2. Nutrients and Sediments
Agricultural development has led to widespread increases in the levels of nitrogen (N) and phosphorus (P) in lowland watercourses and subsequent nuisance growths of algae and other aquatic plants. In New Zealand the introduction of nitrogen-fixing clovers, use of nitrogen fertilisers, including the practice of spreading animal wastes on pastures, and direct addition of stock urine and faeces in pastures have increased the amounts of nitrate leaching from pastoral catchments.
Suspended sediments affect stream habitat and water quality by reducing water clarity for sighted organisms, and reducing light penetration for plant growth. Aesthetic appeal is also reduced by high suspended sediments. Sediments that settle on the substrate fill interstitial spaces and affect the habitat available for invertebrates (Ryan 1991). In agricultural areas, hill slope (sheetwash) erosion, mass movement, stock damage to stream banks, and erosion of tracks and raceways are the key factors that introduce sediment to streams. Increases in sediment delivery to streams in production forestry catchments occur mainly after harvesting, and mass movements and erosion of roads and tracks can be major sources of sediment.
2.1 Modes of particulate and dissolved pollutant transport
The ability of buffer zones to attenuate pollutants will depend upon the mechanisms by which these pollutants reach surface waters. Three transport processes can occur:
-
direct pollution (e.g., stock access to streams, bank erosion);
-
surface runoff;
-
subsurface flow and drainage.
Surface runoff
Surface runoff can occur through several mechanisms. It may result when the surface soil becomes saturated (saturation excess) which is common where flow pathways converge as a result of topography. It may also occur when rainfall intensity exceeds the infiltration capacity of the soil (infiltration-excess) a process that is common in poorly drained clay-rich soils (Muscutt et al. 1993). Surface runoff can be a major transport mechanism for soluble pollutants, particularly when land beside a stream has been grazed, or fertiliser or livestock waste applied to the land during or prior to rain events. Surface runoff can also be a conduit for sediment and particulate pollutants. Sediment transport can occur through sheet erosion in spatially uniform flows over hillslopes, but is most likely to result from areas where flow is concentrated (Fig. 3), and from bare soils (e.g., stock tracks, slips, cultivated soils). Sediment in surface runoff can also carry particulate forms of phosphorus, and a high proportion of total P loss has been found to occur during periods of high flow (Culley & Bolton, 1983, Smith 1987).
Smith (1987) studied the runoff generated from hillslopes in Waikato, New Zealand after rain events between July 1983 and March 1985. This study reported large runoff volumes, and demonstrated that nutrients and sediment were carried downslope, often becoming concentrated into preferred flowpaths. Consequently, nutrients and sediments were not uniformly discharged along stream banks. Almost all of the total phosphorus (TP) and total nitrogen (TN) transport occurred in mid-late winter and early spring. TP and TN were predominantly in particulate forms, suggesting that riparian management that trapped sediment and particulate nutrients would be effective in improving stream water quality at that site.
Figure 3: Water flow concentrated into an ephemeral stream. Photo T. Wilding.
Subsurface flow and drainage
Subsurface flow is frequently the major pathway of N transport in catchment runoff and high concentrations commonly occur in artificial subsurface drains (Muscutt et al. 1993). Intensive agriculture is often accompanied by subsurface drainage especially in clay-based soils. These drains provide routes for the rapid transport of water and pollutants from the soil during high water table conditions and, in many cases, can bypass the riparian zone by directly discharging to the stream.
Subsurface flow paths are influenced by the surrounding topography and soil drainage characteristics. For instance, on land that is free draining, water and associated pollutants may bypass the riparian zone, whereas on poorly drained soils or where the water table is high, pollutants in subsurface water may be carried into the soils of the buffer zone. On occasion, subsurface flows may re-emerge, and discharge downslope as surface runoff.
2.2 Modes of Sediment and Nutrient Removal
Buffer zones can be effective at removing nutrient and sediment inputs to streams by restricting the direct use of land beside the stream and by processing water that has been transported into the riparian zone. The mechanisms of contaminant removal in buffer zones differ according to characteristics of the hydrology, soils, and vegetation as well as the mode of transport to streams.
