| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
School of Natural Resources, The Ohio State University, Ohio Agricultural Research and Development Center, 1680 Madison Ave., Wooster, OH 44691-4096
* Corresponding author (calhoun.2{at}osu.edu)
Received for publication August 29, 1999.
 |
ABSTRACT |
|---|
 TOPNOTES ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
40 mg kg-1. Log-transformed means for Pcu were significantly higher than for Por. The principal determining factors for Por were physiography, soil texture, and soil series. Current P is affected by present tillage practice and drainage class. Change in soluble P in the soil is not as responsive to fertilizer P sales as is SRP in river water. This suggests that as fertilizer P sales decline, a declining percentage of P added as fertilizer is annually dissolved and transported into the drainage system. Soluble P in soil is governed by a combination of fertilizer and tillage management, soil properties, and landscape factors interacting over time.
Abbreviations: OSU, Ohio State University • Pcu, current concentration of Bray-1 phosphorus in the Ap horizon • Por, original concentration of Bray-1 phosphorus in the Ap horizon • SOC, soil organic carbon • SRP, soluble reactive phosphorus in river water
|
INTRODUCTION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Several long-term studies on research plots have demonstrated a parallel relationship between levels of soluble or soil-test P and fertilizer P rates (Webb et al., 1992); however, the soils examined in such studies rarely reflected the range and complexity of soils and landscapes within the total watershed. Furthermore, plot studies have seldom represented the complete range of management practices in a watershed. Fields and small agronomic watersheds have also been studied, with an emphasis on the effect of management on spatial relationships of soluble P (Gburek and Sharpley, 1998), but long-term assessments of soil test P have not been conducted at this scale.
We were able to identify only one study that assessed P levels in the Maumee and Sandusky River watersheds at a given point in time. Logan (1989) analyzed 129 surface soil samples for total P, NaOH-extractable P, and Bray-1 P. Data were reported by soil series and encompassed soil samples collected throughout the Lake Erie basin of northern Ohio. Results indicated that total P, more so than soil test P, was highly correlated with clay content. Mean total P content by soil series ranged from 360 to 930 mg kg-1 with highest values for fine-textured soils. Mean concentration of Bray-1 P ranged from 13 to 50 mg kg-1. Soil sample site locations were not provided; consequently, date of original sampling and cultivation conditions at the time could not be determined.
Another source of chronicled P data is the soil test laboratory files at Ohio State University (OSU). Ohio data can be grouped by year and zip code and should facilitate large watershed summary. Unfortunately, the zip code only identifies the location from which the soil sample was mailed to the laboratory and might not be the zip code in which the field was located. Many farmer field soil samples were consolidated by a fertilizer distributor and sent to the laboratory from a remote location. A second problem with OSU laboratory soil test data is that the number of soil samples submitted annually declined as farm operators increasingly turned to private laboratories for this service. This led to the termination, in 1998, of the OSU soil-testing program. Private labs in the region have only recently begun computer storage of soil test data (R. Warden, personal communication, 1999).
Ideally, benchmark sites for evaluating long-term trends in soil quality encompassing principal soil landscapes and agricultural management would have been established and regularly monitored over the past 25 years. Unfortunately, concerns about soil quality were not a high priority in the 1960s and 1970s. The soil characterization program in support of the progressive soil survey for Ohio (and elsewhere in the USA) offers a viable alternative to assess long-term trends in soil test P. This database, in addition to the archived soil samples, represents an untapped resource for potential use in evaluating long-term soil quality trends.
The objectives of this study were to (i) summarize historical soil test P data for northwest Ohio and (ii) return to the original soil survey sites and resample the Ap horizons for comparison of Bray-1 P levels with those of the archived soil samples. We propose that the soil is the intermediary between fertilizer and manure P and SRP measured in the river water. Furthermore, our hypothesis is that soluble P levels in the soil will reflect changes in fertilizer P management and changes in cultivation practices over time as affected by intrinsic soil and landscape factors. We will show that changes in soil test P levels are related to physiography, drainage class, soil texture, and tillage management. Finally, this study represents an effort to relate soil test P, fertilizer P sales, and soluble reactive P levels in river water over time.
|
MATERIALS AND METHODS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Ohio State University Soil Test Database
Electronically stored soil test data were available for 1977 through 1996. The database was queried for Bray-1 P and number of soil samples submitted by year and zip code. Zip codes were then matched with counties. Counties not within the study area were discarded. The data were then summarized by county and by year. A single yearly average for Bray-1 P for northwestern Ohio was then calculated.
