SSSAJ Journal of Natural Resources and Life Sciences Education
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online 22 August 2006
Published in Soil Sci Soc Am J 70:1807-1816 (2006)
DOI: 10.2136/sssaj2005.0204
© 2006 Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by He, Z. L.
Right arrow Articles by Banks, D. J.
Right arrow Search for Related Content
PubMed
Right arrow Articles by He, Z. L.
Right arrow Articles by Banks, D. J.
Agricola
Right arrow Articles by He, Z. L.
Right arrow Articles by Banks, D. J.
Related Collections
Right arrow Surface Water Quality
Right arrow Nutrient Management
Right arrow Phosphorus

Nutrient Management & Soil & Plant Analysis

Phosphorus Concentrations and Loads in Runoff Water under Crop Production

Z. L. Hea,b,*, M. K. Zhanga, P. J. Stoffellac, X. E. Yanga and D. J. Banksc

a Ministry of Education Key Lab. of Environmental Remediation and Ecological Health, College of Natural Resource and Environmental Sci., Zhejiang Univ., Huajiachi Campus, Hangzhou 310029, P. R. China
b (current address), Univ. of Florida, Institute of Food and Agricultural Sciences, Indian River Research and Education Center, 2199 South Rock Rd., Fort Pierce, FL 34945 USA
c Univ. of Florida, Institute of Food and Agricultural Sciences, Indian River Research and Education Center, 2199 South Rock Rd., Fort Pierce, FL 34945 USA

* Corresponding author (zhe{at}ufl.edu)


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 NOTES
 RESULTS AND DISCUSSION
 REFERENCES
 
Transport of phosphorus (P) through surface runoff from agriculture is suspected to contribute to the eutrophication of surface waters in South Florida and elsewhere. There is minimal quantitative information on the concentrations and loads of various P forms in surface runoff water on a field-scale. The objective of this study was to evaluate the annual loads of various P forms in runoff water from citrus and vegetable crop production systems in sandy soil regions in Florida and their relations to soil P status, fertilizer P input, and environmental conditions. Eleven field sites (four on vegetable farms and seven in citrus groves) were selected for this monitoring study over a 2-yr period. The concentrations of total P (TP) in the runoff water samples varied widely from 0.01 to 22.74 mg L–1, with approximately half of the samples having the TP over 1 mg L–1. Eighty-three percent of the samples had orthophosphate (PO4–P) higher than 0.02 mg L–1. The mean proportion of total dissolved P (TDP) in the TP was higher than that of the total particulate P (TPP). The TDP constituted the major proportion of P in runoff water from most of the sites. The PO4–P accounted for approximately 64% of the TDP. The annual median concentrations of various P forms in the runoff water varied spatially and temporally and were correlated with total and labile P in the soils (water-P, Olsen-P, Mehlich 1-P, and Mehlich 3-P) as well as fertilizer P rate. The vegetable farms had higher concentrations of P in the runoff water than citrus groves due to their more severe soil erosion and higher fertilizer P input, which resulted in higher soil P accumulation and availability. The annual loads of TP, TDP, and PO4–P varied among the field sites and between the 2 yr. The TP loads were significantly correlated with soil labile P estimated by the four extraction procedures, but the Olsen-P was best related to runoff P. Runoff P concentrations and the annual discharge rate accounted for 55 to 64% of the variance in the annual P loads. These results indicate that P transport through surface runoff from agriculture is affected by soil P status and water management, and merits attention in the development of best management practices.

