Content uploaded by Kenneth W Tate
Author content
All content in this area was uploaded by Kenneth W Tate
Content may be subject to copyright.
Reproduced from Journal of Environmental Quality. Published by ASA, CSSA, and SSSA. All copyrights reserved.
Quantifying the Impact of Regular Cutting on Vegetative Buffer Efficacy
for Nitrogen-15 Sequestration
A. Bedard-Haughn,* K. W. Tate, and C. van Kessel
ABSTRACT centrations in runoff from irrigated pasture range from
0.2 to 5 mg L
⫺
1
(Bedard-Haughn, unpublished data, 2002).
This study used the stable
15
N isotope to quantitatively examine Buffer strips are broadly defined as strips of vegeta-
the effects of cutting on vegetative buffer uptake of NO
3
⫺
–N based tion that improve or maintain water quality downslope
on the theory that regular cutting would increase N demand and
of an agriculture or forestry operation (Barling and
sequestration by encouraging new plant growth. During the summer
of 2002, 10 buffer plots were established within a flood-irrigated pas- Moore, 1994). Buffers function to remove pollutants by
ture. In 2003,
15
N-labeled KNO
3
was applied to the pasture area at a reducing or filtering surface runoff and/or by filtering
rate of 5 kg N ha
⫺
1
and 99.7 atom %
15
N. One-half of the buffer plots ground water and stream water (Dosskey, 2001). Atten-
were trimmed monthly. In the buffers, the cutting effect was not uation of NO
3
⫺
by buffers is attributed to a combination
significant in the first few weeks following
15
N application, with both of factors, including denitrification, infiltration, and
the cut and uncut buffers sequestering
15
N. Over the irrigation season, plant uptake (Hill, 1996). The relative importance of
however, cut buffers sequestered 2.3 times the
15
N of uncut buffers, each factor varies according to buffer characteristics
corresponding to an increase in aboveground biomass following cut- such as hydrology, vegetation type (grass vs. forest), soil
ting. Cutting and removing vegetation allowed the standing biomass type (coarse vs. fine), buffer width, and pollutant type
to take advantage of soil
15
N as it was released by microbial mineraliza-
(Bharati et al., 2002; Schmitt et al., 1999). In irrigated
tion. In contrast, the uncut buffers showed very little change in
15
N
sequestration or biomass, suggesting senescence and a corresponding pasture, infiltration and plant uptake appear to have a
decrease in N demand. Overall, cutting significantly improved
15
Ngreater impact on NO
3
⫺
attenuation than denitrification
attenuation from both surface and subsurface water. However, the (Verchot et al., 1997). A recent field study in California
effect was temporally related, and only became significant 21 to 42 d using
15
N-enriched NO
3
⫺
tracers in an irrigated pasture
after
15
N application. The dominant influence on runoff water quality system found that up to 50% of applied
15
N was removed
from irrigated pasture remains irrigation rate, as reducing the rate by by plant uptake the first 10 d following application,
75% relative to the typical rate resulted in a 50% decrease in total making uptake the dominant mechanism for N attenua-
runoff losses and a sevenfold decrease in
15
N concentration. tion (Bedard-Haughn et al., 2004). However, minimal
uptake occurred over the remainder of the growing sea-
son, even with available N in the soil. Consequently,
In California, irrigated pasture provides a relatively
15
N continued to be lost throughout the irrigation season
low-cost source of green forage during the summer via runoff, despite the presence of vegetative buffers. In
months when surrounding rangelands are dry and dor- examining grass buffer trapping efficiency for sediment
mant. Irrigation rates vary by irrigation method, but for and nutrients, Dillaha et al. (1989) reported higher levels
flood irrigation, rates are as high as 70 L s
⫺
1
at the top of soluble nutrients leaving buffers than entering them,
of the slope, applied continuously over an 8- to 14-h which they attributed to low trapping efficiency for solu-
period (up to 12 cm). In the Sierra Nevada foothills, ble nutrients and to release of nutrients previously
with slopes from 5 to 30%, this can generate runoff stored in the buffer. This contributes to concern that
losses of up to 70% of the applied water (Tate et al., buffer efficiency may decrease over time, and that buff-
2000b). Given that irrigated pasture is both fertilized ers will ultimately become a source of N (and other
and grazed, there is concern that runoff water contains nutrients) rather than a sink (Mendez et al., 1999).
toxic levels of pathogens and nutrients. Plant N demand and uptake can be key factors in
Nitrate N is a soluble nutrient commonly cited as a controlling N losses in many ecosystems (Mulholland
source of ground- and surface-water contamination. In et al., 2000). Demand and uptake vary with the N status
the United States, the legal drinking water limit for of the vegetation, NO
3
⫺
availability, plant growth rate,
NO
3
⫺
–N is 10 mg L
⫺
1
, but concentrations as low as 1 mg and plant age or phenology. All other factors remaining
L
⫺
1
can contribute to algal blooms (Mendez et al., 1999). constant, plant N uptake during growth will be greater
Nitrate has been implicated in eutrophication in sea- if the vegetation is N deficient or if there is an abundance
water and fresh water (Cole et al., 2004). Measured con- of available N. Maximum N uptake occurs during the
vegetative growth phase when roots are actively growing
and soil moisture is high (Jackson et al., 1988); increas-
Department of Plant Sciences, University of California, Davis, CA
95616. A. Bedard-Haughn, current address: Department of Soil Sci- ing plant age tends to decrease N uptake (Schenk, 1996).
ence, University of Saskatchewan, 51 Campus Drive, Saskatoon, SK, As Jackson et al. (1988) observed in the Sierra Nevada
S7N 5A8, Canada. Received 28 Jan. 2005. *Corresponding author foothills, even well-watered grasses can senesce within
(bedard.haughn@usask.ca). weeks of anthesis, decreasing N demand. When new
Published in J. Environ. Qual. 34:1651–1664 (2005). sources of N are introduced, microbial immobilization
Technical Reports: Surface Water Quality
doi:10.2134/jeq2005.0033
©ASA, CSSA, SSSA Abbreviations: DON, dissolved organic nitrogen; LME, linear mixed
effects; SFREC, Sierra Foothill Research and Extension Center.677 S. Segoe Rd., Madison, WI 53711 USA
1651
Published online August 9, 2005
Reproduced from Journal of Environmental Quality. Published by ASA, CSSA, and SSSA. All copyrights reserved.
1652 J. ENVIRON. QUAL., VOL. 34, SEPTEMBER–OCTOBER 2005
may compete effectively with plants for N (Jackson et regular cutting and removal of vegetation in buffer strips
al., 1989). Subsequent turnover and mineralization re- would increase plant
15
N uptake and retention, (ii) mea-
leases this previously immobilized N. In annual grass- sure the impact of regular cutting on attenuation of
lands in the Sierra Nevada foothills, turnover of the runoff
15
N, and (iii) determine whether there was a cor-
microbial N pool can occur rapidly (less than one day) responding impact on attenuation of
15
N in the soil solu-
and continuously (Davidson et al., 1990; Jackson et al., tion and on
15
N sequestration in the soil. In addition,
1989), necessitating continual plant demand for N to we considered whether decreasing irrigation rate affects
minimize nutrient losses. buffer efficiency by comparison with results on an adja-
It may be possible to increase plant N uptake via regular cent site. By examining the water quality measurements
cutting, which would increase N demand by encouraging in conjunction with the soil and vegetation results, a
compensatory regrowth. Within two weeks after shoot complete
15
N recovery budget was developed, providing
harvest, uptake of N increases (Ourry et al., 1990). insight into the relative importance of the different N
Matheson et al. (2002) found that although regular cut- sinks and pools in the function of vegetative buffers in
ting of vegetation decreased new shoot production, it irrigated pasture.
increased the NO
3
⫺
assimilation capacity of shoots by a
factor of 5 compared with shoots that were not cut,
MATERIALS AND METHODS
suggesting that even when total plant biomass is reduced
by cutting, the positive effects on N sequestration might Site Description
offset this. The University of California Sierra Foothill Research and
The role of plant uptake in attenuating nutrients is Extension Center (SFREC), located 100 km northeast of Sac-
diminished when nutrients are returned to the soil via ramento, CA, has a xeric climate and hilly terrain. During the
decomposition, therefore periodic harvesting of buffer summer of 2002, 10 adjacent plots were established within an
vegetation might improve the long-term effectiveness existing flood-irrigated pasture at SFREC. Each plot consisted
of buffers (Dosskey, 2001). Mowing alone will increase of a 5-m-wide by 16-m-long (80 m
2
) buffer area immediately
plant N uptake, but removal of the cut vegetation is downslope of a 25-m
2
pasture area (Fig. 1). The pasture-buffer
required to prevent nutrient release via decomposition areas were dominated by orchard grass (Dactylis glomerata
(Barling and Moore, 1994). Although grazing also re- L.), velvet grass (Holcus lanatus L.; also known as Yorkshire
moves vegetation, up to 60 to 90% of the ingested N fog), and dallis grass (Paspalum dilatatum Poir.). Soils
can be returned to the pasture system, mostly as urine (Table 1) were fairly uniform throughout the site and were
classified as fine-loamy, mixed, thermic, Mollic Haploxeralfs
(Di and Cameron, 2002).