Ohio Soil Characterization Database
The archived soil sample set precedes the accelerated adoption of reduced fertilizer P applications in the early 1980s and provides an incremental 30 yr baseline up to that date. The soil characterization program was initiated in northwestern Ohio in the early 1950s and continued until 1982. Some pedon (roughly equivalent to "soil profile") sampling has been reinitiated as soil surveys are modernized. Eighty-two percent of the pedons were collected prior to 1970. Following routine laboratory characterization, a subsample from each horizon was archived. Original pedon descriptions included a detailed description of the site location.
Over 600 pedons were described and soil samples were collected in the Maumee and Sandusky basins of northwestern Ohio as part of the cooperative soil survey program. Roughly 5% of the original soil sample sites have been lost due to nonagricultural development or inadequate descriptions of location. Approximately 65% of the original soil samples have been preserved. For this study, 319 pedon locations were identified that had both an archived soil sample and a viable resample site (Fig. 1) . Routine laboratory characterization initially included particle size distribution analysis (PSDA) using pipette method 3A1 (USDA Natural Resources Convervation Service, 1996), soil organic carbon (SOC) using wet combustion until 1965 and dry combustion afterward, pH in water (1:1), and calcium carbonate equivalent (CCE) on soil samples where pH was greater than 7.2. Later, extractable bases, cation exchange capacity (CEC), and other refinements were added to the routine analyses but extractable P was never included. For the purpose of this paper, only the PSDA and SOC data are presented.
|
The current tillage system was estimated based on crop residues remaining on the soil surface. If the surface was at least 30% covered with crop residues then the site was classified as conservation tillage. If three different crop residues (normally wheat [Triticum aestivum L.), corn [Zea mays L.], and soybean [Glycine max (L.) Merr.]) could be identified then the system was assumed to be no-till. If the site was an abandoned field, pasture, or Conservation Reserve Program it was classed as meadow. A site was classified as conventional tillage when it was obvious that it had been moldboard-plowed in the autumn or if fresh, incorporated residue was noted while sampling the Ap.
Vegetation and land use was reported in the original pedon description using various terms including plowed, cultivated, meadow, pasture, grasses, and weeds. For analysis purposes, these designations were reduced to two categories: cultivated and meadow.
Laboratory Methods
Soil samples were dried at 40°C, ground, and passed through a 2-mm sieve. Material greater than 2 mm in size was discarded. The index of soluble P used in this study was the Bray and Kurtz P-1 test (soil to solution ratio of 1:10) determined from both the current (Pcu) and archived (Por) soil samples as described by the NCR-13 Committee (1998). Determination of soluble reactive P in river water is described in Richards and Baker (2002). Particle size data on the archived soil samples were obtained by pipette method 3A1 (USDA Natural Resources Convervation Service, 1996). Soil texture was assumed not to have changed significantly due to the dominance of level to nearly level slope classes in both watersheds. Data for Bray-1 P are reported in mg kg-1 soil.
Limitations of the Sample Population and Assumptions
Use of historical soil survey sites for sample collection does not assure an unbiased representation of the population of soil test P concentrations that exist in Ap horizons of northwestern Ohio. The rationale for selection of sites for investigation at the time depended on the classification and field correlation needs of the soil survey. In some cases the willingness of the farm operator to allow a pit to be excavated for this purpose was related to the operator's level of innovation in production agriculture and cooperation with soil and water conservation efforts in the county. It is possible that many of the soil survey pits were excavated on better-managed farms. It is also conceivable that this level of management did not continue from initial soil sample collection through present time.
Bray-1 P levels in surface soils are conspicuous for short-range spatial variability in cultivated fields. For this reason, soil samples submitted for testing are normally composited on a field-level basis. The original pedon soil sample was taken from an Ap or A horizon as representative of that horizon over a lateral distance of approximately 1 by 1 m with a total soil sample volume of 2 L. It was agreed that traditional soil test sampling of the site should not be used and that the sampling procedure used during the soil survey should be followed. The original soil sample at the site was intended to be representative of the soil series as it occurred in that county but not of the field in which it was sampled. Our assumption is that the modern soil sample is representative of the specific location at which the original soil survey sample was taken.
Through the use of Mahalanobis analysis to identify outliers for Pcu and Por (SAS Institute, 1996) and subsequent reevaluation of field location descriptions, 17 sites were discarded due to position near turn rows or suspected fertilizer spills. This reduced the sample size of the population to 302.
