| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Turfgrass Management, 932 McCormick Ave. East, State College, PA 16801
b P.O. Box 350, Crystal Beach, FL 34681
* Corresponding author (gordon{at}doctorturf.com)
Received for publication December 13, 2005.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Abbreviations: DAT, days after fertilizer treatment
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Published research has shown that concentrations of nutrients in runoff from turfgrasses are low—in most cases, below minimum water quality standards (Welterlen et al., 1989; Gross et al., 1990, 1991; Harrison et al., 1993; Linde et al., 1994; Linde and Watschke, 1997). Most of the previously published work, however, was conducted on undisturbed established turfgrass sites or bare soil, and focused on NO3––N because of its high solubility and leaching potential (Watschke et al., 2000). Linde et al. (1994) measured PO43––P in runoff from undisturbed turfgrasses and found concentrations in the range of 1.61 to 6.06 mg L–1 after plots were fertilized six times per year, although runoff loading rates of PO43––P generally reflected those detected in the irrigation water.
Mature turfgrasses have been shown to effectively mitigate soil movement in runoff. Krenitsky et al. (1998) reported soil losses of 10 and 180 kg ha–1 from mature sod, which was significantly lower than the 510 and 1620 kg ha–1 soil loss from sloped sites covered with a coconut mat. Gross et al. (1991) reported 15 kg ha–1 soil loss from an established tall fescue site compared with 225 kg ha–1 soil loss for bare soil after a 30-min-duration simulated storm. Linde and Watschke (1997) disturbed turfgrass sites by minimally invasive vertical mowing, which had little effect on sediment transport. The highest potential soil loss reported from one plot was 19.4 kg ha–1, with an average potential soil loss from all the plots as low as 1.5 kg ha–1 and 0.1 kg ha–1 in 1994 and 1995, respectively. Vertical mowing is a cultivation practice where vertically oriented knives, mounted on a rapidly rotating horizontal shaft, slice through the turf canopy and penetrate the soil to varying depths (Turgeon, 1991). No publication was found that assessed how core cultivation and subsequent high P fertilization would impact P movement and soil loss in runoff from turfgrass. Traditional hollow-tine core cultivation is the most intense cultural practice routinely performed by turfgrass managers. Hollow tines are used to extract soil cores from the turf, which are then typically broken up and dragged back into the existing holes on golf course fairways. Core cultivation effectively reduces thatch and compaction, and provides the basis for optimum soil physical and hydraulic properties (Carrow, 1988; Carrow and Petrovic, 1992; Callahan et al., 1998). The effect of core cultivation on soil infiltration rates varies, but largely depends on antecedent soil conditions and the extent of compaction (Murphy et al., 1993). Turfgrass managers may apply high rates of P fertilizers to core-cultivated sites to speed recovery, increasing the potential for P losses in runoff after core cultivation. Core cultivation is a far more intense form of cultivation than vertical mowing, and results in a higher percentage of soil accumulating on the turf canopy, increasing the risk for substantial soil loss in runoff from core-cultivated sites.
Recent studies in agricultural settings have emphasized the importance of grass buffers to mitigate nutrient movement, particularly P (Younos et al., 1998; Heathwaite et al., 1998). The importance of these buffers could have utility in the golf course industry, diminishing sediment and P runoff from sloped golf course fairways adjacent to wetlands, lakes, or other environmentally sensitive areas. Most importantly, these studies and those conducted by Welterlen et al. (1989), Gross et al. (1990), and Linde and Watschke (1997) illustrate the unique capacity of grasses to mitigate contaminated water and soil loss. Our research was conducted, in part, to assist turfgrass managers to develop best management practices suited for the timing of fertilizer applications following core cultivation of sloped fairways adjacent to waterways.
Based on what is known from previous research concerning runoff and nutrient loss from undisturbed and minimally disturbed turfgrass sites, it was hypothesized that the potential existed for substantial sediment loss after core cultivation from turfgrass maintained as a golf course fairway, and that high PO43––P concentrations and loading in runoff could impose an off-site threat. The objectives of this study were to: (i) conduct a hydrologic analysis of runoff from core cultivated turfgrasses; (ii) measure sediment loss, subsequent sediment-bound P loading, and PO43––P concentrations and loading rates in runoff following core cultivation and fertilization; and (iii) compare sediment loss and PO43––P concentrations and loading in runoff between creeping bentgrass and perennial ryegrass.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
The six turfgrass plots, or experimental units, were core cultivated on one date in 1998 (16 September), and three dates in 1999 (12 June, 25 July, and 11 October), resulting in four evaluations. Within 24 to 48 h after core cultivation and fertilization, a simulated rainfall event produced runoff. Two more runoff events followed at weekly intervals. During a runoff event, aqueous runoff and sediment samples were collected and stored for further analysis (see below).
