| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Dep. Agronomy and Soils, Auburn Univ., Auburn, AL 36849
Corresponding author (stoeckel{at}usgs.gov)
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Abbreviations: Pmin, mineral P • Porg, organic P • Ptot, total P • RZ, root zone (0–0.1 m) • SL, slope • SR, sub-root zone (0.2–0.3 m) • SW, swale • %IN, percentage of season inundated
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Microtopography, hydrology, soil texture, and chemical redox state are interrelated in floodplain forests and result in a pattern of mutually dependent plant and microbial communities (Chanway et al., 1991). For this reason, microbial characterization of forested floodplain functions must take into consideration the heterogeneity imparted by microtopography-based effects. Periodic flooding deposits sandy material onto slopes and traps fine clayey and organic material in low areas (Mitsch and Gosselink, 1993). Therefore, slopes are composed of coarser soils that are better drained and less frequently flooded than associated swales. The resulting soil heterogeneity is associated with a mosaic of plant communities that differ in seasonal litter inputs to the soil surface (Conner and Day, 1992; Shure and Gottschalk, 1985), rates of litter degradation (Day, 1983; Lockaby et al., 1996; Shure et al., 1986), and nutrient import–export characteristics (Cooper and Gilliam, 1987; Post and de la Cruz, 1977).
Soil microbial functions in floodplain forests are also influenced by seasonal variations in temperature and hydrology. These influences are manifest in the deposition of organic material to the soil surface with leaf-fall, seasonal fluctuation in root-exuded substrates, and temporal patterns of anaerobic conditions. Aerobic, substrate-rich soil conditions during the growing season contrast sharply with anaerobic conditions and reduced root exudation during dormant season floods. Though plant metabolism slows in winter, soil temperature does not preclude microbial activity during the same time period (Megonigal et al., 1996).
Location in the soil profile can also influence microbial function. Substrate quality, concentration of plant roots, duration and frequency of inundation, and duration of anaerobic conditions all vary with depth in floodplain forests. Substrate quality changes drastically from fresh and partially decomposed surface organic material to partially decomposed organic matter and labile root exudate in rhizosphere soil, and, finally, dissolved organic matter and humified organic material at the subsurface. Rhizosphere microbial communities are demonstrably different from those found in the bulk soil (Chanway et al., 1991; Grayston et al., 1995), and roots do not penetrate deeply in these floodplains (Baker, 1998). Soil in the subsurface is exposed to inundation much more frequently and for greater time periods than is root-zone soil, resulting in prominent periods of anaerobic conditions.
Restoration and remediation success is difficult to gauge in wetland systems, since functioning of these systems is just beginning to be understood (Brinson and Rheinhardt, 1996). Better understanding of microtopography, season, and depth-driven influences on microbially mediated soil nutrient transformations and nutrient import–export dynamics in forested floodplain soils is an important step toward establishment of a reference wetland (Brinson and Rheinhardt, 1996). The objective of this study was to characterize the impacts of microtopography and depth on soil microbial processes related to seasonal nutrient dynamics in a forested floodplain soil. We hypothesize that microbially mediated nutrient cycling activities will vary with microtopographic class and depth. Nutrients are hypothesized to be enriched in low-lying areas and in deeper soils as a result of physical transport and slower microbially mediated transformations in O2-limited conditions. Nutrient import–export characteristics are hypothesized to vary seasonally and with microtopographic class based on system hydrology. Edaphic features are hypothesized to vary with microtopographic class and are expected to exert an influence on hydraulic conductivity, and thereby influence O2 availability and physical transport of organic material in these soils.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
2 m across the floodplain (Murray et al., 2000). Swales are dominated by water tupelo (Nyssa aquatica L.), with minor components of bald cypress [Taxodium distichum (L.) Rich.] and sweetgum (Liquidambar styraciflua L.). Slopes are dominated by laurel oak (Quercus laurifolia Michx.), with lesser representation by red maple (Acer rubrum L.), water tupelo, and sweetgum.
Three pairs of 37-m2 microtopographic plots were chosen across the floodplain for evaluation. One plot of each pair represented concave swales, while the other represented higher slopes that were slightly inclined or convex. The elevation difference between a slope and its associated swale was 1 m or less, and paired plots were 100 m or less apart. Plot pairs were separated by 1 km and were chosen to represent the same hydrologic influences as closely as possible while maintaining statistical independence.
