| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Department of Plant Sciences, University of California, Davis, CA 95616
b Current address: Department of Soil Science, University of Saskatchewan, 51 Campus Drive, Saskatoon, SK, S7N 5A8, Canada
* Corresponding author (bedard.haughn{at}usask.ca)
Received for publication January 28, 2005.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Abbreviations: DON, dissolved organic nitrogen • LME, linear mixed effects • SFREC, Sierra Foothill Research and Extension Center
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Nitrate N is a soluble nutrient commonly cited as a source of ground- and surface-water contamination. In the United States, the legal drinking water limit for NO3––N is 10 mg L–1, but concentrations as low as 1 mg L–1 can contribute to algal blooms (Mendez et al., 1999). Nitrate has been implicated in eutrophication in seawater and fresh water (Cole et al., 2004). Measured concentrations in runoff from irrigated pasture range from 0.2 to 5 mg L–1 (Bedard-Haughn, unpublished data, 2002).
Buffer strips are broadly defined as strips of vegetation that improve or maintain water quality downslope of an agriculture or forestry operation (Barling and Moore, 1994). Buffers function to remove pollutants by reducing or filtering surface runoff and/or by filtering ground water and stream water (Dosskey, 2001). Attenuation of NO3– by buffers is attributed to a combination of factors, including denitrification, infiltration, and plant uptake (Hill, 1996). The relative importance of each factor varies according to buffer characteristics such as hydrology, vegetation type (grass vs. forest), soil type (coarse vs. fine), buffer width, and pollutant type (Bharati et al., 2002; Schmitt et al., 1999). In irrigated pasture, infiltration and plant uptake appear to have a greater impact on NO3– attenuation than denitrification (Verchot et al., 1997). A recent field study in California using 15N-enriched NO3– tracers in an irrigated pasture system found that up to 50% of applied 15N was removed by plant uptake the first 10 d following application, making uptake the dominant mechanism for N attenuation (Bedard-Haughn et al., 2004). However, minimal uptake occurred over the remainder of the growing season, even with available N in the soil. Consequently, 15N continued to be lost throughout the irrigation season via runoff, despite the presence of vegetative buffers. In examining grass buffer trapping efficiency for sediment and nutrients, Dillaha et al. (1989) reported higher levels of soluble nutrients leaving buffers than entering them, which they attributed to low trapping efficiency for soluble nutrients and to release of nutrients previously stored in the buffer. This contributes to concern that buffer efficiency may decrease over time, and that buffers will ultimately become a source of N (and other nutrients) rather than a sink (Mendez et al., 1999).
Plant N demand and uptake can be key factors in controlling N losses in many ecosystems (Mulholland et al., 2000). Demand and uptake vary with the N status of the vegetation, NO3– availability, plant growth rate, and plant age or phenology. All other factors remaining constant, plant N uptake during growth will be greater if the vegetation is N deficient or if there is an abundance 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); increasing plant age tends to decrease N uptake (Schenk, 1996). As Jackson et al. (1988) observed in the Sierra Nevada foothills, even well-watered grasses can senesce within weeks of anthesis, decreasing N demand. When new sources of N are introduced, microbial immobilization may compete effectively with plants for N (Jackson et al., 1989). Subsequent turnover and mineralization releases this previously immobilized N. In annual grasslands in the Sierra Nevada foothills, turnover of the microbial N pool can occur rapidly (less than one day) and continuously (Davidson et al., 1990; Jackson et al., 1989), necessitating continual plant demand for N to minimize nutrient losses.
It may be possible to increase plant N uptake via regular cutting, which would increase N demand by encouraging compensatory regrowth. Within two weeks after shoot harvest, uptake of N increases (Ourry et al., 1990). Matheson et al. (2002) found that although regular cutting of vegetation decreased new shoot production, it increased the NO3– assimilation capacity of shoots by a factor of 5 compared with shoots that were not cut, suggesting that even when total plant biomass is reduced by cutting, the positive effects on N sequestration might offset this.
The role of plant uptake in attenuating nutrients is diminished when nutrients are returned to the soil via decomposition, therefore periodic harvesting of buffer vegetation might improve the long-term effectiveness of buffers (Dosskey, 2001). Mowing alone will increase plant N uptake, but removal of the cut vegetation is required to prevent nutrient release via decomposition (Barling and Moore, 1994). Although grazing also removes vegetation, up to 60 to 90% of the ingested N can be returned to the pasture system, mostly as urine (Di and Cameron, 2002).
