JEQ Grow Your Career With ASA
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online 25 January 2007
Published in J Environ Qual 36:426-439 (2007)
DOI: 10.2134/jeq2006.0232
© 2007 American Society of Agronomy, Crop Science Society of America, and 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 Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Steinke, K.
Right arrow Articles by Thompson, A.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Steinke, K.
Right arrow Articles by Thompson, A.
Agricola
Right arrow Articles by Steinke, K.
Right arrow Articles by Thompson, A.
Related Collections
Right arrow Best Management Practices
Right arrow Phosphorus
Right arrow Turfgrass

TECHNICAL REPORTS

Surface Water Quality

Prairie and Turf Buffer Strips for Controlling Runoff from Paved Surfaces

K. Steinkea,*, J. C. Stierb, W. R. Kussowc and A. Thompsond

a Dep. of Soil and Crop Sciences, Texas A & M Univ., College Station, TX 77843
b Dep. of Horticulture, Univ. of Wisconsin-Madison, Madison, WI 53706
c Dep. of Soil Science, Univ. of Wisconsin-Madison, Madison, WI 53706
d Dep. of Biological Systems Engineering, Univ. of Wisconsin-Madison, Madison, WI 53706

* Corresponding author (ksteinke{at}ag.tamu.edu)

Received for publication June 16, 2006.

   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Eutrophication of surface waters due to nonpoint source pollution from urban environments has raised awareness of the need to decrease runoff from roads and other impervious surfaces. These concerns have led to precautionary P application restrictions on turf and requirements for vegetative buffer strips. The impacts of two plant communities and three impervious/pervious surface ratios were assessed on runoff water quality and quantity. A mixed forb/grass prairie and a Kentucky bluegrass (Poa pratensis L.) blend were seeded and runoff was monitored and analyzed for total volume, total P, soluble P, soluble organic P, bioavailable P, total suspended solids, and total organic suspended solids. Mean annual runoff volumes, all types of mean annual P nutrient losses, and sediment loads were not significantly affected by treatments because over 80% of runoff occurred during frozen soil conditions. Total P losses from prairie and turf were similar, averaging 1.96 and 2.12 kg ha–1 yr–1, respectively. Vegetation appeared to be a likely contributor of nutrients, particularly from prairie during winter dormancy. When runoff occurred during non-frozen soil conditions turf allowed significantly (P ≤ 0.10) lower runoff volumes compared with prairie vegetation and the 1:2 and 1:4 impervious/pervious surface ratios had less runoff than the 1:1 ratio (P ≤ 0.05). In climates where the majority of runoff occurs during frozen ground conditions, vegetative buffers strips alone are unlikely to dramatically reduce runoff and nutrient loading into surface waters. Regardless of vegetation type or size, natural nutrient biogeochemical cycling will cause nutrient loss in surface runoff waters, and these values may represent baseline thresholds below which values cannot be obtained.

Abbreviations: FS, frozen soil • KBG, Kentucky bluegrass • NFS, non-frozen soil • TP, total phosphorus • SP, soluble phosphorus • SOP, soluble organic phosphorus • TSS, total suspended solids • BAP, bioavailable phosphorus


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
ENHANCED aquatic P concentrations have been shown to be the primary cause of surface water eutrophication that may reduce lake water quality through fish kills and algae blooms, and ultimately decrease recreational opportunities (Sharpley et al., 1994; Carpenter et al., 1998). In recent years concerns about nonpoint source (NPS) pollution in lakes bordered by urban environments have led many local, state, and federal officials to pursue regulation of phosphorus (P) use for urban landscapes and to re-examine best management practices used to control storm water runoff (Rosen and Horgan, 2005).

Difficulties in separating anthropogenic eutrophication from natural source enrichment can be problematic when attempting to delineate nutrient loading of runoff waters from urban environments. Turfgrass ecosystems are easily defined plant communities where N and P are often applied to maintain desirable turf quality. The continental United States contains 1.6 x 105 km2 of turfgrass, an area three times larger than any irrigated crop (Milesi et al., 2005). Wisconsin alone has an estimated 1.6 million home lawns encompassing greater than 3.0 x 105 ha of land (Wisconsin Agricultural Statistics Service, 2001). In an effort to address concerns regarding urban water quality and NPS, the Wisconsin Department of Natural Resources (WDNR) proposed legislation in 2001 targeting the reduction of pollution from runoff (WDNR, 2001). One option considered was to use buffer strips comprised of native vegetation (i.e., prairie) around water bodies and construction sites to decrease runoff volumes and nutrient loading. Data from direct comparisons between prairie and turfgrass plant communities for mitigating surface pollution are needed for logical decision-making processes to reduce urban NPS.

Interest is growing in the use of vegetative buffer strips to reduce NPS from urban environments. The use of buffer strips has effectively decreased surface water flows and increased water infiltration in riparian, agricultural, and forested watersheds (Gilliam, 1994; Clausen et al., 2000; Lee et al., 2000). However, urban watersheds have uniquely different characteristics with respect to soil disturbances and presence of impervious surfaces that dramatically modify ecosystem hydrology.

Buffer strip sizes have been specified in agricultural and riparian areas to account for nutrient retention, controlling storm water and sediment, fecal coliform reduction, water temperature control, and improving wildlife habitat (Johnson and Ryba, 1992). Recommended sizes vary according to end goals and may range from 5 to 200 m (Johnson and Ryba, 1992). The type of contaminants, for example sediment or soluble P, entering the buffer strip may also affect the requisite size or vegetation of a buffer strip (Abu-Zreig et al., 2003). Effective buffer strip widths in urban areas have yet to be defined. Urban environments include impervious surfaces which may alter the natural flow path of water by creating channelized surface flow and impact effectiveness of vegetated buffers (Paul and Meyer, 2001). Schuler (1995) discovered that close to 90% of urban surface water runoff concentrated before reaching a buffer strip. The ability of buffer strips to properly function depended on resisting the channelization process.

Plant community composition may play an integral role in the performance of buffer strips as plants vary in dissolved nutrient removal due to differences in nutrient uptake or soil adsorption. All ecosystems are subject to biogeochemical nutrient cycling and may serve as pollutant sources. Native plants are generally praised for their ability to accumulate nutrients and to persist through time (Woodmansee, 1984). For these reasons and due to their inherent low maintenance requirements, native plants have been suggested for buffer strip use in urban areas. However, the ability of native plants to establish, persist, and tolerate large storm water volumes and an array of anthropogenic pollutants compared with turf has not been tested. Riparian buffer strips have functioned well due to little disturbance and human interference once established. In urban areas, buffer strips may have to serve a variety of non-pollution related functions with minimal hazards in terms of visual obstruction or the harboring of nuisance pests and vermin.