of the Auburn–Las Posas–Argonaut rocky loam association
Applying
15
N-enriched techniques in the field pro-
(Herbert and Begg, 1969) and site slope ranged from 15 to
vides a powerful insight into plant–soil N dynamics 18%. The entire study area was fenced to prevent disturbance
(Powlson and Barraclough, 1993), commonly within a by the cattle grazing the surrounding pasture.
single growing season (Bardgett et al., 2003; Di et al.,
1999; Jackson et al., 1989; Mulholland et al., 2000). For
this study,
15
N-enriched isotopes allowed new NO
3
⫺
to Cutting and Irrigation
be distinguished from NO
3
⫺
already present in the sys- Beginning in June 2003, a cutting treatment was randomly
tem and to be quantitatively traced through the buffers allocated to 5 of the 10 buffer areas. Adjacent pasture-buffer
(Bedard-Haughn et al., 2003). areas were separated by landscape edging, which effectively
Given the previously observed abatement in plant N prevented runoff crossover between buffers. Preferential flow
uptake in mature buffers in irrigated pasture (Bedard- along the edging was minimized by typical irrigation manage-
Haughn et al., 2004) and the potential for increasing ment techniques. For the duration of the 2003 irrigation season
plant N uptake via vegetation management, this study (June–October), vegetation in the five cut buffers was trimmed
monthly using nylon-line trimmers to levels corresponding
was designed to: (i) quantitatively determine whether
Fig. 1. Schematic of pasture-buffer plot layout. Not to scale. “Biomass” buffers received no
15
N and were used to get quantitative estimates of
aboveground biomass. Soil samples were taken at the same downslope distances as the soil solution samples.
Reproduced from Journal of Environmental Quality. Published by ASA, CSSA, and SSSA. All copyrights reserved.
BEDARD-HAUGHN ET AL.: VEGETATIVE BUFFER EFFICACY FOR
15
N SEQUESTRATION 1653
to post-grazing height (5–10 cm) in the surrounding pasture. only minimal percolation. To ensure uniform distribution of
both the
15
N fertilizer solution and the additional water, theCutting frequency was based on observed recovery time re-
quired between grazing events in the surrounding pasture. All application area was subdivided into m
2
plots. Natural abun-
dance background levels of
15
N in all N pools were measured10 pasture areas were trimmed at the same intervals as the
cut buffers. Cut residues were collected and removed from before application of
15
N-labeled fertilizer to account for natu-
ral variability and dilution of the applied
15
N fertilizer by
14
N.the site. Uncut buffers were not trimmed.
Plot irrigation was by gated pipe, which delivered water Within a given plant species, the standard deviation of natural
abundance atom % was within ⫾0.0002 atom %.separately to each pasture-buffer area. Irrigation rate was
controlled by a valve and monitored by flow meters (Model Isotopic levels are reported as atom %
15
N excess, which
refers to the amount of
15
N present relative to the averageWT; Netafim, Tel Aviv, Israel) that allowed measurement of
both rate and quantity of water applied. Water was applied naturally occurring background
15
N levels for that particular
N pool. Atom %
15
N excess amounts were extrapolated to obtain5 m upslope from the buffer–pasture interface to maximize
control of water distribution within the study area. During the total amount of
15
N in a given pool by weight and/or volume
and thus to determine a
15
N budget.this project, the irrigation rate was calibrated to 1 L s
⫺
1
per
buffer for approximately 3 h every 9 d. This irrigation rate is
75% lower than the rate applied in Bedard-Haughn et al. Vegetation Sampling and Analysis
(2004) study on an adjacent set of plots (Table 1). A lower
irrigation rate was used in an effort to reduce runoff losses Grab samples of vegetation were collected 3, 11, 21, 42, 60,
and improve irrigation efficiency. On average, 29% of the 79, 98, and 114 d after
15
N application. To determine how far
applied irrigation water was lost as runoff. Total duration of the
15
N fertilizer had moved into the buffers, vegetation sam-
each irrigation event varied according to the volume needed ples were collected along a cross-slope transect within the
to restore soil water content, which was determined using zone of
15
N application and at downslope distances of 1, 4, 8,
evapotranspiration data from the California Irrigation Man- 12, and 16 m from the application area. The uncut buffer
agement Information System weather station located at vegetation samples were separated by the three dominant
SFREC. Climate and plant growth conditions during the 2003 grass species, whereas cut buffer vegetation samples repre-
growing season were normal for the region. sented composites of all species present due to identification
Collection troughs installed across the bottom of each obstacles associated with newly clipped vegetation. All plant
buffer collected surface water runoff. Three pairs of ceramic samples were oven-dried at 65⬚C and analyzed for
15
N isotopic
soil solution samplers (Soilmoisture Equipment, Santa Bar- composition via mass spectrometry (Integra Integrated Stable
bara, CA) were installed in each buffer area at 1, 4, and Isotope Analyzer; Europa Scientific, Crewe, UK) at the Uni-
12 m downslope of the
15
N application (Fig. 1). Samplers were versity of California, Davis, Stable Isotope Facility (van Kessel
installed to depths of 15 and 45 cm, the average depths to the et al., 1994). The current sensitivity of our stable isotope ratio
bottom of the A horizon and the top of the heavy clay Bt mass spectrometers is 0.0002 atom %
15
N.
horizon, respectively. Of the two plots that did not receive
15
N, one received the
same regular cutting as the cut buffers whereas the other was
left to mature the same as the uncut buffers. The species
Nitrogen-15 Application composition, vegetation age, and irrigation rates of these two
In June 2003, four days after the first cutting and three days nonlabeled buffers were equivalent to the labeled buffers.
before the first irrigation,
15
N-labeled KNO
3
was applied in Accurate biomass measurements could not be taken from
solution at a rate of 5 kg N ha
⫺
1
and 99.7 atom %
15
N. The the labeled buffers without compromising results, so on each
rate and atom %
15
N concentration were selected to provide sampling day, representative biomass measurements were
an approximation of post-irrigation fertilizer N levels while taken from the two nonlabeled buffers (Fig. 1). All living
allowing the tracer to be detectable in all N pools throughout biomass within a randomly placed 0.1-m
2
quadrat was col-
the duration of the experiment. The
15
N solution was applied lected, dried, and weighed. For the cut buffer, three composite
across 8 of the 10 plots (4 cut, 4 uncut). The area labeled was quadrat measurements were collected on each day. For the
1 m wide across the width of each plot and located 0.75 m uncut buffer, one representative measurement was taken for
above the buffer areas (Fig. 1). Application rate and area each of the three dominant species. Although this lower num-
were based on Bedard-Haughn et al. (2004). Following appli- ber contributed to greater variability for uncut biomass values,
cation, the
15
N fertilizer was watered in with 18 L of water it allowed for regular sampling over the season without eradi-
per m
2
; under field conditions, this volume was sufficient to cating the less prevalent species. Cover measurements for the
rinse the
15
N solution off of the foliar surfaces but allowed uncut buffers were taken on Days 11, 42, and 114 using the line
intercept method (Canfield, 1941) to determine the relative
dominance of each of the three dominant grass species.
Table 1. Values (mean ⫾SD) for field site properties averaged
Vegetation N content was multiplied by atom %
15
N excess
across all buffers. Soil properties are average values for the
0- to 15-cm layer across all buffers.
values to get the mass (mg) of
15
N in each g of vegetation.
The total mass (mg) of
15
N sequestered in vegetation in a
Bedard-Haughn
given buffer area was determined by multiplying the mg
15
N
Property Current study et al. (2004)
g
⫺
1
vegetation values times biomass values (g m
⫺
2
) and extrap-
C, % 3.3 ⫾0.6 3.0 ⫾0.4
olating to the whole area using cover data.