The (18.9 by 6.4 m or 121 m2) plots contained a plastic border on both sides to prevent overflow. Runoff water accumulated in a concrete weir fitted with a plastic mesh screen to prevent organic material and fertilizer from entering the weir; runoff subsequently moved into a narrow concrete splitting device where it emptied into a basin fitted with a sump pump. This site has been used to assess nutrient transport and surface runoff since 1985 and was the experimental site for Linde's (1993) work assessing nutrient concentrations in surface runoff and leachate from turfgrass and the effects of turfgrass morphology on overland flow.
In April 1998, and again before each runoff event, soil samples from each plot were tested for available P using the Mehlich-3 extraction method to determine fertilizer needs for the experiment, and established base concentrations of total available P within each plot (Mehlich, 1984). Soil P levels were considered low, not exceeding 43 mg kg–1 according to reports from the Agricultural Analytical Services Laboratory (The Pennsylvania State University, University Park, PA).
In case natural rainfall produced runoff, the detection equipment described by Harrison et al. (1993) was calibrated using known flow rates during setup procedures in August 1997 and June 1998. Each experimental unit and its associated collection devices were calibrated and checked for accuracy. Using a flow meter attached to a garden hose, the system was calibrated by applying known flow rates through the collection system and downloading each potentiometer reading to a computer.
The irrigation system for each experimental unit (plot) was calibrated to deliver 152 mm h–1 at a uniform pressure (276 kPa) similar to the intensities previously reported by Linde et al. (1994). Uniformity was achieved by placing 12 shallow pans with a known volume (199 cm3) at random locations in each plot, running the system for 5 min, and measuring the volume of water collected in each pan. The uniformity of the system was checked for accuracy and individual Weathermatic sprinkler heads (Garland, TX) adjusted accordingly.
The runoff collection period lasted at least 10 min after the initial sign of runoff was detected (Linde and Watschke, 1997), or until enough water was collected for the purpose of analysis. In most cases, the irrigation duration lasted 20 min for the perennial ryegrass and 25 min for the creeping bentgrass. Linde et al. (1995) reported consistently higher time-to-runoff values for the creeping bentgrass than the perennial ryegrass; therefore the lengths of the simulated storms for each turfgrass species differed in an attempt to generate similar runoff sample volumes for analysis.
The rationale for choosing 152 mm h–1 as the irrigation intensity was to ensure that sufficient runoff was produced for analysis, and followed those intensities previously reported by Linde and Watschke (1997) on the same experimental units. According to the simulated rainfall intensity–duration–frequency relationship, it was determined that the simulated rainfall had a 50-yr return frequency (Pennsylvania Department of Transportation, 1986). High-intensity, natural storms are likely of a short duration; therefore, the simulated rainfall intensity and duration produced was considered adequate to mimic a natural rainfall event. From our perspective, the greatest limitation of the irrigation system used to simulate rainfall was the size of the water droplets, which were probably finer than natural rainfall. Therefore, it is possible that erosion caused by water droplets impacting the experimental units did not correspond to a natural rainfall event of this intensity. The simulated rainfall intensity and duration was similar to previously reported runoff studies, although it is possible that differences in erosivity from simulated rainfall compared with natural rainfall could lead to underestimation of PO43––P concentrations and soil loss in runoff (Dougherty et al., 2004; Krenitsky et al., 1998; Jennings and Jarrett, 1985).
On the selected core-cultivation dates, 1.6-cm-diameter tines were used, resulting in a moderate amount of soil accumulating on the surface (10–15% of the turfgrass canopy was occupied by soil). On average, the tines penetrated the soil to a depth of 3.8 cm and were spaced 6.4 cm apart. Based on the size and spacing of the tines and the percentage of soil that accumulated on the turf canopy, it was estimated that 0.042 to 0.050 m3 was brought to the surface. Cores brought to the surface of each plot were broken up, shredded using a Ryan Mattaway 2000 (Cushman, Lincoln, NE) vertical mower, and dragged into the existing aerification holes using a large metal mat. Not all of the soil removed as cores was incorporated back into each aerification hole; therefore each macropore accumulated varying levels of soil, typical of a golf course fairway after core cultivation. Rocks and plant materials were collected with a shovel and removed from the site.
