Phosphorus and Nitrate Nitrogen in Runoff Following Fertilizer by ruf23140


									                             Phosphorus and Nitrate Nitrogen in Runoff Following
                                     Fertilizer Application to Turfgrass
                                                                  L. M. Shuman*

                           ABSTRACT                                                volume, which in turn depends on several factors. One
   Intensively managed golf courses are perceived by the public as                 is rainfall intensity. Chichester (1977) found the greatest
possibly adding nutrients to surface waters via surface transport. An              nitrogen runoff during summer months when rainfall
experiment was designed to determine the transport of nitrate N                    intensities were highest. A second determinant of runoff
and phosphate P from simulated golf course fairways of ‘Tifway’                    volume is soil moisture at the time of the rainfall event.
bermudagrass [Cynodon dactylon (L.) Pers.]. Fertilizer treatments                  Using a portable rainfall simulator, Cole et al. (1997)
were 10–10–10 granular at three rates and rainfall events were simu-
                                                                                   found that seven days after a natural rainfall of 165 mm,
lated at four intervals after treatment (hours after treatment, HAT).
Runoff volume was directly related to simulated rainfall amounts and
                                                                                   the nutrient loss to surface runoff for a simulated event
soil moisture at the time of the event and varied from 24.3 to 43.5%               was 10 to 15% of that applied, whereas for dry soil the
of that added for the 50-mm events and 3.1 to 27.4% for the 25-mm                  loss was only 2% of applied. Likewise, Pote et al. (1999)
events. The highest concentration and mass of phosphorus in runoff                 found that soluble phosphorus in runoff water in August
was during the first simulated rainfall event at 4 HAT with a dramatic             when the soil was moist was nearly twice that in May
decrease at 24 HAT and subsequent events. Nitrate N concentrations                 when the soil was dry. Thus, runoff volume depends to
were low in the runoff water (approximately 0.5 mg L 1) for the first              a great extent on rainfall intensity and the soil moisture
three runoff events and highest (approximately 1–1.5 mg L 1) at 168                at the time of the rainfall event.
HAT due to the time elapsed for conversion of ammonia to nitrate.
Nitrate N mass was highest at the 4 and 24 HAT events and step-
                                                                                      There is a dearth of information about nitrogen and
wise increases with rate were evident at 24 HAT. Total P transported               phosphorus transport by runoff water from turfgrass.
for all events was 15.6 and 13.8% of that added for the two non-zero               The information that is available is not definitive. It has
rates, respectively. Total nitrate N transported was 1.5 and 0.9% of that          been shown that phosphorus loss from perennial pasture
added for the two rates, respectively. Results indicate that turfgrass             in Australia where application of up to 1000 kg ha 1
management should include applying minimum amounts of irrigation                   (1.8 lb P per 1000 ft2) resulted in transport of soluble
after fertilizer application and avoiding application before intense               phosphorus, and not particulate phosphorus (Austin et
rain or when soil is very moist.                                                   al., 1996). Significant losses of nitrogen and phosphorus
                                                                                   above control were reported for turfgrass fertilized with
                                                                                   220 kg N ha 1 (4.5 lb per 1000 ft2), but losses were small
M      any turfgrass areas are intensively managed
        with high inputs of fertilizers, herbicides, and
pesticides. Golf courses and commercial landscapes are
                                                                                   compared with losses from agronomic row crops (Gross
                                                                                   et al., 1990). Nitrate in runoff from soluble fertilizer
                                                                                   sources was related to rates of nitrogen and water vol-
especially highly managed, leading to environmental                                ume applied (Brown et al., 1977). In this study, where
concerns about transport of nutrients to surface waters.                           irrigation was kept near the evapotranspiration rate, the
These inputs may lead to eutrophication (Carpenter et                              loss of nitrate was low.
al., 1998), which is often persistent, and the water is                               Because information on nutrient transport from turf-
slow to recover. Research on the fate of fertilizer nutri-                         grass areas is limited, and nutrient transport to surface
ents from turfgrass areas has been somewhat limited                                waters is a current environmental concern, we con-
(Petrovic, 1990). Research results from studies per-                               ducted runoff trials using field runoff plots.
formed mostly in pasture situations have shown that P
does not leach to a great extent, but is readily carried                                       MATERIALS AND METHODS
to surface waters in drains and through macropores
(Gachter et al., 1998; Hooda et al., 1999; Sims et al.,                               Twelve individual plots (7.0     3.6 m) separated by land-
                                                                                   scape timbers were built in a grid with a 5% slope from the
1998). It has been found that for grassed areas, the loss                          back to the front. The topsoil was a Cecil sandy loam (fine,
mechanism in runoff water is by dissolved phosphorus,                              kaolinitic, thermic Typic Kanhapludult) that has a mixed sur-
whereas for cropland it is more by movement with soil                              face horizon (70.2, 18.1, and 11.8% sand, silt, and clay, respec-
particles in the adsorbed state (Sharpley et al., 1994).                           tively; pH 5.8; Mehlich 1–extractable phosphorus 6.15 mg kg 1;
Since grass greatly ameliorates soil erosion, grassed                              total carbon 12 mg kg 1; cation exchange capacity 5.43 cmolc
buffer strips can reduce nitrogen and phosphorus loss                              kg 1; water holding capacity 85 g kg 1). The soil is typical
by runoff from cropland by “catching” the particulate                              of the Piedmont area of the U.S. Southeast. The slope was
forms (Heathwaite et al., 1998).                                                   developed by removing the topsoil, grading the subsoil, and
   Nutrient transport is highly dependent on the runoff                            returning the topsoil over the area. The plots were sprigged
                                                                                   with ‘Tifway’ bermudagrass, and all plots had complete ground
                                                                                   cover. A trough was installed in a ditch at the front of each
Department of Crop and Soil Sciences, Univ. of Georgia, Griffin Cam-               plot to collect the runoff water in a tipping bucket sample
pus, 1109 Experiment Street, Griffin, GA 30223-1797. Received 31 July              collection apparatus. The tipping bucket tips each time that
2001. *Corresponding author (