Direct effects
The removal of stock from streams and riparian areas has obvious benefits for water quality. Belsky et al. (1999) reviewed published information on the effects of livestock grazing in the western US and found that livestock negatively affected water quality, stream channel morphology, hydrology, riparian zone soils, instream and streambank vegetation, and aquatic and riparian wildlife. Livestock contribute nutrients directly to streams and riparian areas in their dung and urine. Faecal material deposited in the riparian zone on soils that have been damaged by treading may be readily washed overland into the stream with little opportunity for filtration of contaminants by vegetation (often reduced, damaged or absent), or infiltration into compacted soil (Nguyen et al. 1998; Trimble and Mendel 1995).
Livestock damage to riparian vegetation and soils destabilises the banks and leads to mobilisation of fine sediment (Trimble 1994), that in turn causes sedimentation in the channel and reduced stream clarity (e.g., Waters 1995). In addition, more runoff of sediment occurs from soils disturbed and compacted by livestock trampling (Nguyen et al. 1998). The resulting increased sediment load is accompanied by particulate nutrients that may contribute to stream enrichment as well as eutrophication of lakes and estuaries downstream (e.g., Williamson et al. 1996). Culley & Bolton (1983) estimated that bank erosion contributed 32% of sediment discharge and 10% to the export of P from an agricultural catchment in Canada. Bank destabilisation can occur as a result of deforestation and conversion to agriculture, as well as from the direct effects of livestock. Planting trees and shrubs in the riparian buffer zone can stabilise stream banks, as long as the rooting depth of the plants are appropriate for the size of the bank.
Line et al. (2000) demonstrated that stock exclusion by fencing reduced TKN, TP and suspended sediment loads in a stream running through a dairy farm in North Carolina, USA by 78, 76, and 82%, respectively. Weekly loading rates of nutrients and sediments were monitored for 81 weeks prior to fencing and 137 weeks after fencing. The effects of installing alternate watering systems only were also evaluated, but these were not as effective as fencing stock from the streams. Nitrate loads were not significantly reduced in the time span of the study and the authors suggested that these were likely to decrease in the future when trees had established in the buffer and mechanisms of nutrient uptake and denitrification developed. This study clearly shows that rapid improvements in water quality can be seen after exclusion of stock (i.e., within 2.5 years), particularly reductions in particulate nutrients and sediment.
Surface pollutant transport
Buffer zones where stock have been excluded and long grass or natural vegetation has been allowed to develop or been planted can reduce diffuse pollutant transport from agricultural land by:
- infiltration within the buffer zone which reduces surface runoff;
- reduction of surface flow velocities from increased hydraulic roughness of the vegetation in the buffer;
- physical filtering effect of dense vegetation.
Much of the research into the effectiveness of buffer zones for removing contaminants from surface runoff has focussed on vegetated filter strips (VFS), usually consisting of rank paddock grasses. Researchers including Young et al. (1980), Dillaha et al. (1989), Magette et al. (1989), and Daniels and Gilliam (1996) have studied the effectiveness of grass filter strips in trapping sediment and nutrient through laboratory or field experiments (see also Table 1). They reported trapping efficiencies exceeding 50% for sediment and nutrients adsorbed to sediments (such as phosphorus), while trapping of dissolved nutrients was less efficient.
The main function of vegetated filter strips for sediment removal is to provide flow resistance (through enhanced hydraulic roughness) that reduces the flow velocity and sediment transport capacity of surface runoff. This leads to an enhanced deposition of particulates (Neibling and Alberts 1979, Gharabaghi et al. 2002). Ponding can also occur at the upslope edge of the buffer zone causing sediment accumulation (Muscutt et al. 1993). Some removal of soluble pollutants also occurs, but infiltration, not deposition is the primary mechanism for removal of soluble pollutants from overland flow (Gharabaghi et al. 2002).
Increased infiltration may occur in buffer zones as a result of the change in soil structure associated with the change in vegetation type (Muscutt et al. 1993). Vegetation can provide root channels for improving infiltration of water into the soil (Collier et al. 1995). Removal of stock grazing increases the infiltration capacity of the soil, because trampling can compact soils.