Soil Properties
Soils in the study area were mapped in 1980 as Argent Association (Stuck, 1980) composed of poorly drained soils formed from coastal plain sediments. This mapping unit contains Argent, Okeetee, Wahee, Santee, and Yonges series soils. A recent reevaluation indicated that the study site was composed primarily of Brookman series soils with smaller areas of Meggett, Elloree, Grifton, Osier, Rutlege, Coosaw, Nakina, and Okeetee soils (Murray et al., 2000). Texture, color, and bulk density of collected samples were compared to official soil series descriptions (USDA-NRCS, 1999) of the 14 series identified in the 1980 and 2000 soil surveys.
Based on comparison with soil series descriptions, four of the plots (all three swales and the gently sloping plot SL-II) were Argent series (fine, mixed, thermic Typic Endoaqualfs) with upper horizons characterized by shallow (0–0.15 m) very dark gray clay loam underlain by gray clay. Soils on these plots had coarse subangular blocky structure and periodic but prominent masses of iron accumulation.
Of the other two slopes, one (SL-I) was a Coosaw series soil (loamy, siliceous semiactive thermic Aquic Arenic Hapludult) and the other (SL-III) was a Rutlege soil (sandy, siliceous, thermic Typic Humaquept). The Coosaw series soil was a loamy sand with weak structure and an upper horizon of dark gray coloration and a lower horizon of pale brown coloration. No mottles were observed, though the soil was inundated annually. The Rutlege soil was a friable sandy loam with an upper layer of black sandy loam interspersed with coarse white sand grains and a lower layer of dark brown, finer loamy sand. The Rutlege soil was inundated annually.
Soil bulk density could not be measured using traditional techniques because of mucky characteristics, dense root mats, and high water contents. Instead, an excavation that measured 0.1 m in diameter and 0.1 m deep was carefully made at each mineral soil depth on the four corners of each plot. The volume of the excavation was measured by addition of incremental sand volumes until fill reached the soil surface. Excavated soils were brought to the laboratory and dried at 105°C to constant mass; bulk density was estimated as dry mass divided by excavated volume. Bulk density was measured in this way during two seasons on slopes, but could only be performed in swales once when the water table was unusually low. Soil pH was measured in a settled 1:1 (w/v) slurry of soil and distilled deionized water with a dual electrode.
Depth to sustained reducing conditions during each season was indicated by the presence or absence of rust on burnished soft steel welding rods (Bridgham et al., 1991). Eight rods were inserted around the perimeter of each plot to a soil depth of 60 cm. Depth of rust formation was measured at each sampling date; then rods were cleaned with steel wool and replaced in the soil. Each observation followed 3 mo of continuous soil contact. Average depth of rust formation per plot was transformed to presence or absence scores. Cases where rust formed through at least one-half of a given soil layer in both years scored positive (+). Cases where no rust was evident or, for SR soils, where the average depth of rust formation was higher than the top of the layer, scored negative (-). Cases where there was year-to-year score variation or where the average depth of rust formation fell within the upper half of a layer scored an ambiguous (+/-) rating. Percentage of season inundated (%IN) was calculated using hydrologic data collected at the site. Depth to the water table was continuously monitored at four locations on the floodplain at USDA Forest Service installations and in the Coosawhatchie main channel at two U.S. Geological Survey gaging stations (Early Branch and Grays, SC). Forest Service personnel performed routine water table measurements at additional wells across the floodplain and then developed regression equations to predict water table levels for all areas of the floodplain based on continuous data (M. Eisenbies, personal communication, 1999). Soil layers were considered inundated when the water table was at or above the surface of the layer.
Experimental Design
Samples were collected and data evaluated according to a split-split plot design with three fixed factors (season, microtopographic class, and depth) and two random factors (year and site). Three microtopographic pairs of plots were established and the study was conducted for two four-season periods (September 1996–August 1998). The first collection date was in December 1996. Plots were defined by microtopographic class (swale or slope). Each block was split by season (autumn, winter, spring, or summer) and depth (surface organic material, RZ, or SR). Subsamples (surface organic material and RZ and SR 10-cm-diam. cores) were collected at five randomly selected grid nodes out of 400 defined for each 37-m2 plot. Subsamples were transported on ice to the laboratory for analysis; all subsamples were analyzed individually within 14 d of collection.