Published in J. Environ. Qual. 31:1710–1715 (2002).                                Abbreviations: HAT, hours after treatment.

                                 SHUMAN: PHOSPHORUS AND NITRATE NITROGEN IN TURF RUNOFF                                              1711

2 L of runoff water is collected, tripping a microswitch attached
to a data collecting device that counts the tips. With each tip
a slot between the buckets collects a subsample of the runoff
water in a stainless steel container. Collected water was ana-
lyzed after each simulated rainfall event. Wobbler off-center
rotary action sprinkler heads (Senninger Irrigation, Orlando,
FL) were mounted 7.4 m apart and 3.1 m above the sod surface.
Operated at 138 kPa, the measured simulated rainfall intensity
was 27 mm h 1 (about 1 inch h 1), which is lower than reported
by Smith and Bridges (1996) for this same facility. Simulated
rainfall water had an electrical conductance of 0.116 S m 1,
nitrate N concentration of 0.14 mg L 1, and a phosphate P
concentration of 0.06 mg L 1.
   Fertilizer treatments were 10–10–10 granular at rates to
give 0, 12, and 24 kg N ha 1 (0, 0.25, and 0.5 lb N per 1000
ft2) and rates of 0, 5, and 11 kg P ha 1 (0, 0.11, and 0.22 lb P
per 1000 ft2). The major nitrogen and phosphorus source was         Fig. 1. Runoff volumes for three rates of fertilizer as averages of two
monoammonium phosphate. This fertilizer source was used                years. Bars with the same letters within a time interval are not
because balanced agricultural-grade fertilizers are often ap-          different according to an LSD at the 5% level.
plied to fairways. Treatments were made through the summer
months from April to September each of two years. The fertil-
izer was spread with a calibrated drop spreader for the first       would be the same. However, as can be seen, there was
year and weighed and spread by hand the second year. Each           some variation for the three treatments within a time
rate was added to every plot so that each rate was replicated
                                                                    period and the time periods were also different (statis-
12 times. Rainfall events were simulated at 24 h (25 mm)
before treatment and at 4 (50 mm), 24 (50 mm), 72 (25 mm),
                                                                    tics not shown for the time periods). The 4 HAT vol-
and 168 (25 mm) hours after treatment (HAT). Samples were           umes were lower than for 24 HAT, probably due to
collected after each simulated rainfall event and also for any      the soil being more saturated at 24 HAT. The soil was
natural rainfall events during the course of the experiment.        brought to field capacity the day before treatment, so
Treatments were spaced to allow natural runoff and incorpora-       after 24 HAT the soil had 50 mm simulated rain for
tion into the soil to lower the potential carryover from one        three days in a row. The volumes for 72 and 168 HAT
treatment to the next. Soil moisture was determined before          simulated rainfalls were lower for two reasons. Only
each simulated rainfall event by oven-drying a 7.5-cm-deep          25 mm of simulated rain was applied, and there was
soil core.                                                          time for the soil to dry out. The volumes for the 168
   Subsamples collected from each rainfall event were stored        HAT simulated rainfall were lower than the volumes
at 4 C prior to analysis. Nitrate N and phosphate P were            for the 72 HAT simulated rainfall because of the four-
determined for samples filtered through 0.45- m filters, which
                                                                    day time period to dry out. Some of the variation in
is considered to be the soluble form (Sharpley et al., 1992).
Nitrate N was analyzed colorimetrically with a Lachat (Mil-         volumes came from weather conditions affecting the
