculate that the retarded decrease in Eh values following saturation in the plot at 7 m from the ditch was caused by the penetration of water from recent rainfall events to at least the 60 cm depth. This water introduced dissolved O2 into the soil and kept the Eh above 500 mv. Rainfall did not penetrate as deeply into the plot found at 60 m from the ditch because the water table remained closer to the surface and fluctuated less as shown in Fig. 4.
The generally higher Eh values and shorter periods for Fe reduction found at 7 m from the ditch than further away indicated that the relatively large influx of Fe(II) detected in the soil water was caused by Fe(II) being reduced in soils further from the ditch and carried toward the ditch by groundwater. Additional evidence for this comes from elevated dissolved Cl concentrations found within 7 m of the ditch. As shown in Table 6, Cl concentrations increased progressively from 60 m toward the ditch. This suggests an enrichment of solutes in the Rains soils at a distance of 7 m from the ditch that is most likely caused by groundwater moving toward the ditch from soils at least as distant as 60 m.
|
Discussion
|
|---|
Our results showed that Fe accumulated in the Bt horizons of soils that were within 30 m of a ditch within the 30-yr period since the east to west portion of ditch B (Fig. 1) had been dug. The effect of the ditch on soils in the landscape is two-fold. First, it creates a discharge point for groundwater in the landscape. Groundwater from below Goldsboro, Lynchburg, and Rains soils is probably moving toward the ditch and bringing Fe(II) into the Bt horizons within 30 m of the ditch. Second, the ditch drains water quickly from the soils within 30 m of the ditch producing a greater frequency of water table fluctuation. A portion of the precipitation falling within 30 m of the ditch penetrates to depths of 60 cm, and the dissolved oxygen in the infiltrated rainfall retards Fe reduction. The Rains soils that were <30 m from the ditch were reduced (i.e., Eh values <500 mv) for the shortest duration of time.
Our concept as to why the Rains and Lynchburg soils within 30 m of the ditch had more Fe masses in their Bt horizons than those further away is shown in Fig. 8
. This zone within 30 m of the ditch is called the "oxidized zone" in the illustration (Fig. 8A), because this zone was where the water table tended to be lower and fluctuated more often than in the plots further away. The width of the oxidized zone extended to at least 7 m away from the ditch, but probably did not go beyond 30 m from the ditch. During the wet season when water table levels were high across the landscape, rainfall infiltrated and delayed the onset of Fe reduction in the oxidized zone by up to 20 d (Fig. 8B). Iron that was reduced in the Lynchburg soils, and possibly below 60 cm in the Goldsboro soils, moved toward the ditch with the groundwater that was discharging into the ditch (Fig. 8C). This movement explains the higher Fe(II) levels found in the Rains groundwater samples that were collected at 7 m from the ditch, and is our explanation for the higher Cl concentrations at these locations. The reduced Fe(II) was oxidized in the Btg horizons that were within the oxidized zone adjacent to the ditch. This most likely occurs in late spring when evapotranspiration increases and water tables drop in all soils and oxygen is introduced into reduced horizons. Our Eh results indicate that Fe(II) oxidation could also occur when Btg horizons of the Rains soils are saturated with oxidized water following rainfall events. This phenomenon apparently occurs in the Lynchburg soils too, especially those within 30 m of the ditch which were reduced for relatively short durations of time.
View larger version (24K):
[in this window]
[in a new window]
|
Fig. 8 Theory explaining why Fe masses develop in soils within 7 to 30 m of the ditch. A: the frequent water table fluctuations and earlier decline of the water table keep this zone oxidized for a longer period during the year than is found further from the ditch. B: the oxidized zone is not saturated for shorter periods because rainfall penetrates to depths of 60 cm and deeper, and this introduces aerated water into the soils. C: the ditch serves as a discharge point for groundwater. Iron(II) that has been reduced further upslope, possibly beneath the Goldsboro (G) or Lynchburg (L) soils is carried toward the ditch in moving groundwater. When the Fe(II) encounters the oxidized zone a portion of it is oxidized in the form of Fe masses
|
|
According to our explanation, the Fe masses accumulated along the ditch because the ditch was a discharge point where groundwater containing Fe(II) moved toward. Other features in a landscape can serve as discharge points such as streams and possibly isolated depressions. Such discharge points could conceivably also be points of Fe accumulation in similar landscapes.