N, % 0.3 ⫾0.1 0.3 ⫾0.04
C to N ratio 10.4 ⫾0.6 10.4 ⫾0.4
Sand, %† 36.8 ⫾4.9 31.7 ⫾3.8
Runoff Sampling and Analysis
Silt, %† 52.5 ⫾4.9 50.6 ⫾1.6
Clay, %† 10.6 ⫾1.0 11.0 ⫾0.9
Runoff samples were collected on the same dates as vegeta-
Slope, % 16.9 ⫾1.1 10.9 ⫾0.8
tion samples. Samples were taken from the collection troughs
Runoff losses, %‡ 29.2 ⫾12.7 56.8 ⫾16.4
15 min following the leading edge of runoff and again just
† Particle size expressed on measured volume % basis for 2003, calculated
before the end of the irrigation event and were stored frozen
volume % basis for 2002 (Eshel et al., 2004).
‡ Runoff volume/irrigation volume.
until analysis. Based on results from an adjacent irrigated site,
Reproduced from Journal of Environmental Quality. Published by ASA, CSSA, and SSSA. All copyrights reserved.
1654 J. ENVIRON. QUAL., VOL. 34, SEPTEMBER–OCTOBER 2005
these two measurements captured the maximum variability Soil Solution Sampling and Analysis
during the irrigation period (Bedard-Haughn et al., 2004). The Immediately before each irrigation event, vacuum was ap-
15-min interval provided a measurement of maximum
15
Nplied to the soil solution sampling tubes and allowed to draw
concentration, whereas the event-end sample reflects the mini- moisture from the soil for 10 d before sample collection
mum
15
N concentration, but the maximum
15
N load. Sample (Bedard-Haughn et al., 2004). Although vacuum was not ap-
collection (500 mL) was as a “grab” sample from the runoff plied continuously over the 10-d period, suction was still pres-
collection trough. Runoff rates were determined at regular ent at sampling in most sampling tubes. After Day 42, the
intervals by measuring the volume of runoff in a 5-s period. time between irrigation and sample collection was shortened
Runoff rate data were used to determine runoff losses (Table 1). from 9 to 3 d, which substantially improved the reliability of the
Runoff
15
N isotope analyses were performed on three N suction in the tubes and the volume of sample collected. Soil
pools: NO
3
⫺
,NH
4
⫹
, and total N. Samples were filtered to re- solution samples were stored frozen until analysis for NO
3
⫺
–
15
N
move sediment and vegetation residues from runoff. The via the TiCl
3
diffusion (25-mL aliquots) as outlined in Bedard-
NH
4
⫹
–
15
N and NO
3
⫺
–
15
N were determined by NH
3
diffusion of Haughn et al. (2004).
a 100-mL aliquot onto polytetrafluoroethylene-encased acid
traps (Stark and Hart, 1996). To measure NO
3
⫺
–
15
N, the Stark Nitrogen-15 Recovery Budget
and Hart (1996) method was modified using TiCl
3
(Titanous
The
15
N recovery budget illustrates the mass of
15
N seques-
Chloride Solution, 20%; Fisher, Hampton, NH) to reduce
tered and/or measured in runoff relative to the mass of
15
N
NO
3
⫺
to NH
3
as outlined in Bedard-Haughn et al. (2004). Total
applied at the beginning of the study. For soil and vegetation
15
N was determined on a separate 20-mL aliquot by performing
samples, atom %
15
N excess amounts were extrapolated using
a persulfate digestion (American Public Health Association,
total N content, soil bulk density (mass/volume), and vegeta-
1989) to convert the dissolved organic nitrogen (DON) and
tion biomass (mass/area) to obtain the total amount of
15
Nin
NH
4
⫹
to NO
3
⫺
, and samples were then diffused for NO
3
⫺
as a given sink by mass and thus to determine the total amount
above. Following diffusion, acid disks were removed from of
15
N stored in the pasture-buffer areas. The total amount of
polytetrafluoroethylene packets and analyzed via mass spec-
15
N lost via runoff (
15
N load) during a given irrigation event
trometry. The DON–
15
N for each sample was calculated using was determined by multiplying runoff volume by
15
N concen-
an isotope mixing model via difference from total
15
N (Shearer trations for each measured interval and integrating over time.
and Kohl, 1993): Summing these values for the measured irrigation events, to-
gether with quantitative estimates for intervening irrigation
15
N
DON
⫽
15
N
NT
m
NT
⫺
15
N
NH
4
m
NH
4
⫺
15
N
NO
3
m
NO
3
m
NT
[1] events where runoff was not measured, provided a value for
the total amount of
15
N lost as runoff from the pasture-buffer
areas. There were no total flux measurements for soil solution,
where
15
N
x
refers to the atom %
15
N value for a given N form
so total subsurface
15
N load could not be calculated.
(NT ⫽total dissolved
15
N) and m
x
refers to the quantity of N
in g. For irrigations where samples were not collected (Days
30, 51, 70, 80, and 106), runoff, concentration, and
15
N values Statistical Analysis
were estimated by calculating the linear relationship between The results were analyzed using linear mixed effects model
adjacent sampling dates. analysis (S-PLUS; Insightful Corporation, 2001). Linear mixed
effects analysis can be applied to both structured and observa-
Soil Sampling and Analysis tional studies (Pinheiro and Bates, 2000) and was used here
to account for the influence of fixed (cutting) effects on buffer
Soil samples were taken at 0, 1, 4, and 12 m from the
15
N
15
N uptake levels and for the repeated measures (group effect–
application at 3 and 114 d following
15
N application. On both plot identity) embedded in the data structure. Treating time
dates, samples were taken to a 15-cm depth in two increments as a fixed effect provided a test of how response varied over
(0–7 and 7–15 cm) using a slack hammer (Ben Meadows Com- the duration of the study. The magnitude and direction (⫾)
pany, Janesville, WI), corresponding to the depth of the A of the coefficient for treatment and time effects was used to
horizon. Soil texture was determined on the 114-d samples define the relationship between
15
N uptake and runoff
15
N load
using laser diffraction and reported in volume percent (Eshel and cutting effects. This flexible model also allowed within-
et al., 2004). On Day 114, soil samples were also taken to a group variance and correlation structures for handling within-
1-m depth in two increments (0–40 and 40–100 cm) to allow group (plot) heteroscedasticity and temporally correlated er-
for a more complete estimate of the final
15
N budget. Soil rors (irrigation series within year) (Pinheiro and Bates, 2000).
samples were oven-dried at 40⬚C and analyzed for total N This approach has been used in modeling other complex longi-
and
15
N via mass spectrometry. Soil C was analyzed by mass tudinal datasets (Atwill et al., 2002; Tate et al., 2000a, 2003).
spectrometry in conjunction with soil N. Bulk density measure- The soil and soil solution data were analyzed using the nonpar-
ments for all depth increments were done on oven-dried intact ametric Wilcoxon rank sum test (S-PLUS; Insightful Corpora-
cores and were within the range of values measured by Dahl- tion, 2001), which is not restricted by assumptions of normality.
gren et al. (1997) in SFREC grazed pasture.
Soil microbial
15
N was measured using fumigation–extrac- RESULTS
tion method (Brookes et al., 1985), with fumigation for 48 h
with chloroform vapor and extraction with 0.5 MK
2
SO
4
.Ex- Aboveground Vegetation
tract
15
N was determined by persulfate digestion (American Cut or uncut, there was a general decrease in atom
Public Health Association, 1989) to convert the DON and %
15
N excess (i.e., %
15
N present in excess of background
NH
4
⫹
to NO
3
⫺
, and diffusion using a modification of the Stark
15
N levels) with increasing distance from the
15
N applica-
and Hart (1996) method, as outlined in Bedard-Haughn et al. tion zone. However, there was
15
N present in vegetation
(2004). Microbial
15
N was determined by difference between at the 16-m distance even after a single irrigation event
fumigated and nonfumigated samples for both dates (0–15 cm
only) for the 0- and 1-m distances. (11 d; Fig. 2). At the first vegetation sampling following
Reproduced from Journal of Environmental Quality. Published by ASA, CSSA, and SSSA. All copyrights reserved.
BEDARD-HAUGHN ET AL.: VEGETATIVE BUFFER EFFICACY FOR
15
N SEQUESTRATION 1655
Fig. 2. Atom %
15
N excess in buffer vegetation by distance from
15
N application (averages, standard error bars). From top to bottom, days after
15
N application ⫽11 d, 60 d, 114 d. Note log yaxis.
15
N application (11 d), vegetation atom %
15
N excess was had the lowest. The total biomass values for a given
uncut buffer varied according to the cover distributionhigher for uncut buffers within 1 m of the
15
N application
zone, whereas further downslope, vegetation atom % of the species within that area.