waukee, WI) flow analyzer. The instrument first reduces ni-         rate of soil drying. The soil would naturally dry out
trate to nitrite with a copper–cadmium column and the nitrite       more on hot, sunny days than on cool, cloudy days. The
color is developed with a sulfanilamide and N-(1-naphthyl)          experiments were performed during a time span of April
EDTA reagent. The magenta color is read at 520 nm. Phos-            to September, so drying would be different in the spring
phate was also determined colorimetrically (Murphy and              and fall as compared with summer. These simulated
Riley, 1962). The first year the samples were analyzed without      rainfall intensities and amounts were lower than those
the aid of the flow analyzer. We developed the color in 50-mL       used by Cole et al. (1997), who used 51 to 64 mm h 1
volumetrics and measured absorbance with a spectrophotome-          for 75 to 140 min within 24 HAT.
ter at a wavelength of 880 nm. The second year they were               Since soil moisture has an effect on runoff volume,
analyzed with the Lachat flow system.
                                                                    soil moisture was sampled just before each runoff event.
   Means of nitrate N and phosphorus concentration and mass
were separated with analysis of variance with an LSD at the
                                                                    Regression analyses were performed on the averages
5% level of significance. A General Linear Model was used           of the 12 replications for each event for each year (not
to test years to determine if there were interactions. The          averages of years). The data are presented for the 50-
Year      Rate interactions were tested with Rep          Year as   and 25-mm simulated rainfall events separately (Fig. 2).
the error term. If the interaction was significant, then the data   Some of the soil moisture data sets are missing, which
are presented for each year. If it was nonsignificant, then the     accounts for the number of points presented. Most of the
data are presented as an average of the two years.                  data are from the second year. For the 50-mm simulated
                                                                    rainfalls there is a good linear relationship with only
                                                                    one outlier. The R2 value is significant. For the 25-mm
           RESULTS AND DISCUSSION                                   simulated rainfall events, there is a linear pattern, but
   The volumes of runoff water for the various experi-              the points are more scattered. This scatter, and the fact
mental runs were enough alike for the two years statisti-           that there are only six points, led to a nonsignificant R2
cally that they could be averaged (Fig. 1). Ideally, all            value. The longer times between rainfall events allowed
the volumes for 4 and 24 HAT (50-mm simulated rain-                 more variation in soil drying than for the first two rain-
falls) and 72 and 168 HAT (25-mm simulated rainfalls)               fall events. Cole et al. (1997) likewise found that the
1712                                       J. ENVIRON. QUAL., VOL. 31, SEPTEMBER–OCTOBER 2002

                                                                         Fig. 3. Phosphorus concentrations in runoff water for three fertilizer
                                                                            rates at 4 hours after treatment (HAT) and 24 HAT for each of
                                                                            two years. Bars with the same letters within a time interval are
                                                                            not different according to an LSD at the 5% level.