Nutrient Analyses
Soils and organic material were dried to constant mass at 105°C and ground by rotation for 12 to 16 h in glass jars containing steel rods. Subsamples were analyzed for total C and N (CHN600, Leco Corp., St. Louis, MO). Plant standards (29.2 g N kg-1 and 406 g C kg-1 nominal values) were used in the place of soil calibration standards to encompass the range of organic contents found in mineral soil samples.
Soil microbial biomass C was measured following a modified fumigation–extraction method (Vance et al., 1987). Twenty-gram subsamples (wet weight) of each fumigated and control soil were extracted with four volume-equivalents (80 mL) of 0.5 mol K2SO4 L-1. Soil water content was not adjusted to a constant value since soils never contained <250 g water kg-1 (Tate et al., 1988). Extracts were clarified through 0.45-µm polycarbonate filters and dissolved C in soil extracts was measured with an automated dissolved organic C analyzer (Dohrman DC80, Rosemount Analytical, Santa Clara, CA). A constant factor of 2.22 g biomass g-1 C extracted (kEC = 0.45) was used as described by Joergensen (1996) and Sparling et al. (1990) for forest soils. Subsamples of each soil were ashed at 450°C for 4 to 6 h in the presence of strong base to oxidize organic P forms (O'Halloran, 1992). Base was added to these acidic soils to prevent P loss as volatilized phosphoric acid at temperatures >213°C (Weast et al., 1984). Mineral phosphates were dissolved from subsamples of matched soil and ash samples into 0.5 mol H2SO4 L-1 (Tiessen and Moir, 1992). Phosphorus concentrations in clarified extracts were determined colorimetrically as molybdate-reactive P using a microplate reader (Olsen and Sommers, 1982). Phosphate concentrations in ash extracts were used to estimate total soil P (Ptot). Phosphate concentrations in the extracts of dry soil (RZ and SR) and surface organic material were used to indicate the sum of free and occluded forms of mineral P (Pmin). Organic phosphorus (Porg) was calculated as the difference between Ptot and Pmin.
Statistical Analyses
Analyses were performed on both nutrient concentration data and, for mineral soils, nutrient concentration data transformed to a mass content basis. Bulk density estimates were used to convert analytical results from concentration to mass basis, and then soil volumes were converted to floodplain area by defining a 0.1-m layer. Dry season bulk density estimates were used to convert slope concentration data from summer and autumn, and wet season bulk density estimates were used for winter and spring. Since swales were always wet, the same bulk density estimate was used for all seasons.
Analysis of variance was performed using the proc mixed routine (Littell et al., 1996) in the SAS system. Site and year were treated as random variables; hydrology, season, and depth were treated as fixed variables. Means separation was performed by least square means analysis with the significance level adjusted by the Sidak option to guard against type III error in multiple comparisons. The rejection criterion for the null hypothesis was set at P < 0.10 as recommended for this type of biological sampling (Parkin, 1993).
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
Frequency and duration of flooding comprise the characteristic hydroperiod of a site. Hydroperiod has been shown to exert an influence over floodplain nutrient dynamics (Lockaby et al., 1996). On the Coosawhatchie River floodplain, autumn and spring were characterized by water table rise then fall; the water table dipped below the swale surface soil only briefly after the onset of autumnal flooding. Swales experienced 3-wk flood events in late September and early October of each year, followed by water table fluctuation near the soil surface. Long-term flooded conditions ensued in early to mid November. Winter flooding was more intense in 1998 than 1997 though swales were continuously flooded both years. In 1997, but not 1998, floodwaters receded briefly in March. Spring draw-down began immediately before the spring (mid May) sampling date both years (Table 1). The wetter summer of 1997 resulted in water table fluctuation at or near the swale soil surface for nearly the entire season. In contrast, during summer 1998 the water table only rose to the surface for two brief (3–12 d) events.
Slopes were rarely saturated except during intense flooding during winter 1998. No substantial summer or fall inundation events were observed on slopes. The two lower slopes (SL-II, SL-III; Table 1) experienced two brief (<2 wk) flood events in winter 1997 and five to eight flood spikes (1–5 d duration) in winter and spring 1998. Though slope plots were rarely flooded the SR soil was frequently saturated, especially during winter 1998.