When biomass (Fig. 3) and percent cover distribution
15
N excess was higher for cut buffers (Fig. 2). As the
irrigation season progressed (60 d, 114 d), atom %
15
N data were used to determine mass of
15
N for each domi-
nant species in a given uncut buffer, orchard grassexcess values remained higher in uncut buffers for the
1-m sampling distance, but there were no downslope tended to sequester the majority of the
15
N, whereas
velvet grass sequestered the least (Table 2). Althoughdifferences between cut and uncut buffers.
There were differences in seasonal biomass trends dallis grass had the highest biomass per m
2
(Fig. 3), it
was intermediate in its
15
N storage. The mass of
15
Nbetween the cut and uncut buffers (Fig. 3). The cut buffer
biomass values reflect the effects of regular cutting, with sequestered by a given species did not change signifi-
cantly over the course of the season.increasing biomass values between cuttings and sharp
drops in biomass on the actual cutting dates. For the The amount of
15
N sequestered by each species was
summed to get the total mass (mg) of
15
N sequestereduncut buffers, biomass values varied nonlinearly through-
out the season, but generally, of the three species, dallis per uncut buffer (Fig. 4). Values for uncut buffers reflect
the
15
N in the standing biomass on a given date, whereasgrass had the highest biomass (per m
2
) and velvet grass
Reproduced from Journal of Environmental Quality. Published by ASA, CSSA, and SSSA. All copyrights reserved.
1656 J. ENVIRON. QUAL., VOL. 34, SEPTEMBER–OCTOBER 2005
Fig. 3. Vegetation biomass (g m
⫺
2
) by time after
15
N application for each of the three dominant species within the uncut buffers and for a
composite of all species present per m
2
within the cut buffers.
values for cut buffers are cumulative, reflecting the
15
N an order of magnitude, despite a much smaller reference
in the standing biomass as well as the
15
N removed from area. Unlike the cut buffers, the total cumulative mass
the plots by cutting. Overall, the uncut buffers had a of
15
N sequestered in the standing vegetation of the
constant mass of
15
N sequestered over the course of the application zone did not increase over the season; it
season, regardless of biomass fluctuations. In contrast, increased from Day 11 to Day 42 and then decreased
the cut buffers had a lower mass of
15
N sequestered to a new lower level, suggesting
15
N losses from the
immediately following
15
N application, indicative of the standing vegetation (Fig. 4).
lower biomass in these buffers on Day 11. Over the A similar decrease in
15
N mass within the zone of
15
N
course of the season, however, there was a linear in- application was observed in the soil microbial biomass
crease in the mass of
15
N in the cut buffers such that by (Fig. 5). In both the 0- to 7- and 7- to 15-cm depth
the end of the season there was nearly double the increments, the amount of microbial
15
N decreased be-
amount of
15
N sequestered in the cut buffers compared tween Days 3 and 114. In contrast, just 1 m downslope,
with the uncut. the amount of microbial
15
N increased between Days
The linear mixed effects (LME) model confirms that 3 and 114 in both depth increments. There were no
cutting effect on
15
N uptake was time dependent (Table 3). significant differences in microbial
15
N content between
Cutting alone resulted in a decrease in
15
N uptake by the cut and uncut buffers, regardless of date.
buffer vegetation (coefficient ⫽⫺13.2, P⫽0.1); how- Within 114 d of
15
N application, 14 to 16% of the total
ever, if the interaction with time is taken into consider- amount of
15
N applied was taken up by the pasture
ation, cutting substantially increased the amount of
15
Nvegetation within the zone of
15
N application (Table 2).
sequestered, with the most significant differences be- However, the observed differences in recovery between
tween cut and uncut buffers occurring at the end of the the cut and uncut buffers were most notable in the buffer
season (coefficient ⫽⫹46.6, P⫽⬍0.0001). vegetation, where the cut buffers recovered an average
The majority of
15
N sequestration by vegetation oc- of 59 mg (2.4%) of the applied tracer compared with
curred within the
15
N application zone (Fig. 4). As for 26 mg (1%) in the uncut buffers.
the cut buffers, the
15
N application zone was cut regularly
and the
15
N contained in the vegetation was removed, Surface Runoff
so there was a steady increase over the season of
15
N
Runoff rates averaged 0.4 L s
⫺
1
plot
⫺
1
(SD ⫾0.1)
removed. The difference in mass of
15
N removed from
the application zone versus the cut buffers was nearly within 15 min of the start of runoff and leveled off at
Reproduced from Journal of Environmental Quality. Published by ASA, CSSA, and SSSA. All copyrights reserved.
BEDARD-HAUGHN ET AL.: VEGETATIVE BUFFER EFFICACY FOR
15
N SEQUESTRATION 1657
Table 2. The
15
N budget for runoff, soil, and vegetation after final
time was taken into consideration, with the cut buffers
irrigation (Day 114). Percent recovery refers to the mass of
having less
15
N load in runoff as shown by the negative
15
N recovered in a given pasture-buffer area relative to the
regression coefficients. For the NO
3
⫺
and NH
4
⫹
pools,
total mass applied (2500 mg) in the zone of
15
N application.
the effect of cutting only became statistically significant
Average
15
N recovery per
(LME Pⱕ0.05) on Day 42, whereas for total dissolved N,
pasture-buffer area (⫾SD)
cutting had a significant effect by Day 21 (LME P⫽0.05).
Cut buffer Uncut buffer
For the DON pool, cutting reduced the
15
N load (LME
mg
P⫽0.08) regardless of time since
15
N application; adding
Runoff
time as a fixed effect improved the significance slightly,
Depth
but not enough to warrant its inclusion in the model.
NH
4
ⴙ
1⫾0.2 1 ⫾0.2
NO
3
⫺
3⫾0.6 4 ⫾1.3
Dissolved organic nitrogen (DON) 3 ⫾0.3 4 ⫾0.7
Subsurface: Soil Solution and Soil
Total dissolved N 7 ⫾0.6 9 ⫾2.5
Soil
The NO
3
⫺
–
15
N concentration of the soil solution
Depth
15
N zone
(Fig. 7) was similar in range to the
15
N concentration of
0–7 cm 571 ⫾223 419 ⫾102
the NO
3
⫺
–
15
N in runoff (Fig. 6), but the soil solution
7–15 cm 79 ⫾31 77 ⫾37
15–40 cm 513 ⫾460 384 ⫾349
NO
3
⫺
–
15
N concentrations tended to be much more vari-
40–100 cm 52 ⫾35 81 ⫾35
able. This was particularly true in the first 42 d after
Total (0–100 cm) 1215 ⫾429 961 ⫾293
15
N application during which time the samples were col-
Buffer
lected 10 d after irrigation, versus after 3 d.
0–7 cm 91 ⫾18 95 ⫾25
7–15 cm 23 ⫾14 18 ⫾2
In the cut buffers, the solution samplers at the 15-cm
15–40 cm 157 ⫾61 224 ⫾33
depth tended to have decreasing NO
3
⫺
–
15
N concentra-
40–100 cm 260 ⫾115 220 ⫾94
tions with increasing distance from the zone of
15
N appli-
Total (0–100 cm) 531 ⫾144 557 ⫾135
cation (Fig. 7). In contrast, those samplers at the same
Vegetation
Vegetation type
15
N zone
depth in the uncut buffers developed a pattern of in-
Grass 395 ⫾52 338 ⫾50
creasing NO
3
⫺
–
15
N concentration with increasing dis-
Buffer
tance by Days 101 and 116. The samplers at the 45-cm
Grass (composite) 59 ⫾10 26 ⫾5
depth did not demonstrate any clear patterns associated
Orchard grass NA† 17 ⫾6
Velvet grass NA 1 ⫾2
with distance from the zone of
15
N application. Regard-
Dallis grass NA 7 ⫾5
less of sampler depth and distance from the zone of
15
N
Total recovery
application, the NO
3
⫺
–
15
N concentrations were signifi-
15
N recovered 2207 ⫾522 1891 ⫾444
cantly higher for soil solution in the uncut buffer than
Total recovery, %† 88 ⫾20 76 ⫾18
in the cut buffer (P⫽0.002, Wilcoxon rank sum test).
† Cut buffers have composite values only.
This difference between the cut and uncut buffers for
NO
3
⫺
–
15
N concentrations in the subsurface water was
approximately 0.7 L s
⫺
1
plot
⫺
1
(SD ⫾0.2) by the end not reflected in the 0- to 15-cm soil atom %
15
N excess
of the 3-h irrigation event. When
15
N concentrations (Fig. 8). There was no significant difference in soil atom
were multiplied by runoff volume to calculate the total %
15
N excess between the cut and uncut buffers on
load of
15
N in runoff over a given irrigation event, the either sampling date (P⫽0.7, Wilcoxon rank sum test).