                                                                         These authors used a rainfall intensity of 51 or 64 mm
                                                                         h 1, adding from 63 to 149 mm total within 24 h of
                                                                         adding nutrients. As shown in Fig. 2 there was a direct
                                                                         relation between soil moisture and amount of runoff.
                                                                         Also, the time of year, the temperature, cloud cover,
                                                                         and humidity all play a part in how moist the soil is at
                                                                         the time of the rainfall event. In some cases we received
                                                                         natural rain during the course of an experiment, which
                                                                         caused the later runoff volumes at 72 and 168 HAT to be
                                                                         higher than normal. At the 168 HAT simulated rainfall
Fig. 2. Runoff volume versus soil moisture for 25- and 50-mm simu-       event, the soil was usually dry and the runoff volume
   lated rainfall events.                                                from the 25-mm rainfall was very low. These data indi-
                                                                         cate that dry soil conditions when applying fertilizer are
higher the antecedent soil moisture, the higher the run-                 beneficial for reducing runoff volume, and thus nutrient
off volumes from simulated rainfall.                                     transport from turf areas.
  The amount of water that ran off the plots is expressed                   The phosphorus concentrations in the runoff water
as a percent of that added in Table 1. These data are                    for the first two runoff events are presented in Fig. 3. The
essentially the same as for Fig. 1, but are calculated with              individual years gave slightly different data resulting
the amounts that were theoretically added with each                      in a significant interaction, so the years are presented
simulated rainfall event. These data were not statisti-                  separately. Nonetheless, the general patterns were simi-
cally different for the two years and are presented as                   lar for the two years. At the zero control rate the phos-
averages of the years. For the 50-mm simulated rainfall                  phorus concentrations were about 0.5 to 1 mg L 1 in the
events, there was a high of 43.5% and a low of 24.3%                     runoff water. Even this concentration, if undiluted, could
on soil that was near field capacity. For the 25-mm                      lead to eutrophication, since the threshold is usually given
simulated rainfall events, the range was from 17.7 to                    at less than 0.1 mg L 1 (USEPA, 1976; Schindler, 1977).
27.4%. These variations had much to do with the mois-                    There were step-wise increases in phosphorus concen-
ture status of the soil at the time of the rainfall event.               trations for both 4 and 24 HAT, but concentrations were
In a simulated rainfall experiment with bermudagrass,                    much higher in the water from the first runoff event
when the soil was relatively dry, runoff was 4 to 16%                    than the second. Austin et al. (1996) found a similar
of that applied, whereas when the soil was moist, the                    linear relationship between soluble phosphorus concen-
runoff was 49 to 80% of that added (Cole et al., 1997).                  tration in runoff from flood-irrigated pastures and phos-
                                                                         phorus rates as superphosphate. Phosphorus concentra-
Table 1. Percent of applied water in runoff averaged over two            tions in the water from the second runoff event were
                                                                         still above the zero control level. The second runoff
Rate (N, P)      4 HAT†         24 HAT        72 HAT        168 HAT      event at 24 HAT produced the highest runoff volumes
kg ha  1
                                          %                              (Fig. 1).
0.00, 0.00        24.3b‡         38.2a          17.7b         15.3a         The second two runoff events resulted in much lower
0.25, 0.11        31.5b          36.6a          23.6a          3.1b
0.50, 0.22        35.3a          43.5a          27.4a          5.5b      phosphorus concentrations in the runoff water (Fig. 4).
                                                                         Note that to aid comparisons, the y axis is the same
† HAT, hours after treatment.
‡ Values followed by the same letter within a column are not different   scale as for Fig. 3. The concentrations for treated plots
  according to the LSD at the 5% level.                                  were generally not significantly different from the zero
                                    SHUMAN: PHOSPHORUS AND NITRATE NITROGEN IN TURF RUNOFF                                               1713

Fig. 4. Phosphorus concentrations in runoff water for three fertilizer   Fig. 5. Nitrate N concentrations in runoff water for three fertilizer
   rates at 72 hours after treatment (HAT) and 168 HAT for each             rates at 4 hours after treatment (HAT) and 24 HAT for each of
   of two years. Bars with the same letters within a time interval are      two years. Bars with the same letters within a time interval are
   not different according to an LSD at the 5% level.                       not different according to an LSD at the 5% level.