Percentage of time inundated was expected to influence both redox condition and soil pH. Though 100% seasonal inundation (%IN) prevented rust formation on test rods, data indicated an inconsistent relationship between %IN <90 and redox condition (Table 1). For example, although the SR layer of plot SL-I was saturated for >80% of autumn, sufficient O2 was present to form rust on test rods throughout the layer. Sub-root soil at plot SW-III, on the other hand, was only saturated for 25% of winter, yet rust did not form on test rods below the root zone. These contradictory findings may be explained by seasonal differences in metabolic activity of surface or subsurface soils, soil temperatures, or residual O2 from the long summer dry periods (Atlas and Bartha, 1993). Information related to water table levels does not appear sufficient to consistently predict redox status of soils except in cases of long-term continuous dry-downs or flooding. Direct evaluation of soil redox status in floodplain forests is advised when O2 availability is expected to be an important factor in data interpretation.
Soil saturation and associated redox potential decline was expected to increase soil pH (Ponnamperuma, 1984). Although soil pH did not decrease when anoxic soil dried in the spring, long-term winter inundation brought swale SR soil pH closer to neutrality (Ponnamperuma, 1984). Analysis of variance on pH measurements of Coosawhatchie floodplain soils indicated a season x depth interaction (P < 0.01) as well as a hydrology x depth interaction (P < 0.01). Across all sites, seasonal soil pH averages ranged from 4.60 to 4.92 in swales and from 3.96 to 4.65 on slopes (Table 1). In all cases, average swale soil pH values were higher than those of slopes. This trend was significant for both soil layers following spring and summer (P < 0.01), but only in the RZ layer after autumn and winter (P < 0.05).
Surface Organic Material Nutrient Pools
Surface organic material from the Coosawhatchie floodplain contained 6.2 and 21.5 g N kg-1. Previously published nutrient ratios indicated that the N content of partially decomposed surface organic matter on the Coosawhatchie floodplain was between the N contents of fresh litter and partially decomposed litter (Table 2). Average Ptot concentrations in surface organic material at the Coosawhatchie were 1030 mg kg-1 in swales and 681 mg kg-1 on slopes. Between 46% (winter) and 82% (autumn) of this Ptot was in organic forms. Low Ptot concentrations measured in partially decomposed surface organic material (Table 2) relative to fresh litter at the same site (Baker, 1998) may reflect rapid leaching of P following litterfall on the Coosawhatchie floodplain. Baker (1998) found that fresh Coosawhatchie litter contained 1300 to 1700 mg kg-1 Ptot, more than the 170 to 1000 mg kg-1 Ptot measured at similar forested areas (Table 2).
|
|
Phosphorus seasonal dynamics in surface organic material were not similar to either C or N dynamics (Fig. 1). The average Ptot concentration of surface organic material collected in autumn was slightly higher in swales than on slopes, possibly a result of accumulated P from previous years. A separate study on the Coosawhatchie floodplain during the same time period revealed that C/P ratios in freshly fallen leaf litter at swale and slope communities did not differ significantly (Baker, 1998), so the observed microtopographic difference in Ptot concentration of surface organic material collected in autumn was not explained by different P concentrations in fresh litterfall. Since autumn Porg concentrations were not significantly different (P > 0.1) between microtopographic classes (755 mg kg-1 in swales and 597 mg kg-1 on slopes; Fig. 1), the observed difference was the result of lower levels of both Porg and Pmin on slopes. With the onset of winter flooding, Porg concentrations in surface organic material declined precipitously for both microtopographic classes (Fig. 1). This decline was unexpected since other studies have found that, as with N, decomposition of surface organic material frequently results in seasonal accumulation of P relative to C (Baker, 1998; Lockaby et al., 1996; Lockaby and Walbridge, 1998).
Multiple wetting and drying cycles on slopes in late fall and winter may have resulted in physical transport of P-rich fibric material from elevated slopes into puddled swales. This transport could have caused the decline in slope Porg concentration from 640 mg kg-1 in autumn to the winter low of 180 mg kg-1. Organic P has been reported to be associated with fine soil particles that accumulate in depositional floodplains (Pinay et al., 1992). The employed sampling method, which collected intact solid material but did not capture light debris that suspended in floodwater, precluded confirmation of this suspended P flux into swales (Porg concentrations in swales were observed to fall from 790 to 310 mg kg-1; Fig. 1). With subsidence of floodwaters, however, P-enriched material may have settled back onto the soil surface and increased swale Porg to 760 mg kg-1 organic material. At the same time, increased temperatures and aerobic conditions accelerated decomposition in spring and summer; the result was the expected enrichment in both Porg and N (nutrient ratios decreased; Table 3). Under this scenario, the Coosawhatchie floodplain would be expected to function as a source of Porg to surface water during periods of overbank flooding. This hypothesized mechanism for P (but not N) redistribution on the floodplain is supported by the research of Vought et al. (1994). Their work in Swedish streams indicated that P was much more likely to be transported with fine particulate matter (>50% of stream Ptot load) than was N (20% of stream N load).
