| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Soil and Water Science Dep., 106 Newell Hall, P.O. Box 110510, Univ. of Florida, Gainesville, FL 32611-0510
b Everglades Dep., Watershed Research and Planning Division, South Florida Water Management District, West Palm Beach, FL 33416-4680
* Corresponding author (wdebusk{at}ufl.edu)
Received for publication June 2, 2000.
 |
ABSTRACT |
|---|
 TOPNOTES ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Abbreviations: TP, total phosphorus • WCA, Water Conservation Area
|
INTRODUCTION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
The main focal point of ecological impact has been Water Conservation Area 2A (WCA-2A), where extensive documentation exists for P-related impacts on ecosystem structure and function (Davis, 1991; Koch and Reddy, 1992). Among the more noticeable ecological changes in WCA-2A have been the transformation of native sawgrass (Cladium jamaicense Crantz) marsh and openwater sloughs to dense stands of cattail (Typha domingensis Pers. and T. latifolia L.) (Davis, 1991; South Florida Water Management District, 1992; Jensen et al., 1995) and replacement of endemic periphyton communities by algal species typically associated with more eutrophic waters (McCormick and O'Dell, 1996; McCormick et al., 1996). Soil P concentration in WCA-2A has been strongly linked to productivity and community structure of macrophytes (Newman et al., 1998; Miao and Sklar, 1998; Miao and DeBusk, 1999; Richardson et al., 1999; Vaithiyanathan and Richardson, 1999) and periphyton (McCormick and Stevenson, 1998; McCormick et al., 1998), organic carbon turnover (DeBusk and Reddy, 1998), nitrogen cycling (White and Reddy, 1999, 2000), microbial community structure (Drake et al., 1996), and diatom assemblages (Cooper et al., 1999).
Due to the absence of significant gaseous losses of P via metabolic pathways (in contrast to nitrogen), preferential retention of P has occurred in WCA-2A soils, manifested as increases in both P storage (P mass per unit area of land) and enrichment (P concentration in soil, i.e., mg P kg-1 soil). Increased biological production, stimulated by loading of P from external sources, has resulted in accelerated rates of peat accretion, and consequently P storage, in impacted areas near surface water inflows (Reddy et al., 1993; Craft and Richardson, 1993, 1998). Perhaps more significant with respect to biological productivity is the fact that increased sorption and bioconcentration of P in the soil has resulted in widespread P enrichment of the soil (DeBusk et al., 1994). Soil characterization studies of the P gradient in WCA-2A have shown a roughly proportional increase in concentrations of both inorganic (Ca- and Mg-bound) and organic (associated with humic and fulvic acids) P near the major surface water inflow points (Qualls and Richardson, 1995; Reddy et al., 1998).
The overall objective of our study was to evaluate recent (1990 to 1998) changes in the spatial extent and patterns of soil P enrichment (total P concentration) in Everglades WCA-2A. Of particular interest was the progression of soil P enrichment in the northeast quadrant into the relatively unimpacted sawgrass marsh of the interior region. The study was accomplished by conducting a comprehensive sampling and analysis of surface soil across WCA-2A, performed in a similar manner to an earlier soil characterization study (DeBusk et al., 1994). Comparison of results of the initial and follow-up studies formed the basis of our analysis of recent spatio–temporal patterns of soil P enrichment.
|
MATERIALS AND METHODS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
We have previously documented a distinct pattern of widespread P enrichment of soil in WCA-2A, associated with long-term nutrient loading from nearby agricultural areas (DeBusk et al., 1994). Progressive replacement of sawgrass marsh by cattails in the more highly P-enriched areas has also been documented (Jensen et al., 1995; Rutchey and Vilchek, 1999) (Fig. 1).
Soil Sampling and Analysis
Surface peat and overlying sediment were sampled at 62 locations within WCA-2A, during October 1998 (Fig. 2). Selection of the sampling sites was based on the sampling grid used for soil P characterization in WCA-2A in 1990 (DeBusk et al., 1994). Some modification of the original sampling grid was made for the 1998 sampling event to provide higher sampling density for areas previously characterized by high spatial variability of soil P concentration. Typically, a relatively evenly distributed grid of sampling points is the most appropriate sampling design, and provides the most efficient use of the data, where there is a high degree of spatial covariance for that variable (Isaaks and Srivastava, 1989). Since spatially distributed sampling of soil chemical and physical properties typically does not yield independent outcomes, a random sampling design is less appropriate and, moreover, is often inefficient due to clustering of sampling points.
|
Soil samples extruded from the coring tube were placed in water-tight plastic bags and transported on ice to the laboratory. Samples were analyzed for total phosphorus (TP) using a dry-ashing procedure (Andersen, 1976) followed by determination of dissolved reactive P by an automated ascorbic acid method (Method 365.4; USEPA, 1983). Overall, the sampling and analytical procedures for the 1998 sampling event were the same as those used for the 1990 WCA-2A study (DeBusk et al., 1994).