Abbreviations: TP, total P • TDP, total dissolved P • TPP, total particulate P


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 NOTES
 RESULTS AND DISCUSSION
 REFERENCES
 
PHOSPHORUS LOSSES from agricultural soils to water can pose a threat to water quality. The concern over P with respect to water quality is the stimulation of algae and other aquatic plant growth. High P inputs in chemical fertilizers and organic manure to agricultural soils can exceed crop requirements. This practice often results in excessive P accumulation in the soil, and such soils represent a potential diffuse source of pollution as P plays a key role in the eutrophication of surface waters (Sims et al., 1998). Consequently, control of P in drainage water is usually recommended as the best way to minimize the eutrophication process (Sharpley et al., 2000). Recent studies concluded that annual outputs of P in surface runoff from agricultural soils were generally <2 kg P ha–1 (Ritter, 1988; Sims et al., 1998). Although negligible from an agronomic point of view, such P loads may well have a significant impact on aquatic ecosystems. Phosphorus movement in soils is present in both dissolved and particulate forms (Haygarth and Sharpley, 2000). Particle-bound P includes P associated with soil minerals and large molecular weight organic matter eroded during flow events and constitutes a significant proportion of P transported from most cultivated lands. The proportion of particulate P in the TP in surface runoff varies with soil types and hydrological conditions (Culley et al., 1983; Jordan and Smith, 1985; Pietilainen and Rekolainen, 1991; Heckrath et al., 1995; Beauchemin et al., 1998; Haygarth et al., 1998; Stamm et al., 1998; Hooda et al., 1999; Simard et al., 2000; Uusitalo et al., 2001).

Sandy soils make up the dominant soil types for vegetable and citrus production in Florida. The sandy soils usually contain low levels of clay and organic matter for holding P. There is, therefore, a good reason for concern about P losses from the sandy soils in field-drains as the leaching potential of P in sandy soil is usually high (Stanley et al., 1995; Zhang et al., 2002). However, there is minimal quantitative information on the concentration and forms of P in the surface runoff from field-scale studies. This information is needed to better assess the environmental impact and the possible mechanisms of P losses to surface waters, because different forms of P have very different availability for algal growth (Labry et al., 2005). This study examined the forms, concentration and loads of P over a 2-yr period at four vegetable and seven citrus production sites in Florida.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
NOTES
 RESULTS AND DISCUSSION
 REFERENCES
 
Field Sites
Eleven field sites (seven in commercial citrus groves and four on vegetable production farms) in the Indian River area, South Florida were selected for this monitoring study in 2001–2002 (Table 1). Annual rainfall for the 11 sites ranged from 1203 to 1572 mm and 1002 to 1362 mm for the Years 2001 and 2002, respectively. Rainfall and water discharge varied seasonally, with most of them occurring from May to October (Fig. 1 and 2) . All sites were in a flat landscape (<5% slope) with shallow water table where the dominant hydrological pathways were an extensive network of artificial drainage ditches. In the Indian River area, the dominant soils under agricultural production are Alfisol and Spodosol (locally called flatwoods soils). These soils are characterized by sand of surface soil often in excess of 90% underlain by an argillic (for Alfisol) or a spodic horizon (located at about the 60- to 120-cm depth) that contained relatively high clay content. The clayey horizon restricts the downward transport of leachate and P. Citrus production in this area requires drainage of these soils by bedding the field and planting the trees on the raised beds. The soil solution above the clayey layer is drained into the water furrow (approximately 80 cm deep and 100 cm wide), which in turn is disposed into the surface water through ditches and canals. Because of surface fertilization and irrigation, the roots of citrus are mostly distributed within the 0- to 60-cm depth (>98%) (Zhang et al., 1996). Therefore, the water and nutrients that may have leached to the top of the argillic or spodic horizon from the surface of the bed are not available to the plants and can be transported laterally into the water furrow that drains one direction to a ditch or canal. Citrus groves in the Indian River area are generally double-row bedded with a water furrow between beds. The citrus beds are typically 6 m wide, 1 m high, and of variable length, with a crown in the middle of the bed, between the two tree rows. The area of the citrus grove used for water collection and sampling is the area between the crowns of two neighboring beds, running the length of the bed. Water from this area enters the water furrow, which drains in only one direction to a ditch or a canal. The total discharge from this area can be quantified by a flow meter integrated to an autosampler, which is placed at the end of the water furrow. For vegetable sites, the drainage is similar except that there are more beds and they are covered with plastic mulch during the cropping season. During the cropping season, surface runoff and seepage water are drained to canals alongside the beds using shovel ditches and drainage pipe. During the fallow period, the mulch is removed, the fields are plowed and surface runoff is directly drained through the drainage pipe. The water sampling site area is the total water collecting area that drains into a particular shovel ditch or drainage pipe.