Several studies have indicated minimal nutrient and sediment contributions from turf ecosystems (Gross et al., 1990, 1991; Linde et al., 1995) but data in comparison with alternative landscapes are limited. Erickson et al. (2001) recorded no runoff volume differences between turf and a mixed grass, tree, and shrub plant community on sandy soils in Florida.

The amount and form of P present may dictate what effect over-enrichment has on a water body. Total phosphorus (TP) consists of soluble P (SP) and sediment-attached P (particulate P) present in runoff water. Because SP concentrations can change every few minutes, the available quantities of SP can lead to inappropriate conclusions concerning water quality (Correll et al., 1975). Total phosphorus has a slower turnover rate and is regarded as the critical parameter to measure when predicting responses to P additions (Correll, 1998). Soluble P does not include sediment P contributions but does have an immediate impact on plant growth and is the most available form of P for plant uptake. Bioavailable P (BAP) is an estimate of long-term P availability and includes all SP and the portion of particulate P that can be made available in solution.

Turner and Haygarth (2000) demonstrated that a substantial portion of P lost in surface or subsurface runoff from grasslands occurred in organic forms. Despite representing small agronomic losses, the P concentrations were sufficient to cause water eutrophication. McDowell and Sharpley (2001) confirmed this finding and stated that the importance of soluble organic P should not be overlooked even at low soil test P levels. Dalal (1982) indicated that high levels of inorganic P inhibit the mineralization of organic P resulting in a larger organic P pool remaining in the soil environment, important because organic P can also contribute to eutrophication.

The relative ability of prairie and turf plant communities to affect runoff water quality and quantity are not known. Because data from a variety of independent studies suggested nutrient loading and runoff may be greatly affected by vegetation density more than by type of vegetation, we hypothesized that prairie and turf vegetation would have similar effects on urban runoff water volume and quality.

The objectives of this study were to (i) compare the ability of prairie and turf buffer strips to reduce surface water runoff, P, and sediment losses, (ii) measure the naturally occurring levels of P that could be expected to run off from a nonfertilized prairie ecosystem, and (iii) compare the effect of three impervious/pervious surface area ratios to mitigate storm water runoff.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Research site preparation began in August 2002 with renovation of pre-existing runoff plots located on a Batavia silt loam soil (fine-silty, mixed, superactive, mesic, Fluvaquentic Endoaquoll) at the University of Wisconsin O.J. Noer Turfgrass Research and Education Facility in Verona, WI. The soil had a 7.1 pH and contained 82 mg kg–1 P and 197 mg kg–1 K (values represent plant-available P and K as determined by the Bray method). Soil bulk density averages were 6.58 and 6.53 g cm–3 for turf and prairie vegetation with a 0.20 and 0.12 standard deviation, respectively. Glyphosate [N-(phosphonmethyl)glycine] was applied according to label recommendations on 28 Aug. 2002 to eliminate existing turf vegetation from all plot areas. Plots measuring 2.44 m in width by 9.75 m in length were stripped of all existing turf vegetation on 3 Sept. 2002 and individually graded by hand to a uniform slope of 5.5 ± 0.2%. Impervious surfaces were then installed at the upslope end of each plot to provide 1:1, 1:2, or 1:4 impervious/pervious surface ratios. The impervious surfaces consisted of poured concrete (465 mix, Wingra Cement Co., Madison, WI) 10.2 cm thick that was placed between 27 Sept. and 9 Oct. 2002. Concrete was allowed to set for 28 d before traffic. To achieve a smooth, continuous surface from head to bottom slope, 10.2 cm of topsoil was first removed before the concrete was poured. Each plot was bounded by heavy duty contractor-grade plastic landscape edging (CobraCo Manufacturing Inc., Lake Zurich, IL) measuring 12.7 cm in height and set vertically 10.2 cm deep into the soil to prevent surface water flow between plots and to direct flow toward the runoff collectors.

Runoff collectors were constructed from galvanized steel following the design of Brakensiek et al. (1979). Collectors were positioned along the down slope end of individual runoff plots and clamped to steel cutoff plates driven vertically into the soil flush with the soil surface as pictured in Stier et al. (2005). Landscape edging was fitted to the runoff collectors and silicone cement used to form a water-tight seal between the edging and collection trough. Collection troughs channeled runoff water into galvanized steel three-way sample splitters. Two of the three channels were allowed to empty into a layer of pea gravel and exit through drain tile placed beneath the runoff collection bins. Water in the center channel flowed into a 121-L plastic runoff collection container placed into the ground and above the layer of pea gravel and drain tile. A 15-L plastic bin placed within the larger collection bin collected water from smaller runoff events. The 121-L collection bins were held in place with two 1.2 m long pieces of steel rebar to prevent frost heaving. Fiberglass screening was wired onto the collection troughs to prevent runoff water contamination from organic debris and vermin. A 2.44-m section of galvanized sheet metal was clipped into place at a 30° angle from horizontal on top of the collection troughs to prevent rain and snow from falling directly into the collection troughs.

The experimental design was a randomized complete block with three replications of a 2 by 3 factorial treatment arrangement. Treatments randomly assigned to the plots were: (i) concrete + turfgrass buffer strip, 1:1 ratio; (ii) concrete + prairie buffer strip, 1:1 ratio; (iii) concrete + turfgrass buffer strip, 1:2 ratio; (iv) concrete + prairie buffer strip 1:2 ratio; (v) concrete + turfgrass buffer strip, 1:4 ratio; (vi) concrete + prairie buffer strip, 1:4 ratio.

Glyphosate was applied to all plots on 5 Oct. 2002 to eliminate weed regrowth and to prepare the seed bed. Dormant seeding of the plots occurred between 4 Nov. and 8 Nov. 2002. In forb-dominated prairie seed mixes, autumn prairie seeding encourages earlier germination and plant development (Prairie Nursery, 2005). Because dormant seeding results in lower germination rates than non-dormant seeding in turfgrass, both prairie and turf plantings were seeded at twice their recommended rates. Plots were seeded using a shaker jar method. The seeds were mixed with sand to uniformly distribute the seed. Prairie plots were seeded at a rate of 22.2 kg ha–1 with a shortgrass, medium soil mixture (Table 1) (Prairie Nursery Inc., Westfield, WI, Stock # 50002). Turf plots were seeded at a rate of 195.1 kg ha–1 with equal percentages of ‘Odyssey,’ ‘Arcadia,’ ‘Cynthia’, ‘Cannon’, and ‘Showcase’ Kentucky bluegrass (Poa pratensis L.) (Olds Seed Solutions, Madison, WI). All plots were covered with biodegradable Futerra wood-fiber mulch blankets to prevent soil and seed erosion over the winter months (Profile Products LLC, Buffalo Grove, IL).