Applying 15N-enriched techniques in the field provides a powerful insight into plant–soil N dynamics (Powlson and Barraclough, 1993), commonly within a single growing season (Bardgett et al., 2003; Di et al., 1999; Jackson et al., 1989; Mulholland et al., 2000). For this study, 15N-enriched isotopes allowed new NO3– to be distinguished from NO3– already present in the system and to be quantitatively traced through the buffers (Bedard-Haughn et al., 2003).
Given the previously observed abatement in plant N uptake in mature buffers in irrigated pasture (Bedard-Haughn et al., 2004) and the potential for increasing plant N uptake via vegetation management, this study was designed to: (i) quantitatively determine whether regular cutting and removal of vegetation in buffer strips would increase plant 15N uptake and retention, (ii) measure the impact of regular cutting on attenuation of runoff 15N, and (iii) determine whether there was a corresponding impact on attenuation of 15N in the soil solution and on 15N sequestration in the soil. In addition, we considered whether decreasing irrigation rate affects buffer efficiency by comparison with results on an adjacent site. By examining the water quality measurements in conjunction with the soil and vegetation results, a complete 15N recovery budget was developed, providing insight into the relative importance of the different N sinks and pools in the function of vegetative buffers in irrigated pasture.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
|
Plot irrigation was by gated pipe, which delivered water separately to each pasture-buffer area. Irrigation rate was controlled by a valve and monitored by flow meters (Model WT; Netafim, Tel Aviv, Israel) that allowed measurement of both rate and quantity of water applied. Water was applied 5 m upslope from the buffer–pasture interface to maximize control of water distribution within the study area. During 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. (2004) study on an adjacent set of plots (Table 1). A lower irrigation rate was used in an effort to reduce runoff losses and improve irrigation efficiency. On average, 29% of the applied irrigation water was lost as runoff. Total duration of each irrigation event varied according to the volume needed to restore soil water content, which was determined using evapotranspiration data from the California Irrigation Management Information System weather station located at SFREC. Climate and plant growth conditions during the 2003 growing season were normal for the region.
Collection troughs installed across the bottom of each buffer collected surface water runoff. Three pairs of ceramic soil solution samplers (Soilmoisture Equipment, Santa Barbara, CA) were installed in each buffer area at 1, 4, and 12 m downslope of the 15N application (Fig. 1). Samplers were installed to depths of 15 and 45 cm, the average depths to the bottom of the A horizon and the top of the heavy clay Bt horizon, respectively.
Nitrogen-15 Application
In June 2003, four days after the first cutting and three days before the first irrigation, 15N-labeled KNO3 was applied in solution at a rate of 5 kg N ha–1 and 99.7 atom % 15N. The rate and atom % 15N concentration were selected to provide an approximation of post-irrigation fertilizer N levels while allowing the tracer to be detectable in all N pools throughout the duration of the experiment. The 15N solution was applied across 8 of the 10 plots (4 cut, 4 uncut). The area labeled was 1 m wide across the width of each plot and located 0.75 m above the buffer areas (Fig. 1). Application rate and area were based on Bedard-Haughn et al. (2004). Following application, the 15N fertilizer was watered in with 18 L of water per m2; under field conditions, this volume was sufficient to rinse the 15N solution off of the foliar surfaces but allowed only minimal percolation. To ensure uniform distribution of both the 15N fertilizer solution and the additional water, the application area was subdivided into m2 plots. Natural abundance background levels of 15N in all N pools were measured before application of 15N-labeled fertilizer to account for natural variability and dilution of the applied 15N fertilizer by 14N. Within a given plant species, the standard deviation of natural abundance atom % was within ±0.0002 atom %.
Isotopic levels are reported as atom % 15N excess, which refers to the amount of 15N present relative to the average naturally occurring background 15N levels for that particular N pool. Atom % 15N excess amounts were extrapolated to obtain the total amount of 15N in a given pool by weight and/or volume and thus to determine a 15N budget.