In an experimental study of grass filter strip (perennial rye grass) efficiencies, Gharabaghi et al. (2002) found that the first 5 m of the filter strip were critical for sediment removal. Almost all of the easily removable particles (larger than 40 microns in diameter) were captured with the first few meters of the filter strip. However, the remaining small size particles were very difficult to remove by filtering as they stayed in suspension. The only mechanism that helped in removal of small size sediments was infiltration. During experimental runs with low to moderate flow rates on longer plot lengths (20 m wide filter strips), 90% removal efficiencies of sediment could be achieved because fine sediments were able to infiltrate into the soil. Gharabaghi et al. also concluded that sediment removal efficiency did not increase much beyond 10 m filter strip widths, although the potential for the buffer to become clogged with fine sediment over time should be considered when establishing optimum buffer widths.
Nutrients that are sediment-bound can also be effectively removed in VFS. Dillaha et al. (1989) applied manure and fertiliser to bare fallow plots with different filter strip widths of 0, 4.6 and 9.1 m. The 9.1 and 4.6 m filter strips with shallow uniform flow (11% and 16% slopes) removed an average of 84 and 74% of the incoming solids, 79 and 61 % of the incoming P, and 73 and 54% of N. The removal of P and N from the runoff was nearly as effective as sediment removal because a large proportion of the nutrients were sediment-bound.
In a New Zealand study, Smith (1989) found that retired pasture buffers of 10-13 m were capable of reducing suspended sediment and particulate nutrients in channelised surface runoff by over 80%. Dissolved N and P removal was less (67, 55%; Table 2).
The proportion of surface runoff that infiltrates into the buffer soil is likely to reduce the load of soluble pollutants that are transported through the buffer, at least in the short term (Muscutt et al. 1993). In other words, improving the infiltration capacity of the buffer will also improve the efficiency of buffers for soluble nutrient removal.
Subsurface pollutant transport
Reduction in soluble nutrients, largely N, from buffer zones and headwater and riparian wetlands has been demonstrated in a number of studies and the two main mechanisms for nutrient removal are:
- uptake by vegetation;
- denitrification.
The relative importance of these two processes may differ between buffer zone types, however, most researchers agree that riparian zones can be highly effective for soluble nitrate removal. Many studies have shown >90% reductions in nitrate concentrations in subsurface flows as water passes through riparian areas or wetlands (Gilliam 1994, Fennesey & Cronk 1997, Table 3.). Buffers are consistently reported to reduce nitrate to below 2 mg/L, often throughout the year and even when nitrate inputs are extremely high (Muscutt et al. 1993). Biological denitrification is the most desirable means of nitrate attenuation as the microbial conversion of NO3 removes nitrate from the system in the form of N gases (Martin et al. 1999). Plant uptake can eventually return the nitrogen to the system as the plants die and decompose. Denitrification is a microbially mediated process whereby bacteria convert nitrate to N2 gas when there is a plentiful carbon source, such as wetland or riparian soils rich in organic matter, and when conditions are anoxic or low in oxygen.
Wetlands
Wetland areas and seeps that intercept drainage before the flow enters streams have been clearly identified as areas causing loss of NO3, with denitrification being the most important mechanism. Cooper (1990) found that the majority of nitrate loss occurred in riparian organic soils, despite these soils occupying only 12% of the border of a small pasture stream in New Zealand. He attributed this to characteristics of the catchment hydrology, as a disproportionately high percentage of groundwater flowed through these small wetlands in the base of hollows, and also to their high capacity for denitrification (anoxic, high in denitrifying enzymes and available carbon). Stream channel nitrate removal was largely through plant uptake (watercress) and was much more variable. Cooper (1990) also identified that the capacity for denitrification in these soils was under-utilised.
Nguyen et al. (1999a) found 27% removal of phosphorus and 54% removal of nitrogen over a 6-month period in a wetland at the head of a small stream at Whatawhata, Waikato. Measurements and modelling in two contrasting wetlands showed the importance of hydrology and contact time in determining the effectiveness of riparian wetlands in removal of nitrate (Nguyen et al. 1999b); the longer the contact time, the greater the removal.
Wetlands can also be effective at phosphorus removal depending on the physical-chemical-microbiological processes that affect P uptake (McDowell et al. 2004). These methods include sorption-precipitation of dissolved phosphorus by wetland substrate, sedimentation-deposition of particulate P, and P assimilation by microbial and plant biomass. Plant P, unless it is removed by harvesting, can be released back to wetlands via decomposition of plant litter. Thus, the most important processes are sorption and sedimentation (Cooke et al. 1992, Nguyen 2000). However, P removal by wetlands generally declines after a period of years or decades depending on loading rates, hydraulic retention time, wastewater characteristics, wetland substratum, and accumulation of organic solids (McDowell et al. 2004).