15
N load in runoff in all N pools was greater from the There was also no difference between sampling dates.
uncut buffer than from the cut buffer after Day 42 The only general pattern was a decrease in atom %
15
N
(Fig. 6). The NO
3
⫺
–
15
N load decreased to a steady state excess with increasing distance from the zone of
15
N appli-
by Day 42, NH
4
⫹
–
15
N load increased to a steady state cation.
by Day 42, and DON–
15
N load remained relatively level
throughout the study. Maximum NO
3
⫺
–
15
N was lost in Nitrogen-15 Recovery Budget
the first 21 d after
15
N application, and maximum differ-
ences in NO
3
⫺
–
15
N load between the cut and uncut buff- The
15
N lost via runoff was relatively small compared
with the amount applied: 0.3% of the applied
15
N wasers appeared after Day 60. For the NH
4
⫹
and DON
pools, significant differences between the cut and uncut lost in runoff from the cut buffers and 0.4% of the
applied
15
N was lost in runoff from the uncut buffersbuffers started to appear as early as Day 42. The data
gap on Day 60 is due to the occurrence of an isolated (Table 2). Maximum recovery occurred in the soil, where
approximately 38 to 49% of the applied
15
N was mea-precipitation event on that sampling day; the total vol-
ume of precipitation was comparable with the volume sured as total soil
15
N within the zone of
15
N application.
A further 21 to 22% was measured in the soil withinduring a typical irrigation event. Over one-half of the
precipitation fell within 1 h; the total duration of the the buffers. The vegetation within the zone of
15
N appli-
cation recovered 14 to 16% of the applied
15
N over theevent was 8 h. For Day 60, vegetation and soil solution
samples could be collected, but there was no measur- course of the study. Only a small amount was recovered
by the buffer vegetation itself: 2% in the cut buffers andable runoff.
According to the runoff LME model, for the NO
3
⫺
–, 1% in the uncut buffers. The difference in
15
N recovery
between the cut and uncut buffers was not significantNH
4
⫹
–, and total dissolved
15
N pools, cutting alone did
not have a significant effect on the
15
N load. There was, for any pool except for the within-buffer vegetation
(Table 2). Total
15
N recovery from the cut buffers washowever, a significant effect when the interaction with
Reproduced from Journal of Environmental Quality. Published by ASA, CSSA, and SSSA. All copyrights reserved.
1658 J. ENVIRON. QUAL., VOL. 34, SEPTEMBER–OCTOBER 2005
Fig. 4. Total mass (mg) of
15
N sequestered in aboveground vegetation within a given area by time after
15
N application for
15
N application zone,
uncut buffer areas, and cut buffer areas, where
15
N application zone and cut buffer areas are cumulative, including
15
N removed by clipping
during the irrigation season.
88% (SD ⫾20) and from the uncut buffers was 76% tion (Table 3); higher
15
N sequestration in the uncut
(SD ⫾18). vegetation reflects the greater initial biomass available
for N uptake. Soon thereafter, however,
15
N sequestra-
tion increased in the cut buffers as regular cutting main-
DISCUSSION tained plant N demand.
Effects of Cutting on Plant Nitrogen-15 Uptake In comparing the N capacity of cut and uncut buffers
over the irrigation season, the cut buffers sequestered
The effect of cutting on plant
15
N uptake was not
twice the
15
N of the uncut buffers (Fig. 4). Given that
significant in the first few weeks following
15
N applica-
the cut and uncut buffers had very similar atom %
Table 3. Linear mixed effects model estimating
15
N uptake by
15
N excess values for much of the irrigation season, the
buffer vegetation over time by treatment (uncut versus cut).
difference in sequestration can be attributed primarily to
Coefficients quantify the expected effect of cutting and time
increases in biomass in the cut buffers. The increase in
on mg
15
N sequestered per buffer area relative to the refer-
ence level.
biomass following each cutting (Fig. 3) was a typical com-
pensatory response to defoliation (Ferraro and Oester-
Model term Coefficient 95% CI† P
held, 2002). This period of growth should be a period
Intercept 20.2 10.2, 30.2 0.0002
of high N demand (Jackson et al., 1988). Cutting of
Treatment
Uncut‡ 0.0 – –
aboveground vegetation can increase shoot N assimila-
Cut ⫺13.2 ⫺30.3, 3.9 0.1
tion by more than 5 times (Matheson et al., 2002). Cut-
Days since
15
N application
11 d‡ 0.0 – –
ting and removing vegetation from the buffers allowed
21 d ⫺0.3 ⫺10.7, 10.2 1.0
the standing biomass to take advantage of immobilized
42 d 9.1 ⫺1.3, 19.6 0.1
soil
15
N as it was released by microbial mineralization
60 d 5.1 ⫺5.4, 15.6 0.3
79 d 9.5 ⫺1.0, 19.9 0.1
(Bardgett et al., 2003). Removal of the cut vegetation
98 d 6.1 ⫺4.3, 16.6 0.2
is essential, otherwise decomposition will simply return
114 d 5.6 ⫺4.8, 16.1 0.3
nutrients to the system, increasing the potential for
Treatment ⫻days after
15
N
Cut ⫻11 d‡ 0.0 – –
losses via runoff or leaching (Dosskey, 2001). In con-
Cut ⫻21 d 6.4 ⫺8.4, 21.2 0.4
trast, the uncut buffers showed very little change in
15
N
Cut ⫻42 d 3.5 ⫺11.3, 18.3 0.6
Cut ⫻60 d 19.9 5.1, 34.7 0.01
sequestration throughout the irrigation season (Fig. 4),
Cut ⫻79 d 26.7 11.9, 41.5 0.0008
suggesting the occurrence of senescence and a corre-
Cut ⫻98 d 38.8 24.0, 53.6 ⬍0.0001
sponding decrease in N demand (Jackson et al., 1988)
Cut ⫻114 d 46.6 31.8, 61.4 ⬍0.0001
or the absence of net growth during the study.
† 95% confidence interval for coefficient (lower, upper).
‡ Reference category for variable.
Examining interspecific differences in
15
N was ex-
Reproduced from Journal of Environmental Quality. Published by ASA, CSSA, and SSSA. All copyrights reserved.
BEDARD-HAUGHN ET AL.: VEGETATIVE BUFFER EFFICACY FOR
15
N SEQUESTRATION 1659
Fig. 5. Soil microbial biomass
15
N(g
15
Ng
⫺
1
soil) by depth, distance from
15
N application, and time after
15
N application. Note log yaxis.
pected to provide insight into the functioning of the buffers had greater root biomass, the vegetation imme-
uncut buffers, and to help determine whether one spe- diately downslope of the zone of application may have
cies might be better suited for buffers in irrigated pas- been better able to take advantage of
15
N moving down-
ture systems; however, the primary determinant of
15
Nslope via lateral movement, or, particularly if there was
sequestration was total aboveground biomass within the significant lateral root development, drawn on the much
buffer. Dallis grass had greater biomass per m
2
than higher concentrations of
15
N available within the zone
orchard grass (Fig. 3), but orchard grass was by far more of application itself.
prevalent with the buffers (averaging 60% of buffer
area, compared with 20% of buffer area for dallis grass), Effects of Cutting on Surface Runoff Nitrogen-15
and so served to sequester the most
15
N of the three For the first 21 d following application of the
15
N tracer,
species. Although the cut buffers were not examined there were no differences in surface runoff NO
3
⫺
–
15
N
by species, orchard grass appeared to be the dominant between the cut and uncut buffers (Fig. 6). Regardless
species within the cut buffers (and in the surrounding of cutting treatment, there was excess
15
N measured in
pasture), attributable in large part to its rapid regrowth the surface runoff during the first irrigation event after
following grazing or cutting, compared with moderate the
15
N was applied. Similarly, there was
15
N measured
and slow regrowth for dallis grass and velvet grass, re- in the soil solution (Fig. 7), soil (Fig. 8), and vegetation
spectively (USDA, 2005). Rapid regrowth following cut- (Fig. 2) at the furthest distance from the zone of
15
N
ting and extent of ground cover are likely the best pre-
application following the first irrigation event. This sug-
dictors of plant uptake ability in managed buffers.
gests that both the cut and uncut buffers were attenuat-
Cut or uncut, maximum plant uptake of
15
N occurred
ing some
15
N, but during this period, the NO
3
⫺
–
15
N tracer
within the first 4 m of the buffer (Fig. 2). The higher
was extremely mobile and its redistribution via surface
vegetation atom %
15
N excess observed in the first meter
runoff was identical for both the cut and uncut buffers.
downslope of the application area for the uncut buffers
As Di and Cameron (2002) observed, maximum NO
3
⫺
compared with the cut buffers may be attributable in
leaching tends to occur whenever NO
3
⫺
is present in the
part to differences in root biomass. Although soil mois-
soil profile during periods of significant drainage, as
ture is a major factor controlling fine root production
would be associated with irrigation events. It is interest-
in annual grasslands (Cheng and Bledsoe, 2002), cutting
ing to note, however, that the sharp decrease in surface
may have reduced the root biomass (Williams et al.,
runoff NO
3
⫺
–
15
N between Days 21 and 42 (Fig. 6) corre-
2003), increasing root turnover, or inhibited the produc-
tion of new roots (Matheson et al., 2002). If the uncut sponds to the first post–
15
N application cutting of the
Reproduced from Journal of Environmental Quality. Published by ASA, CSSA, and SSSA. All copyrights reserved.