Vegetation Sampling and Analysis
Grab samples of vegetation were collected 3, 11, 21, 42, 60, 79, 98, and 114 d after 15N application. To determine how far the 15N fertilizer had moved into the buffers, vegetation samples were collected along a cross-slope transect within the zone of 15N application and at downslope distances of 1, 4, 8, 12, and 16 m from the application area. The uncut buffer vegetation samples were separated by the three dominant grass species, whereas cut buffer vegetation samples represented composites of all species present due to identification obstacles associated with newly clipped vegetation. All plant samples were oven-dried at 65°C and analyzed for 15N isotopic composition via mass spectrometry (Integra Integrated Stable Isotope Analyzer; Europa Scientific, Crewe, UK) at the University of California, Davis, Stable Isotope Facility (van Kessel et al., 1994). The current sensitivity of our stable isotope ratio mass spectrometers is 0.0002 atom % 15N.
Of the two plots that did not receive 15N, 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 composition, vegetation age, and irrigation rates of these two nonlabeled buffers were equivalent to the labeled buffers. Accurate biomass measurements could not be taken from the labeled buffers without compromising results, so on each sampling day, representative biomass measurements were taken from the two nonlabeled buffers (Fig. 1). All living biomass within a randomly placed 0.1-m2 quadrat was collected, dried, and weighed. For the cut buffer, three composite quadrat measurements were collected on each day. For the uncut buffer, one representative measurement was taken for each of the three dominant species. Although this lower number contributed to greater variability for uncut biomass values, it allowed for regular sampling over the season without eradicating the less prevalent species. Cover measurements for the 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.
Vegetation N content was multiplied by atom % 15N excess values to get the mass (mg) of 15N in each g of vegetation. The total mass (mg) of 15N sequestered in vegetation in a given buffer area was determined by multiplying the mg 15N g–1 vegetation values times biomass values (g m–2) and extrapolating to the whole area using cover data.
Runoff Sampling and Analysis
Runoff samples were collected on the same dates as vegetation samples. Samples were taken from the collection troughs 15 min following the leading edge of runoff and again just before the end of the irrigation event and were stored frozen until analysis. Based on results from an adjacent irrigated site, these two measurements captured the maximum variability during the irrigation period (Bedard-Haughn et al., 2004). The 15-min interval provided a measurement of maximum 15N concentration, whereas the event-end sample reflects the minimum 15N concentration, but the maximum 15N load. Sample collection (500 mL) was as a "grab" sample from the runoff collection trough. Runoff rates were determined at regular intervals by measuring the volume of runoff in a 5-s period. Runoff rate data were used to determine runoff losses (Table 1).
Runoff 15N isotope analyses were performed on three N pools: NO3–, NH4+, and total N. Samples were filtered to remove sediment and vegetation residues from runoff. The NH4+–15N and NO3––15N were determined by NH3 diffusion of a 100-mL aliquot onto polytetrafluoroethylene-encased acid traps (Stark and Hart, 1996). To measure NO3––15N, the Stark and Hart (1996) method was modified using TiCl3 (Titanous Chloride Solution, 20%; Fisher, Hampton, NH) to reduce NO3– to NH3 as outlined in Bedard-Haughn et al. (2004). Total 15N was determined on a separate 20-mL aliquot by performing a persulfate digestion (American Public Health Association, 1989) to convert the dissolved organic nitrogen (DON) and NH4+ to NO3–, and samples were then diffused for NO3– as above. Following diffusion, acid disks were removed from polytetrafluoroethylene packets and analyzed via mass spectrometry. The DON–15N for each sample was calculated using an isotope mixing model via difference from total 15N (Shearer and Kohl, 1993):
 | [1] |
Soil Sampling and Analysis
Soil samples were taken at 0, 1, 4, and 12 m from the 15N application at 3 and 114 d following 15N application. On both dates, samples were taken to a 15-cm depth in two increments (0–7 and 7–15 cm) using a slack hammer (Ben Meadows Company, Janesville, WI), corresponding to the depth of the A horizon. Soil texture was determined on the 114-d samples using laser diffraction and reported in volume percent (Eshel et al., 2004). On Day 114, soil samples were also taken to a 1-m depth in two increments (0–40 and 40–100 cm) to allow for a more complete estimate of the final 15N budget. Soil samples were oven-dried at 40°C and analyzed for total N and 15N via mass spectrometry. Soil C was analyzed by mass spectrometry in conjunction with soil N. Bulk density measurements for all depth increments were done on oven-dried intact cores and were within the range of values measured by Dahlgren et al. (1997) in SFREC grazed pasture.