Gilliam (1994) called for an effort to protect ephemeral and intermittent stream channels as well as wetlands, as these are areas that initially receive surface runoff and where shallow groundwater seeps into surface water, and thus may be some of the most important areas for preserving water quality.
2.3 Buffer Strip Design and Efficiency for Sediment and Nutrient Removal
Much of the variability in studies of nutrient or sediment removal efficiencies can be explained by site specific differences in characteristics of the buffer zone or in characteristics of the surrounding land. Some of these factors are outlined below.
Width
Data from studies comparing multiple width buffers in the same location (Young et al. 1980, Dillaha et al. 1988, Dillaha et al. 1989, Magette et al. 1989, Peterjohn and Correll 1984, Vought et al. 1994) have shown that sediment and total phosphorus removal rates (between 53 and 98%) increase with increasing buffer width (4.6 m to 27m). Where a grass buffer strip has been designed sensibly to treat sheet rather than channelised flow, many researchers report substantial sediment removal within a few metres of the upslope boundary (Barling & Moore 1994, Fennessy & Cronk 1997). Grass filter strips in particular have been shown to be very effective at trapping sediment particles. Neibling & Alberts (1979) found that 91% of incoming sediment to a grass filter strip was deposited in the first 0.6 m. Much of the larger particles of sediment may be removed in 5 m of grass buffer, but finer particles may require 10 m (Gharabaghi et al. (2002)
The width required to optimise nutrient removal has been debated with little systematic study of the issue. Fennessy and Cronk (1997) reviewed studies of RBZ effectiveness for the removal of contaminants, particularly soluble nitrate. Almost 100% of nitrate can be removed by buffers 20-30 m wide, while examples of forested buffers of 10 m achieved over 70% retention of N. Table 3 lists a range of buffer widths that have been assessed for nitrate removal (from Fennessy & Cronk 1997). Many of the buffers were forested, and N uptake by plants and denitrification were believed to have been an important factor in removing soluble N. However, one problem in assessing minimum widths is that many studies have had to use existing buffer widths, rather than deriving it experimentally.
Because of the different modes of particulate and dissolved contaminant transport, multi-tier or combination buffers are often advocated. A narrow combination buffer consisting of 5 m of grass filter strip and a 1m wide row of deciduous trees significantly reduced nitrate in subsurface flows beneath cropland in Italy (Borin & Bigon 2002). A substantial reduction in nitrate (average 81%) was observed at the field/grass buffer boundary and the authors concluded that the roots of the trees were extending beyond the combined 6 m buffer so that the zone of influence was larger than the land that was retired from use. Further reductions in nitrate were measured through the buffer and discharge to the stream had concentrations that were less than 2 ppm.
The width required for nutrient and sediment removal can be quite variable and the Auckland Regional Council suggested another method of determining optimal buffer width, which was based on the width needed to develop a self-sustaining buffer of native vegetation. Parkyn et al. (2000) recommended a buffer width of 10-20 m as the minimum necessary for the development of sustainable indigenous vegetation with minimal weed control, and to achieve many aquatic functions.
Vegetation
The consensus in the literature is that grass buffer strips are effective at filtering sediment and sediment-associated pollutants (particulate P and N) from surface runoff. However they are less effective in removing soluble nutrients such as nitrate, ammonia, and dissolved P. Nitrate removal from subsurface flows is considered to be greater in forested buffers, partly through uptake by plants (Fennessy & Cronk 1997, Martin et al. 1999). However, the main mechanism by which nitrate is removed from groundwater is thought to be biological denitrification. Wetlands and soils in riparian zones have been shown to have a high capacity for denitrification compared to terrestrial and aquatic soils (Cooper 1990). Vegetation in the riparian zone can contribute to denitrification through root exudates and plant decomposition in some ecosystems, and organic matter status of the soil may have a major effect on N removal efficiency (Cooper 1990, Muscutt et al. 1993)
Riparian carbon inputs to streams (i.e., leaf litter and wood) can also increase the potential for stream bed denitrification. This may be particularly important for systems where groundwater inflows bypass the riparian zone or where there are tile drains. In a study comparing buffer effectiveness in well-drained and poorly drained settings in North Carolina, USA, Spruill (2004) showed that buffers were effective at reducing nitrate in the groundwater of both. Most of the nitrate removal occurred through denitrification in the buffer soils and streambeds. Thus, even though nitrate in ground water passed beneath the buffer, the relatively high organic carbon in the discharge zone of those sites (derived largely from riparian vegetation) provided an environment conducive to denitrification (Spruill 2004).
