Published online 1 March 2006
Published in J Environ Qual 35:575-583 (2006)
DOI: 10.2134/jeq2005.0364
© 2006 American Society of Agronomy, Crop Science Society of America, and Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
TECHNICAL REPORTS
Landscape and Watershed Processes
Phosphorus and Sediment Loss in a Catchment with Winter Forage Grazing of Cropland by Dairy Cattle
R. W. McDowell*
AgResearch Ltd., Invermay Agricultural Centre, Private Bag 50034 Mosgiel, Otago, New Zealand
* Corresponding author (richard.mcdowell{at}agresearch.co.nz)
Received for publication September 20, 2005.
 |
ABSTRACT
|
|---|
The loss of phosphorus and sediment to surface waters can impair their quality. It was hypothesized that the practice of winter grazing dairy cattle on cropland of moderate slope (5–20%) would exacerbate the loss of P and suspended sediment (SS) from land to water. In a small (4.3 ha) catchment two flumes were installed, upstream and downstream of one field (about 2 ha) that had been cropped for 2 yr and grazed in winter (June–July) by dairy cattle. Flow proportional samples were taken and measured for dissolved reactive phosphorus (DRP), particulate phosphorus (PP), total phosphorus (TP), and SS. During the 2002 hydrologic year (March–February) loads of SS increased per hectare downstream (1449 kg ha–1) compared to upstream (880 kg ha–1). The same increase from upstream (873 kg ha–1) to downstream (969 kg ha–1) happened in 2003. However, while in 2003 TP increased downstream by 1.64 kg ha–1 compared to upstream (0.24 kg ha–1), in 2002 an increase of only 0.006 kg ha–1 at the downstream flume occurred compared to upstream (0.98 kg ha–1). Investigation of P transport pathways suggested that overland flow contributed <0.1 kg P ha–1 to stream flow, 10 and 5% of TP load in 2002 and 2003, with the greater load in 2002 reflecting more rainfall in that year. The contribution to stream flow by subsurface flow was estimated at 0.3 kg P ha–1. Stream bed sediments showed an increase in total P concentration in summer when no flow occurred due to the admission by the farmer of 10 cattle upstream of the cropped paddock in summer 2001–2002 and 20 cattle between the two flumes in 2003 to graze stream banks. This action was calculated to contribute via dung at least, the remaining P lost: about 0.5 kg P in 2002 and 1.0 kg P in 2003. Clearly, not allowing animals to "clear-up" stream banks is a priority if good surface water quality is to be achieved. Furthermore, compared to stock access the impact of winter grazing cropland on P losses was minimal, but SS load was increased by an average of 75%.
Abbreviations: DRP, dissolved reactive phosphorus • DURP, dissolved unreactive phosphorus • PP, particulate phosphorus • SS, suspended sediment • TDP, total dissolved reactive phosphorus • TP, total phosphorus
|
INTRODUCTION
|
|---|
THE LOSS OF PHOSPHORUS (P) and suspended sediment (SS) from land is a significant cause of impaired surface water quality (Carpenter et al., 1998). Much of these losses are attributed to diffuse agricultural sources and dairying has been highlighted as a contributor (Parkyn and Wilcock, 2004). In cooler areas of New Zealand, such as Otago and Southland, the lack of winter pasture growth requires an alternative strategy for feed supply to stock, which is to graze cows on paddocks of forage crops (e.g., Brassica rutabaga L.) in June and July. This protects remaining pastures on the farm from pugging damage, but results in dung deposition and soil disturbance of the cropped fields when the soil is typically wet (Drewry and Paton, 2005). This can result in increased P and SS transfer in overland flow and impaired surface water quality.
Several studies have shown that grazing and dung deposition enhances the loss of nutrients, and sediment in overland flow and stream flow compared to areas without regular grazing. For instance, Doran et al. (1981) showed a 1.1- to 1.8-fold increase in nutrient (e.g., NH4–N, NO3–N, soluble, and total P) loss in overland flow from grazed pasture compared to when pasture was not grazed. Numerous studies have also shown that losses of sediment and P are much greater from arable land and catchments than those dominated by pasture. For example, Wallbrink et al. (2003) used radionuclides to show that sediment and P losses were 84 and 42 times greater from cultivated than pasture areas within a catchment in Australia. However, very few have investigated P and SS losses from catchments with grazed winter forage crops.