View this table:
[in this window]
[in a new window]
  		Table 1. Soil and management characteristics of experimental field sites showing average and standard deviations for measured parameters.

 

Figure 1
View larger version (14K):
[in this window]
[in a new window]
  		Fig. 1. (a) Mean monthly rainfall and (b) runoff water discharge of 11 sites, 2001–2002. Error bars represent one standard deviation.

 

Figure 2
View larger version (25K):
[in this window]
[in a new window]
  		Fig. 2. Annual rainfall of the 11 field sites, 2001–2002.

 
Fertilization
Fertilizers of blended dry water-soluble granular (P as NH4H2PO4/KH2PO4) were surface-applied for the citrus groves using a mechnical spreader and three equally divided doses were applied in February, June, and October, respectively. Fertilization of the vegetable fields consisted of basal (fertilizers applied before bedding, which accounted for 90% of TP and 20% of total N and K for the cropping season), and banded application (applied in the middle of the bed, which amounted to 10% of TP and 60% of total N and K), and fertigation (applied through irrigation system for the rest of the fertilizers, 20% of total N and K).

Soils and Sampling
The soils of the experimental sites were representative for commercial citrus and vegetable production systems in the Indian River area. They included Wabasso sand (sandy, siliceous, hyperthermic Alfic Haplaquods), Waveland fine sand (sandy, siliceous, hyperthermic, ortstein Arenic Haplaquods), Ankona sand (sandy, siliceous, hyperthermic, ortstein Arenic Haplaquods), Winder variant sand (fine-loamy siliceous, hyperthermic Typic Glossaqualfs), and Nettles sand (sandy, siliceous, hyperthermic, ortstein Alfic Arenic Haplaquods). General characteristics of the study sites are presented in Table 1.

For each field site, three composite soil samples were taken across each field site before the experiment. Each composite sample was composed of a mixture of four samples taken at a depth of 0 to 15 cm from four locations within each field site. All soil samples were air-dried and ground to <2 mm before chemical analysis.

Water Sampling and Data Collection
Portable autosamplers (SIGMA 900MAX, American Sigma, Loveland, CO1) were installed at a drainage outlet for each field site. Each autosampler has a Doppler sensor and a water sampling strainer, which were installed inside a plastic drainage pipe (20-cm diam.) that leads drainage water from field furrow to the ditch or canal. The autosampler starts to collect the water sample whenever there is significant flow through the pipe. Runoff samples from each field site were collected in 1-L bottles placed inside each autosampler during each rainfall event. The autosamplers were programmed so that six individual composited runoff samples were taken every 24 h if runoff was occurring. The first three samples were collected in the first 2 h after the sampling process was triggered, each for 40 min in sequence. During the first 40 min the sampler collected three subsamples (one subsample every 13 min and 20 s) that were combined as Sample 1. Samples 2 and 3 were collected in a similar way. The remaining three combined samples were collected into another three bottles, each for 7 h and 20 min in sequence.

The autosamplers were checked daily to ensure proper performance and to collect water samples if available. Rainfall and runoff flow rate were recorded every 10 min and the data in the autosamplers were transferred to a laboratory computer weekly using a data logger. Water samples collected from the autosamplers were immediately transported to the laboratory in an ice chest.

Chemical Analysis
Soil pH was measured in water at a soil/water ratio of 1:1 using a pH/ion/conductivity meter (Accumet Model 50, Fisher Scientific, Norcross, GA). Total organic carbon (C) was determined using a CN-Analyzer (Vario MAX CN Macro Elemental Analyzer, Elemental Analysensystem GmbH, Hanau, Germany). Particle composition of soil was determined using the micropipette method (Miller and Miller, 1987). Soil pH varied greatly across the 11 sites, and ranged from 4.4 to 8.1 (Table 1). Clay content of the soil samples ranged from 19 to 81 g kg–1. Soil TP, water-P, Olsen-P, Mehlich 1-P, and Mehlich 3-P were measured using standard methods (Kuo, 1996), and ranged from 100 to 430, 0.94 to 23.9, 3.18 to 68.2, 9.42 to 241, and 11.4 to 269 mg kg–1, respectively.