The type of vegetation planted in buffer zones can also influence nutrient removal. For instance, James et al. (1990) (cited in Gilliam 1994) noted the failure of a riparian buffer to reduce NO3 in Maryland, USA, was due to leguminous trees that actually increased the NO3 in groundwater.
Nutrient removal efficiencies in buffers may also be affected by the age of the vegetation. Mander et al. (1997) studied N and P budgets in four riparian forests of varying age in Estonia and USA. While the buffers were able to remove both nitrogen and phosphorus, even when the input concentrations were very high, young forest stands, bushes, and wet grasslands showed the most intensive nutrient removal. This was due to intensive nutrient uptake by plants as they were in an active growth phase, and high microbiological activity and adsorption capacity of the soils.
Phosphorus accumulates in riparian soils and can be taken up by plants but there is no process similar to denitrification that removes P to the atmosphere. Therefore, buffer zones could potentially become saturated and their ability to trap P may decline with age unless sediments or organic matter are removed from the buffer zone (Barling & Moore 1994).
Harvesting production trees or plants, or fruit and nuts from trees in riparian zones can provide a mechanism where P can be removed from the riparian zone. Examples of this include indigenous systems of tropical agroforestry where non-timber products (fruits, nuts and ornamentals) can be harvested (Robles-Diaz-de-León & Kangas 1999). There may be scope in New Zealand to use riparian buffers as zones for flax harvesting, medicinal plant growth, manuka honey, etc.
Combination buffer systems in the USA often consist of an upslope grass buffer, a managed forest zone and an undisturbed forest zone next to the stream. Because no studies had assessed the impact of forest harvesting and management on these riparian systems, Hubbard & Lowrance (1997) studied the nitrate removal from shallow groundwater where the forest zone was either mature forest, clear cut, or selectively thinned. All three forest management treatments were effective in assimilating nitrate and there were no differences between treatments. Concentrations of nitrate in shallow groundwater were highest at the field – grass buffer interface and dropped most dramatically (by factor of 10) within the managed forest zone.
Slope
Slope angle is a key factor in determining sediment entrapment within RBZ (Young et al. 1980, Peterjohn and Correll 1984, Dillaha et al. 1989, Magette et al. 1989, Phillips 1989). Dillaha et al. (1988) compared sediment removal under differing slopes with all other factors constant, deriving an inverse relation between slope angle (6º-9º) and sediment entrapment (50-90 %). In general, many review articles of buffer zone studies conclude that buffers need to be wider when the slope is steep, generally to give more time for the velocity of surface runoff to decrease (Barling & Moore 1994, Collier et al. 1995).
Soils and drainage
Soil drainage properties influence RBZ performance. Free draining soils minimise the generation of surface runoff, both on the hillside and within a buffer. Better paddock management may increase buffer effectiveness. For example, in a grazing system, a reduction in stocking rates and longer times between paddock rotations may be sufficient to alleviate surface compaction problems and enhance infiltration, while establishment of a good groundcover will slow incoming water to the buffer (Herron & Hairsine 1998).
Lowrance et al. (1997) use existing information and their best professional judgement to provide expected levels of pollutant control by riparian forest buffer systems (RFBS) in 9 different physiographic provinces of Chesapeake Bay, USA. They stress the importance of the hydrologic connection between pollution sources, the buffer zone and the stream, stating that water quality improvements will be most likely in areas where most of the excess precipitation moves across, in, or near the root zone of riparian forest buffers. Several studies in this region had shown that high rates of nitrogen removal occurred in areas with high water table conditions and shallow groundwater movement near the root zone. In regions with deeper soils (i.e., aquiclude or bedrock 10-30 m below surface) or where water drained into aquifers or large rivers, the removal potential of RFBS were expected to be much less.
Topography
The effectiveness of grass buffer strips as filters for nutrients and sediment is less in steep hilly terrain than rolling land, as overland flow is concentrated in channelised natural drainage-ways giving rise to high flow velocities. As a result buffer effectiveness is minimal, or at best, patchy along the stream length.