Among the exceptions, McDowell et al. (2003a) examined the losses of sediment and P from small (2 m2) bounded plots at two positions on a hillslope within the catchment studied in this paper. Results showed that losses from grazed cropland were about four times that lost from ungrazed pasture and twice that lost from ungrazed cropland. Furthermore, losses from steep (20%) plots near the stream channel were about twice that from plots on gently rolling (<5%) land at the top of the slope. If slope is excluded as a cause of enhanced losses from steep plots, another reason is due to more flow from near-stream steep plots as saturated areas expand out from the stream channel in response to rainfall resulting in saturation-excess overland flow (Needleman et al., 2004). In this catchment, rainfall intensities are consistently low (<5 mm h–1) and flow is controlled by the prevailing soil type, which has a B horizon 25 to 30 cm below the surface and a hydraulic conductivity of <1 mm h–1. In a second study, plots near the stream channel were kept (McDowell et al., 2005a), and it was shown that the fecal indicator bacteria Escherichia coli was lost in proportion to the number of dung depositions. The E. coli data were then used to determine the contribution of P from dung and revealed that the number of dung pats on grazed cropland was a better indicator of particulate phosphorus (PP) loss than SS, which was lost in similar quantities to ungrazed cropland. The minimal increase in sediment loss was attributed to the deposition and retention of sediment in hoof imprints up to 20 cm deep, while lighter dung particles were easily lost in overland flow (McDowell et al., 2003b).
By taking the mean of grazed cropland plots in the studies of McDowell et al. (2005a, 2005b) and, scaling-up to a hectare, it was estimated that the quantity of P lost in overland flow would be less than 0.1 kg per year: surprisingly small considering most dairy grazed pastures in New Zealand will lose on average around 1.0 kg P ha–1 annually, half of which is estimated to be from overland flow (Wilcock, 1986; Gillingham and Thorrold, 2000). This raises a number of questions. If we assume that P and sediment loss from grazed cropland is greater than grazed pasture, then overland flow may not be the main mechanism for P loss in grazed cropland, or plots in the study of McDowell et al. (2005a) may not have been positioned in the right place to detect areas of greatest overland flow, or plots suffered from a lack of scale. For instance, it is well known that sediment losses change as a result of scale (Le Bissonnais et al., 1998). Alternatively, it may be that grazed cropland does not cause increased P losses compared to grazed pasture.
To answer these questions and to address a research gap surrounding P and SS loss from grazed winter forage crops, flow, P, and sediment data were collected in a stream draining a catchment (approximately 4.3 ha) in South Otago where one field (approximately 2 ha) was used for grazed winter forage crops for 2 yr (2002–2004). Flumes were located in the stream receiving flow from pastures before and after the cropped paddock. Additional samples were taken and analyzed for E. coli to indicate the presence of dung. The objective of this paper was to determine the losses in stream flow of P and SS at points upstream and downstream of the cropped paddock and elucidate the relative sources and mechanisms involved.
|
MATERIALS AND METHODS
|
|---|
Site Description
The site was located in a 2-ha field within the 4.3-ha Dull Burn catchment (NZ Map Grid East 2256630, North 5449720) near Balclutha, South Otago, New Zealand (Fig. 1). The soil, a Waitahuna silt loam (NZ Classification, Mottled Fragic Pallic soil; USDA Taxonomy, Fragiochrept) had, until 1998, been used for sheep and beef cattle grazing. Conversion to dairy farming has resulted in increased fertilizer inputs, grazing frequency, and winter cropping with Brassica spp. In November 2001, one paddock was plowed while the rest of the catchment remained in pasture that was not used for grazing between June and July. A seedbed was prepared, sown with a mixture of swede (Brassica rutabaga L.) and kale (Brassica oleracea L.), and 32 kg N, 58 kg P, and 18 kg K ha–1 applied. In December, this was followed by 46 kg N and 12 kg B ha–1, and an application of 46 kg N ha–1 in May. Cultivation, resowing, and fertilization was repeated at the same time the following year. On the first week of June a 240 cow herd stocked at 80 cows ha–1 was allowed to roam the site for 3 wk until all winter forage was consumed. Over the 3 wk of grazing the herd was supplied with sufficient feed to maintain condition (approximately 85 MJ ME cow–1 d–1). All paddocks, cropped or in pasture, were fenced to prevent cattle access to the stream. However, against advice, the farmer grazed the upper part of the stream channel with about 10 calves in autumn 2002 and all the channel with 20 heifers in early summer 2003 for 2 d to remove rank pasture from stream banks.
Steep topography (5 to 20%) within the catchment has resulted in a layer of stream sediment 1 m deep above schist bedrock. Rainfall during the duration of the trial was about 870 mm in 2002 and 650 mm in 2003 (30-yr mean 820 mm). Mean annual temperature, taken at a meteorological site 4 km away, is about 9°C. Characteristically there is no flow between January and April, with the stream bed drying to approximately 30 cm depth by April. During the rest of the year flow is characterized by low baseflow (<1.5 L s–1) and periodic storm events (up to 25 L s–1) (Fig. 2). In February 2002, H-flumes were installed approximately 50 and 200 m from the head of the stream, flow (L s–1) data collected every 20 min, and samples (1 L) taken at upstream and downstream flumes in proportion to flow during flood events triggered when flow reached >1.5 L s–1. Additional grab samples were taken during baseflow at the outlet of each flume on a monthly basis. Monitoring of flow stopped at the site in April 2004.