1660 J. ENVIRON. QUAL., VOL. 34, SEPTEMBER–OCTOBER 2005
Fig. 6. Runoff
15
N load over the course of the irrigation season. Values are averaged by buffer treatment and time; error bars represent standard
error. Note log yaxis.
buffer vegetation, suggesting a very strong initial cutting cycled into other N pools, as shown by the parallel
decrease in runoff NO
3
⫺
–
15
N load and increases in runoffeffect on runoff water quality.
During Days 3 to 42, some of the NO
3
⫺
–
15
N appears DON– and NH
4
⫹
–
15
N (Fig. 6). One possible pathway for
the movement between the NO
3
⫺
– and NH
4
⫹
–
15
N poolsto have been immobilized by microbial biomass and
Reproduced from Journal of Environmental Quality. Published by ASA, CSSA, and SSSA. All copyrights reserved.
BEDARD-HAUGHN ET AL.: VEGETATIVE BUFFER EFFICACY FOR
15
N SEQUESTRATION 1661
Fig. 7. Soil solution NO
⫺
3
–
15
N concentrations by time and distance from
15
N application. Values are averaged by buffer treatment, time, and
distance; error bars represent standard error. Note log yaxis.
is dissimilatory nitrate reduction to ammonium. Al- out the experiment is indicative of a rapid N turnover;
Davidson et al. (1990) observed a turnover time of 0.3though it is unlikely here given the inherently low soil
NH
4
⫹
levels, this could not be confirmed with our field to 1.6 d in grassland soils at SFREC.
tracer study design. In a microcosm study, Matheson et
al. (2002) found that within 32 d, up to 49% of applied Effects of Severe Cutting on
NO
3
⫺
–
15
N was reduced to NH
4
⫹
–
15
N and up to 25% was Nitrogen-15 Retention
immobilized in the microbial biomass. After Day 42, The buffer areas that were cut regularly exhibited
there were significantly lower
15
N loads in runoff from good
15
N retention due to the continual plant demand
the cut buffers compared with the uncut buffers for both for N as it was released by the microbial biomass (Fig. 4).
the NO
3
⫺
– and NH
4
⫹
–
15
N pools (Fig. 6). This corresponds The uncut buffers also had good N retention within
with the observed increase in aboveground plant bio- the time frame of the irrigation season, but previous
mass and plant
15
N storage in the cut buffers (Fig. 4). research in irrigated pasture (Bedard-Haughn et al.,
However, the difference between cut and uncut buffers 2004) suggests that plant decomposition during the win-
did not increase substantially after 42 d despite contin- ter months would ultimately contribute to N losses from
ued growth in the cut buffers, indicating that even though the uncut buffers. The rate and amount of new growth
there was a continual demand for
15
N, mineral N was and hence new N demand within uncut buffers will
available for
15
N runoff losses, albeit at extremely low determine how much of the recycled
15
N will be retained
concentrations (⬍1g total dissolved N L
⫺
1
runoff). over the long term.
Unlike the runoff NH
4
⫹
–
15
N pool, there was no lag Even during the course of the irrigation season, there
time between the application of
15
N and the leveling off were
15
N losses observed within the application zone
in runoff DON–
15
N load. This may reflect the observa- vegetation (Fig. 4), despite regular cutting and removal
tion of Davidson et al. (1990) that these grassland soils of vegetation
15
N. This may be due to unintended severe
have a significant heterotrophic microbial sink for cutting (i.e., too short) of the vegetation in the applica-
NO
3
⫺
, particularly when NH
4
⫹
availability is low. There tion zone compared with the buffer areas; vegetative
was also minimal temporal effect on the efficacy of growth and vigor in this zone after 42 d was limited.
cutting for reducing runoff DON–
15
N load: cut buffers Severe cutting has been found to contribute to elevated
had consistently lower runoff DON–
15
N load than uncut
rates of root death (Jarvis and Macduff, 1989). Cutting
buffers throughout the experiment (P⫽0.08; Fig. 6).
can also give rise to increased partitioning of N to the
The higher DON–
15
N in runoff from the uncut buffers
belowground biomass and increased rhizodeposition
may reflect slightly greater partitioning of mineral
15
N
(Paterson and Sim, 2000). In a study using
15
N-enriched
to the microbial pool in the absence of significant plant
synthetic sheep urine, Williams et al. (2003) observed
demand. For example, Jackson et al. (1989) observed
more
15
N in the soil when vegetation was subject to
that microbial immobilization of NO
3
⫺
and NH
4
⫹
was
regular cutting. This belowground partitioning associ-
greater than plant uptake regardless of plant growth
ated with aboveground cutting and the increased poten-
stage, and for NH
4
⫹
, the relative dominance of immobili-
tial for root death, coupled with the high C levels already
zation was even more pronounced after plant senes-
present in the rhizosphere, provide optimum conditions
cence. In this study, however, the partitioning was not
for microbial
15
N uptake (Jackson et al., 1989). As ob-
significant enough to be reflected in microbial biomass
15
N (Fig. 5). The constancy of the DON–
15
N load through- served, soil microbes can compete effectively for both
Reproduced from Journal of Environmental Quality. Published by ASA, CSSA, and SSSA. All copyrights reserved.
1662 J. ENVIRON. QUAL., VOL. 34, SEPTEMBER–OCTOBER 2005
Fig. 8. Atom %
15
N excess in soils by distance from
15
N application at 3 d (top) and 114 d (bottom) after
15
N application. Values are averaged
by treatment and distance; error bars represent standard error.
NH
4
⫹
and NO
3
⫺
(Bardgett et al., 2003; Davidson et al., buffers established the spatial pattern of decreasing
1990). This immobilization of inorganic N can be an NO
3
⫺
–
15
N with increasing distance from the zone of
15
N
effective mechanism for minimizing leaching losses in application; at the same time, the difference in soil solu-
agroecosystems (Di et al., 1999), but given the rapid tion NO
3
⫺
–
15
N concentration between the cut and uncut
turnover of the microbial N pool, the positive effects buffers became much more pronounced (Fig. 7). These
may be temporary (Jackson et al., 1989). This may be patterns complement the decrease in
15
N load in the
of particular importance where frequent wet–dry cycles surface runoff from the cut buffers after 42 d (Fig. 6).
occur, as is the case in irrigated pasture, because wetting Given that maximum root concentrations in California
cycles can cause significant pulses of N mineralization grasslands tend to occur in the top 10 to 20 cm of the
(Fierer and Schimel, 2002). soil profile (Cheng and Bledsoe, 2002; Jackson et al.,
Analysis of the microbial biomass
15
N (Fig. 5) does 1988), these soil solution patterns likely reflect the in-
show high microbial immobilization of the applied
15
Ncreased root uptake associated with increased vegeta-
in the first few days following application. There is then tion growth in the cut buffers (Ourry et al., 1990).
a decrease in microbial
15
N in the soil within the area In the cut buffers, the 15-cm soil solution NO
3
⫺
–
15
N
of application over the course of the irrigation season, concentrations remained at a maximum closest to the
but just 1 m downslope, the microbial
15
N increases. If zone of
15
N application, but decrease with distance due
this increase were due to decomposition of
15
N-enriched to vegetative buffer uptake (Fig. 7). In the uncut buffers,
vegetation within the buffers themselves, there would 15-cm soil solution NO
3
⫺
–
15
N concentrations were vari-
likely be a difference between the cut and uncut buffers able or increased with distance due to increased down-
because the cut buffers did not show any evidence of slope movement via surface runoff and subsurface lat-
senescence during the irrigation season. Instead, the eral flow (Bedard-Haughn et al., 2004) and due to lower
increase in microbial
15
N at the 1-m distance may be plant N demand associated with senescence of the ma-
attributable to losses via root exudation and/or decom- ture vegetation (Jackson et al., 1988). In the 45-cm soil
position in the rhizosphere of the zone of
15
N application, solution samples (Fig. 7), the differences between the
uptake by microbial biomass, and subsequent mineral- cut and uncut buffers likely reflect leaching from the
ization and lateral movement of inorganic N. root zone because the lack of significant root density at
The differences in
15
N retention between the regularly this depth (Cheng and Bledsoe, 2002) makes it unlikely
cut buffers and the severely cut application zone (Fig. 4) that differences in plant uptake are the cause of differ-
highlight the importance of responsible buffer manage- ences in concentration.