We assume that the samples are representative of the population of soil series found in northwestern Ohio (Richards et al., 2002a). In 1962, as part of a systematic evaluation of surface horizon and parent material texture, 45 sites of the Hoytville series (fine, illitic, mesic Mollic Epiaqualfs) were sampled in Henry County. Inclusion of all of these sites heavily biased the sample (26% of the total) toward glacial till–derived, very poorly drained soils collected during a short period of time. These soil samples were collected in transects along section-line roads in the lake plain area of the county and included only a topsoil sample and a subsoil sample. For these Hoytville sites the minimum, mean, and maximum values, respectively, were as follows: Pcu (12, 34, 88); Por (8, 23, 68); and clay (30.9, 39.8, 48.6). We consolidated these sites into one averaged observation. The predominant current tillage was conservation as earlier defined in this paper. This reduced the sample to 259 observations, of which Hoytville now represented 16% of the total. This also brought the remaining soil series representation in the sample proportionately closer to that of the extent of soil series in the two watersheds. Predominant soil series in the two watersheds are Blount (fine, illitic, mesic Aeric Epiaqualfs; 19%), Hoytville (16%), Pewamo (fine, mixed, active, mesic Typic Argiaquolls; 10%), Paulding (very-fine, illitic, nonacid, mesic Typic Epiaquepts; 4%), and Glynwood (fine, illitic, mesic Aquic Hapludalfs; 4%). Approximately 125 series constitute the remaining 47% of the area. The sample population consists of the following proportions of soil series: Hoytville, 38 sites (16%); Blount, 30 sites (12%); Pewamo, 15 sites (6%); Paulding, 12 sites (5%); and Glynwood, 7 sites (3%). The remaining 157 sites represent 65 soil series.
Distributions of Pcu and Por are exponential and the data were consequently log10–transformed. The exponential nature is explained by the effect of accelerated and nonuniform P fertilization over time. The effect is illustrated in Table 1, where the mean is greater than the median for Pcu and Por by 29 and 18%, respectively. Generally, as the mean approaches the median, sample distribution approaches normality. Log-transformation resulted in a distribution much closer to normality (Fig. 2) . All tests for statistical significance were done on log-transformed data and the means were then back-transformed. Nontransformed basic statistics for the overall sample population are shown in Table 1 in addition to surface soil samples collected from 22 forested sites that are provided for comparison with the cultivated sites. All references to statistical means are for those back-transformed from log10 transformation unless otherwise noted.
|
|
|
|
RESULTS AND DISCUSSION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
The untransformed mean Pcu for all sites is significantly higher than it was at the time of original sampling between 1953–1982 (Table 1). Thirty percent of the Pcu values and 17% of the Por are 40 mg kg-1. This is a weighted average concentration for the complete Ap horizon. Median values for Pcu and Por are 8 and 4 mg kg-1 less than the mean, reflecting the exponential distribution of the untransformed data. The present mean thickness of the Ap is 26% greater than it was at the time of the collection of the original soil survey sample. At the same time, Pcu is 38% higher than Por and the mean concentration of P in the included soil below the original Ap is 14 mg kg-1. Assuming no change in soil bulk density, there is 53% more soluble P now stored in the Ap horizon compared with that when it was originally sampled. If bulk density were greater now then this would result in an even higher estimate of stored soluble P.
As tillage is reduced, soluble P concentrations increase in the upper few millimeters of the Ap (Dick, 1983). Our data show that the mean concentrations (untransformed) of Pcu in the upper 10 cm of the Ap horizons of conventional and conservation tilled sites in this database are 40 and 49 mg kg-1, respectively. The mean concentrations of Pcu in the 10- to 24-cm zone are 37 and 34 mg kg-1, respectively. No-till sites are similar to the conservation till sites but have lower mean Pcu values of 37 and 22 mg kg-1 in the upper and lower Ap horizon, respectively.
Soil and Landscape Factors
Soil test P is nonuniformly distributed across the landscape. In addition to fertilizer and animal manure P applications, the intrinsic properties of the soil, slope, and vegetative cover affect the magnitude of soluble P.
Physiography
The land surface in northwestern Ohio is the result of geomorphic processes related to glacial till deposition, natural lake drainage and shoreline modification, littoral deposition, melt-water outwash from glacier edges, and modern erosion. These processes can have an effect on the initial pool of P in the underlying soil because they control chemical, physical, and mineralogical properties of the parent material. The Pcu is significantly higher than Por in the Ap horizons of soils located on lake plains, ground moraines, and ridge moraines (Table 2). Mean separation, using the Tukey–Kramer Honestly Significant Difference test (SAS Institute, 1996), shows that Por was significantly (p 
0.05) higher in the Ap horizons of soils on beach ridges and outwash plains. Following 15 to 45 yr of P fertilization, there is no significant difference in Pcu between physiographic classes.