Soil microbial 15N was measured using fumigation–extraction method (Brookes et al., 1985), with fumigation for 48 h with chloroform vapor and extraction with 0.5 M K2SO4. Extract 15N was determined by persulfate digestion (American Public Health Association, 1989) to convert the DON and NH4+ to NO3–, and diffusion using a modification of the Stark and Hart (1996) method, as outlined in Bedard-Haughn et al. (2004). Microbial 15N was determined by difference between fumigated and nonfumigated samples for both dates (0–15 cm only) for the 0- and 1-m distances.
Soil Solution Sampling and Analysis
Immediately before each irrigation event, vacuum was applied to the soil solution sampling tubes and allowed to draw moisture from the soil for 10 d before sample collection (Bedard-Haughn et al., 2004). Although vacuum was not applied continuously over the 10-d period, suction was still present at sampling in most sampling tubes. After Day 42, the time between irrigation and sample collection was shortened from 9 to 3 d, which substantially improved the reliability of the suction in the tubes and the volume of sample collected. Soil solution samples were stored frozen until analysis for NO3––15N via the TiCl3 diffusion (25-mL aliquots) as outlined in Bedard-Haughn et al. (2004).
Nitrogen-15 Recovery Budget
The 15N recovery budget illustrates the mass of 15N sequestered and/or measured in runoff relative to the mass of 15N applied at the beginning of the study. For soil and vegetation samples, atom % 15N excess amounts were extrapolated using total N content, soil bulk density (mass/volume), and vegetation biomass (mass/area) to obtain the total amount of 15N in a given sink by mass and thus to determine the total amount of 15N stored in the pasture-buffer areas. The total amount of 15N lost via runoff (15N load) during a given irrigation event was determined by multiplying runoff volume by 15N concentrations for each measured interval and integrating over time. Summing these values for the measured irrigation events, together with quantitative estimates for intervening irrigation events where runoff was not measured, provided a value for the total amount of 15N lost as runoff from the pasture-buffer areas. There were no total flux measurements for soil solution, so total subsurface 15N load could not be calculated.
Statistical Analysis
The results were analyzed using linear mixed effects model analysis (S-PLUS; Insightful Corporation, 2001). Linear mixed effects analysis can be applied to both structured and observational studies (Pinheiro and Bates, 2000) and was used here to account for the influence of fixed (cutting) effects on buffer 15N uptake levels and for the repeated measures (group effect– plot identity) embedded in the data structure. Treating time as a fixed effect provided a test of how response varied over the duration of the study. The magnitude and direction (±) of the coefficient for treatment and time effects was used to define the relationship between 15N uptake and runoff 15N load and cutting effects. This flexible model also allowed within-group variance and correlation structures for handling within-group (plot) heteroscedasticity and temporally correlated errors (irrigation series within year) (Pinheiro and Bates, 2000). This approach has been used in modeling other complex longitudinal datasets (Atwill et al., 2002; Tate et al., 2000a, 2003). The soil and soil solution data were analyzed using the nonparametric Wilcoxon rank sum test (S-PLUS; Insightful Corporation, 2001), which is not restricted by assumptions of normality.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
|
|
|
|
A similar decrease in 15N mass within the zone of 15N application was observed in the soil microbial biomass (Fig. 5) . In both the 0- to 7- and 7- to 15-cm depth increments, the amount of microbial 15N decreased between Days 3 and 114. In contrast, just 1 m downslope, the amount of microbial 15N increased between Days 3 and 114 in both depth increments. There were no significant differences in microbial 15N content between the cut and uncut buffers, regardless of date.
|
Surface Runoff
Runoff rates averaged 0.4 L s–1 plot–1 (SD ± 0.1) within 15 min of the start of runoff and leveled off at approximately 0.7 L s–1 plot–1 (SD ± 0.2) by the end of the 3-h irrigation event. When 15N concentrations were multiplied by runoff volume to calculate the total load of 15N in runoff over a given irrigation event, the 15N load in runoff in all N pools was greater from the uncut buffer than from the cut buffer after Day 42 (Fig. 6)
. The NO3––15N load decreased to a steady state by Day 42, NH4+–15N load increased to a steady state by Day 42, and DON–15N load remained relatively level throughout the study. Maximum NO3––15N was lost in the first 21 d after 15N application, and maximum differences in NO3––15N load between the cut and uncut buffers appeared after Day 60. For the NH4+ and DON pools, significant differences between the cut and uncut buffers started to appear as early as Day 42. The data gap on Day 60 is due to the occurrence of an isolated precipitation event on that sampling day; the total volume of precipitation was comparable with the volume during a typical irrigation event. Over one-half of the precipitation fell within 1 h; the total duration of the event was 8 h. For Day 60, vegetation and soil solution samples could be collected, but there was no measurable runoff.