ment; cutting must be managed to allow for maximum A similar pattern of higher
15
N levels in the uncut
compensatory regrowth, otherwise any benefits associ- buffers was expected for the 0- to 15-cm soil atom %
15
N
ated with cutting may be lost. excess (Fig. 8), but there were no significant differences
between the cut and uncut buffers on either sampling
Effects of Cutting on Subsurface Nitrogen-15 date. However, the soil atom %
15
N excess values re-
flected a combination of the soil and the root biomass;
The primary effect of cutting in the subsurface envi-
roots were not analyzed separately. The cut buffers are
ronment is lower NO
3
⫺
–
15
N concentrations in the soil
likely to have greater belowground partitioning of
15
N
solution within the cut buffers. Within 45 d of
15
N appli-
cation, the 15-cm soil solution samples from the cut into the root biomass due to stress effects of cutting
Reproduced from Journal of Environmental Quality. Published by ASA, CSSA, and SSSA. All copyrights reserved.
BEDARD-HAUGHN ET AL.: VEGETATIVE BUFFER EFFICACY FOR
15
N SEQUESTRATION 1663
(Paterson and Sim, 2000; Williams et al., 2003). Given to 7 to 9 mg N buffer
⫺
1
in this study (Table 2). Vegetative
that the uncut buffers have higher NO
3
⫺
–
15
N concentra- growth and vigor were comparable with surrounding
tions in the 15-cm soil solution samples (Fig. 7), uncut pasture irrigated at the typical higher rate. This empha-
solution
15
N and cut root
15
N may balance each other sizes the critical importance of managing irrigation rates
out, resulting in similar total soil
15
N values. to minimize runoff as a primary method for reducing
nutrient loading in surface water (Tate et al., 2000b).
Vegetative buffers still have a significant impact on nu-
Nitrogen-15 Recovery Budget
trient loading, but must remain secondary measures
The majority of the applied
15
N (59–71%) was recov- (Barling and Moore, 1994).
ered in the soil beneath the pasture and buffer areas, Reducing the irrigation rates by 75% also appears to
indicating that for this study, infiltration was the domi- have substantially increased the relative importance of
nant mechanism for minimizing
15
N losses in surface run- infiltration, particularly within the zone of
15
N applica-
off. A further 17 to 19% of the applied
15
N was recovered tion. Verchot et al. (1997) also found infiltration to be
by vegetation uptake. Although the majority of the infil- a major mechanism for minimizing N losses in surface
tration and uptake occurred within the zone of
15
N appli- runoff under unsaturated conditions. At the end of this
cation itself, the buffers attenuated approximately 25% study, the amount of soil
15
N stored in the A horizon
of the applied
15
N, mostly within the soil, indicating (0–15 cm) within the zone of
15
N application (Table 2)
that the buffers themselves were effective, regardless of was approximately 10 times that stored when the higher
cutting treatment. Runoff losses represented less than 1% irrigation rate was applied (Bedard-Haughn et al., 2004).
of the applied
15
N (Table 2). Note that the only permanent However, some of this greater soil storage is related to
sink for the applied
15
N was that removed in the cut vege- belowground
15
N losses from the vegetation within the
tation; all other
15
N could potentially be re-released at a zone of application (Fig. 4).
later point and become available for leaching and runoff.
Although the amount of
15
N recovered within the
buffer vegetation was low compared with the overall N CONCLUSIONS
pool, there was a significant difference in vegetation
Although both the cut and uncut buffers served to
recovery between the cut and uncut buffers, with the
attenuate
15
N in surface runoff, regular cutting of vegeta-
cut buffers recovering approximately twice as much
15
N
tion in upland buffer areas contributed to a significant
as the uncut buffers (Table 2). The absence of a signifi-
increase in plant
15
N uptake and a corresponding de-
cant difference in total
15
N recovery between the cut
and uncut buffers does not reflect the temporal improve- crease in
15
N concentration of both the surface runoff
ment in runoff water quality or vegetative uptake (Ta- and the subsurface water, indicating that cutting is a
ble 3). It does, however, reflect the absence of significant viable management technique for improving both the
differences in vegetation, runoff, soil solution, and soil capacity and effectiveness of vegetative buffers in irri-
15
N concentrations between the cut and uncut buffers gated pasture. Monthly cutting of buffer vegetation dou-
in the first 21 d of the experiment, when
15
N concentra- bled
15
N uptake compared with uncut buffers. Although
tions in all N pools were at their highest. mineralization of microbially immobilized
15
N provided
The applied
15
N that was not recovered in the runoff, an ongoing source of
15
N over the course of the irrigation
soil, or vegetation likely reflects losses due to denitrifica- season, vegetation in the cut buffers had greater N de-
tion, volatilization, or leaching within the soil profile to mand due to increased growth and potential for shoot
depths greater than 1 m. Note that runoff losses may assimilation.
be higher under the more typical granular fertilizer ap- The dominant factor affecting
15
N concentration in
plication. surface runoff from irrigated pasture is the irrigation
rate itself. Reducing the irrigation rate by 75% substan-
Runoff and Nitrogen Losses tially decreased both the volume of runoff and the con-
centration of
15
N within the runoff. This appears to be
Reducing the irrigation rate from 4 to1Ls
⫺
1
plot
⫺
1
primarily due to greater infiltration within the zone of
decreased the runoff losses by approximately 50% (Ta-
15
N application. Given this increase in infiltration with
ble 1) compared with Bedard-Haughn et al. (2004), a the lower irrigation rate, consideration must be given
buffer study on an adjacent set of plots that used the to the long-term effectiveness of infiltration as a mecha-
more typical irrigation rates for the region. The initial nism for attenuating
15
N. Nitrogen storage within the
runoff rate of 0.4 L s
⫺
1
plot
⫺
1
was identical to that ob- soil may be ephemeral and could eventually be leached
served in Bedard-Haughn et al. (2004), but the maxi- to ground water unless removed by plant uptake and
mum level of 0.7 L s
⫺
1
plot
⫺
1
was considerably lower cutting or denitrification.
than the previously measured 3 L s
⫺
1
plot
⫺
1
. This smaller However, total buffer vegetation uptake was rela-
range of runoff rates was reflected in a smaller range
tively small in this irrigated pasture, so the importance
of
15
N loads between the beginning and the end of a
of the cutting effect needs to be considered under a
given irrigation event.
broader range of N inputs and in other agroecosystems.
By reducing the irrigation rate by 75%, the total
Within-pasture fertilizer timing and irrigation manage-
amount of dissolved N lost from a given buffer de-
ment must still be considered the primary techniques
creased by six- to eightfold, from 55 mg N buffer
⫺
1
at
the 4 L s
⫺
1
irrigation rate (Bedard-Haughn et al., 2004) for minimizing NO
3
⫺
losses in irrigated pasture.
Reproduced from Journal of Environmental Quality. Published by ASA, CSSA, and SSSA. All copyrights reserved.
1664 J. ENVIRON. QUAL., VOL. 34, SEPTEMBER–OCTOBER 2005
Ferraro, D.O., and M. Oesterheld. 2002. Effect of defoliation on grass
ACKNOWLEDGMENTS growth. A quantitative review. Oikos 98:125–133.
This research was funded by the UC Water Resource Cen- Fierer, N., and J.P. Schimel. 2002. Effects of drying-rewetting fre-
ter. The authors acknowledge the support of the Sierra Foot- quency on soil carbon and nitrogen transformations. Soil Biol.
Biochem. 34:777–787.
hill Research and Extension Center and the UC Davis Stable Herbert, F.W., and E.L. Begg. 1969. Soils of the Yuba area, California.
Isotope Facility. A. Bedard-Haughn also wishes to recognize Dep. of Soils and Plant Nutr., Univ. of California, Davis, and
diligent field and laboratory assistance from D. Bedard- County of Yuba, California.
Haughn, C. Peterson, and D. Sok and fellowship/scholarship Hill, A.R. 1996. Nitrate removal in stream riparian zones. J. Environ.
support from the UC Davis Department of Agronomy and Qual. 25:743–755.