Dosskey et al. (2002) studied four farms in Nebraska, USA, to develop a method for assessing the extent of concentrated flow in riparian buffers and for evaluating the impact that this has on sediment trapping efficiency. Riparian buffers averaged 9-35 m wide and 1.5-7.2 ha in area, but the effective buffer area that actually contacted runoff water was only 0.2-1.3 ha due to the patterns of topography preventing uniform distribution of runoff. Using mathematical relationships, it was estimated that between the four farms, buffers could theoretically remove 41-99% of sediment, but because of non-uniform distribution it was estimated that only 15-43% would actually be removed.
Grass buffers may need to extend further inland following a drainage way, resulting in a non-uniform buffer width along the length of the stream.
Longevity
There are indications that buffer zones may have a limited life span where they are effective. For example they may become saturated with P, pore spaces in soils may clog with sediments, or dissolved nutrient uptake by plants may be greatest during early growth phases and decline as vegetation matures. Some evidence of P saturation of a riparian zone was shown by Cooper et al. (1995) who studied riparian soils in native scrub (manuka), grazed pasture, and 12yr old retired pasture (tussock dominated) near Taupo, New Zealand. Retired pasture soils had extremely high hydraulic conductivity indicating that surface runoff water transported into the zone would infiltrate, fill soil pores and emerge as subsurface flow at the stream edge. The runoff that emerged from the buffer was depleted in sediment-bound nutrients and dissolved N, but enriched in dissolved P.
To optimise the long term value of riparian zones as nutrient filters, a number of strategies could be employed: (1) riparian retirement needs to be accompanied by improved land use practices over the broader landscape so that nutrient influx to the riparian zone is reduced. (2) Periodic harvesting of plant material could ensure plant uptake remains a continued net nutrient removal mechanism (3) Buffer widths should be established on the basis that a sustainable nutrient removal can be achieved with regard to the nutrient influx it would receive.
Methods to remove P could include selective harvesting for wood or fruits as mentioned earlier, or in the case of grass buffers, light grazing with sheep for a short time during summer may be acceptable providing that temporary fences are used immediately beside the stream to keep stock out. Alternatively, the strip could be mown for haymaking.
Predictive tools for designing RBZ
Phillips (1989) employed mathematical models to estimate the relative importance of soil hydrologic properties, topography, and surface roughness in determining the effectiveness of water quality buffers in North Carolina, USA. He found that slope gradient was the most critical factor for effective removal of sediment or particulate pollutants transported in surface runoff. However, buffer width was by far the most important factor for effective removal of dissolved pollutants in surface or subsurface flow.
The effectiveness of buffers can be greatly affected by its design and site-specific factors such as slope, clay content of the soil, drainage patterns, etc. In 1995, DoC and NIWA published a set of guidelines (Collier et al. 1995) that provided practical measures to improve the design of RBZ to manage bank stability, light climate, water temperature, carbon supply, habitat diversity, flood flows, and contaminants. For contaminants, the guidelines can be used to calculate the optimal filter strip width for attenuating overland flow. These calculations were based on the modified CREAMS model (Chemical, Runoff, and Erosion from Agricultural Management Systems), and they require information on topography, slope, soil types for drainage and clay content categories, and hillslope length. Generally, buffer widths will need to widen as the slope length, angle and clay content of the adjacent land increase and as soil drainage decreases. For nitrate removal in subsurface flow, the guidelines recommend protection of existing riparian wetlands, based on their proven effectiveness for nitrate removal.
Herron & Hairsine (1998) used time independent equations to assess the effectiveness of riparian buffer zones in reducing overland flow to streams. Buffer widths, expressed as a proportion of total hillslope length, were calculated based on a variety of Australian rainfall environments and varying topographic convergence. From these scenario results, the authors proposed a riparian width not exceeding 20% of the hillslope length as a practical management option, although larger buffer widths may be required where riparian areas or slopes are degraded.
High rates of denitrification are known to occur in water-logged soils, anoxic conditions, and with a source of organic carbon. Recent research in the USA has used soil map data to identify area with wet soils as a planning tool for riparian management to enhance denitrification (Gold et al. 2001).