Location (latitude and longitude) of each of the sampling sites was determined by global positioning system (GPS), using post-processing data correction techniques to achieve an accuracy of 1 to 2 m from the true position.
Geostatistical Analysis
Spatial patterns of soil TP concentration for both 1998 and 1990 data sets were determined using geostatistical analysis. Variograms were calculated for each data set to establish the degree of spatial continuity of soil total P among data points (autocorrelation) and to estimate the range (i.e., distance) of spatial dependence for the total P variable. Information generated through the variogram is used to calculate sample weighting factors for spatial interpolation by a kriging procedure (Isaaks and Srivastava, 1989). In contrast to nongeostatistical interpolation methods, kriging accounts for the spatial arrangement (clustering) of sample points as well as the distance of each point from other sample points. An ordinary block kriging procedure (Goovaerts, 1997) was used to create a continuous grid of interpolated data points, or blocks, from the original data points. The block size used for this procedure was 300 x 300 m. Both the variogram and kriging procedures were performed using the GS+ software package (Gamma Design Software, 2000).
We also used a variation of ordinary kriging, known as indicator kriging (Goovaerts, 1997), to calculate the conditional probability of the occurrence of soil P concentrations in excess of arbitrarily specified thresholds, or cutoff values, that were selected to facilitate comparison over time. The procedure involves substitution of variables that indicate either occurrence (1) or non-occurrence (0) of values greater than the threshold, in place of the actual concentration values. The resulting sets of indicator variables for each threshold value underwent ordinary kriging analysis as described above. The resulting interpolated values for the indicator variables represented the probability, expressed as a value between 0 and 1, that the soil total P concentration at specific locations exceeded the threshold value.
Interpolated grid data generated by geostatistical methods were exported to the ArcView GIS program (Environmental Systems Research Institute, 1999), then converted to grid themes with cell size corresponding to the kriged output (300 m).
|
RESULTS AND DISCUSSION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Site-by-site comparisons of 1990 vs. 1998 soil TP concentrations were not performed due to the likelihood that the 1998 site locations did not coincide precisely with the corresponding 1990 sites. The significance of even a relatively small error in relocating the 1990 sampling sites in 1998 was demonstrated by an analysis of local site variability performed in highly impacted and minimally impacted areas of WCA-2A (unpublished data, 1998). Our analysis revealed that the coefficient of variation for soil total P concentration in triplicate samples increased from 9.6% at 0.5-m spacing to 21.6% at 25-m spacing in the impacted area, and from 7.9 to 12.9% at 0.5- and 25-m spacing in the unimpacted area. Based on the sampling scheme used during 1990 sampling of WCA-2A, and repeated for the 1998 characterization study, the most appropriate procedure was to use a spatial interpolation procedure to create high-resolution grids of soil P concentration for both data sets, which could then be directly compared.
Spatio–Temporal Trends in Soil Total Phosphorus
Sample variograms constructed from both the current (1998) and 1990 data sets revealed comparable patterns of spatial continuity for soil total P concentration (Fig. 3). A spherical model provided the best fit for the semivariance function, as determined by both the software's least squares fitting routine and visual comparison of the model output with data points. Model estimates of the range of spatial continuity for soil total P were 11.3 km for the 1998 data set, compared with 9.3 km for the 1990 data. The most conspicuous difference between the 1990 and 1998 variograms was the overall magnitude of the semivariance, which was approximately twice as great for the 1998 data set than for the 1990 data. The maximum semivariance, indicated by the sill of the variogram, is roughly equivalent to the total sample variance. Increased sample variance in the 1998 soil total P data would be consistent with the observed increase in the range of total P concentration in 1998 vs. 1990.