View larger version (19K):
[in this window]
[in a new window]
|
Fig. 2. Flow at the upstream and downstream flumes. Gray bands indicate when the cropped paddock was grazed.
|
|
Sample Analysis
Flow samples were collected from flumes and analyzed for dissolved reactive phosphorus (DRP) within 24 h and total dissolved reactive phosphorus (TDP) after persulfate digestion (Eisenreich et al., 1975) within 48 h. An unfiltered sample was also digested and total phosphorus (TP) measured within 7 d. Fractions defined as dissolved unreactive (largely organic) phosphorus (DURP) and particulate phosphorus (PP) were determined as the difference between TDP and DRP, and TP less TDP, respectively. All P analyses used the colorimetric technique of Watanabe and Olsen (1965). Suspended sediment (SS) was determined by weighing the residue left after filtration through a Whatman (Maidstone, UK) GF/F glass fiber filter paper of a shaken 250-mL subsample. Escherichia coli was measured as the preferred fecal indicator bacteria for freshwater in New Zealand (Ministry for Environment, 2003). Live E. coli were counted in 100-mL samples of flow using the Colilert media and the Quanti-Tray enumeration system (IDEXX Laboratories, Westbrook, ME). Unfortunately, due to equipment failure, monitoring of E. coli ceased at the end of September 2003.
The annual loads of P fractions, SS, and E. coli were calculated via interpolation of measurements taken during baseflow and flood events. For the purpose of loads a year is defined as from March to February, in line with the normal flow regime. Before interpolation, measurements taken at the same time for upstream and downstream flumes were tested for normality. Since data were not normally distributed comparisons were made with a nonparametric Wilcoxon test for paired data.
The proportion of quickflow was estimated by linear separation of storm hydrographs by drawing a line across from baseflow before the flood event to baseflow after the event; all flow above this line was defined as quickflow.
|
RESULTS
|
|---|
Flow Data
Flow within the catchment during 2002 began in mid-April at both flumes and ceased at the start of January at the upstream flume and February at the downstream flume. In 2003, rainfall was about 25% less than in 2002 and as a result flow began in mid-May and ceased at the upstream site at the end of December and at the downstream site in early January, although a small event was noted in late February. The decreased rainfall also translated to less flow in 2003 compared to 2002 (Table 1).
View this table:
[in this window]
[in a new window]
|
Table 1. Total flow and loads per hectare of P fractions, E. coli, and suspended sediment (SS) in baseflow, quickflow, and total flow during the 2002 and 2003 hydrological years (March to February) upstream and downstream of the cropped paddock.
|
|
Typically flow was characteristic of upland headwaters where flood events occurred rapidly in response to rainfall (Newson, 1994). The deep layer of stream sediments acted like a sponge adsorbing water in late autumn and early winter until saturated, at which point flow was in proportion to the volume of rain that fell. During late summer and autumn, drier conditions coupled with lesser rainfall and macrophyte growth gradually desiccated sediment to a depth of about 30 cm.
During storm events, quickflow and baseflow were determined via linear separation of hydrographs. Quickflow was flow in excess of baseflow and defined here as sources such as overland flow and rapid through flow via macropores. In the 2002 hydrologic year (March–February), quickflow made up 36% of total flow, whereas the decreased rainfall in 2003 resulted in only 26% of total flow (Table 1).
Additional analysis of storm events likely to produce overland flow (peak flow > 2 L s–1) was made by assuming variable source area hydrology (Ward, 1984): by dividing quickflow (m3) during the event by the length of stream (m), the minimum area (m2) around the stream channel contributing to saturation-excess overland flow is given (Table 2). Previous work in the catchment has indicated the widespread existence of a near impermeable B horizon (saturated hydraulic conductivity <1 mm h–1) leading to saturated conditions in winter and part of spring (Drewry and Paton, 2005). These authors also found that even in heavily treaded and grazed cropland, infiltration rates in the A horizon were high enough and rainfall intensities low enough that infiltration-excess overland flow was unlikely. Only nine events occurred during monitoring with sufficient peak flow and volume for overland flow to occur (Table 2). Of these events the maximum contributing distance from the stream, assuming equal distances either side of the stream channel, was only 19 m with the majority <3 m. Although this is the mean distance and certain contributing distances along the stream may be longer or shorter than the mean, data in Table 1 and 2 show that overall the contribution to stream flow from overland flow was small.