In summary, we found that the Lynchburg and Rains soils within 30 m of a ditch contained significantly more Fe masses than soils further from the ditch. The increase in masses was believed to have occurred over a 30-yr period because of the effect of the ditch on soil hydrology. Ditching caused Fe masses to increase in the soils adjacent to the ditch by (i) lengthening periods of oxidation where Eh values >500 mv, and (ii) serving as a discharge point for Fe(II)-bearing groundwater. Both processes had to occur for the Fe masses to accumulate. The ditch did not significantly affect durations of saturation below depths of 15 cm.
Auger borings in additional soils along ditches in North Carolina have confirmed that concentrations of Fe masses decrease away from the ditch. These results can be used to quickly estimate the distance that a ditch is influencing soil water table fluctuation and redox potential perpendicular to the ditch. As an on-site evaluation technique, we suggest collecting a series of soil cores to a depth of approximately 1 m with a bucket auger at varying distances from the ditch being examined. By laying the cores next to one another, their morphology can be compared, and the change in Fe mass abundance can be noted. The point at which the oxidized masses decrease in abundance by roughly one-half may indicate the limit of the ditch's effectiveness on soil redox potential and water table fluctuation.
These results will be landscape dependent. Different results may be found on landscapes where (i) there are low levels (e.g., <5 mg/L) of Fe(II) in groundwater due to little reduction occurring or small amounts of reducible Fe being in the soils, (ii) a shallow ditch was installed to handle surface water but not to discharge groundwater, (iii) ditches are spaced closely and do not discharge groundwater that originates at distances >30 m from the ditch, (iv) hydraulic conductivity and or hydraulic gradients are so small that little water movement occurs, and (v) insufficient time has elapsed since ditch installation for substantial changes in the soil to occur. Different results would also be expected where Fe masses are present but not visible due to masking of soil organic material as may happen in Mollisols.Gee Bauder 1986; SAS Institute 1985; U.S. EPA. 1979
|
ACKNOWLEDGMENTS
|
|---|
Funding for the project was obtained from the U.S. Environmental Protection Agency (contract no. CR 824735-01-0) and the Water Resources Research Institute of the University of North Carolina (WRRI Project no. 70175). Their assistance is greatly appreciated.
Received for publication September 6, 1999.
|
REFERENCES
|
|---|
- Childs C.W. Field testing for ferrous iron and ferric-organic complexes (on exchange sites or in water-soluble forms) in soils. Aust. J. Soil Res. 1981;19:175-180.
- Coffin D.E. A method for the determination of free iron in soils and clays. Can. J. Soil Sci. 1963;43:7-17.
- Culik M.N., Doran J.W., Richards K.A. Construction of soil thermocouples for the novice. Soil Sci. Soc. Am. J. 1982;46:882-884.[Abstract/Free Full Text]
- Dahl T.E. Wetlands losses in the United States, 1790 to 1980s. Washington, DC: U.S. Dep. Interior, Fish and Wildlife Serv, 1990.
- Daniels R.B., Gamble E.E. The edge effect in some Ultisols in the North Carolina Coastal Plain. Geoderma 1967;1:117-124.
- Daniels R.B., Gamble E.E., Buol S.W., Bailey H.H. Free iron sources in an Aquult-Udult sequence from North Carolina. Soil Sci. Soc. Am. Proc. 1975;39:335-340.
- Daniels, R.B., H.J. Kleiss, S.W. Buol, H.J. Byrd, and J.A. Phillips. 1984. Soil systems in North Carolina. NC Agric. Res. Serv. Bull. 467.
- Evans C.V., Franzmeier D.P. Saturation, aeration, and color patterns in a toposequence of soils in north-central Indiana. Soil Sci. Soc. Am. J. 1986;50:975-980.[Abstract/Free Full Text]
- Franzmeier D.P., Yahner J.E., Steinhardt G.C., Sinclair H.R., Jr. Color patterns and water table levels in some Indiana soils. Soil Sci. Soc. Am. J. 1983;47:1196-1202.[Abstract/Free Full Text]
- Gee G.W., Bauder J.W. Particle-size analysis. In: Klute A., ed. Methods of soil analysis. Part 1, 2nd ed Madison, WI: ASA and SSSA, 1986:383-411.