|
0.05) on Day 42, whereas for total dissolved N, cutting had a significant effect by Day 21 (LME P = 0.05). For the DON pool, cutting reduced the 15N load (LME P = 0.08) regardless of time since 15N application; adding time as a fixed effect improved the significance slightly, but not enough to warrant its inclusion in the model.
Subsurface: Soil Solution and Soil
The NO3––15N concentration of the soil solution (Fig. 7)
was similar in range to the 15N concentration of the NO3––15N in runoff (Fig. 6), but the soil solution NO3––15N concentrations tended to be much more variable. This was particularly true in the first 42 d after 15N application during which time the samples were collected 10 d after irrigation, versus after 3 d.
|
This difference between the cut and uncut buffers for NO3––15N concentrations in the subsurface water was not reflected in the 0- to 15-cm soil atom % 15N excess (Fig. 8) . There was no significant difference in soil atom % 15N excess between the cut and uncut buffers on either sampling date (P = 0.7, Wilcoxon rank sum test). There was also no difference between sampling dates. The only general pattern was a decrease in atom % 15N excess with increasing distance from the zone of 15N application.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
In comparing the N capacity of cut and uncut buffers over the irrigation season, the cut buffers sequestered twice the 15N of the uncut buffers (Fig. 4). Given that the cut and uncut buffers had very similar atom % 15N excess values for much of the irrigation season, the difference in sequestration can be attributed primarily to increases in biomass in the cut buffers. The increase in biomass following each cutting (Fig. 3) was a typical compensatory response to defoliation (Ferraro and Oesterheld, 2002). This period of growth should be a period of high N demand (Jackson et al., 1988). Cutting of aboveground vegetation can increase shoot N assimilation by more than 5 times (Matheson et al., 2002). Cutting and removing vegetation from the buffers allowed the standing biomass to take advantage of immobilized soil 15N as it was released by microbial mineralization (Bardgett et al., 2003). Removal of the cut vegetation is essential, otherwise decomposition will simply return nutrients to the system, increasing the potential for losses via runoff or leaching (Dosskey, 2001). In contrast, the uncut buffers showed very little change in 15N sequestration throughout the irrigation season (Fig. 4), suggesting the occurrence of senescence and a corresponding decrease in N demand (Jackson et al., 1988) or the absence of net growth during the study.
Examining interspecific differences in 15N was expected to provide insight into the functioning of the uncut buffers, and to help determine whether one species might be better suited for buffers in irrigated pasture systems; however, the primary determinant of 15N sequestration was total aboveground biomass within the buffer. Dallis grass had greater biomass per m2 than orchard grass (Fig. 3), but orchard grass was by far more prevalent with the buffers (averaging 60% of buffer area, compared with 20% of buffer area for dallis grass), and so served to sequester the most 15N of the three species. Although the cut buffers were not examined by species, orchard grass appeared to be the dominant species within the cut buffers (and in the surrounding pasture), attributable in large part to its rapid regrowth following grazing or cutting, compared with moderate and slow regrowth for dallis grass and velvet grass, respectively (USDA, 2005). Rapid regrowth following cutting and extent of ground cover are likely the best predictors of plant uptake ability in managed buffers.
Cut or uncut, maximum plant uptake of 15N occurred within the first 4 m of the buffer (Fig. 2). The higher vegetation atom % 15N excess observed in the first meter downslope of the application area for the uncut buffers compared with the cut buffers may be attributable in part to differences in root biomass. Although soil moisture is a major factor controlling fine root production in annual grasslands (Cheng and Bledsoe, 2002), cutting may have reduced the root biomass (Williams et al., 2003), increasing root turnover, or inhibited the production of new roots (Matheson et al., 2002). If the uncut buffers had greater root biomass, the vegetation immediately downslope of the zone of application may have been better able to take advantage of 15N moving downslope via lateral movement, or, particularly if there was significant lateral root development, drawn on the much higher concentrations of 15N available within the zone of application itself.