View this table:
[in this window]
[in a new window]

  		Table 1. Prairie plant composition of a commonly available shortgrass, medium soil prairie mixture from Prairie Nursery, Westfield, WI. Vegetation type categories are comprised of legumes (L), grasses (G), and forbs (F).

 
Per standard recommendations, prairie plantings received no fertilizer during the study (Prairie Nursery, 2005). Turf plots were fertilized at recommended rates (Stier, 2001) with 48 kg N ha–1 during May, September, and November of each year using a 21–1.3–10 (N-P-K) granular fertilizer (8.0% water-insoluble N, 9.9% slowly available water-soluble N, 2.5% urea N, and 0.6% ammoniacal N, Spring Valley, Jackson, WI). At this fertilizer rate, the monoammonium phosphate fertilizer resulted in three applications of 3.06 kg P ha–1 yr–1. A drop-spreader was used to confine fertilizer to the turf plots. All plots showed signs of germination by 23 Apr. 2003. Prairie plantings were trimmed to a height of 20.3 cm on 16 June, 15 July, and 16 Aug. 2003 and once at 30.5 cm in June 2004 to reduce weed encroachment (Prairie Nursery, 2005). Burning is often prohibited in urban areas and mowing has been identified as a viable alternative to burning in the maintenance of upper Midwest prairies (Diboll, 1986), thus prairie plots were mowed to 1.3 cm in April 2004 and 2005 to simulate controlled burns. All prairie mowings were conducted according to recommendations from the seed supplier (Prairie Nursery, 2005), with residue removed from the soil surface. Turf plots were mowed as needed with a rotary mulching mower with clippings left in place to maintain a height of 6.4 cm without removing more than one-third of the leaf tissue at any one time (Stier, 2001). No irrigation was applied to prairie or turf during the study. Winter snow accumulation was shoveled from the impervious surfaces onto the buffer strips to simulate an urban environment. The delineation between frozen and non-frozen soil conditions was determined according to soil temperature probes that were placed at a 5-cm depth in close proximity to the study (Campbell Scientific, Logan, UT).

Runoff water samples were collected with sufficient redundancy during events to prevent overflow within 12 h of all rain or snowmelt events and total volumes recorded. A tipping bucket collected precipitation data which was stored in a datalogger (Campbell 21X, Campbell Scientific, Logan, UT). Runoff water was manually mixed to resuspend sediment and organic matter and 250-mL samples removed and stored at –10°C until analyzed.

Phosphorus water analysis required two steps: (i) converting the P form of interest to solution orthophosphate (H2PO4– or HPO42–) through acid hydrolysis, and (ii) the colorimetric determination of solution orthophosphate (Greenberg et al., 1985). Runoff samples were analyzed for TP using a modified version of the ammonium persulfate and sulfuric acid digestion method (USEPA, 1993). In place of autoclaving, 10-mL aliquots were placed into 30 mL Teflon digestion vials and heated at 120°C for 4 h. Runoff samples analyzed by both methods produced similar results (data not shown). Upon cooling, samples were filtered through Whatman #40 filter paper and transferred to scintillation vials. Samples were then analyzed for orthophosphate using the ascorbic acid method to determine TP (Murphy and Riley, 1962). Soluble phosphorus was determined using 10-mL aliquots centrifuged for 15 min at 2500 rpm (1326 x g). Centrifugation removed solids >0.45 µm more efficiently than filtration through 0.45-µm filters, and experimentation between the two methods found no difference in solution P. A 5-mL aliquot from the centrifuged SP sample was heated to 500°C in an ashing furnace with the residue being analyzed according to the ascorbic acid method (Murphy and Riley, 1962). The orthophosphate value obtained from the SP sample was subtracted from the ashed residue orthophosphate value to determine the soluble organic phosphorus (SOP) concentration. Bioavailable phosphorus was determined using the iron oxide paper strip method of shaking 50 mL of runoff sample with one 55-mm diam. filter strip for 16 h (Sharpley, 1993). Filter strips were analyzed with the methods of Murphy and Riley (1962). Total suspended solids (TSS) were determined by evaporating 50 mL of runoff sample in a tared beaker for 12 h at 105°C and re-weighing to determine the TSS content. The dried residue was heated to 500°C in a muffle furnace, and weight reduction assumed to be a measure of organic suspended solids content.

Biomass was collected each autumn before snowfall and each spring after snowmelt to determine dry matter and changes in N, P, and K nutrient concentrations for aboveground vegetation. Samples were collected: 10 Oct. 2003, 9 Mar. and 6 Oct. 2004, 2 Apr. and 18 Oct. in 2005, and 29 Mar. 2006. Sampling was in accord with the methods described by Diboll (1986) and Rogers and Hartnett (2001). One 0.1 m2 sample ring was randomly tossed into each plot and all vegetation removed to the soil surface. Samples were oven-dried at 60°C for 48 h, and weighed and ground to pass through a 20-mesh sieve. Dry matter was analyzed for total N by the Kjeldahl digestion procedure (Bremner, 1996; Jones, 2001). Two hundred mg samples were dissolved in sulfuric acid and digestion mixture and heated to 350°C on a digestion block for 60 min after solution clearing. Ammonium was determined through steam distillation (Bremner, 1996). One hundred mg samples were analyzed for P and K by weighing into 50-mL beakers and heating for 2 h at 500°C in a muffle furnace. The ash was dissolved in 3 mL of 2 N HCl and diluted with deionized water to determine P and K concentrations. Phosphorus concentration was determined by the vanadate-molybdate-yellow method (Chapman and Pratt, 1961). Absorbance was measured at 440 nm using a Spectronic 21D spectrophotometer (Milton-Roy, Rochester, NY). Potassium detection was completed by flame photometry (Buck Scientific, East Norwalk, CT, Model PFP-7). Standard curves were developed from 500 mL of potassium phosphate (K2HPO4) primary standard. Secondary standards were produced using 0-, 1-, 2-, 3-, 4-, and 5-mL aliquots of primary standard and diluted in a 100-mL volumetric flask. Aliquots of 2.5 mL were removed from the secondary standards, mixed with 0.5 mL Murphy-Riley solution B, and analyzed for orthophosphate on the spectrophotometer to determine the final standard curve.