Solid concentrations of the water samples were measured using a gravimetric method with oven drying. Total P in the unfiltered surface runoff sample was determined by the molybdenum-blue method after digestion with acidified ammonium persulfate (Greenberg et al., 1992). Portions of the subsamples were filtered through a 0.45-µm syringe filter for measurement of TDP and PO4–P. The TDP was determined using an inductively coupled plasma atomic emission spectrometry (ICP–AES, Ultima, JY Horiba Inc. Edison, NJ). Total particulate P was calculated from the difference between the TP and TDP. The concentrations of PO4–P were measured within 24 h after sample collection using an ion chromatograph (DX 500; Dionex Corporation Sunnyvale, CA). The TP, TDP, and PO4–P loads in the runoff water were determined as a product of the P concentrations in runoff samples and each runoff water discharge:

Formula 1[1]
where 10–3 is a conversion factor from mg L–1 to kg m–3, 104 is an area (m2) of 1 ha, and site area is the area for collecting runoff of each site. Annual loads were determined as a summation of loads from each rainfall event.

Statistical Analysis
Separation of means for the concentrations of various P forms in surface runoff was conducted by the Duncan multiple range test procedure of the Statistical Analysis System release 8.2 (SAS Institute, 2001). Each variable of P forms in surface runoff was subjected to an analysis of variance (ANOVA) using the SAS for each site. An orthogonal contrast was partitioned from the main effect of field site to determine the differences in the concentrations of various P forms in runoff water between the two crop production systems (vegetable vs. citrus) and the differences in the annual loads of TP, TDP, and PO4–P in surface runoff between 2001 and 2002. The correlation and regression procedures of the SAS were used to evaluate the relationships between soil test P values and median runoff P concentrations or loads, and among the various P forms in runoff water. Multiple regression analyses (forward procedure) were performed to evaluate the relationships between the annual loads of various P and runoff P concentration, soil labile P (Olsen P), fertilizer P rate, or annual water discharge rate of each field site.


   RESULTS AND DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 NOTES
 RESULTS AND DISCUSSION
REFERENCES
 
Phosphorus Concentration and Forms in Runoff Water
Total P concentration in the 1272 runoff water samples collected from the 11 sites during the 2-yr period varied widely, ranging from 0.01 to 22.74 mg L–1 (detailed data were not shown due to a large volume, the same with TDP, PO4–P, and TPP in the following), with a median concentration of approximately 1.0 mg L–1 (Fig. 3 ). The TDP in the runoff water ranged from 0.01 to 11.88 mg L–1, with a median value of 0.65 mg L–1. Both TP and TDP were higher than TP concentrations (0.03–0.14 mg L–1) in water of the Indian River Lagoon in Florida (Sigua et al., 2000). The concentrations of PO4–P in the runoff varied from <0.01 to 9.85 mg L–1, with a median value of 0.43 mg L–1. Eighty-three percent of the samples had PO4–P higher than 0.02 mg L–1, the critical concentration for lake eutrophication (Schindler, 1977). The TPP in the runoff water also varied widely, ranging from <0.01 to 19.94 mg L–1, with a median of 0.23 mg L–1 (Fig. 3).


Figure 3
View larger version (16K):
[in this window]
[in a new window]
  		Fig. 3. The concentrations of total P (TP), total dissolved P (TDP), orthophosphate (PO4–P), and total particulate P (TPP) in runoff water samples (n = 1272). The triangles represent 1 and 99 percentile of the data and the error bars represent the 5 and 95 percentile of the data. The middle line is the median value of the data range. The upper value of the box is the 75 percentile and the lower value of the box is the 25 percentile.