Effects of Cutting on Surface Runoff Nitrogen-15
For the first 21 d following application of the 15N tracer, there were no differences in surface runoff NO3––15N between the cut and uncut buffers (Fig. 6). Regardless of cutting treatment, there was excess 15N measured in the surface runoff during the first irrigation event after the 15N was applied. Similarly, there was 15N measured in the soil solution (Fig. 7), soil (Fig. 8), and vegetation (Fig. 2) at the furthest distance from the zone of 15N application following the first irrigation event. This suggests that both the cut and uncut buffers were attenuating some 15N, but during this period, the NO3––15N tracer was extremely mobile and its redistribution via surface runoff was identical for both the cut and uncut buffers. As Di and Cameron (2002) observed, maximum NO3– leaching tends to occur whenever NO3– is present in the soil profile during periods of significant drainage, as would be associated with irrigation events. It is interesting to note, however, that the sharp decrease in surface runoff NO3––15N between Days 21 and 42 (Fig. 6) corresponds to the first post–15N application cutting of the buffer vegetation, suggesting a very strong initial cutting effect on runoff water quality.
During Days 3 to 42, some of the NO3––15N appears to have been immobilized by microbial biomass and cycled into other N pools, as shown by the parallel decrease in runoff NO3––15N load and increases in runoff DON– and NH4+–15N (Fig. 6). One possible pathway for the movement between the NO3–– and NH4+–15N pools is dissimilatory nitrate reduction to ammonium. Although it is unlikely here given the inherently low soil NH4+ levels, this could not be confirmed with our field tracer study design. In a microcosm study, Matheson et al. (2002) found that within 32 d, up to 49% of applied NO3––15N was reduced to NH4+–15N and up to 25% was immobilized in the microbial biomass. After Day 42, there were significantly lower 15N loads in runoff from the cut buffers compared with the uncut buffers for both the NO3–– and NH4+–15N pools (Fig. 6). This corresponds with the observed increase in aboveground plant biomass and plant 15N storage in the cut buffers (Fig. 4). However, the difference between cut and uncut buffers did not increase substantially after 42 d despite continued growth in the cut buffers, indicating that even though there was a continual demand for 15N, mineral N was available for 15N runoff losses, albeit at extremely low concentrations (<1 µg total dissolved N L–1 runoff).
Unlike the runoff NH4+–15N pool, there was no lag time between the application of 15N and the leveling off in runoff DON–15N load. This may reflect the observation of Davidson et al. (1990) that these grassland soils have a significant heterotrophic microbial sink for NO3–, particularly when NH4+ availability is low. There was also minimal temporal effect on the efficacy of cutting for reducing runoff DON–15N load: cut buffers had consistently lower runoff DON–15N load than uncut buffers throughout the experiment (P = 0.08; Fig. 6). The higher DON–15N in runoff from the uncut buffers may reflect slightly greater partitioning of mineral 15N to the microbial pool in the absence of significant plant demand. For example, Jackson et al. (1989) observed that microbial immobilization of NO3– and NH4+ was greater than plant uptake regardless of plant growth stage, and for NH4+, the relative dominance of immobilization was even more pronounced after plant senescence. In this study, however, the partitioning was not significant enough to be reflected in microbial biomass 15N (Fig. 5). The constancy of the DON–15N load throughout the experiment is indicative of a rapid N turnover; Davidson et al. (1990) observed a turnover time of 0.3 to 1.6 d in grassland soils at SFREC.
Effects of Severe Cutting on Nitrogen-15 Retention
The buffer areas that were cut regularly exhibited good 15N retention due to the continual plant demand for N as it was released by the microbial biomass (Fig. 4). The uncut buffers also had good N retention within the time frame of the irrigation season, but previous research in irrigated pasture (Bedard-Haughn et al., 2004) suggests that plant decomposition during the winter months would ultimately contribute to N losses from the uncut buffers. The rate and amount of new growth and hence new N demand within uncut buffers will determine how much of the recycled 15N will be retained over the long term.