Soil samples (2.0 cm diam. x 1.3 cm depth) were collected below the thatch layer at the rate of three plugs every 1.5 m2 from all buffer strips on 16 June 2003, 3 June 2004, and 2 June 2005 and analyzed for pH, organic matter, and soil test P (Bray #1) and K at the University of Wisconsin Soil and Plant Analysis Laboratory. Prairie and turf plots were visually rated monthly for ground cover (0–100% ground cover) and categorized as % intended species, % weeds, and % bare soil. Cover was placed into one of six categories: <1%, 1 to 4%, 4 to 10%, 10 to 25%, 25 to 50%, and 50 to 100%. At less than 10% cover, the median value within the category was used because 1 to 4% cannot be distinguished (Leps et al., 2001). Starting with 10%, a precise percent cover was determined.

Analysis of variance was used to separate the means of runoff volumes, all forms of P loading, sediment loading, plant dry matter yield, plant nutrient concentrations, soil test P, organic matter, pH, density measurements, and vegetative cover ratings (Statistix, 2003) (Analytical Software, Tallahassee, FL). A General Linear Model was used to determine if year x vegetation or year x buffer ratio interactions existed. Since year interactions were significant, data are presented for each year. When appropriate, significant differences among treatment means were determined using Fisher's protected least significant difference (LSD) test.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Runoff Volume Measurements
Impervious/Pervious Ratio
The impervious/pervious surface ratio significantly affected runoff volumes during non-frozen soil conditions in 2004 (P = 0.04) but not from frozen soils (Table 2). Eleven of the 15 runoff events in 2004 occurred during frozen soil conditions with the remainder occurring during the growing season. Nine of the 11 winter (December through March) runoff events were due to snowmelt over frozen soil whereas the two remaining events were due to rainfall events over frozen, non-snow covered soil. The two winter rainfall events contributed 13, 17, and 5% of the total winter runoff volume for the 1:1, 1:2, and 1:4 impervious/pervious ratios, respectively (data not shown). Runoff volumes in 2004 were significantly greater from the 1:1 surface ratios compared with the 1:2 and 1:4 ratios during non-frozen soil conditions. During the growing season, the 1:1 treatment ratios allowed 2.7 and 3.6 times more runoff than the 1:2 and 1:4 surface ratios, respectively. Annual winter runoff accounted for 87, 91, and 96% of the total 2004 runoff load for the 1:1, 1:2, and 1:4 ratios, respectively. This resulted in approximately seven to 23 times more runoff occurring from all buffer ratios during the 2004 winter season when compared with the 2004 growing season.


View this table:
[in this window]
[in a new window]

  		Table 2. Effects of buffer ratio and vegetation type on seasonal and annual runoff volumes 2004 through 2005, Madison, WI.

 
Surface type ratios did not influence frozen soil, non-frozen soil, or cumulative runoff volumes in 2005. Runoff volumes were four times greater in 2005 compared with 2004 despite 1.7 times more precipitation and five additional rainfall events greater than 25.4 mm in 2004 (Fig. 1). In spite of the greater runoff volumes in 2005 compared with 2004 (Table 2), no surface water runoff occurred during the 2005 growing season (Table 2). All 12 winter runoff events in 2005 occurred between January and March in the form of snowmelt or rainfall over frozen soil. Peak winter runoff volumes in 2005 occurred during a period where above average temperatures happened to coincide with above average precipitation resulting in rainfall over frozen soil (Table 3). These conditions produced a precipitation rate that exceeded the soils infiltration capabilities as discussed by Petrovic (1990). The majority of this precipitation occurred in January 2005 as rainfall over a 15.3 to 22.9 cm snow cover. Two of the twelve runoff events resulted from rainfall over snow-covered ground in January and accounted for 24, 23, and 22% of the total annual volumes for the 1:1, 1:4, and 1:2 buffer ratios, respectively (data not shown).


Figure 1
View larger version (16K):
[in this window]
[in a new window]

  		Fig. 1. Rain and snowfall precipitation data at the O.J. Noer Research Facility, Madison, WI, 2004 through 2005.

 

View this table:
[in this window]
[in a new window]

  		Table 3. Weather data displaying departures of monthly temperature and precipitation from 30-yr averages, Madison, WI, 2004 through 2005 (National Climatic Data Center, Asheville, NC).

 
Vegetation Type
Runoff depths differed between vegetation types only during non-frozen soil conditions in 2004. In 2004, runoff occurred four times (23 May, 31 May, 4 August, and 1 September) during non-frozen soil conditions. Turf runoff combined over these four dates was 54% less (P = 0.08) compared with prairie (Table 2). Only 5% of runoff from turf occurred during non-frozen soil conditions compared with 17% of the annual runoff from prairie. On an annual basis, however, the effect of vegetation type was negligible as >80% of runoff occurred during frozen soil conditions during which vegetation type had no effect (Table 2). Runoff depths in 2004 ranged from 1.3 to 21.5 mm for frozen soil and 0.0 to 7.0 mm for non-frozen soil (Fig. 2).


Figure 2
View larger version (31K):
[in this window]
[in a new window]

  		Fig. 2. Vegetation treatment effects on cumulative total P (TP) load and mean annual runoff volumes 2004 through 2005, Madison, WI. Mean runoff volumes were significantly different (P < 0.05) on 20 Feb. and 28 Feb. 2004 and 7, 16, and 19 Mar. 2005. Asterisks indicate significant treatment differences (P = 0.05) in mean TP load on specified dates. Soil freezing (5 cm) occurred on 12 Dec. 2003 and 14 Dec. 2004. Soil thawing took place on 24 Mar. 2004 and 28 Mar. 2005.

 
During frozen soil conditions vegetation type did not influence runoff in 2005 and no runoff events occurred outside of this time period. Runoff depths in 2005 ranged from 3.7 mm to 64.7 mm (Fig. 2) resulting in 284 to 302 mm of total runoff for prairie and turf, respectively. Despite no growing season runoff events, runoff volumes in 2005 were fourfold greater than 2004.