Phosphorus and Sediment Losses
Summary statistics for the concentrations of P fractions and SS are given in Fig. 3. Data for E. coli are also given but monitoring was stopped in September 2003 and hence data are incomplete. In New Zealand, guideline concentrations commonly used to assess risk of adverse effects to lowland surface water quality due to DRP and TP inputs are 0.010 mg DRP L–1 and 0.033 mg TP L–1 in stream flow (Australian and New Zealand Environment and Conservation Council, 2000). The guideline for SS in freshwater fisheries is 25 nephelometric turbidity units (about 13 mg L–1) (Davies-Colley et al., 1992). In both 2002 and 2003, the median concentration of DRP and SS at the upstream flume was at or just below the recommended guideline, while median TP concentrations were well in excess (Fig. 3). With the exception of DRP in 2002, all guidelines were exceeded in both years at the downstream site. The presence of E. coli in samples is used as an indicator of fecal material entering stream flow. Data for E. coli showed large variation (note log scale in Fig. 3), but the median concentration was less than the guideline for bathing water (127 cfu 100 mL–1, Ministry for Environment, 2003). Much of this variation could be due to the farmer putting 10 young stock near the upstream flume in the autumn of 2001–2002 and 20 cattle into the stream channel between the two flumes in December 2003.
View larger version (21K):
[in this window]
[in a new window]
|
Fig. 3. Box (median and 25 and 75% quartile) and whisker (5 and 95% of range) plots of P fractions, suspended sediment, and E. coli concentrations at upstream and downstream flumes in 2002 and 2003.
|
|
The inclusion of stock to the stream channel was also evident in the pattern of P loss during the trial. A comparison of cumulative DRP and TP load against cumulative flow indicates, by the slope of the lines, times when P supply is either generous (steep) or nearly exhausted (flat) (Fig. 4). Two periods, outside the grazing of the cropped paddock, indicate steep increases in DRP or TP loss to stream flow, which corresponded to periods when stock were, or had been recently, introduced into the stream channel (marked with arrows in Fig. 4).
View larger version (15K):
[in this window]
[in a new window]
|
Fig. 4. Cumulative loads of dissolved reactive phosphorus (DRP), total phosphorus (TP), and suspended sediment (SS) per hectare versus cumulative flow per hectare. Gray bands indicate when the cropped paddock was grazed, and arrows when stock were let into the stream channel.
|
|
Additional evidence for stock being responsible for the rapid increase in DRP loss at the upstream site is given in Fig. 5. This plots E. coli concentration at the two sites along with flow for two events: one a couple of weeks after stock had been removed from the stream channel, and one a month later when grazing of the cropped paddock occurred. During the first event, E. coli concentration at the upstream site was much greater at the upstream site compared to the downstream site presumably due to the presence of stock in the stream channel upstream of the cropped paddock. However, in the second event the situation was reversed presumably by the flushing of E. coli during the first event and the contribution of E. coli from overland flow from the cropped paddock. During cool conditions (June–July) in winter grazing only occurred on the cropped paddock. Thus, as a result of decreased biological activity, significant P inputs from surrounding pastures were unlikely. Increases in DRP loss in stream flow due to fertilizer addition were also unlikely since superphosphate was applied in summer when no or low flow was occurring and P would have had sufficient time to sorb to soil: generally P losses in overland flow from fertilizer are only greater than unfertilized soils for about 30 d after application (McDowell et al., 2003c).
View larger version (14K):
[in this window]
[in a new window]
|
Fig. 5. Flow and concentration during two events in early winter 2002. The first on the fourth of May trailed a grazing event in the stream by 4 wk, after which cattle were not allowed to graze the stream bed until the following autumn.
|
|
After grazing (gray bands in Fig. 4), slight increases were noted in DRP, TP, and SS load. However, these changes were small in comparison to stock grazing the stream channel. Irrespective of grazing, SS loss (kg ha–1) at the downstream site was elevated due to the presence of a winter forage crop, but TP losses (kg ha–1) were similar.
|
DISCUSSION
|
|---|
Flow Regime
In general, with increasing catchment scale the proportion of flow that can be identified as quickflow decreases. This suggests that in headwaters the proportion should be the greatest, provided sufficient rainfall falls to saturate soils (saturation-excess overland flow) or falls at an intensity greater than the soil's infiltration rate (infiltration-excess overland flow) (Nash et al., 2002). In this catchment, only 36% in 2002 and 26% in 2003 of total flow was noted as quickflow, despite the general consensus that pugging soil with cattle traffic decreases the soil infiltration rate (Drewry and Paton, 2005). However, infiltration-excess overland flow is not the main mechanism of overland flow in this catchment as rainfall intensity is generally low: a return period of a 20-min event with 15 mm h–1 is once a year. Furthermore, cultivated and pasture saturated hydraulic conductivities within the catchment have been measured at various times of the year ranging from about 100 mm h–1 in late winter to >1000 mm h–1 at other times of the year for cropped soil, and over 200 mm h–1 for pasture soils year-round (Drewry and Paton, 2005; McDowell et al., 2005b). Consequently, we can hypothesize that overland flow occurs as saturation-excess and thus may apply to the variable source area concept of hydrology (Ward, 1984).