Even during the course of the irrigation season, there were 15N losses observed within the application zone vegetation (Fig. 4), despite regular cutting and removal of vegetation 15N. This may be due to unintended severe cutting (i.e., too short) of the vegetation in the application zone compared with the buffer areas; vegetative growth and vigor in this zone after 42 d was limited. Severe cutting has been found to contribute to elevated rates of root death (Jarvis and Macduff, 1989). Cutting can also give rise to increased partitioning of N to the belowground biomass and increased rhizodeposition (Paterson and Sim, 2000). In a study using 15N-enriched synthetic sheep urine, Williams et al. (2003) observed more 15N in the soil when vegetation was subject to regular cutting. This belowground partitioning associated with aboveground cutting and the increased potential for root death, coupled with the high C levels already present in the rhizosphere, provide optimum conditions for microbial 15N uptake (Jackson et al., 1989). As observed, soil microbes can compete effectively for both NH4+ and NO3– (Bardgett et al., 2003; Davidson et al., 1990). This immobilization of inorganic N can be an effective mechanism for minimizing leaching losses in agroecosystems (Di et al., 1999), but given the rapid turnover of the microbial N pool, the positive effects may be temporary (Jackson et al., 1989). This may be of particular importance where frequent wet–dry cycles occur, as is the case in irrigated pasture, because wetting cycles can cause significant pulses of N mineralization (Fierer and Schimel, 2002).
Analysis of the microbial biomass 15N (Fig. 5) does show high microbial immobilization of the applied 15N in the first few days following application. There is then a decrease in microbial 15N in the soil within the area of application over the course of the irrigation season, but just 1 m downslope, the microbial 15N increases. If this increase were due to decomposition of 15N-enriched vegetation within the buffers themselves, there would likely be a difference between the cut and uncut buffers because the cut buffers did not show any evidence of senescence during the irrigation season. Instead, the increase in microbial 15N at the 1-m distance may be attributable to losses via root exudation and/or decomposition in the rhizosphere of the zone of 15N application, uptake by microbial biomass, and subsequent mineralization and lateral movement of inorganic N.
The differences in 15N retention between the regularly cut buffers and the severely cut application zone (Fig. 4) highlight the importance of responsible buffer management; cutting must be managed to allow for maximum compensatory regrowth, otherwise any benefits associated with cutting may be lost.
Effects of Cutting on Subsurface Nitrogen-15
The primary effect of cutting in the subsurface environment is lower NO3––15N concentrations in the soil solution within the cut buffers. Within 45 d of 15N application, the 15-cm soil solution samples from the cut buffers established the spatial pattern of decreasing NO3––15N with increasing distance from the zone of 15N application; at the same time, the difference in soil solution NO3––15N concentration between the cut and uncut buffers became much more pronounced (Fig. 7). These patterns complement the decrease in 15N load in the surface runoff from the cut buffers after 42 d (Fig. 6). Given that maximum root concentrations in California grasslands tend to occur in the top 10 to 20 cm of the soil profile (Cheng and Bledsoe, 2002; Jackson et al., 1988), these soil solution patterns likely reflect the increased root uptake associated with increased vegetation growth in the cut buffers (Ourry et al., 1990).
In the cut buffers, the 15-cm soil solution NO3––15N concentrations remained at a maximum closest to the zone of 15N application, but decrease with distance due to vegetative buffer uptake (Fig. 7). In the uncut buffers, 15-cm soil solution NO3––15N concentrations were variable or increased with distance due to increased downslope movement via surface runoff and subsurface lateral flow (Bedard-Haughn et al., 2004) and due to lower plant N demand associated with senescence of the mature vegetation (Jackson et al., 1988). In the 45-cm soil solution samples (Fig. 7), the differences between the cut and uncut buffers likely reflect leaching from the root zone because the lack of significant root density at this depth (Cheng and Bledsoe, 2002) makes it unlikely that differences in plant uptake are the cause of differences in concentration.
A similar pattern of higher 15N levels in the uncut buffers was expected for the 0- to 15-cm soil atom % 15N excess (Fig. 8), but there were no significant differences between the cut and uncut buffers on either sampling date. However, the soil atom % 15N excess values reflected a combination of the soil and the root biomass; roots were not analyzed separately. The cut buffers are likely to have greater belowground partitioning of 15N into the root biomass due to stress effects of cutting (Paterson and Sim, 2000; Williams et al., 2003). Given that the uncut buffers have higher NO3––15N concentrations in the 15-cm soil solution samples (Fig. 7), uncut solution 15N and cut root 15N may balance each other out, resulting in similar total soil 15N values.