Runoff Water Phosphorus Characterization
No significant vegetation type by buffer ratio interactions on P loading occurred throughout the study. Buffer ratio main effects were significant during non-frozen soil conditions in 2004 for the TP, SP, and BAP fractions. The 1:1 surface ratios resulted in a 3.5-fold greater TP, SP, and BAP load than the 1:4 ratios during 2004 non-frozen soil conditions. No significant differences occurred between the 1:2 ratios compared with the 1:1 and 1:4 surface ratios. Total P losses from prairie were 4 times greater than turf during the 2004 growing season, but the growing season accounted for only 10% of the TP losses for the year (P = 0.02; Table 4). The majority of TP losses occurred during frozen soil (winter) conditions and were not significantly different between vegetation types. No significant treatment effects on annual TP losses occurred in 2005 despite greater TP loads from turf plots on nine of 12 runoff events, of which four events were statistically significant (P ≤ 0.05; Fig. 2). Cumulative TP losses were 3.8 and 5.4 times greater in 2005 vs. 2004 for prairie and turf, respectively. This effect was primarily due to increased TP losses from three runoff events during frozen soil conditions in 2005 compared with one large TP loss in 2004. The 23 Feb. 2004 snowmelt runoff event accounted for 36 and 32% of the winter and mean annual TP nutrient load, respectively, for the prairie treatment and for 36% of the winter and mean annual TP nutrient load for the turf treatment. Three large rainfall or snowmelt-induced TP nutrient flushes occurred on 12 Jan., 5 Feb., and 6 Mar. 2005 accounting for 66 and 40% of the respective prairie and turf annual TP loading in 2005.


View this table:
[in this window]
[in a new window]

  		Table 4. Effects of vegetation selection on runoff water P loads from frozen soil (FS) and non-frozen soil (NFS) conditions during 2004–2005, Madison, WI.

 
Soluble P loading in runoff followed the trends of TP loading with fourfold greater SP losses from prairie compared with turf during the 2004 growing season (P = 0.01; Table 4). However, these loads were minor as they only represented 6.6 and 2.0% of the cumulative 2004 SP load for the prairie and turf vegetations, respectively. Regardless of plant community, SP represented between 63 and 70% of the TP load for both the winter runoff and cumulative nutrient loads across both years. Soluble organic P generally represented less than 10% of the SP load regardless of vegetation type or season of runoff (Table 4). Bioavailable P represented a consistent portion of TP within both winter (61–64%) and runoff during the growing season (47%) for both vegetative treatments. Bioavailable P losses followed a similar pattern to other P forms with BAP in snowmelt/frozen ground runoff exceeding growing season rainfall BAP by 12 and 50 times for prairie and turf, respectively.

Vegetative Production and Nutrient Content
Turf plots had significantly less bare soil than prairie plots during 2004 and 2005 (Table 5). In 2004 through 2005, turf density was usually 78% or greater. Turf cover had relatively little seasonal fluctuation while prairie plant density ranged from 4 to 42%, with least density in early spring and maximum density in summer. Weeds covered much of the surface area in prairie plots, so that >50% of the surface was always covered with vegetation. Weeds were primarily annuals such as Chenopodium album L., Kochia scoparia L., and Erigeron annuus L. Weed cover in prairie plots averaged 48% in 2004 and 32% in 2005, whereas weed cover in turf rose from a mean of 4.5% in 2004 to 19% in 2005.


View this table:
[in this window]
[in a new window]

  		Table 5. Qualitative vegetative cover ratings according to vegetation type, 2004 through 2005, Madison, WI.{dagger}

 
Prairie averaged three to five times greater aboveground biomass in the autumn and two times greater biomass in the spring as compared with turf (P ≤ 0.01; Table 6). Biomass decreased 48 and 46% from autumn to spring for prairie during winter 2003 through 2004 (–803 kg ha–1) and 2004 through 2005 (–2435 kg ha–1), compared with a 17% decrease (–89 kg ha–1) and 35% increase (395 kg ha–1) in turf biomass during these time periods (Tables 6, 7).


View this table:
[in this window]
[in a new window]

  		Table 6. Vegetation type dry matter yield, nutrient concentrations, and nutrient loads in the autumn and spring of 2003 through 2006, Madison, WI.

 

View this table:
[in this window]
[in a new window]

  		Table 7. Changes in biomass, nutrient content, and runoff water losses from vegetated buffer strips between autumn and spring 2003 through 2005, Madison, WI.

 
Prairie and turf vegetation both contained between 0.2 and 0.4% P on a dry weight basis in autumn (Table 6). Phosphorus concentration in prairie vegetation decreased by 42 and 67% between the winter and spring of 2003 through 2004 and 2004 through 2005, respectively, whereas turf P concentration decreased by 49 and 5%. Based on aboveground biomass, prairie held an estimated four times greater P amount in foliage compared with turf on a land area basis in autumn but turf had greater amounts by spring. Aboveground prairie biomass decay during the winter resulted in P losses of approximately 3 and 12 kg ha–1 in both years (Table 7). Turf foliage lost an estimated 0.8 kg P ha–1 during the first winter and gained 0.6 kg P ha–1 in the second year due to increased biomass production between autumn and the spring sampling.

Beginning in 2004, 1 yr after establishment, nitrogen concentrations in fertilized turf foliage were always significantly (P ≤ 0.01) greater than the nonfertilized prairie, averaging 2.3% compared with 1.3%. Nitrogen amounts on a land area basis varied by season with prairie vegetation containing greater amounts of N compared with turf in two of three autumnal seasons but similar or lesser amounts by spring (P ≤ 0.01; Table 6). Prairie vegetation lost an average of 28.3 kg N ha–1 between autumn and spring compared with a loss of 2.0 kg N ha–1 in turf (Tables 6, 7). Turf N losses on a land area basis decreased 17%, increased 74%, and decreased 77% during the winters of 2003 through 2004, 2004 through 2005, and 2005 through 2006, respectively. Prairie N losses on a land area basis decreased 64, 59, and 81% during the same period.

Turf maintained higher K concentrations in spring compared with prairie (P ≤ 0.01). No significant differences in K concentrations were evident during autumn (Table 6). The amount of K on a land area basis was greater in prairie vegetation during the autumn season but the reverse was true in spring with turf vegetation containing K loads equal to or greater than prairie.

Soil Characteristics
According to University of Wisconsin-Madison soil test recommendation guidelines (University of Wisconsin Soil and Plant Analysis Laboratory, 2006), soil test P values remained excessively high (>20 mg kg–1) for turf throughout the study (Table 8). Turfgrass soil P levels were significantly greater than prairie in 2003 and 2004 but not in 2005. Soil test P levels fluctuated over time in both systems, decreasing approximately 12 to 14% from 2003 to 2004 then increasing approximately 3 to 7% between 2004 and 2005 (Table 8). Soil pH values also fluctuated, increasing from 2003 through 2004 then declining by 2005. Prairie had statistically greater pH values than turf though the differences (0.05 to 0.15) were close to neutral (7.0) and likely not biologically important (Table 8). Organic matter was similar for turf and prairie systems on each date though it rose steadily from 3.3 to 3.9% for prairie and 3.4 to 4.0% for turf between 2003 and 2005 (Table 8).