Phosphorus and Sediment Loads and Sources
A survey of Olsen P concentration in the top 7.5 cm of soil was conducted on the experimental site in November 2001 on a 20-m square grid. The data in Fig. 1 show that the mean concentration up to 20 m away from the stream channel was about 20 to 30 mg kg–1, with small areas near the head water and grass strip up to 70 mg kg–1. Data from McDowell et al. (2003a) using the same site showed that 2-m2 plots near the stream channel (downslope position) exhibited an Olsen P concentration of 20 mg kg–1. By scaling loads estimated by McDowell et al. (2003a) to a per hectare basis for 2002 and 2003, TP losses from overland flow were under 0.1 kg ha–1, well below the minimum load lost at either flume during 2002 or 2003 (Table 1). Similarly, SS losses from plots scaled-up to a hectare were small (<35 kg ha–1) in comparison to that lost in stream flow (873–1499 kg ha–1).
Stream flow loads at both flumes were not out of range for agricultural catchments in New Zealand. Wilcock (1986) and Vant (2001) established the range at between 0.04 and 0.3 kg ha–1 for DRP, between 0.3 and 1.7 kg ha–1 for TP, and between 600 and 2000 kg ha–1 for SS. Considering the small load of P and SS lost from the small bounded plots (2 m2) measured by McDowell et al. (2003a), this infers that overland flow in this catchment was not the major mechanism of P or sediment loss to stream flow or that the short flow path of plots in McDowell et al. (2003a) yield less erosion and P losses than long flow paths on hillslopes. Factors such as scaling (i.e., length of flow path) or overland flow from areas near the stream channel with elevated soil Olsen P concentration may affect the load. However, work has shown that the quantity of P and sediment loss generally decreases during storm events with increasing catchment size (Le Bissonnais et al., 1998; Haygarth et al., 2005). Furthermore, soil in the small plots did not have an Olsen P concentration great enough to cause the loss of P measured at either flume. A curvilinear relationship has been established between Olsen P for a Pallic (Fragiochrept) soil, similar to that studied here, and DRP loss in overland flow generated by simulated rainfall (McDowell et al., 2003d). Assuming the same volume of flow, an Olsen P of >200 mg kg–1 would be required to yield 1 kg P ha–1.
In stream flow, loads during 2002 and 2003 of suspended sediment (SS) upstream of the cropped paddock were 880 and 873 kg ha–1 and 1499 and 969 kg ha–1 downstream. This inferred that SS was lost from the cropped paddock. For total P, estimates of loads were 0.98 and 0.24 kg ha–1 upstream and 0.99 and 1.89 kg ha–1 downstream. Comparing upstream with downstream sites, SS increased 70% in 2002 and 10% in 2003, whereas TP increased 2% in 2002 and 688% in 2003. The SS results are in-line with the decreased proportion of overland flow in 2003. However, TP data suggest that either the cropped paddock was not contributing much P or that alternative sources other than overland flow were responsible.
The Waitahuna soil within the catchment was poor in P-retaining Fe and Al oxides and thus would be more prone to P leaching than other soil types. However, it was unlikely that subsurface flow could account for the 0.8 and 1.7 kg P ha–1 not lost by overland flow. McDowell and Condron (2004) have developed an equation that relates DRP in overland or subsurface flow to the quotient of soil Olsen P and anion storage capacity. Using this equation and assuming the soil over the whole catchment had a mean Olsen P of 35 mg kg–1, a P retention of 35%, and 400 mm of drainage occurred (annual rainfall = about 1000 mm), then the cropped paddock would have contributed about 0.3 kg P ha–1 to stream flow via subsurface flow. Clearly there was another main source.
Investigation of stream bed sediments within the top 20 cm via sequential P fractionation during autumn when dry and in spring when inundated showed that the proportion of total P as bioavailable P increased via mineralization during desiccation and could be released as DRP when inundated in spring (McDowell and Stewart, 2005a). However, it was also noted that total P concentration increased by 10% when sediment was dry (largely in bioavailable forms) compared to inundated, but could not have received inputs from overland flow. Inputs via subsurface flow would also have been minimal during the summer–autumn dry period. Of greater likelihood was input via dung from the 10 young cattle allowed to graze the upper stream channel in autumn 2001–2002 and the 20 heifers allowed to graze the channel between the flumes in early summer 2003. Davies-Colley et al. (2004) showed that considerable quantities of contaminants were associated with cows crossing streams, and that cows can defecate up 50 times more in the stream bed than elsewhere. One cattle dung pat of about 2 kg can contain 2 g of total P (McDowell and Stewart, 2005b). Assuming a conservative mean deposition rate of 13 dung patches per day per cow (range 10–16; Williams and Haynes, 1990), and cows were allowed to graze the stream bed for 4 d then at least the remaining 0.5 kg P not accounted for by overland or subsurface flow may have been deposited directly onto the bed sediments by 10 cattle: logically more P is deposited by 20 cattle.