|
Mineral Soil Nutrient Pools
Total Carbon
Soil C content of slopes was greater in the RZ than the SR layer, but for swales the C content was greater in SR than RZ soil (Table 4). Root zone soil acquires C by surface organic matter burial and soil faunal activity (Maxwell and Coleman, 1995), leaching of dissolved and particulate compounds from surface organic material (Dosskey and Bertsch, 1994; Grant et al., 1996; Marcus et al., 1998), and mortality of and exudation from plant roots (Cheng et al., 1994; Jones et al., 1996; Martens, 1990). Factors that contributed to microtopographic differences in RZ organic C content may include greater biomass and activity of plant roots on slopes, greater infiltration of fine particulate organic material into the sandier soil of slopes, and greater profile mixing activity by soil fauna during longer nonflooded periods. Carbon inputs to SR soil were probably limited to leaching of dissolved organic material from RZ soil (Marcus et al., 1998) and lateral subsurface flow (Dosskey and Bertsch, 1994). Subsurface (SR) microtopographic differences in organic C content may have been a result of protection of organic material by accumulated swale clay minerals (Torn et al., 1997), slower degradation of organic material under anaerobic conditions commonly found in swale SR soils (DeLaune et al., 1981; Schipper et al., 1994), or more favorable conditions for polymerization into recalcitrant humic material in swales (Atlas and Bartha, 1993).
|
280-g (dry weight) soil sample; it seemed unlikely that such a large quantity of C washed out in flood water during collection of the sample, as was earlier hypothesized to explain winter P losses from surface organic material. Extended saturation and water table fluctuations may have resulted in leaching of labile C from the RZ layer of swales. This labile C pool could be partially replenished by root production and exudates in spring and fully replenished by the end of summer. Slopes were not subjected to long-term saturation so their labile C was not mobilized. This hypothesis is supported by the work of Dosskey and Bertsch (1994), who reported that during flood events, soils of the Fourmile Branch floodplain contributed dissolved organic C to the stream representative of up to 10% of aboveground net primary productivity.
Microbial Biomass Carbon
Microbial biomass C contents in mineral soils of the Coosawhatchie floodplain ranged from 0.65 to 1.4 Mg ha-1 (Table 4); these values corresponded to concentrations of 0.56 to 4.7 g C kg-1 of soil. Groffman et al. (1996) reported soil microbial biomass C concentrations of 1.5 to 4.0 g kg-1 in red maple swamps and 2.0 to 3.5 g kg-1 in woodland pools. Díaz-Raviña et al. (1993) found that for organic soils in Galicia, Spain, microbial biomass C ranged from 0.6 to 1.5 g kg-1, and there was a strong correlation between soil C content and microbial biomass C concentration (r = 0.88); microbial biomass C trends were similar to seasonal and spatial total C trends on the Coosawhatchie floodplain (Table 4). Where differences with depth were observed, the SR layers had greater microbial biomass C contents in swales, while RZ microbial biomass contents were greater on slopes. Seasonal effects were only observed in the RZ layer of swales, where total C and microbial biomass C fluctuated in tandem (r = 0.490, P < 0.001). In the other three soils (SR of swales and both depths of slopes) the correlation was lower (average r = 0.308, P < 0.05).
Total Nitrogen
Average soil nitrogen contents on the Coosawhatchie floodplain ranged from 1.5 to 4.4 Mg ha-1 (Table 4); these contents corresponded to concentrations of 0.98 to 10.5 g N kg-1. Carbon/N ratios ranged from 13.2 to 20.0, with higher values observed in soils from slope plots (Fig. 2)
. Coosawhatchie floodplain soil may be somewhat N-poor compared with other types of mid-Atlantic wetland areas. The Coosawhatchie River floodplain had lower N concentrations than a Michigan fen, a Wisconsin marsh, a Maryland bog and forested swamp, and even a North Carolina pocosin (14–22 g kg-1), but similar values to a North Carolina swamp (5.7 g kg-1; Faulkner and Richardson, 1989). Soil N contents were also lower than those found on other forested floodplains (Faulkner and Richardson, 1989; Nelson et al., 1987). Microtopographic and seasonal trends for total soil N were similar to those for C (Table 4 and Fig. 2). Slope SR layers were lower in N relative to swales during autumn, spring, and summer; N content in the RZ was not different between microtopographic classes (Table 4).