View this table:
[in this window]
[in a new window]

  		Table 8. Vegetation type effects on soil test P, organic matter, and pH values (0- to 1.3-cm depth) 2003 through 2005, Madison, WI.

 

   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
CONCLUSIONS
 REFERENCES
 
Runoff Volumes
Over 80% of the annual runoff occurred during the winter because the frozen ground conditions caused vegetated surfaces to act as an impervious surface. Both snowmelt and rainfall over frozen ground were the primary runoff sources in 2004 with rainfall over snow-covered ground the primary runoff source in 2005. In 2004, episodes of above-freezing temperatures melted snow cover earlier in the winter runoff period than in 2005 and may have allowed some precipitation to infiltrate the thawing soil of the buffer strips. Few natural runoff studies have documented complete 12-mo runoff cycles where soil freezing can impact runoff volumes and be a dominant component of the hydrologic cycle (White and Williamson, 1973; Timmons and Holt, 1977; Easton and Petrovic, 2004). The type (solid or liquid) and timing of precipitation (frozen vs. non-frozen soil) along with the runoff source (snowmelt or rainfall) influenced the amount of runoff more than total precipitation as winter precipitation amounts were identical in both years (15 mm) yet over four times more runoff occurred in winter 2005 (24.4 mm) compared with winter 2004 (4.8 mm) (Fig. 2). Previous results from separate turf and prairie studies showed frozen conditions to be the major contributor to annual runoff volumes. Timmons and Holt (1977) reported 80% of annual runoff occurred due to snowmelt on a native prairie area with 5% slope in west-central Minnesota. In New York, 62% of annual runoff from a mowed lawn-type turf on a 7 to 9% slope occurred during snowmelt or runoff over frozen soils (Easton and Petrovic, 2004).

Previous studies have indicated the importance of buffer length in controlling runoff pollution (Dillaha et al., 1987; Magette et al., 1989; Patty et al., 1997; Abu-Zreig et al., 2003). However, these data were collected in the absence of frozen soil conditions. In areas subject to soil freezing, runoff models using data from non-frozen soil conditions may be inaccurate because winter runoff dominated in our study and overwhelmed the effects of vegetation type and buffer strip size during the growing season. Data demonstrating the importance of the type and timing of winter precipitation compared with snowmelt as a water source for runoff are needed to refine runoff models and create best management practices to reduce urban runoff.

Individual runoff events are often not significant due to high variability of the data (Ebeling et al., 2002). Runoff over non-frozen ground was highly variable (Table 2) and thus runoff was separated into non-frozen and frozen soil runoff periods. The coefficients of variation (CV) for cumulative and frozen soil runoff volumes were between 50 and 67%. The CV for all non-frozen soil data was commonly greater than 80%, thus demonstrating the high variability in individual runoff events. Combining data into FS and NFS achieved a normal distribution and highlighted the importance of winter runoff. As anticipated, runoff volumes during non-frozen soil conditions decreased as the proportion of pervious/impervious surface area increased from 1:1 to 4:1. Though the increase in pervious area concomitantly decreased, the impervious surface area causing less run-on water from the impervious surfaces, our objective was to study the ratio of impervious/pervious surfaces within a given land area. Our results agreed with Arnold and Gibbons (1996) who also reported that runoff totals increased in proportion to the cover of impervious surfaces. The lack of runoff during the 2005 growing season was primarily due to below average precipitation from June through October (Table 3). Abnormally large precipitation amounts (including one episode of 102 mm within 24 h) in May 2004 (Table 3) caused soil saturation which resulted in two runoff events during this time period.

The size of urban buffer strips may depend on a combination of soil type, percentage impervious surface, slope, and size of area to be buffered. Buffer ratios may be of greater importance in milder climates where runoff during non-frozen conditions dominates the hydrologic cycle. The current investigation was the first known study to directly compare runoff from prairie and turfgrass ecosystems. The closest similar study, conducted by Erickson et al. (2001), revealed no significant differences in surface runoff between a predominantly native mixed-species landscape and turf of St. Augustinegrass (Stenotaphrum secundatum (Walt.) Kuntze). Timmons and Holt (1977) found runoff volumes from a native Minnesota prairie ranged from 2 to 80 mm per year over a 5-yr period. Assessments of six watersheds composed of native grassland in Oklahoma over 4 yr showed mean annual runoff ranged from 7 to 154 mm though these represented a range of soil types, slopes, and other conditions (Sharpley et al., 1992). In our study values for mean prairie runoff were 67 mm in 2004, but a high of 284 mm in 2005 was due to intense winter rainfall over snow. In turf situations, Easton and Petrovic (2004) reported 68 mm of runoff the first 18 mo after seeding KBG and mowing at lawn height. Gross et al. (1990) recorded a runoff range of 1 to 14 mm ha–1 from a mixed stand of KBG and tall fescue (Festuca arundinacea Schreb.) on a 5 to 7% slope maintained at lawn height. Turfgrass reduced the runoff volume compared with prairie vegetation during non-frozen conditions in 2004 by rapidly establishing a dense vegetative cover. Dense turf vegetation reduces runoff by creating "tortuous" pathways which reduce runoff rate and enhance infiltration (Linde et al., 1995). Additionally, evapotranspiration of KBG turf removes water from the upper soil surface providing pore space for water infiltration to reduce runoff (Ebdon et al., 1999). Annual weed growth during prairie establishment likely played a similar role, preventing even vaster differences between turf and prairie plots. Our results indicate that vegetation selection between prairie and turfgrass may play a significant role in mitigating surface water runoff during the growing season when precipitation occurs over non-frozen soil during the first 3 yr of buffer strip establishment. However, the enhanced runoff volumes that occurred over the winter period overrode any significant differences in runoff that occurred during the growing season. Therefore, vegetation selection between prairie and turfgrass may be a moot point for mitigating runoff volumes for the first 3 yr after buffer installation in regions where soil freezing and snowfall commonly occur for long periods.

Nutrient Losses
Potential P sources in runoff include sediment, atmospheric deposition, and vegetative losses. Sediment and TP losses in runoff from conventional row crop systems usually exceed those of perennial (unmowed) grass systems, with particulate P the main P form (Sharpley et al., 1992). Mean annual atmospheric P deposition in the U.S. was estimated at 0.4 kg ha–1 in 1988 (United Nations Environment Programme, 1999) though the amount even within a watershed can vary greatly (Cole et al., 1990). In urban situations both impervious and pervious surfaces can contribute to runoff. Pervious (vegetated) surfaces allow runoff when precipitation rate exceeds infiltration capacity which primarily occurs when soil moisture exceeds field capacity or when the ground is frozen. Soluble P is of primary importance, with sediment and particulate P losses substantive only during land development in which case up to 20 times more soil erosion may occur compared with agricultural land use (Daniel et al., 1979).