Maintenance Practices
During both growing seasons, the plots were maintained as similar to a golf course fairway as possible by mowing three times a week with a reel mower at a height of 1.9 cm. All turfgrass clippings were collected and removed from the site. Fertilizers used were urea (46–0–0) and an immediately soluble N–P–K (19–25–5, N–P2O5–K2O) granular starter fertilizer (O.M. Scott & Sons, Marysville, OH), with the P derived from diammonium phosphate. The grasses were fertilized with foliar urea at an N rate of 48.8 kg ha–1 on nine selected dates to maintain adequate turfgrass quality. The granular starter fertilizer was applied only after core cultivation, and just before forced runoff, to each experimental unit at a P rate of 42 kg ha–1. The high rate of P was based, in part, on the original soil test report, but primarily used to mimic the potentially high rates of P applied to renovated turfgrass sites to promote turf recovery.
Pesticides were applied at the recommended label rate when needed during the 2-yr study to control broadleaf and grassy weeds, and dollar spot (Sclerotinia homoeocarpa F.T. Bennett) disease using 2, 4-D [(2,4-dichlorophenoxy)acetic acid], fenoxaprop-ethyl (2-[4-[(6-chloro-2-benzoxazolyl)oxy]phenoxy]propanoic acid), and chlorothalonil (2,4,5,6-tetrachloro-1,3-benzenedicarbonitrile), respectively. The pesticide formulations did not contain P. Irrigation, other than that used to produce sufficient runoff, was applied only when the turfgrass reached its wilting point, but not enough to produce runoff.
Experimental Procedures
The experimental procedures closely followed Linde and Watschke (1997), but were not confined to those methodologies. Approximately 24 h after core cultivation, the starter fertilizer was applied at a P rate of 42 kg ha–1. Following fertilization, runoff was produced by irrigating each experimental unit 1, 7, and 14 d after the fertilizer application (DAT). Irrigation was applied between runoff events only when the turf showed signs of wilt. The time period between the onset of irrigation and the first sign of runoff was determined visually and recorded. Runoff was determined to have been initiated when the first drops of water entered the detection equipment.
Water Sampling and Nutrient Analysis Procedures
Aqueous runoff samples were collected from each forced runoff event for chemical analysis. Runoff samples were collected on a flow-paced basis using an ISCO portable sampler and ISCO portable flow meter (ISCO, Lincoln, NE). During the course of each runoff event, 20-mL subsamples were drawn into the sampler from the splitting device, a polyethylene chamber used to collect runoff subsamples from the weir, for every 20 L of runoff volume moving as overland flow. The 20-mL samples were composited into one bottle, frozen at –20°C, and analyzed for PO43––P using the molybdate blue Murphy–Riley method (Murphy and Riley, 1962). Phosphate-P has been reported to contain the largest percentage of algal-available P (Sharpley et al., 1994).
Sediment Sampling Procedures
Sediment from each plot was collected using a coarse, cloth mesh filter designed to trap sediment but still allow the unobstructed flow of water through the splitter. The filters were inserted into each splitter within the collection stations. At the end of each runoff event, the filters were removed, dried, and weighed. The unused filter weight was subtracted from the dry weight of the filter containing trapped sediment and the total sediment was recorded. Distinguishable plant and other organic material was carefully removed by sieve before weighing. Sediment samples were taken to the Agricultural Analytical Services Laboratory (The Pennsylvania State University, University Park, PA) and tested for total bound P using microwave acid digestion (USEPA, 1992, 1986b).
Statistical Designs and Procedures
The two treatments, perennial ryegrass and creeping bentgrass turf species, were arranged in a randomized complete block design with three replications and analyzed as a split plot in time arrangement. The blocks were arranged based on the slope of the experimental units. This design followed the model for repeated measures for whole plots (Steel et al., 1997). An analysis of variance was conducted using the general linear model in SAS Institute (2000). Where significant effects were noted, differences were separated using Fisher's LSD t-test (P 
0.05).
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
Presumably, as the core-cultivation holes filled with soil and plant biomass, infiltration would decrease, increasing overland flow. Therefore it was expected that, as the turf recovered, runoff volumes would increase, but this was not the case. The variation in mean runoff volumes that occurred between dates was attributed to differences in antecedent soil conditions between experimental units (Murphy et al., 1993), variable slope, and any number of uncontrollable environmental factors, including evapotranspiration rates between the turf species, wind speeds, temperatures, or different rates of infiltration, although these variables were not measured during the experimental period.