                                                                         slightly above the zero control plots, but were statisti-
control plots with the exception of the second year for                  cally significant. At the 168 HAT runoff event, the am-
the 72 HAT simulated rainfall event, where the highest                   monia form added in the 10–10–10 fertilizer had seven
phosphorus rate resulted in a concentration level higher                 days to revert to the nitrate form and was thus elevated.
than the zero rate. However, it still was not higher than                The highest rate did not give the highest nitrate N con-
the control and middle rate for the first year for the 168               centrations. There was some variability in this data that
HAT runoff event. Thus, the significance is not believed                 may have led to this unusual result, which was consistent
to be consequential. These results show that the major                   across the two years. Again, the nitrate N concentrations
flush of fertilizer phosphorus would be with initial rain-               were well below the 10 mg L 1 drinking water standard
fall events, and very little treatment fertilizer phos-                  (USEPA, 1976).
phorus would be transported in later rainfall events.                       Another way to look at amounts of nutrients in runoff
A similar exponential decrease in soluble phosphorus                     water is to calculate the mass of nutrient by multiplying
concentrations with successive irrigation events was re-                 the concentrations by the runoff volumes. The phospho-
ported by Austin et al. (1996).                                          rus mass data are not shown, because they are almost
   As for phosphorus concentrations, the nitrate N con-                  the same as the concentration data as far as the pattern
centrations in the runoff water were different for the                   of transport. The mass of nitrate N, however, had a
two years. Nitrate N concentrations were very low for                    different pattern than the concentration (Fig. 7). These
the first two runoff events, and treatments caused values                data were produced by using data in Fig. 1, 5, and 6.
only slightly above the zero rate control plots (Fig. 5).                The mass data were not statistically different for the
The y axis is higher than the data may seem to warrant,                  two years, so the two years were averaged. In these data
so that comparisons with the nitrate N concentrations                    we see increases in nitrate N mass for treatments over
for the second two simulated rainfall events are possible.               control at each simulated rainfall event with the excep-
The treatments did produce high enough nitrate N levels                  tion of 168 HAT, where concentration was the highest.
to be significantly different, but did show some step-
wise increase for all but the 4 HAT simulated rainfall
event for the first year. Nitrate N concentrations in the
runoff water were a bit higher for the 24 HAT than
the 4 HAT runoff event. This is most likely because the
ammonia form had 20 h to be transformed to the nitrate
form. These data show that when the ammonia form of
nitrogen is added, the nitrate concentrations in the run-
off water are very low, and much below the 10 mg
L 1 drinking water standard (USEPA, 1976). Runoff of
nitrate N from Kentucky bluegrass (Poa pratensis L.)
fertilized with urea was lower than for agricultural crops
(Gross et al., 1990). In that study, the nitrogen would
have to have been converted to the nitrate form.
   Contrary to the results for phosphorus concentra-
tions, the nitrate N concentrations were higher in the
                                                                         Fig. 6. Nitrate N concentrations in runoff water for three fertilizer
runoff water for the 72 and 168 HAT runoff events than                      rates at 72 hours after treatment (HAT) and 168 HAT for each
the earlier events (Fig. 6). At the 72 HAT runoff events,                   of two years. Bars with the same letters within a time interval are
the treated plots had nitrate N concentration values just                   not different according to an LSD at the 5% level.
1714                                            J. ENVIRON. QUAL., VOL. 31, SEPTEMBER–OCTOBER 2002

                                                                           1997). These authors also found that buffer strips, buffer
                                                                           mowing height, and length of buffer did not affect
                                                                           amounts of nutrient transport.