Studies have implicated the importance of buffer width in removing both P and sediment from runoff waters. Phosphorus removal in runoff has been shown to have a linear relationship with buffer strip width when the primary source of P was sediment during non-frozen conditions (Abu-Zreig et al., 2003). According to Lim et al. (1998), 75% of incoming TP, ortho-P, and TSS were removed within the first 6.1 m or 14.6 m2 of vegetative filter strip when buffering manured cropland area 29.3 m2 in size on a 3% slope. When buffering an impervious surface, our data showed a 4:1 pervious/impervious ratio on a 5.5% slope was needed to significantly reduce TP, SP, and BAP compared with a 1:1 ratio during non-frozen conditions. Differences in P runoff among buffer widths were inconsequential during frozen conditions.

Runoff losses of nutrients and sediment in our study represent values that could be expected from a fertilized turf and a nonfertilized prairie within an urban environment. The mean annual quantities of nutrient and sediment losses were similar between both vegetation types though it differed between years as a function of precipitation event rate and timing. Sediment loss from prairie tended to be greatest when foliage was dead/dormant resulting in reduced aboveground biomass. Increased runoff volumes during the winter of 2005 were responsible for the larger nutrient loads compared with 2004. Soluble P comprised the majority of P in runoff as expected since vegetative cover inhibited soil erosion and corresponding particulate P (Olness et al., 1975; Nickolaichuk and Read, 1978; Gross et al., 1990; Sharpley et al., 1992).

In our study TP and SP losses from turf during non-frozen conditions were within the ranges reported by Gross et al. (1990) and TSS were less than those reported during non-frozen conditions (6.4 x 10–3 to .033 kg ha–1 TP, 3.1 x 10–3 to 0.02 kg ha–1 SP, and 5.5 to 14.4 kg ha–1 TSS). Runoff from frozen soil conditions caused mean annual losses of TP, SP, and TSS to be one or two orders greater than those reported by Gross et al. (1990), though similar to the 0.2 to 1.3 kg SP ha–1 reported from turf subjected to freezing in New York (Easton and Petrovic, 2004). Prairie TP, SP, and TSS losses were of similar magnitude to those reported for mature native prairie (White and Williamson, 1973; Timmons and Holt, 1977).

In climates subjected to freezing conditions the predominant time for P losses in runoff from both turf and prairie occur during frozen soil conditions/snowmelt (Timmons and Holt, 1977; Easton and Petrovic, 2004). Freeze-thaw events increase the amount of SP loss from vegetation as cell lysis occurs in susceptible vegetation (Bechmann et al., 2005). The C-4 plants which dominate prairie mixtures may be especially subject to freeze-thaw nutrient losses as their foliage dies in autumn with the onset of cooler temperatures while leaves of relatively cold-tolerant KBG will often stay alive during the winter, particularly if snow-covered. Our prairie biomass decreased by approximately 45 to 50% between autumn and winter of each year, presumably releasing 3 kg P ha–1 the winter of 2003 through 2004 and nearly 12 kg P ha–1 the following year. Mean prairie litterfall and plant N, P, and K concentrations in our study were in their expected ranges according to Seastedt (1988) and Chapman and Pratt (1961). Turfgrass mean tissue N, P, and K concentrations were slightly lower than their expected tissue levels (Beard, 2002) and may be due to our moderate N rates used as N fertility drives P uptake (Barber, 1995). Turf aboveground biomass ranged between 20 and 50% that of prairie throughout the study but only declined 17% the winter of 2003 through 2004 and actually increased 34% between autumn and early spring the second year. KBG resumes growth in early to mid-spring, at least 4 to 8 wk before C-4 prairie vegetation renews growth, which accounts for the increased biomass in early spring of 2005. Although KBG had greater nutrient concentrations than prairie foliage, calculated nutrient losses were less than prairie as mowing restricted KBG aboveground biomass. The amount of aboveground biomass and nutrient concentration may be important for determining baseline levels of sp. as nutrients can be leached from both living and dead vegetation (Timmons et al., 1970; Sharpley, 1981). However, phosphorus recovery in runoff between autumn and spring did not relate to changes in calculated P losses from foliage, indicating that either the P from prairie and turf vegetation was retained in the ecosystem and/or other P sources such as soil or atmospheric deposition masked the contribution from vegetation. Further work is needed to determine potential for P losses from living, green and dead turf and prairie vegetation, and to determine if P is shunted to a belowground sink before foliage going dormant.

Phosphorus loads from both prairie and turf in our study were at least one order of magnitude lower than those observed for agronomic row crops even when those areas were not subjected to freezing conditions (Burwell et al., 1975; Olness et al., 1975). This result was not surprising as agronomic crops such as wheat (Triticum aestivum L.) and alfalfa (Medicago sativa L.), which form more contiguous plant communities similar to prairie and turf, have also demonstrated reduced nutrient contamination from surface runoff compared with row crops which have less vegetative cover (Olness et al., 1975). As expected, soluble P was the primary P form in runoff due to relatively dense vegetative cover in buffer strips of both prairie and turf. Nutrient runoff in our study, particularly from prairie plots, may change over time as perennial prairie cover matures.

Quantities of contaminants in our study were similar to or below those expected from other urban land uses. The Avco Economic Systems Corporation (1970) found SP loads ranging from 1.1 to 3.3 kg ha–1 yr–1 for urban areas in Tulsa, OK with the largest loading occurring in areas of dense tree cover over impervious surfaces. Dorney (1986) found 9% of the TP available in tree leaves was leached over a 2-h test period. This was significant because leaves remain attached to trees and exposed to precipitation for a large portion of the year. Data from Waschbusch et al. (1999) indicated that even moderate tree canopies (<35%) overhanging streets could be important sources of runoff P in urban environments. However, many studies including Bannerman et al. (1993) estimate or model runoff volumes instead of physically collecting them which may lead to faulty conclusions. As pointed out by Leavitt (1998), models can help gain an understanding of natural processes but they are based on simplified assumptions which may not accurately model field conditions or runoff patterns. The nutrient loading data from our study show that when runoff volumes are physically collected and not estimated there was no difference between the turf and prairie plant communities in terms of annual P loading.