Following deposition, dung decomposes via physical breakdown by macrofauna (e.g., worms and dung beetles), rainfall, temperature, and the invasion of plant roots (Williams and Haynes, 1990). However, the rate of decomposition depends largely on climate: if dung is moist then it breaks down quickly but if it is dry then decomposition is slow (MacDiarmid and Watkin, 1972). For instance, Rowarth et al. (1985) showed that sheep dung decomposed on the soil surface completely within 28 d in winter but took 75 d in summer. During the first stream bed grazing (2001–2002), 10 cattle were let into the stream channel by the farmer 4 wk before flow at the upstream site, but not let into the rest of the bed. Although no flow was occurring and sediments were dry to 20 cm, rainfall and treading during grazing had enabled dung to decompose and become incorporated with stream bed sediment. While >80% of P in cattle dung is orthophosphate (McDowell and Stewart, 2005b), it is associated with colloids transportable in water, but no flow occurred, allowing orthophosphate to interact and sorb onto stream bed sediment. This caused the 10% increase in sediment total P and the large release of DRP when flow resumed in May (Fig. 4).
In fresh dung, P can be lost largely as PP if flow occurs soon after deposition. For example, Smith et al. (2001) found about half of P lost in overland flow from a field with cattle slurry applied in autumn was PP, while Smith and Monaghan (2003) found that DRP only accounted for 30% of P lost in overland flow, the majority was lost as PP from fresh dung. As mentioned above, the reason for this is probably that P in dung was associated with flocs or colloids that are mobile enough to be easily transported in flow but large enough not to go through a 0.45-µm filter and hence be defined as PP (McDowell and Sharpley, 2002). As dung ages the proportion of colloids available to flow deceases. During a trial examining the loss of P and SS from the cropped soil taken from the Dull Burn catchment, McDowell et al. (2005b) showed that of the SS lost in overland flow 1 d after a 1.5-kg dung pat had been applied, >90% were <20-µm particles, whereas after a couple of weeks this had decreased to 50% (Fig. 6). During 2003, a sharp increase in P loss was noted when the stream bed near the cropped paddock was grazed in December (Fig. 4). The majority (90%) of this was not DRP and most likely PP associated with dung deposition before flow had ceased at the downstream flume in January. Unfortunately, no E. coli data were available to confirm this. However, during this period about 1.4 kg P was lost, commensurate with inputs via the 20 cattle grazing at this time.
View larger version (12K):
[in this window]
[in a new window]
|
Fig. 6. Particle size distribution in overland flow from cropped soil 1 and 14 d after cattle dung had been applied (redrawn from McDowell et al., 2005b).
|
|
Grazing of stream beds is not a recommended practice in New Zealand. Although conspicuous slumping, pugging, or erosion is prohibited by law, current rules do not prohibit grazing by young stock (<2 yr old) (Otago Regional Council, 2005). The data presented here suggest that irrespective of age, stock access to waterways should not occur if surface water quality is to be maintained or improved.
|
CONCLUSIONS
|
|---|
During the 2002 and 2003 hydrologic years (March–February), loads were estimated from flumes upstream and downstream of a cropped paddock used for winter forage grazing. Loads of SS increased per hectare downstream compared to upstream, but TP only increased in 2003 and not 2002. Investigation of the sources of P loss in the catchment suggested that overland flow accounted for more P loss in 2002 than 2003 commensurate with rainfall and quickflow in the stream. However, the overall contribution via overland flow was <0.1 kg P ha–1: only 10% of the TP load in 2002 and 5% in 2003. The maximum contributing distance away from the stream channel was estimated to be 18 m with the majority at <3 m. Even by accounting for the effect of scale or differing Olsen P concentrations, the contribution to P loss from overland flow was deemed minimal. Similarly, the contribution of subsurface flow was estimated to be a maximum of 0.3 kg P ha–1. An indication of the main source of P loss was given by an increase in sediment total P concentration by 10% during autumn when no flow occurred. Against the experimental plan and regional authority advice, the farmer had allowed for grazing of the stream channel by 10 cattle upstream of the cropped paddock in 2001–2002 and 20 cattle between the two flumes in 2003. This grazing corresponded with a flush of P and E. coli lost at the beginning of 2002 at the upstream flume and a large amount of PP lost in 2003 at the downstream flume, and was attributed to the deposition of at least the remaining 0.5 kg P lost in 2002, and 1.0 kg P lost in 2003. Clearly, not allowing animals on to "clear-up" pasture in the stream channel is a priority if good surface water quality is to be achieved. Furthermore, from this study it would appear that compared to stock access the impact of winter grazing cropland on P losses was minimal.
|
REFERENCES
|
|---|
- Australian and New Zealand Environment and Conservation Council. 2000. Australian and New Zealand guidelines for fresh and marine water quality. ANZECC–Agric. and Resour. Manage. Council of Australia and New Zealand, Canberra, ACT.