 
The proportions of PO4–P, TDP, and TPP in the TP of the runoff samples varied with the samples, ranging from 2 to 97, 2 to 100, 0 to 98%, with median values of 40, 73, and 27%, respectively (Fig. 4 ). The proportion of the TDP in the TP was higher than those of the TPP, suggesting that most of P in the runoff water from the sandy soils was soluble. The median proportion of TDP in the TP in runoff water from four vegetable farms (53.1%) was lower than that from seven citrus groves (77.0%), which was probably due to the higher solids concentration in runoff water from the vegetable farms (3.83 g L–1) than from the citrus groves (0.99 g L–1). The proportion of TDP in the TP decreased with increasing TP, with a correlation coefficient of 0.21 (p < 0.001, n = 1272). The PO4–P was a dominant P form in the TDP, and had a median value of 64% in the TDP. However, the proportion of PO4–P in the TDP varied from 0 to 100% (Fig. 4). The median value of PO4–P in the TDP for the four vegetable farms (72.9%) was higher than that for seven citrus groves (50.1%). The TDP, PO4–P, or TPP concentrations were positively correlated with the TP (r = 0.79, 0.76, and 0.72, respectively, P < 0.01, n = 1272). Factors affecting the proportions of various P in runoff were not well understood. Soil P status, fertilizer P rate and type, and other soil properties may have significant effects on the concentrations and distribution of various P forms in the runoff water.


Figure 4
View larger version (15K):
[in this window]
[in a new window]
  		Fig. 4. Proportions of total P (TP) as PO4–P, total dissolved P (TDP), and total particulate P (TPP) in runoff water samples collected from 11 sites during the 2-yr experiment (n = 1272). The triangles represent the 1 and 99 percentile of the data and the error bars represent the 5 and 95 percentile of the data. The middle line is the median value of the data range. The upper value of the box is the 75 percentile and the lower value of the box is the 25 percentile.

 
Spatial Variation
There were significant differences in the concentrations of each P form in runoff water among the 11 sites within a year (Tables 2 and 3). Seven of the 11 sites had median TP higher than 1.0 mg L–1. The highest annual median concentrations of TP in the runoff water were found at two vegetable sites and one citrus site (Sites 7 and 11 in 2001 and Sites 7 and 8 in 2002), whereas the lowest annual occurred at three citrus sites (Sites 1 and 10 in 2001 and Sites 2 and 10 in 2002). The highest TP concentration was approximately eightfold of the lowest. The highest median concentrations of TDP (3.16 mg L–1) and PO4–P (2.31 mg L–1) in the runoff water were found at Site 11 whereas the lowest (0.33 mg L–1 for TDP and 0.11 mg L–1 for PO4–P) occurred at Site 4. Similar to TP, the TPP were highest at Sites 7 and 8, and lowest at Site 1 (Table 2). The mean values of quarterly median concentrations (the median concentration of each P form in the samples collected from each field site during a period of each quarter or 3-mo) of TP and TPP in the runoff water from vegetable farms were generally higher than those from citrus groves (P < 0.05), whereas those of TDP, and PO4–P had similar trend, but were not significantly different between the two crop production systems (Tables 2 and 3). Many factors may affect concentration of different P forms in surface runoff water. The generally higher concentrations of P in runoff water from the vegetable farms than from citrus groves are likely related to their higher soil P accumulation and greater annual fertilizer P input, as compared with the citrus groves (Table 1). In addition, the vegetable farm soils are subjected to more severe erosion, especially at the end of a growing season when the plastic mulch is removed and the land is tilled and exposed to heavy rains, which is evident by the significantly higher TP and TPP concentrations in runoff water from the vegetable farms than from the citrus groves.


View this table:
[in this window]
[in a new window]
  		Table 2. Annual median concentrations of various P forms in runoff water from each field site.