Drainage Class
Drainage class connotes the length of time that a soil profile is saturated with water each year. This is closely related to landscape position in northwestern Ohio, and all drainage classes can occur on any physiographic surface. Very poorly and poorly drained soils are saturated for long periods each year and are generally located on the lowest landscape positions where internal drainage is restricted. Nearly 47% of the sites in this database are very poorly drained. For all drainage classes, Pcu is significantly higher than Por (Table 2). In contrast to physiography, Por was not significantly different among any of the four drainage classes. Somewhat poorly drained soils are significantly lower in Pcu than are well and poorly drained soils, indicating that P fertilizer application rates have been lower on somewhat poorly drained soils between the two sampling periods.
Soil Texture
Soil texture is an important factor in determining the magnitude of soil test P, and Pcu is significantly higher than Por for all texture classes. Sandy Ap horizons have significantly greater levels of soil test P than the other texture groups that are higher in clay content (Table 2). Phosphorus application over the years has had the effect of decreasing differences in Pcu between soil texture groups. Clayey Ap horizons were typically fertilized at a higher rate so that Pcu is no longer significantly different from that of coarse-loamy Ap horizons. The rank of the means for both Pcu and Por suggests that surface area is important in the buffering and magnitude of soil test P. Sandy and coarse-loamy Ap horizons are most common on beach ridges and outwash plains. Fine-loamy and clayey surface soils are prevalent on ground moraines and lake plains. Commercial vegetable production, accompanied by heavier fertilization than on grain crops, has traditionally been more prevalent on coarse-loamy and sandy soils in northwestern Ohio. In northwestern Ohio, both SOC and the clay content of Ap horizons increase as internal drainage decreases (Table 3). The SOC, in addition to clay, affects the proportion of labile P in soils. Soluble P is strongly adsorbed on both organic and inorganic colloid surfaces. Organic colloids are also strongly adsorbed on clay surfaces (Logan, 1989). The joint effect is, under equal rates of fertilizer P application, for very poorly drained soils to be more resistant to change in soil test P than better-drained, coarser-textured soils.
|
60%), and Hoytville soils are nearly always tile-drained and have a higher SOC content. Based on this sample of the population of Paulding soils, it is evident that the inability to efficiently drain these soils has resulted in lower P fertilization rates as shown by a mean Pcu to Por ratio of less than one. Correspondingly, the Hoytville series has a corn yield rating 40% greater than Paulding (Soil Survey Staff, 1997).
Tillage Management
The results in Table 2 indicate that fertilizer P application rates have been closely related to changes in tillage management. Tillage practices, especially current tillage, can be an indicator of fertilizer management. When categorized by original tillage practice, Por is significantly lower than Pcu for both cultivated and meadow sites (Table 2). For the two original tillage practices, neither Por nor Pcu are significantly different. When originally collected, between 1953 and 1982, 46% of the sites were meadow, reflecting a greater animal (primarily dairy and beef cattle and horses) population (Richards et al., 2002b). The 54% of sites classed as cultivated can be assumed to be the same as that currently identified as conventional tillage.
The Pcu is not significantly greater than Por for sites currently classed as meadow (Table 2). The Por is not significantly different between current tillage classes, indicating that the sites in each class represented a balanced set of physiographic positions, drainage, and soil texture. For Pcu, conventional and conservation tillage are significantly higher than no-till. There is no statistically significant difference between conservation and conventional tilled sites. Conservation tilled sites reflect a conventional tilled past and an assumption is made that initiation of reduced tillage was also accompanied by reduced fertilizer P rates. This appears to be the case.