Range Science, Natural Sciences and Engineering Research Insightful Corporation. 2001. S-PLUS Version 6.0. Insightful Corp.,
Council of Canada, UC Davis Graduate Studies, UC Davis Seattle.
Jackson, L.E., J.P. Schimel, and M.K. Firestone. 1989. Short-term
Soil Science Graduate Group, and Jastro-Shields Graduate
partitioning of ammonium and nitrate between plants and microbes
Research Awards. in an annual grassland. Soil Biol. Biochem. 21:409–415.
Jackson, L.E., R.B. Strauss, M.K. Firestone, and J.W. Bartolome.
REFERENCES 1988. Plant and soil-nitrogen dynamics in California annual grass-
land. Plant Soil 110:9–17.
American Public Health Association. 1989. Standard methods for the Jarvis, S.C., and J.H. Macduff. 1989. Nitrate nutrition of grasses from
examination of water and wastewater. 17th ed. APHA, Washing- steady-state supplies in flowing solution culture following nitrate
ton, DC. deprivation and or defoliation: I. Recovery of uptake and growth
Atwill, E.R., L.L. Hou, B.A. Karle, T. Harter, K.W. Tate, and R.A. and their interactions. J. Exp. Bot. 40:965–975.
Dahlgren. 2002. Transport of Cryptosporidium parvum oocysts Matheson, F.E., M.L. Nguyen, A.B. Cooper, T.P. Burt, and D.C. Bull.
through vegetated buffer strips and estimated filtration efficiency. 2002. Fate of N-15-nitrate in unplanted, planted and harvested
Appl. Environ. Microbiol. 68:5517–5527. riparian wetland soil microcosms. Ecol. Eng. 19:249–264.
Bardgett, R.D., T.C. Streeter, and R. Bol. 2003. Soil microbes compete Mendez, A., T.A. Dillaha, and S. Mostaghimi. 1999. Sediment and
effectively with plants for organic-nitrogen inputs to temperate nitrogen transport in grass filter strips. J. Am. Water Resour.
grasslands. Ecology 84:1277–1287. Assoc. 35:867–875.
Barling, R.D., and I.D. Moore. 1994. Role of buffer strips in manage- Mulholland, P.J., J.L. Tank, D.M. Sanzone, W.M. Wollheim, B.J.
ment of waterway pollution—A review. Environ. Manage. 18: Peterson, J.R. Webster, and J.L. Meyer. 2000. Nitrogen cycling in a
543–558. forest stream determined by a
15
N tracer addition. Ecol. Monogr.
Bedard-Haughn, A., K.W. Tate, and C. van Kessel. 2004. Using
15
N70:471–493.
to quantify vegetative buffer effectiveness for sequestering N in Ourry, A., J. Boucaud, and M. Duyme. 1990. Sink control of nitrogen
runoff. J. Environ. Qual. 33:2252–2262. uptake and assimilation during regrowth after cutting of ryegrass
Bedard-Haughn, A., J.W. van Groenigen, and C. van Kessel. 2003. (Lolium perenne L). Plant Cell Environ. 13:185–189.
Tracing N-15 through landscapes: Potential uses and precautions. Paterson, E., and A. Sim. 2000. Effect of nitrogen supply and defolia-
J. Hydrol. (Amsterdam) 272:175–190. tion on loss of organic compounds from roots of Festuca rubra. J.
Bharati, L., K.H. Lee, T.M. Isenhart, and R.C. Schultz. 2002. Soil- Exp. Bot. 51:1449–1457.
water infiltration under crops, pasture, and established riparian Pinheiro, J.C., and D.M. Bates. 2000. Mixed effects models in S and
S-Plus. Springer, New York.
buffer in Midwestern USA. Agrofor. Syst. 56:249–257.
Powlson, D.S., and D. Barraclough. 1993. Mineralization and assimila-
Brookes, P.C., A. Landman, G. Pruden, and D.S. Jenkinson. 1985.
tion in soil-plant systems. p. 209–239. In R. Knowles and T.H.
Chloroform fumigation and the release of soil nitrogen: A rapid
Blackburn (ed.) Nitrogen isotope techniques. Academic Press, San
direct extraction method to measure microbial biomass nitrogen Diego, CA.
in soil. Soil Biol. Biochem. 17:837–842. Schenk, M.K. 1996. Regulation of nitrogen uptake on the whole plant
Canfield, R. 1941. Application of the line interception method in level. Plant Soil 181:131–137.
sampling range vegetation. J. For. 39:388–394. Schmitt, T.J., M.G. Dosskey, and K.D. Hoagland. 1999. Filter strip
Cheng, X.M., and C.S. Bledsoe. 2002. Contrasting seasonal patterns of performance and processes for different vegetation, widths, and
fine root production for blue oaks (Quercus douglasii) and annual contaminants. J. Environ. Qual. 28:1479–1489.
grasses in California oak woodland. Plant Soil 240:263–274. Shearer, G., and D.H. Kohl. 1993. Natural abundance of
15
N. p. 89–126.
Cole, M.L., I. Valiela, K.D. Kroeger, G.L. Tomasky, J. Cebrian, C. In R. Knowles and T.H. Blackburn (ed.) Nitrogen isotope tech-
Wigand, R.A. McKinney, S.P. Grady, and M.H. Carvalho da Silva. niques. Academic Press, San Diego, CA.
2004. Assessment of a ␦
15
N isotopic method to indicate anthro- Stark, J.M., and S.C. Hart. 1996. Diffusion technique for preparing
pogenic eutrophication in aquatic ecosystems. J. Environ. Qual. 33: salt solutions, Kjeldahl digests, and persulfate digests for nitrogen-
124–132. 15 analysis. Soil Sci. Soc. Am. J. 60:1846–1855.
Dahlgren, R.A., M.J. Singer, and X. Huang. 1997. Oak tree and grazing Tate, K.W., E.R. Atwill, M.R. George, M.K. McDougald, and R.E.
impacts on soil properties and nutrients in a California oak wood- Larsen. 2000a. Cryptosporidium parvum transport from cattle fecal
land. Biogeochemistry 39:45–64. deposits on California rangelands. J. Range Manage. 53:295–299.
Davidson, E.A., J.M. Stark, and M.K. Firestone. 1990. Microbial pro- Tate, K.W., E.R. Atwill, N.K. McDougald, and M.R. George. 2003.
duction and consumption of nitrate in an annual grassland. Ecol- Spatial and temporal patterns of cattle feces deposition on range-
ogy 71:1968–1975. land. J. Range Manage. 56:432–438.
Di, H.J., and K.C. Cameron. 2002. Nitrate leaching in temperate Tate, K.W., G.A. Nader, D.J. Lewis, E.R. Atwill, and J.M. Connor.
agroecosystems: Sources, factors and mitigating strategies. Nutr. 2000b. Evaluation of buffers to improve the quality of runoff from
Cycling Agroecosyst. 64:237–256. irrigated pastures. J. Soil Water Conserv. 55:473–478.
Di, H.J., K.C. Cameron, S. Moore, and N.P. Smith. 1999. Contribution USDA. The PLANTS database. Version 3.5 [Online]. Available at
to nitrogen leaching and pasture uptake by autumn-applied dairy http://plants.usda.gov (verified 20 Apr. 2005). USDA, Washing-
effluent and ammonium fertilizer labeled with
15
N isotope. Plant ton, DC.
Soil 210:189–198. Van Kessel, C., R.E. Farrell, and D.J. Pennock. 1994. Carbon-13 and
Dillaha, T.A., R.B. Reneau, S. Mostaghimi, and D. Lee. 1989. Vegeta- nitrogen-15 natural abundance in crop residues and soil organic
tive filter strips for agricultural nonpoint source pollution-control. matter. Soil Sci. Soc. Am. J. 58:382–389.
Trans. ASAE 32:513–519. Verchot, L.V., E.C. Franklin, and J.W. Gilliam. 1997. Nitrogen cycling
Dosskey, M.G. 2001. Toward quantifying water pollution abatement in piedmont vegetated filter zones: 1. Surface soil processes. J.
in response to installing buffers on crop land. Environ. Manage. Environ. Qual. 26:327–336.
28:577–598. Williams, B.L., L.A. Dawson, S.J. Grayston, and C.A. Shand. 2003.
Eshel, G., G.J. Levy, U. Mingelgrin, and M.J. Singer. 2004. Critical Impact of defoliation on the distribution of N-15-labelled synthetic
evaluation of the use of laser diffraction for particle size distribution sheep urine between shoots and roots of Agrostis capillaris and
soil N pools. Plant Soil 251:269–278.analysis. Soil Sci. Soc. Am. J. 68:736–743.