 

View this table:
[in this window]
[in a new window]
  		Table 3. Mean values of quarterly median concentrations of various P forms in runoff water from each field site, 2001–2002 (n = 71).{ddagger}

 
The annual median concentrations of various P forms in the runoff water for each field site were significantly affected by soil P status and fertilizer P inputs (Table 4). The median concentrations of TP in runoff were significantly correlated with each soil test P (Olsen-P, water-P, Mehlich 1-P, and Mehlich 3-P), soil TP or fertilizer P rate. However, the median concentrations of TDP and PO4–P were significantly correlated only with Olsen-P and soil TP (Table 4). Among the soil test P indexes, Olsen-P provided the best prediction of runoff P concentrations, followed by water P. These results indicate that P loading in runoff water is related to the accumulation and availability of soil P and the amount of fertilizer P input.


View this table:
[in this window]
[in a new window]
  		Table 4. Correlations (r) between annual median concentrations of various P forms in the runoff water and soil test P indexes, soil total P (TP), and fertilizer P rate (n = 21).

 
Temporal Variations
The median concentrations of various P forms in the runoff water across all the sites were similar between Year 2001 and 2002 (Tables 2 and 3). Representative patterns of seasonal changes in the TP, TDP, and PO4–P concentrations in runoff water at the same site were demonstrated with Sites 7, 9, and 11 (Fig. 5 ). The concentrations of various P forms in the runoff water appeared to be higher from July to October for both years, as compared with other times. July to October is the summer season with a high temperature and a high rainfall in Florida (Fig. 1 and 2). The high temperature enhances the release of P from organic matter decomposition and the desorption of P from soil surfaces. Consequently, more soil P is readily subjected to runoff losses. For most sites, the concentrations of TP and TPP in the runoff water were significantly affected by solids contents (Table 5), indicating that increased losses of solid materials in the runoff water could significantly increase losses of the TP and TPP. Quinton et al. (2001) also reported that TP concentration in runoff was significantly affected by the amount of P in the sediment. This is particularly important during the storm season when heavy rains not only leach P out of the soil but also move P-enriched fine soil particles into runoff waters.


Figure 5
View larger version (26K):
[in this window]
[in a new window]
  		Fig. 5. Seasonal variations of total P (TP), total dissolved P (TDP), and PO4–P in runoff water at field Sites 7, 9, and 11.

 

View this table:
[in this window]
[in a new window]
  		Table 5. Correlations (r) for total P (TP) vs. solids and total particulate P (TPP) vs. solids in runoff water.

 
Discharge and Phosphorus Loads of Runoff Water
Although differences in annual rainfall among the 11 sites were moderate (Fig. 2), annual runoff discharge rate ranged from 47 to 5268 m3 ha–1, and varied greatly among the sites within a sampling year and between the Year 2001 and 2002 (Fig. 6 ). The highest discharge was about 112 times the lowest. The spatial and temporal variations of runoff water discharge are affected by several factors, including variation in precipitation, soil texture and structure, land slope, vegetation coverage, and soil permeability (Table 1). For instance, the greatest discharge rate was found at Site 10 where the highest rainfall occurred in the last 2 yr. In addition, the soil on this site had an impermeable layer (argillic horizon or hard pan) close to the surface (60–80 cm); all leachate from the overlaid sand layer would seep laterally into the furrow when it reached the top of the impermeable layer. Site 6 that had the smallest discharge rate had a relatively lower rainfall, and the soil on this site had a deep sand layer (>100 cm) overlaid on a spodic horizon, and therefore, the rainfall or irrigated water could penetrate deep into the soil, with less available for surface runoff losses. Except for Sites 1, 2, and 5, the annual discharges were generally higher in 2001 than 2002, which was probably related to the difference in rainfall (Fig. 2). The annual mean loads of TP, TDP, and PO4–P in the runoff were higher in 2001 than 2002 (P < 0.05) (Table 6). However, the annual loads of various P forms in the runoff water varied with the sites, ranging from 0.20 to 12.80 kg ha–1 for TP, 0.08 to 6.36 kg ha–1 for TDP, and 0.07 to 5.46 kg ha–1 for PO4–P. Higher P loads occurred at Sites 7, 8, and 9 in 2001, and Site 7 in 2002, which was likely related to their higher P saturation in the soils as evidenced by their higher soil test P values (Table 1).