Table 1: Contaminant removal efficiencies from references within Castelle et al. (1994) review of U.S. vegetated buffers. VFS = vegetated filter strip.
|
Contaminant |
Buffer width |
Removal (%) |
Slope (%) |
Farm type |
Buffer type |
Reference |
|
Sediment |
30.5 |
90 |
2 |
|
|
Wong & McCuen (1982) |
|
Sediment |
61 |
95 |
2 |
|
|
Wong & McCuen (1982) |
|
Sediment |
24.4 |
92 |
|
|
Veg. |
Young et al. (1980) |
|
Sediment |
22.9 |
33 |
|
dairy |
Filter strip |
Schellinger & Clausen (1992) |
|
Sediment |
61 |
80 |
|
|
Grassy swale |
Horner & Mar (1982) |
|
Sediment |
30 |
75-80 |
|
Logging activity |
|
Lynch et al. (1985) |
|
Sediment |
9.1 |
85 |
7 and 12 |
|
Grass VFS |
Ghaffarzadeh et al. (1992) |
|
NO3-N, NH4-N, PO4-P |
4.6 |
90% |
|
|
Grass VFS |
Madison et al. (1992) |
|
NO3-N, NH4-N, PO4-P |
9.1 |
96-99.9 |
|
|
Grass VFS |
Madison et al. (1992) |
|
Sediment, N, P |
9.1 |
84, 79, 73 |
11-16 |
|
Grass VFS |
Dillaha et al. (1989) |
|
Sediment, N, P |
4.6 |
70, 61, 54 |
11-16 |
|
Grass VFS |
Dillaha et al. (1989) |
|
NO3-N |
10 |
99.9% |
|
|
forested |
Xu et al. (1992) |
|
N, P |
19 |
89, 80 |
|
|
forested |
Shisler et al. (1987) |
Table 2: Some New Zealand studies of efficiency.
| Contaminant | Buffer width | Removal (%) | Farm type | Buffer type | Reference |
| Nitrate | c. 3-4m | 88-97 | pasture | Riparian organic soils - wetland | Cooper 1990 |
| Nitrate | c. 3-4m | 0-62 | pasture | Riparian mineral soils - wetland | Cooper 1990 |
| Nitrate | –140-91 | pasture | streambed | Cooper 1990 | |
| Nitrate | 32-100% | Waste water treatment | wetland | Cooper 1994 | |
| Nitrate | 10-13m | 67 | pasture | Retired pasture | Smith 1989 |
| Dissolved P | 10-13m | 55 | pasture | Retired pasture | Smith 1989 |
| Particulate P, N | 10-13m | 80, 85 | pasture | Retired pasture | Smith 1989 |
| Total Suspended solids | 10-13m | 87 | pasture | Retired pasture | Smith 1989 |
Table 3: Experimental studies of buffer widths required for sub-surface and surface nitrate removal (from Fennessy & Cronk 1997).
| Flow type | Buffer type |
Buffer width (m) |
N retention (%) |
N inflow (mg/l) |
| Subsurface | Forest | 9 | 61-97 | 180 |
| Forest | 10 | 70-90 | 13.5 | |
| Forest | 10 | Up to 77 | 0.6-2.5 | |
| Forest | 19 | 93 | 7.4 | |
| Forest | >20 | 90 | 7.4 | |
| Forest | >20 | 99 | 6.8 | |
| Forest | 20 | Up to 87 | 0.6-2.5 | |
| Forest | 25 | 68 | ~2-6 | |
| Forest | 26 | ~100 | 2-9 | |
| Forest | 30 | ~100 | 5 | |
| Forest | >10-50 | 94 | 1.8 | |
| Forest | 50 | 95 | 8 | |
| Forest | 60 | ~100 | 10 | |
| Herbaceous | 22 | 84 | 2-12 | |
| Sub- and surface | Forest | 16 | 90 | 10 |
| Surface | Forest | 19 | 60 | 4.5 |
| Forest | >20 | 79 | 4.5 | |
| Herbaceous | 5 | 54 | - | |
| Herbaceous | 8 | 20 | 20 | |
| Herbaceous | 9 | 73 | - | |
| Herbaceous | 16 | 50 | 20 | |
| Herbaceous | 27 | 84 | - | |
| Herbaceous | 30 | 11 | 20 |
Contact for Enquiries
MAF Information Services
Pastoral House
25 The Terrace
PO Box 2526
Wellington, NEW ZEALAND
Fax: +64 4 894 0721
Contact this person