Nitrogen-15 Recovery Budget
The majority of the applied 15N (59–71%) was recovered in the soil beneath the pasture and buffer areas, indicating that for this study, infiltration was the dominant mechanism for minimizing 15N losses in surface runoff. A further 17 to 19% of the applied 15N was recovered by vegetation uptake. Although the majority of the infiltration and uptake occurred within the zone of 15N application itself, the buffers attenuated approximately 25% of the applied 15N, mostly within the soil, indicating that the buffers themselves were effective, regardless of cutting treatment. Runoff losses represented less than 1% of the applied 15N (Table 2). Note that the only permanent sink for the applied 15N was that removed in the cut vegetation; all other 15N could potentially be re-released at a later point and become available for leaching and runoff.
Although the amount of 15N recovered within the buffer vegetation was low compared with the overall N pool, there was a significant difference in vegetation recovery between the cut and uncut buffers, with the cut buffers recovering approximately twice as much 15N as the uncut buffers (Table 2). The absence of a significant difference in total 15N recovery between the cut and uncut buffers does not reflect the temporal improvement in runoff water quality or vegetative uptake (Table 3). It does, however, reflect the absence of significant differences in vegetation, runoff, soil solution, and soil 15N concentrations between the cut and uncut buffers in the first 21 d of the experiment, when 15N concentrations in all N pools were at their highest.
The applied 15N that was not recovered in the runoff, soil, or vegetation likely reflects losses due to denitrification, volatilization, or leaching within the soil profile to depths greater than 1 m. Note that runoff losses may be higher under the more typical granular fertilizer application.
Runoff and Nitrogen Losses
Reducing the irrigation rate from 4 to 1 L s–1 plot–1 decreased the runoff losses by approximately 50% (Table 1) compared with Bedard-Haughn et al. (2004), a buffer study on an adjacent set of plots that used the more typical irrigation rates for the region. The initial runoff rate of 0.4 L s–1 plot–1 was identical to that observed in Bedard-Haughn et al. (2004), but the maximum level of 0.7 L s–1 plot–1 was considerably lower than the previously measured 3 L s–1 plot–1. This smaller range of runoff rates was reflected in a smaller range of 15N loads between the beginning and the end of a given irrigation event.
By reducing the irrigation rate by 75%, the total amount of dissolved N lost from a given buffer decreased by six- to eightfold, from 55 mg N buffer–1 at the 4 L s–1 irrigation rate (Bedard-Haughn et al., 2004) to 7 to 9 mg N buffer–1 in this study (Table 2). Vegetative growth and vigor were comparable with surrounding pasture irrigated at the typical higher rate. This emphasizes the critical importance of managing irrigation rates 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 nutrient loading, but must remain secondary measures (Barling and Moore, 1994).
Reducing the irrigation rates by 75% also appears to have substantially increased the relative importance of infiltration, particularly within the zone of 15N application. Verchot et al. (1997) also found infiltration to be a major mechanism for minimizing N losses in surface runoff under unsaturated conditions. At the end of this study, the amount of soil 15N stored in the A horizon (0–15 cm) within the zone of 15N application (Table 2) was approximately 10 times that stored when the higher irrigation rate was applied (Bedard-Haughn et al., 2004). However, some of this greater soil storage is related to belowground 15N losses from the vegetation within the zone of application (Fig. 4).
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
The dominant factor affecting 15N concentration in surface runoff from irrigated pasture is the irrigation rate itself. Reducing the irrigation rate by 75% substantially decreased both the volume of runoff and the concentration of 15N within the runoff. This appears to be primarily due to greater infiltration within the zone of 15N application. Given this increase in infiltration with the lower irrigation rate, consideration must be given to the long-term effectiveness of infiltration as a mechanism for attenuating 15N. Nitrogen storage within the soil may be ephemeral and could eventually be leached to ground water unless removed by plant uptake and cutting or denitrification.
However, total buffer vegetation uptake was relatively small in this irrigated pasture, so the importance of the cutting effect needs to be considered under a broader range of N inputs and in other agroecosystems. Within-pasture fertilizer timing and irrigation management must still be considered the primary techniques for minimizing NO3– losses in irrigated pasture.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
15N isotopic method to indicate anthropogenic eutrophication in aquatic ecosystems. J. Environ. Qual. 33:124–132.