Figure 6
View larger version (25K):
[in this window]
[in a new window]
  		Fig. 6. Annual discharge of runoff water at each field site.

 

View this table:
[in this window]
[in a new window]
  		Table 6. Annual loads of total P (TP), total dissolved P (TDP) and PO4–P in the runoff water from each field site.

 
The annual loads of TP, TDP, and PO4–P were significantly correlated with the annual mean concentrations of various P forms in surface runoff water except for TPP with TDP or PO4–P loads (Table 7). The TP load was also significantly correlated with each of the soil test P values (water P, Olsen-P, Mehlich 1-P, and Mehlich 3-P), but the best correlation occurred with the Olsen-P. The PO4–P load was significantly related only to Olsen-P. The relationships between TP, TDP, or PO4–P load and soil TP, fertilizer rate, or discharge rate were positive, but did not reach significant levels (Table 7).


View this table:
[in this window]
[in a new window]
  		Table 7. Correlation (r) between annual P loads and soil test P, discharge rate, and P concentrations in runoff water (n = 21).

 
Multiple regression analysis was conducted by entering the best correlated factors from each category (including the corresponding runoff P concentration, Olsen-P, annual discharge rate, and fertilizer P rate). The results indicated that runoff P concentrations accounted for 43 to 55% of the variance in the annual P loads, and the addition of discharge rate to the model significantly improved the accountability (55 to 66%), and soil Olsen-P had a modest contribution to the loads of TP and TDP (Table 8). The effects of runoff P concentration and runoff discharge rate on annual P loads are apparent as they are the primary components of loads. Olsen-P was better correlated with P loads than annual discharge rate (Table 7), but did not explain as much in the variance of P loads as discharge rate, probably because Olsen-P is indirectly related to loads and has been integrated into runoff P concentration factor in the model. However, if runoff P concentration was removed from the model, then Olsen-P explained a larger proportion of the variance in P loads than discharge rate (Table 9). Olsen-P, fertilizer P rate and discharge together accounted for approximately 53 to 59% of the variance in the loads of TP, TDP, and PO4–P in runoff waters from the two crop production systems.


View this table:
[in this window]
[in a new window]
  		Table 8. Regression models for total P (TP), total dissolved P (TDP), and PO4–P loads (Y, kg ha–1yr–1) as related to annual median P concentrations (TP, TDP, or PO4–P) in runoff water (X1, mg L–1), annual discharge rate (X2, m3 ha–1), soil Olsen P (X3, mg kg–1), and annual fertilizer P rate (X4, kg ha–1) (n = 21).

 

View this table:
[in this window]
[in a new window]
  		Table 9. Regression models for total P (TP), total dissolved P (TDP), and PO4–P loads (Y, kg ha–1) as related to soil Olsen P (X1, mg kg–1), annual fertilizer P rate (X2, kg ha–1), and annual discharge rate (X3, m3 ha–1) (n = 21).

 
General Discussion
Phosphorus export from agricultural production systems has been considered to contribute to the accelerated eutrophication of fresh water bodies, as algal booming is often related to increased P inputs (Phlips et al., 2002). The concentrations of TP measured in the 1272 surface runoff water samples collected from the 11 field sites in our study ranged from 0.01 to 22.74 mg L–1, with a median concentration of approximately 1.0 mg L–1. The value is lower than those (2.2–11 mg L–1) observed in surface runoff from a dairy farm (Barlow et al., 2005) or those (9.55 mg L–1) from soils amended with farm manure (Michaud and Laverdiere, 2004), but greater than those from other upland watershed (0.05–1.0 mg L–1) (Gelbrecht et al., 2004; Weld et al., 2001). The high end likely resulted from the effects of dairy waste water and farm manure, since the former irrigated with treated wastewater from the dairy farm (Barlow et al., 2005) or received farm manure (Michaud and Laverdiere, 2004) and therefore, increased soil P saturation and P export potential, whereas the low end is probably due to the fact that TP was measured in drainage from the whole watershed with a mix of soil types and crops/forests rather than in runoff water directly from the field as in this study. However, this value (1.0 mg L–1) is much greater than the ecoregional critical levels (0.008–0.038 mg L–1) of TP for surface waters (streams, rivers, and lakes) established by the United States Environmental Protection Agency (USEPA) (USEPA, 2005). The P in surface runoff from agricultural fields can be a potential nonpoint source for receiving waters.