                                                                              Although these experiments were performed under
                                                                           somewhat severe conditions where runoff was forced,
                                                                           the results yield some interesting observations about
                                                                           nutrient runoff from turfgrass, and some recommenda-
                                                                           tions can be made. The runoff volume is related to
                                                                           rainfall amounts and soil moisture. Thus, irrigation after
                                                                           fertilization should be held to a minimum and certainly
                                                                           low enough to prevent runoff. Fertilizer should not be
                                                                           applied when soil moisture is near or above field capac-
Fig. 7. Mass of nitrate N in runoff water for three fertilizer rates at    ity and not applied when intense rainfall is expected.
   four times after treatment as averages of two years. Bars with the      Phosphorus concentrations and mass in the runoff water
   same letters within a time interval are not different according to
   an LSD at the 5% level.
                                                                           varied directly with fertilizer rate. Therefore, several
                                                                           small applications of phosphorus throughout the year
                                                                           are preferable to one large application once a year from
There were obvious step-wise increases in nitrate N                        an environmental standpoint. The percent of phospho-
mass with treatment, especially at 24 HAT. Since vol-                      rus of that added was near 14% for the conditions of
umes were not significantly different (Fig. 1) at 24 HAT,                  these experiments for both rates applied. Nitrate N will
these differences were caused by nitrate N. Differences                    initially be low in runoff water when the ammonia form
in nitrate N mass were low among treatments for the                        is applied. This amount increases with time as the ammo-
simulated rainfalls at 72 and 168 HAT. This result may                     nia is converted to nitrate. However, the concentration
have been due to the conversion of ammonia to nitrate.                     of nitrate N in the runoff water for this experiment
The conversion rate may not be the same for each rate                      never exceeded the 10 mg L 1 drinking water standard.
of ammonia nitrogen.                                                       As for phosphorus, it is always a good practice to apply
   The amounts of nitrate N and phosphorus that were                       low amounts of nitrogen of any form to prevent nitrogen
transported from the runoff plots are expressed as per-                    loading of surface waters through transport.
centages of that applied in Table 2. For phosphorus, the
most came at the 4 HAT rainfall event. The amounts                                              ACKNOWLEDGMENTS
as percentages were very similar for the two fertilizer                       The author gratefully acknowledges Ray Pitts, Kathy Ev-
rates, indicating that the amounts in the runoff will be                   ans, and Garland Layton for technical assistance. This research
proportional to that added. The total phosphorus trans-                    as supported by the United States Golf Association, the Geor-
ported was 13.8 and 15.6% for the two rates. The per-                      gia Turfgrass Foundation Trust, and by State and HATCH
centages of nitrate N were, of course, much lower at                       funds allocated to the Georgia Experiment Stations.
around 1 to 1.5% for the totals. These data reflect the
fact that the ammonia form was added, which had to                                                  REFERENCES
be converted before any nitrate N would appear. The                        Austin, N.R., J.B. Prendergast, and M.D. Collins. 1996. Phosphorus
percent nitrate N recovered in the runoff was lower for                      losses in irrigation runoff from fertilized pasture. J. Environ. Qual.
the higher rate for the 4 and 24 HAT runoff events.                          25:63–68.
In a similar runoff experiment where phosphorus and                        Brown, K.W., R.L. Duble, and J.C. Thomas. 1977. Influence of man-
                                                                             agement and season on fate of N applied to golf greens. Agron.
nitrogen each was added at 49 kg ha 1 (1 lb per 1000                         J. 69:667–671.
ft2) as urea or sulfur-coated urea and superphosphate,                     Carpenter, S.R., N.F. Caraco, D.L. Correll, R.W. Howarth, A.N.
the highest percent phosphorus of added recovered in                         Sharpley, and V.H. Smith. 1998. Nonpoint pollution of surface
runoff was 10% and for nitrogen it was 4% (Cole et al.,                      waters with phosphorus and nitrogen. Ecol. Applic. 3:559–568.
                                                                           Chichester, F.W. 1977. Effects of increased fertilizer rates on nitrogen
                                                                             content of runoff and percolate from monolith lysimeters. J. Envi-
Table 2. Percent of applied phosphorus and nitrate N in runoff               ron. Qual. 6:211–217.
  averaged over two years.                                                 Cole, J.T., J.H. Baird, N.T. Basta, R.L. Huhnke, D.E. Storm, G.V.
                                                                             Johnson, M.E. Payton, M.D. Smolen, D.L. Martin, and J.C. Cole.
Rate        4 HAT†       24 HAT          72 HAT       168 HAT    Total       1997. Influence of buffers on pesticide and nutrient runoff from
                                                                             bermudagrass turf. J. Environ. Qual. 26:1589–1598.
kg ha                                    %
                                                                           Gachter, R., J.M. Ngatiah, and C. Stamm. 1998. Transport of phos-
                                Phosphorus                                   phate from soil to surface waters by preferential flow. Environ.
0.11          10.9          3.8               0.8       0.2       15.6       Sci. Technol. 32:1865–1869.
0.22           9.7          3.2               0.7       0.2       13.8     Gross, C.M., J.S. Angle, and M.S. Welterlen. 1990. Nutrient and sedi-
                                  Nitrate N                                  ment losses from turfgrass. J. Environ. Qual. 19:663–668.
                                                                           Heathwaite, A.L., P. Griffiths, and R.J. Parkinson. 1998. Nitrogen and
0.25           0.6          0.6               0.2       0.1        1.5
0.50           0.3          0.4               0.1       0.1        0.9
                                                                             phosphorus in runoff from grassland with buffer strips following
                                                                             application of fertilizers and manures. Soil Use Manage. 14:142–
† HAT, hours after treatment.                                                148.
                                  SHUMAN: PHOSPHORUS AND NITRATE NITROGEN IN TURF RUNOFF                                              1715