- Carpenter, S.R., N.F. Caraco, D.L. Correll, R.W. Howarth, A.N. Sharpley, and V.H. Smith. 1998. Nonpoint pollution of surface waters with phosphorus and nitrogen. Ecol. Appl. 8:559–568.[CrossRef]
- Davies-Colley, R.J., C.W. Hickey, J.M. Quinn, and P.A. Ryan. 1992. Effects of clay discharges on streams: 1. Optical properties and epilithon. Hydrobiologia 248:215–234.
- Davies-Colley, R.J., J.W. Nagels, R.A. Smith, R.G. Young, and C.J. Phillips. 2004. Water quality impact of a dairy herd crossing a stream. N. Z. J. Mar. Freshwater Res. 38:569–576.
- Doran, J.W., J.S. Schepers, and N.P. Swanson. 1981. Chemical and bacteriological quality of pasture runoff. J. Soil Water Conserv. 36:166–171.
- Drewry, J.J., and R.J. Paton. 2005. Soil physical quality under cattle grazing of a winter-fed brassica crop. Aust. J. Soil Res. 43:525–531.[CrossRef]
- Eisenreich, S.J., R.T. Bannerman, and D.E. Armstrong. 1975. A simplified phosphorus analysis technique. Environ. Lett. 9:43–53.[Web of Science][Medline]
- Gillingham, A.G., and B.S. Thorrold. 2000. A review of New Zealand research measuring phosphorus in runoff from pasture. J. Environ. Qual. 29:88–96.[Abstract/Free Full Text]
- Haygarth, P.M., F.L. Wood, A.L. Heathwaite, and P.J. Butler. 2005. Phosphorus dynamics observed through increasing scales in a nested headwater-to-river channel study. Sci. Total Environ. 344:83–106.[Medline]
- Le Bissonnais, Y., H. Benkhadra, V. Chaplot, D. Fox, D. King, and J. Daroussin. 1998. Crusting, runoff and sheet erosion on silty loamy soils at various scales and upscaling from m2 to small catchment. Soil Tillage Res. 46:69–80.[CrossRef]
- MacDiarmid, B.N., and B.R. Watkin. 1972. The cattle dung patch. 3. Distribution and rate of decay of dung patches and their influence on grazing behaviour. J. Br. Grassland Soc. 27:48–54.
- McDowell, R.W., and L.M. Condron. 2004. Estimating phosphorus loss from New Zealand grassland soils. N. Z. J. Agric. Res. 47:137–145.
- McDowell, R.W., J.J. Drewry, R.W. Muirhead, and R.J. Paton. 2003a. Cattle treading and phosphorus and sediment loss in overland flow from grazed cropland. Aust. J. Soil Res. 41:1521–1532.[CrossRef]
- McDowell, R.W., J.J. Drewry, R.W. Muirhead, and R.J. Paton. 2005a. Restricting the grazing time of cattle to decrease phosphorus, sediment and E. coli losses in overland flow from cropland. Aust. J. Soil Res. 43:61–66.[CrossRef]
- McDowell, R.W., J.J. Drewry, R.J. Paton, P.L. Carey, R.M. Monaghan, and L.M. Condron. 2003b. Influence of soil treading on sediment and phosphorus losses in overland flow. Aust. J. Soil Res. 41:949–961.[CrossRef]
- McDowell, R.W., R.M. Monaghan, and P.L. Carey. 2003c. Phosphorus losses in overland flow from pastoral soils receiving long-term applications of either superphosphate or reactive phosphate rock. N. Z. J. Agric. Res. 46:329–337.
- McDowell, R.W., R.M. Monaghan, and J. Morton. 2003d. Soil phosphorus concentrations to minimise potential P loss to surface waters in Southland. N. Z. J. Agric. Res. 46:239–253.
- McDowell, R.W., R.W. Muirhead, and R.M. Monaghan. 2005b. Nutrient, sediment and bacterial losses in overland flow from pasture and cropping soils following cattle dung deposition. Commun. Soil Sci. Plant Anal. 36 (in press).
- McDowell, R.W., and A.N. Sharpley. 2002. Phosphorus transport in overland flow in response to position of manure application. J. Environ. Qual. 31:217–227.[Abstract/Free Full Text]
- McDowell, R.W., and I. Stewart. 2005a. Effect of wetting and drying on phosphorus forms in upland stream sediments. RMZ Mater. Geoenviron 52:279.