|
|
Mapping based on the interpolated grids developed from 1990 and 1998 soil total P data clearly showed an increase in the areal extent of soil P enrichment during the period 1990–1998 and, most notably, confirm the expansion of the P-enriched area south of the S-10 inflows (Fig. 5). The 1998 map also indicates that soil TP concentrations in a substantial portion of the interior sawgrass marsh have increased since the 1990 sampling event. From our 1990 soil characterization data for WCA-2A we estimated that the historical background TP concentration (i.e., the concentration not impacted by anthropogenic P loading) for surface soil (0–10 cm depth) in WCA-2A was approximately 500 mg kg-1 (DeBusk et al., 1994). Under this assumption, using our spatial interpolation procedure, 48% of the total area of WCA-2A would have been considered P-enriched in 1990, compared with 73% of the total area by late 1998 (Fig. 4). In a similar fashion, changes in the spatial extent of moderately to highly enriched areas during the 1990–1998 period can be calculated. For example, the area for which soil P concentration exceeded 1000 mg kg-1 increased from 10.5 to 18.0% of the total area of WCA-2A, an increase of 393 ha yr-1 (assuming a linear increase over time). Moreover, the proportion of land area with a soil total P concentration greater than 1500 mg kg-1 (i.e., highly P-enriched), increased from 2.6 to 7.3% of the total area, representing an annual increase of 246 ha yr-1.
|
Perhaps more striking is the increase in the total area considered as P-enriched, that is, the area for which soil P concentration is greater than 500 mg kg-1, which has been suggested as the upper limit of historical background soil P concentrations (McCormick et al., 1999). Using those criteria, about 73% of the total area (31777 ha) would be considered P-enriched in 1998, compared with 48% of the land area (20829 ha) in 1990. Thus, the rate of increase of the P-enriched area, or alternatively, the rate of decrease of the unimpacted area during the period 1990–1998, was estimated to be 1327 ha yr-1, or 5.25% yr-1.
Indicator Kriging
The interpolated grids created by the ordinary kriging procedure represent an optimized estimate of soil TP concentration for points where no data (measurements) are available. In theory, if not in practice, ordinary kriging (i) provides an unbiased estimator, because it drives the mean residual toward zero and (ii) minimizes the overall variance of the estimation error (Isaaks and Srivastava, 1989). However, there is no straightforward method to provide a measure of the uncertainty in individual estimates. One measure of uncertainty associated with kriged values is the error variance (Isaaks and Srivastava, 1989). The value of this parameter is dependent on the covariance model used in the kriging procedure and the spatial configuration of the data points. However, the kriging error variance does not depend on the actual data values, a major shortcoming for determining the local spread of estimation errors.
A more useful and statistically valid procedure for conveying the uncertainty associated with the interpolated values of soil total P concentration is the use of indicator kriging (Goovaerts, 1997). This nonparametric kriging technique can be used to approximate a conditional cumulative probability distribution function (ccdf) for the parameter under study, by calculating for each grid cell the probability of exceeding each of a number of threshold values (e.g., deciles). Rather than construct a detailed ccdf, we used a series of five evenly distributed threshold values for soil TP concentration to visually evaluate changes in the distribution of soil total P. The threshold values selected were 500, 750, 1000, 1250, and 1500 mg P kg-1.
The series of plots presented in Fig. 6 provides a visual evaluation of changes in the spatial extent and patterns of soil P enrichment at several levels (thresholds) of total P concentration. The probabilistic approach used here allows the observer to assign any level of uncertainty to the interpolated values in the kriged grids, and thus may be considered to be analogous to risk assessment. The use of indicator kriging provides clear evidence of the differences in soil TP enrichment between 1990 and 1998. For example, these data highlight the considerable expansion of low-level P enrichment (>500 mg kg-1) throughout most of WCA-2A. Also noteworthy is an apparent shift of the more highly P-impacted area (i.e., TP > 1500 mg kg-1) toward the south. As previously discussed, this could be attributed to a breakthrough of P from saturated soils close to the S-10 structures. Alternatively, this may be a manifestation of a substantial diversion of flow from S-10C toward the east and south along the interior canal, then back toward the west along numerous well-used airboat trails. Field observations in the vicinity of S-10C suggest that the accelerated peat accretion and increased density of herbaceous and woody vegetation in the highly P-enriched area along the interior canal has resulted in increased hydraulic resistance along the southerly flow path into the WCA-2A marsh.
|
Recent efforts have been made to estimate the long-term, or sustainable, P assimilation capacity of WCA-2A using empirical models based on historical P loading and removal data (Walker, 1995; Richardson et al., 1997; Richardson and Qian, 1999; Kadlec, 1999). Short-term rates of P removal in newly P-impacted wetlands, including treatment wetlands in the startup phase, are governed principally by sorption reactions (adsorption and precipitation) within the soil and vegetation and by plant uptake. Long-term removal of P in WCA-2A, and in wetlands as a rule, is a function of accretion of recalcitrant organic or inorganic P compounds, for example peat formation.