Hooda, P.S., M. Moynagh, I.F. Svoboda, A.C. Edwards, H.A. Ander-      Sharpley, A.N., S.C. Chapra, R. Wedephol, J.T. Sims, T.C. Daniel,
  son, and G. Sym. 1999. Phosphorus loss in drainflow from inten-       and K.R. Reddy. 1994. Managing agricultural phosphorus for pro-
  sively managed grassland soils. J. Environ. Qual. 28:1235–1242.       tection of surface waters: Issues and options. J. Environ. Qual.
Murphy, J., and J.P. Riley. 1962. A modified single-solution method     23:437–451.
  for the determination of phosphate in natural waters. Anal. Chim.   Sharpley, A.N., S.J. Smith, O.R. Jones, W.A. Berg, and G.A. Coleman.
  Acta 27:31–36.                                                        1992. The transport of bioavailable phosphorus in agricultural run-
Petrovic, A.M. 1990. The fate of nitrogenous fertilizers applied to     off. J. Environ. Qual. 21:30–35.
                                                                      Sims, J.T., R.R. Simard, B.C. Joern, and A. Sharpley. 1998. Phosphorus
  turfgrass. J. Environ. Qual. 19:1–14.
                                                                        loss in agricultural drainage: Historical perspective and current
Pote, D.H., T.C. Daniel, D.J. Nichols, A.N. Sharpley, P.A.J. Moore,     research. J. Environ. Qual. 27:277–293.
  D.M. Miller, and D.R. Edwards. 1999. Relationship between phos-     Smith, A.E., and D.C. Bridges. 1996. Movement of certain herbicides
  phorus levels in three Ultisols and phosphorus concentrations in      following application to simulated golf greens and fairways. Crop
  runoff. J. Environ. Qual. 28:170–175.                                 Sci. 36:1439–1445.
Schindler, D.W. 1977. Evolution of phosphorus limitation in lakes.    USEPA. 1976. Quality criteria for water. USEPA Rep. 440/9-76-023.
  Science 195:260–262.                                                  U.S. Gov. Print. Office, Washington, DC.

To top