Changes with Time
Between 1960 and 1979 fertilizer P sales for the USA increased from 0.55 to 1.21 million metric tons in 1979 (Commercial Fertilizers, 1997). Nationally, the sale of fertilizer P for 1996 was 19% less than the amount sold in 1979 (Fig. 4)
. For northwestern Ohio, the decline has been much greater (44%) for the same time period. Neither county-level nor state-level fertilizer P sales records exist for Ohio prior to 1971. These were estimated from the national data by assuming that fertilizer P consumption in northwestern Ohio was proportionately the same as that of the USA between 1960 and 1970. In 1971, fertilizer P sales in northwestern Ohio accounted for 3% of the national sales, and this figure was used to calculate estimated annual amounts. Both nationally and in northwestern Ohio the greatest rate of decline in fertilizer P sales occurred between 1982–1984. In fact, fertilizer P sales in northwestern Ohio had declined 18% prior to the implementation of the P reduction program in 1982. It appears that operators began reducing fertilizer P applications in response to documented P buildup in the soil, energy economics at the time, and increasing concern about pollution of Lake Erie. National sales appear to be gradually increasing since 1990, but sales in northwestern Ohio have remained near the same level of consumption that occurred in 1965. The question is, do the Bray-1 P concentrations of Ap horizons in northwestern Ohio reflect these historical shifts?
|
Mean Por declines for 1953–1961, rises slightly for 1962–1969, and increases sharply for 1970–1982 (Fig. 5a) . This trend roughly approximates the buildup of fertilizer P sales from 1960–1970 and the maximum sales from 1974–1981 (Fig. 4). After 1982, there is a continued, but gentler slope to mean Pcu for the resampled period (1996–1998). The difference between Pcu and Por steeply declines between 1962–1969 and 1970–1982 because the buildup of P was nearly complete in 1982. By 1982, Por was approaching the same magnitude as Pcu. Mean Por for 1970–1982 is significantly greater than the means for 1955–1969 but is not significantly greater than the mean for 1953–1954. The mean for 1996–1998 is that for Pcu and includes all sites.
|
Not all of the fertilizer P will be fixed by soil colloids or assimilated by plants. The high correlation between fertilizer P sales and SRP in river samples (Fig. 3a) suggests that an annual percentage of added P is dissolved and moves primarily by surface flow into the drainage system. Tile drainage may also contribute a portion of the SRP. Soluble reactive phosphorus is declining more rapidly than total phosphorus in river water (Baker and Richards, 2002), and this suggests that as added P has declined, this has freed remaining fixation capacity of the soil. The slightly greater negative slope for SRP compared with fertilizer P (Fig. 3a) supports this conclusion, although sorbed P was not measured in our study. The effects of changing fertilizer P application methods and tillage practices and seasonal runoff are further discussed as contributing factors by Baker and Richards (2002). This trend supports the conclusion of Baker and Richards (2002) that net accumulation of P has declined in these soils over the past 20 yr and is reflected in soil test P.
|
CONCLUSIONS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
This study shows that soluble P in the soil (indirectly measured as Bray-1 P) is controlled by a combination of fertilizer and tillage management, soil properties, and landscape factors interacting over time. Former soil survey sampling sites are an underutilized resource for reconstructing environmental history and for observing temporal change in soil properties and environmental quality. Soil survey sites are important because they are referenced in both space and time, and can therefore serve as benchmarks. The archived soil samples from these sites are a source of baseline chemical data collected across time. The value of these data is enhanced by access to accessory information including detailed morphological and site descriptions and a variety of other standardized analytical data, permitting assessment of environmental changes that affect both soil and water quality.
|
NOTES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
REFERENCES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  R. P. Richards, F. G. Calhoun, and G. Matisoff The Lake Erie Agricultural Systems for Environmental Quality Project: An Introduction J. Environ. Qual., January 1, 2002; 31(1): 6 - 16. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. P. Richards, D. B. Baker, and D. J. Eckert Trends in Agriculture in the LEASEQ Watersheds, 1975-1995 J. Environ. Qual., January 1, 2002; 31(1): 17 - 24. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. G. Calhoun, D. B. Baker, and B. K. Slater Soils, Water Quality, and Watershed Size: Interactions in the Maumee and Sandusky River Basins of Northwestern Ohio J. Environ. Qual., January 1, 2002; 31(1): 47 - 53. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. B. Moog and P. J. Whiting Climatic and Agricultural Factors in Nutrient Exports from Two Watersheds in Ohio J. Environ. Qual., January 1, 2002; 31(1): 72 - 83. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. P. Richards and D. B. Baker Trends in Water Quality in LEASEQ Rivers and Streams (Northwestern Ohio), 1975-1995 J. Environ. Qual., January 1, 2002; 31(1): 90 - 96. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. B. Baker and R. P. Richards Phosphorus Budgets and Riverine Phosphorus Export in Northwestern Ohio Watersheds J. Environ. Qual., January 1, 2002; 31(1): 96 - 108. [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 |
Soil Science Society of America Journal |