|
Organic P exhibited similar trends to Ptot (Table 4). On average, Porg accounted for 77% of Ptot; treatment averages were consistently between 72 and 82%. No hydrologic, seasonal, or depth differences in the percentage of Ptot present as Porg were detected (data not shown). Average treatment Pmin was always between 43 and 150 kg ha-1 (data not shown). No hydrologic, seasonal, or depth differences were detected in the pool of Pmin except that, in spring, the swale SR pool of Pmin (128 kg ha-1) was greater than the RZ pool (75 kg ha-1). For these reasons, treatment effects discussed above for Ptot were assumed to have been caused by fluctuation in the dominant Porg pool.
Soil C/P and N/P ratios are presented in Fig. 2 to build a concept of soil nutrient fluctuations on this floodplain type. Swale RZ soils had narrower C/Ptot ratios (relatively more P) than did slope RZ soils. The C/Porg ratio was more variable, and, although trends were the same as those for the C/Ptot ratio, the only seasonal or depth difference was in slope SR soils where the ratio was narrower in autumn than in spring or summer. Both layers showed equivalent N/P ratios in swale soils, but slope RZ soil N/P ratios were higher than SR soils in summer and autumn. Microtopographic effects were especially pronounced in the SR soils, although the effects were not significant since RZ data were highly variable. It seems reasonable to hypothesize that the swale RZ layer had narrower N/P ratios as a result of surface enrichment of P with floodwater recession. Slope RZ soil experienced annual export of surface P and, therefore, had wider N/P ratios.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Soil chemical and biological processes varied with microtopography and the concomitant differences in soil structure and hydroperiod. Swale soils were more frequently flooded and had finer textures. These characteristics resulted in longer and more frequent periods of inundation and pH values closer to neutral. The high frequency of flooding at swales resulted in greater N and P enrichment (narrower C/N and C/P ratios) and led to C, N, and P accumulation in swale SR soil. Seasonal fluctuations in soil C contents indicated that forested floodplain swale SR layers may serve as C sinks, while seasonal losses may reflect C-source status from RZ soil to the stream.
Nutrient pool information was presented as an addition to the body of knowledge covering this wetland type. Too little information is available in the current literature to fully interpret the functional significance of soil nutrient ratios and patterns of nutrient pool fluctuation on the Coosawhatchie floodplain. Nutrient concentration dynamics of surface organic material suggested seasonal accumulation of N and especially P into a fraction of light, partially decomposed organic material at the mineral soil surface. The result was hypothesized to be mass transfer of the Porg present in this fraction from slopes to swale surfaces and infiltration into slope RZ soil during wetting and drying cycles in winter. Preferential transfer of Porg but not C and N with this light organic material would explain the accumulation of P into slope RZ soil relative to C and N. With spring dry-down, however, transported P-rich organic material appeared to add to decomposition-based P enrichment and resulted in much higher P concentrations (and lower P-based ratios) in swales than on slopes. Some of the transported material was incorporated into the RZ soil; the result was P accumulation in swale soil relative to slope soil. In both mineral soil and surface organic material, P fluctuated independently of N and C.
|
ACKNOWLEDGMENTS |
|---|
Received for publication March 13, 2000.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  B. G. Lockaby, R. Governo, E. Schilling, G. Cavalcanti, and C. Hartsfield Effects of Sedimentation on Soil Nutrient Dynamics in Riparian Forests J. Environ. Qual., January 1, 2005; 34(1): 390 - 396. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Crop Science | |||
| Vadose Zone Journal | Journal of Plant Registrations | ||||
| Journal of Natural Resources and Life Sciences Education |
Journal of Environmental Quality |
||||
B), and average seasonal redox-associated characteristics [pH, percentage of season inundated (%IN), and maximum depth to sustained anaerobic conditions (O2)] for two depths in experimental swale (SW) and slope (SL) plots at three sites near Ridgeland, SC, December 1996 to August 1998