Bioavailable P represents P forms that can be used by algae and is an estimation of SP and particulate P coming into solution (Sharpley, 1993). However, runoff BAP was consistently lower than SP during 2004 through 2005, averaging between 91 and 95% of SP. Our results could have been influenced by natural conditions and/or test procedures. Sharpley et al. (1992) noted that runoff from pasture often contains so little sediment that sorption of SP by sediment does not occur causing SP losses to exceed values from fields with relatively greater erosion rates. Ebeling et al. (2002) found SP values to exceed BAP values from a dairy manure-treated corn (Zea mays L.) ecosystem. The researchers speculated that a large portion of the SP reverted to other forms of P not detected by the BAP method during the 7-mo frozen sample storage period in between collection and processing. This may be a function of P adsorption during the sample shaking period, or the shaking process creating newly exposed surfaces for P adsorption. Sharpley et al. (1981) indicated that surface runoff from fertilized agricultural soils could increase SP concentrations by greater than 10% if waters were not filtered within 30 min of collection, though in our study unfertilized prairie plots still produced sp. exceeding BAP. Conversely, Nelson and Romkens (1972) documented a 2 to 27% decrease in SP if runoff samples were not filtered before freezing. The decreased BAP values measured in our study were likely influenced by the small amount of soil erosion from prairie and turf ecosystems compared with agricultural ecosystems. Refinement of the techniques used to estimate BAP may be useful.

Nitrogen is often recognized as an undesirable contaminant in runoff for human health reasons and because it is usually the second most limiting nutrient for surface water eutrophication and total ocean productivity (Tyrrell, 1999). Turf ecosystems are fertilized according to quality, not dry matter yield, and consequently are often N-deficient. The N requirements for turf may be greater than prairie as mowing stimulates growth which requires N. Nitrogen is recycled in both turf and prairie ecosystems. Seastedt (1988) found N inputs to the soil surface from prairie litterfall were approximately 1.94 g N m–2 yr–1. Kaye et al. (2005) reported 11 g N m–2 yr–1 were recycled within a mowed KBG receiving two applications of 5.5 g N m–2 yr–1. In our study turf vegetation had less N loading on a per area basis despite having higher N concentrations. The low N loading was likely due to the lower quantity of aboveground biomass present during the October through March period compared with prairie ecosystems and to living if not actively growing turf during the critical late winter/early spring runoff period.

Since nutrient losses via surface runoff were similar between a nonfertilized prairie and a fertilized turf, fertilizer was either an insignificant source of P in runoff or turf has less inherent P losses from natural sources than prairie. Fertilized turf may have slower runoff initiation and hence lower nutrient loading than forb-dominated prairie, at least during the first few years after planting, due to a greater evapotranspiration rate increasing the soil water holding capacity and increased hydraulic resistance due to a dense turf canopy and thatch production (Linde et al., 1995). Key studies of P runoff from turf have simulated worst-case scenarios by either applying excessive amounts of P and/or artificially forcing runoff immediately following application (Cole et al., 1997; Shuman, 2002). Although P applied to turf in our study was water-soluble monoammonium phosphate, runoff P from water-soluble fertilizers decreases with time after application as P binds to soil or is otherwise utilized. The upper 2.5 cm alone of KBG sod can retain up to 88% of applied organic P due to the dense root system and affinity of P for soil (Vietor et al., 2002). Furthermore our fertilizer applications only coincided with runoff events during May 2004 and P losses were still similar to those from prairie. Since the majority of P lost in runoff occurred during winter in both years, and did not always relate to concomitant losses in vegetation, much of the SP may have been lost from soil solution. Soil solution P maintains a range of 1 to 50 µmol L–1 regardless of soil test P values (Barber, 1995) and surface losses during periods of inactive or reduced plant growth could have occurred. Atmospheric deposition might also have accounted for some of the P found in runoff (Cole et al., 1990; United Nations Environment Programme, 1999).


   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
REFERENCES
 
Urban runoff is often facilitated by conjoining areas of pavement which serve to funnel runoff into drains and often ultimately to surface waters. Buffer strips of mowed and fertilized KBG turf provided similar levels of runoff control plus sediment and nutrient loading compared with forb-dominated prairie vegetation 2 and 3 yr after seeding. Although turf may receive P fertilization, it may not pose any greater risk to environmental surface water quality than a nonfertilized prairie environment. Ultimately, even low-maintenance landscapes can be a source of P in urban environments due to natural and necessary P uptake by leaves and subsequent leaching from vegetation.

Sediment and P losses from both immature prairie and turf vegetation are an order of magnitude less compared with agricultural crops and construction sites. Buffer strips of 4:1 pervious/impervious surface ratio were significantly superior to 1:1 buffer strips for reducing runoff volume, sediment, P, and N loss but only during non-frozen conditions. Most or all of the annual runoff occurred during winter conditions and negated the effects of both vegetation type and buffer size. Our data help quantify the ambient, background levels of P that may naturally occur from prairie and turf areas and should be considered when establishing nutrient and storm water management plans. Urban runoff models need to account for winter runoff when determining potential for nutrient loading in areas subject to freezing and regular snow cover. Urban designs should prevent runoff during winter conditions from directly entering surface waters to reduce nutrient and sediment loading.


   ACKNOWLEDGMENTS
 
This research was supported by the Univ. of Wisconsin College of Agricultural And Life Sciences and Federal Hatch funds (Project 5232).


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 




This article has been cited by other articles:


Home page
 Crop Sci.Home page
K. Steinke, J.C. Stier, and W.R. Kussow
Prairie and Turfgrass Buffer Strips Modify Water Infiltration and Leachate Resulting from Impervious Surface Runoff
Crop Sci., March 17, 2009; 49(2): 658 - 670.
[Abstract] [Full Text] [PDF]

Home page
 Journal of Soil and Water ConservationHome page
D.F. Webber, S.K. Mickelson, T.L. Richard, and H.K. Ahn
Effects of a livestock manure windrow composting site with a fly ash pad surface and vegetative filter strip buffers on sediment, nitrate, and phosphorus losses with runoff
Journal of Soil and Water Conservation, March 1, 2009; 64(2): 163 - 171.
[Abstract] [PDF]

Home page
 Crop Sci.Home page
D. J. Soldat and A. M. Petrovic
The Fate and Transport of Phosphorus in Turfgrass Ecosystems
Crop Sci., November 24, 2008; 48(6): 2051 - 2065.
[Abstract] [Full Text] [PDF]

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 Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Steinke, K.
Right arrow Articles by Thompson, A.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Steinke, K.
Right arrow Articles by Thompson, A.
Agricola
Right arrow Articles by Steinke, K.
Right arrow Articles by Thompson, A.
Related Collections
Right arrow Best Management Practices
Right arrow Phosphorus
Right arrow Turfgrass

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