Phosphate-Phosphorus Transport in Runoff
The aqueous portion of runoff was analyzed for PO43––P, or dissolved P, which is immediately algal available and most threatening as a potential contaminant that might lead to subsequent eutrophication of surface waters (Sharpley et al., 1994). During this study, the highest concentration of PO43––P found in runoff water was 6.1 mg L–1, which occurred from one perennial ryegrass plot 1 DAT on 12 Oct. 1999. This concentration, as well as other high values, occurred in the first simulated rainfall event following core cultivation and fertilization, but only constituted 9% of the total applied P. Concentrations of PO43––P detected in runoff were generally <2 mg L–1. An ANOVA for combined mean PO43––P concentrations is summarized in Table 2.
|
The initial flush of reactive P reported should be considered a threat to surface waters, but it was temporary and concentrations decreased with time for each evaluation conducted (P < 0.0001; Table 3). Phosphate-P concentrations in runoff decreased for each runoff event and, by 14 DAT, reached concentrations slightly above those found in the irrigation water (Fig. 1 ), and at or below the suggested threshold limit of 1 mg L–1 (USEPA, 1986a). For the 28 June 1999 runoff event, only 1% of the applied P was detected in runoff.
|
|
|
Another factor to consider is that the P loading rates were determined in water collected immediately as it exited the sloped turf plots. In "real world" situations, runoff from similar sites would most likely enter a receiving body of water, where dilution occurs. The high application rates of P in this study probably contributed to the high PO43––P concentrations found in each of the first runoff events after core cultivation and should be considered a worst-case scenario. These factors lessen the potential threat of contamination, but in no way minimize the relative importance of the high PO43––P concentrations found in this study.
Sediment and Bound Phosphorus Transport
Sediment was collected during each runoff event, and no differences were detected between turfgrass species with respect to sediment yield (Fig. 3
) or its bound P (data not shown). Linde and Watschke (1997) reported minimal sediment losses from vertically mowed turf. Sediment losses from core-cultivated turfgrass were considered low for the intensity and duration of the storm simulated. The largest amount of sediment collected in any runoff event was 0.70 kg ha–1 from a perennial ryegrass plot during the first evaluation on 18 Sept. 1998, and the average potential soil loss during each evaluation ranged from 0.1 to 0.35 kg ha–1 (Fig. 3). The highest soil loss rate detected by Linde and Watschke (1997) was 19.4 kg ha–1 for one plot, and soil loss rates averaged between 0.1 and 1.5 kg ha–1 from vertically mowed turf.
|
During this study, not enough sediment was collected in runoff to perform the Mehlich-3 extraction soil test. As a result, the EPA 3051 method was used to test collected sediment for total bound P. Bound P found in this study did not exceed 0.70 g ha–1 and averaged 0.35 g ha–1. The values associated with sediment-bound P should be considered a worst-case scenario since much of this P is insoluble (USEPA, 1992).
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Perennial ryegrass produced higher runoff volumes than creeping bentgrass after core cultivation, corroborating earlier research conducted by Linde et al. (1995). The species differences with respect to runoff volumes were evident after core cultivation and may best be attributed to differences in morphology and growth habit, or, more specifically, differences in thatch accumulation, stolon production, and tiller density (Linde et al., 1995). After core cultivation, PO43––P concentrations found in runoff did not differ between turfgrass species, although P export was typically higher for perennial ryegrass than creeping bentgrass, and included one statistically significant difference after one core-cultivation date (Evaluation 2). Interestingly, Linde and Watschke (1997) found no differences in P export between undisturbed or vertically mowed perennial ryegrass and creeping bentgrass turf. Phosphorus export did not exceed 0.12 kg ha–1 and, in general, P losses from core-cultivated plots were equal to or less than those found on undisturbed (Linde et al., 1994; Easton and Petrovic, 2004; Shuman, 2004) and vertically mowed turfgrass sites (Linde and Watschke, 1997).
Soil loss from core-cultivated turf sites remained below values observed on agricultural fields, low-density turfgrass stands, and vertically mowed turfgrass sites (Daniel et al., 1979; Welterlen et al., 1989; Gross et al., 1990, 1991; Linde and Watschke, 1997; Krenitsky et al., 1998; Gillingham and Thorrold, 2000). Sediment-bound P never exceeded 0.70 g ha–1, and was considered not threatening for the duration and intensity of the storm simulated.
Based on low soil test P levels, the small amount of sediment lost in runoff, and low sediment-bound P values, contamination of surface waters from sediment was considered minimal from core-cultivated turfgrass sites in this study. Careful monitoring of soil P levels can help to further minimize this threat. The disturbed turfgrass condition in this experiment did not greatly increase the amount of sediment and dissolved P concentrations in runoff compared with undisturbed, high-density or low-density turfgrass stands. In addition, infiltration probably increased, overland flow declined, and a smaller portion of water ran off, thus minimizing potential threats to surface waters. This research provides a better understanding of the capacity of different turfgrass species to mitigate sediment and dissolved P in runoff after core cultivation.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Crop Science | Vadose Zone Journal | |||
| Journal of Natural Resources and Life Sciences Education |
Soil Science Society of America Journal | ||||
| Journal of Plant Registrations | Journal of Environmental Quality |
The Plant Genome | |||
Means for a single evaluation followed by a different letter are significantly different (P < 0.05) using Fisher's LSD t-test.