The annual P load in surface runoff is a measure of P lost from agricultural production systems to the environment. Load is often more meaningful than concentration alone as it takes account of both concentration and discharge. The TP loads (0.2–12.8 kg ha–1 yr–1) in the surface runoff from citrus and vegetable fields are smaller than those (2.5–23 kg ha–1) from dairy farms (Barlow et al., 2005), but greater than those observed from some arable land watershed (0.03–1.09 kg ha–1 yr–1, Djodjic et al., 2004; 0.04–0.25 kg ha–1 yr–1, Gelbrecht et al., 2004). In addition to difference in climate that influences runoff discharge rate, other factors such as soil P status, fertilizer application rate, soil type, tillage, and vegetation coverage also affect P loads in runoff waters (Weld et al., 2001; Daverede et al., 2003; Andraski et al., 2003).

Phosphorus loads in surface runoff have been reported to be related to soil extractable P measured by Bray P1 (Daverede et al., 2003), Olsen P (Djodjic et al., 2004; Turner et al., 2004), and Mehlich-3 P (Tarkalson and Mikkelsen, 2004). The results from this study indicate that TP load is highly correlated with each of the Olsen-P, water-P, Mehlich 1-P, and Mehlich 3-P, but the correlation with Olsen-P was the best (Table 7). Similar conclusion is also obtained from the correlation analysis between the concentrations of various P forms in runoff water and each of the soil test P values (Table 4). The Olsen-P method was developed for calcareous soils. However, this method is also adequate for other type of soils based on the good correlations of Olsen-P with other parameters of P availability such as plant P uptake, A value, or crop yield (Kamperath and Watson 1980). The Olsen-P extractant (0.5 M NaHCO3) is a mild reagent, which extracts water soluble P and a portion of adsorbed P with moderate binding force. These P forms are largely plant-available and readily subjected to losses by surface runoff or leaching. Multiple regression analysis also indicates that Olsen-P can explain a larger proportion of the variance in P loads than fertilizer P or discharge rate (Table 9). Based on previous results (Zhang et al., 2002) and the results from this study, it seems that Olsen-P is a good potential soil P test for predicting P loss in surface runoff from agricultural fields.


   ACKNOWLEDGMENTS
 
This study was, in part, supported by a section 319 Nonpoint Source Management Program grant (DEP contract # WM746) from the U.S. Environmental Protection Agency (US EPA) through a contract with the Nonpoint Source Management/Water Quality Standard Section of the Florida Department of Environmental Protection (FDEP) and by grants (DEP contract # SP566, G0018, and S0132) from the FDEP. Florida Agricultural Experiment Station Journal Series No. R-10956.


   NOTES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 NOTES
RESULTS AND DISCUSSION
 REFERENCES
 
1 Mention of particular companies or commercial products does not imply recommendations or endorsement by the Zhejiang University, Hangzhou, China or the University of Florida, Gainesville, USA over other companies or products not mentioned. Back

Received for publication June 27, 2005.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 NOTES
 RESULTS AND DISCUSSION
 REFERENCES
 





This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by He, Z. L.
Right arrow Articles by Banks, D. J.
Right arrow Search for Related Content
PubMed
Right arrow Articles by He, Z. L.
Right arrow Articles by Banks, D. J.
Agricola
Right arrow Articles by He, Z. L.
Right arrow Articles by Banks, D. J.
Related Collections
Right arrow Surface Water Quality
Right arrow Nutrient Management
Right arrow Phosphorus

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
Journal of Plant Registrations Journal of
Environmental Quality		 The Plant Genome