- Gosselink J.G., Maltby E. Wetland losses and gains. In: Williams M., ed. Wetlands: A threatened landscape. Oxford, England: B. Blackwell Ltd, 1990:296-322.
- James H.R., Fenton T.E. Water tables in paired artificially drained and undrained soil catenas in Iowa. Soil Sci. Soc. Am. J. 1993;57:774-781.[Abstract/Free Full Text]
- Light T.S. Standard solution for redox potential measurements. Anal. Chem. 1972;44(6):1038-1039.
- McBride M.B. Environmental chemistry of soils. New York: Oxford Univ. Press, 1994.
- Megonigal J.P., Faulkner S.P., Patrick W.H., Jr. The microbial activity season in southeastern hydric soils. Soil Sci. Soc. Am. J. 1996;60:1263-1266.[Abstract/Free Full Text]
- Nelson D.W., Sommers L.E. Total carbon, organic carbon, and organic matter. In: Page A.L., et al. , ed. Methods of soil analysis. Part 2, 2nd ed Madison, WI: ASA and SSSA, 1982:539-577 Agron. Monogr. 9..
- Pickering E.W., Veneman P.L.M. Moisture regimes and morphological characteristics in a hydrosequence in central Massachusetts. Soil Sci. Soc. Am. J. 1984;48:113-118.[Abstract/Free Full Text]
- SAS Institute. SAS user's guide: Statistics, 5th ed Cary, NC: SAS Institute, 1985.
- Skaggs R.W. A water management model for shallow water table soils. Raleigh: NC State Univ, 1978 NC Water Resources Research Instit. Rep. No. 134..
- U.S. Dep. of Agriculture, Natural Resources Conservation Service. 1998. Field indicators of hydric soils in the United States. G.W. Hurt, P.M. Whited, and R.F. Pringle (ed.) USDA, NRCS, Fort Worth, TX.
- U.S. Dep. of Agriculture, Natural Resources Conservation Service. 1997. Hydrology tools for wetland determination, Chapter 19. In D.E. Woodward (ed.) Engineering field handbook, USDA, NRCS, Fort Worth, TX.
- U.S. EPA. Methods for chemical analyses of water and wastes. Washington, DC: USEPA, 1979 USEPA Rep. 600-4-79-020..
- Vepraskas, M.J. 1994. Redoximorphic features for identifying aquic conditions. NC Agric. Res. Serv. Tech. Bull. 301, Raleigh.
- Vepraskas M.J., Richardson J.L., Tandarich J.P., Teets S.J. Dynamics of hydric soil formation along a created deep marsh. Wetlands 1999;19(1):78-89.
- Vepraskas M.J., Wilding L.P. Aquic conditions in soils with and without low chroma colors. Soil Sci. Soc. Am. J. 1983;47:280-285.[Abstract/Free Full Text]
This article has been cited by other articles:
|
|
|
|
 
R. E. Vaughan, B. A. Needelman, P. J. A. Kleinman, and M. C. Rabenhorst
Morphology and Characterization of Ditch Soils at an Atlantic Coastal Plain Farm
Soil Sci. Soc. Am. J.,
May 1, 2008;
72(3):
660 - 669.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
X. He, M. J. Vepraskas, D. L. Lindbo, and R. W. Skaggs
A Method to Predict Soil Saturation Frequency and Duration from Soil Color
Soil Sci. Soc. Am. J.,
May 1, 2003;
67(3):
961 - 969.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
X. He, M. J. Vepraskas, R. W. Skaggs, and D. L. Lindbo
Adapting a Drainage Model to Simulate Water Table Levels in Coastal Plain Soils
Soil Sci. Soc. Am. J.,
September 1, 2002;
66(5):
1722 - 1731.
[Abstract]
[Full Text]
[PDF]
|
|
|
TOP
ABSTRACT
INTRODUCTION