Both Richardson and Qian (1999) and Kadlec (1999) used a similar approach, but contrasting assumptions regarding model parameterization, to arrive at an approximate location, in terms of distance from the S-10 inflow, of the boundary between the region of nonsustainable ("impacted" area) and sustainable P assimilation ("unimpacted" area). The former study yielded an estimated distance of 5.1 km from S-10C and the latter generated a value of approximately 16 km. Both studies are based on the assumption, drawn from historical surface water quality data, that a steady-state condition exists between P loading and assimilation in WCA-2A (i.e., that the P enrichment front is stationary). A recent study suggests that the surface water P front may not be stationary and marsh TP concentrations vary in response to inflow TP concentrations, water depth, and flow (Smith and McCormick, 2001). While surface water TP concentrations appear to have declined in the 1990s (Smith and McCormick, 2001), our results indicate that the soil P enrichment front has been dynamic during recent years. These results may reflect a certain degree of downstream redistribution of soil P, for example by way of upstream P desorption and/or mineralization and downstream assimilation. Regardless of the mechanism, our results are indicative of a progressive P impact, in the form of soil P enrichment, to the south of the S-10A and S-10C inflows.
|
CONCLUSIONS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Geostatistical analysis and comparison of current spatial patterns of soil P enrichment with spatial patterns circa 1990 have indicated that the extent of soil P enrichment increased over an 8.25-yr period from 1990 through 1998. The change in spatial distribution of soil P enrichment between 1990 and 1998 suggest that the soil P concentration gradient south of S-10C has become more diffuse as the overall extent of soil P enrichment has increased.
Soil P enrichment above historical background concentrations can be considered an indicator of ecological change. Dynamic patterns of soil P enrichment have been closely aligned with documented changes in vegetation patterns, most notably the change from native sawgrass marsh to cattail marsh. Results of this soil characterization study, showing a substantial increase in the extent of moderate soil P-enrichment, is indicative of the magnitude of recent increases in the ecologically impacted area of WCA-2A.
|
ACKNOWLEDGMENTS |
|---|
|
NOTES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
REFERENCES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  S. Grunwald and K. R. Reddy Spatial Behavior of Phosphorus and Nitrogen in a Subtropical Wetland Soil Sci. Soc. Am. J., June 18, 2008; 72(4): 1174 - 1183. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. L. Bruland, S. Grunwald, T. Z. Osborne, K. R. Reddy, and S. Newman Spatial Distribution of Soil Properties in Water Conservation Area 3 of the Everglades Soil Sci. Soc. Am. J., August 3, 2006; 70(5): 1662 - 1676. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Corstanje, S. Grunwald, K. R. Reddy, T. Z. Osborne, and S. Newman Assessment of the spatial distribution of soil properties in a northern everglades marsh. J. Environ. Qual., May 1, 2006; 35(3): 938 - 949. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Grunwald, R. Corstanje, B. E. Weinrich, and K. R. Reddy Spatial Patterns of Labile Forms of Phosphorus in a Subtropical Wetland J. Environ. Qual., January 5, 2006; 35(1): 378 - 389. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. J. Anderson, W. J. Mitsch, and R. W. Nairn Temporal and Spatial Development of Surface Soil Conditions at Two Created Riverine Marshes J. Environ. Qual., October 12, 2005; 34(6): 2072 - 2081. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. F. DeBusk and K. R. Reddy Nutrient and Hydrology Effects on Soil Respiration in a Northern Everglades Marsh J. Environ. Qual., March 1, 2003; 32(2): 702 - 710. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. L. Childers, R. F. Doren, R. Jones, G. B. Noe, M. Rugge, and L. J. Scinto Decadal Change in Vegetation and Soil Phosphorus Pattern across the Everglades Landscape J. Environ. Qual., January 1, 2003; 32(1): 344 - 362. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Castro, K. R. Reddy, and A. Ogram Composition and Function of Sulfate-Reducing Prokaryotes in Eutrophic and Pristine Areas of the Florida Everglades Appl. Envir. Microbiol., December 1, 2002; 68(12): 6129 - 6137. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Crop Science | |||
| Journal of Natural Resources and Life Sciences Education |
Vadose Zone Journal | ||||
| Soil Science Society of America Journal | Journal of Plant Registrations | The Plant Genome | |||