- McDowell, R.W., and I. Stewart. 2005b. Phosphorus in fresh and dry dung of grazing dairy cattle, deer, and sheep: Sequential fractionation and phosphorus-31 nuclear magnetic resonance analysis. J. Environ. Qual. 34:598–607.[Abstract/Free Full Text]
- Ministry for Environment. 2003. Microbiological water quality guidelines for marine and freshwater recreational areas. Available at http://www.mfe.govt.nz/publications/water/microbiological-quality-jun03/ (verified 14 Dec. 2005). Ministry for the Environ., Wellington, New Zealand.
- Nash, D., D. Halliwell, and J. Cox. 2002. Hydrological mobilization of pollutants at the field/slope scale. p. 225–242. In P.M. Haygarth and S.C. Jarvis (ed.) Agriculture, hydrology and water quality. CABI Publ., Wallingford, UK.
- Needleman, B.A., W.J. Gburek, G.W. Peterson, A.N. Sharpley, and P.J.A. Kleinman. 2004. Surface runoff along two agricultural hillslopes with contrasting soils. Soil Sci. Soc. Am. J. 68:914–923.[Abstract/Free Full Text]
- Newson, M. 1994. Hydrology and the river environment. 1st ed. Oxford Univ. Press, Oxford.
- Otago Regional Council. 2005. Environmental considerations for clean streams: A guide to managing waterways in Otago. Otago Regional Council, Dunedin, New Zealand.
- Parkyn, S., and R. Wilcock. 2004. Impacts of agricultural land use. p. 34.1–34.16. In J. Harding, P. Mosley, C. Pearson, and B. Sorrell (ed.) Freshwaters of New Zealand. New Zealand Hydrol. Soc., New Zealand Limnol. Soc., Caxton Press, Christchurch, New Zealand.
- Rowarth, J.S., A.G. Gillingham, R.W. Tillman, and J.K. Syers. 1985. Release of phosphorus from sheep faeces on grazed, hill country pastures. N. Z. J. Agric. Res. 28:497–504.
- Smith, C., and R.M. Monaghan. 2003. Nitrogen and phosphorus losses in overland flow from a cattle-grazed pasture in Southland. N. Z. J. Agric. Res. 46:225–237.
- Smith, K.A., D.R. Jackson, and P.J.A. Withers. 2001. Nutrient losses by surface un-off following the application of organic manures to arable land. 2. Phosphorus. Environ. Pollut. 112:53–60.
- Vant, W.N. 2001. New challenges for the management of plant nutrients and pathogens in the Waikato River, New Zealand. Water Sci. Technol. 43:137–144.[Web of Science][Medline]
- Wallbrink, P.J., C.E. Martin, and C.J. Wilson. 2003. Quantifying the contributions of sediment, sediment-P and fertiliser-P from forested, cultivated and pasture areas at the landuse and catchment scale using fallout radionuclides and geochemistry. Soil Tillage Res. 69:53–68.[CrossRef]
- Ward, R.C. 1984. On the response to precipitation of headwater streams in humid areas. J. Hydrol. 74:171–189.[CrossRef]
- Watanabe, F.S., and S.R. Olsen. 1965. Test of an ascorbic acid method for determining phosphorus in water and NaHCO3 extracts from soil. Soil Sci. Soc. Am. Proc. 29:677–678.
- Wilcock, R.J. 1986. Agricultural runoff: A source of water pollution in New Zealand? N. Z. Agric. Sci. 20:98–103.
- Williams, P.H., and R.J. Haynes. 1990. Influence of improved pastures and grazing animals on nutrient cycling within New Zealand soils. N. Z. J. Ecol. 14:49–57.
This article has been cited by other articles:
|
|
|
|
 
R. W. McDowell, Z. Dou, J. D. Toth, B. J. Cade-Menun, P. J. A. Kleinman, K. Soder, and L. Saporito
A Comparison of Phosphorus Speciation and Potential Bioavailability in Feed and Feces of Different Dairy Herds Using 31P Nuclear Magnetic Resonance Spectroscopy
J. Environ. Qual.,
May 1, 2008;
37(3):
741 - 752.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
R. W. McDowell and R. J. Wilcock
Sources of Sediment and Phosphorus in Stream Flow of a Highly Productive Dairy Farmed Catchment
J. Environ. Qual.,
March 1, 2007;
36(2):
540 - 548.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
W. J. Chardon, G. H. Aalderink, and C. van der Salm
Phosphorus Leaching from Cow Manure Patches on Soil Columns
J. Environ. Qual.,
January 9, 2007;
36(1):
17 - 22.
[Abstract]
[Full Text]
[PDF]
|
|
|
TOP
ABSTRACT
INTRODUCTION
