Figure 1 serves as a conceptual ‘model’ to anchor the following discussion of transmission pathways (Section 3) and mitigation measures (Section 4) for faecal microbes of concern. We distinguish ‘direct’ pathways where faecal matter is deposited directly into waterways or so close that the potential for wash-in is very high, and ‘indirect’ pathways, via surface and subsurface flows, from faeces deposited on land. This distinction is important because direct deposition is very fresh with no opportunity for die-off of microbes in the faecal matter before they reach water. Furthermore, unlike most indirect pathways, direct deposition is obviously not dependent upon rainfall to drive the transfer process. Indirect pathways move faecal matter that is ‘aged’ to a greater or lesser extent, with consequent immobilisation e.g., by ‘skin’ formation or desiccation of the surface of pats, and die-off of some microbes. Furthermore, there is also considerable opportunity during surface and subsurface flows for attenuation of faecal microbes.
Figure 1: Key pathogen transmission pathways and associated mitigation measures. The effectiveness of some of the (potential) mitigation measures shown has yet to be fully evaluated.
Faecal contamination of freshwaters can arise through the deposition of faeces by grazing animals directly into waterways. This ‘direct’ deposition can occur when dairy cows cross a stream on the way to or from the milking shed and, through sporadic incursions by grazing cattle into waterways. The ‘waterway’ in this context could include the channel and riparian zone, not just the water. This is because deposition ‘near-water’ may, at least on occasions, be nearly as damaging as direct deposition into water because of the high potential for wash-in by surface runoff and entrainment by rising flows. It should be noted that, in addition to cattle, deer are also attracted to waterways.
A study of the water quality impacts of a dairy herd crossing the Sherry River in New Zealand (Davies-Colley et al. 2004) showed very high levels of faecal contamination, with concentrations of the faecal bacterial indicator E. coli temporarily elevated to more than 100X background levels and more than 100X the guidelines for contact recreation. During one crossing (of four each day) the herd was estimated to have deposited about 1011 E. coli directly into the river, associated with 25 individual defecation events.
Studies of cattle behaviour in New Zealand have been conducted on both hill-country (dry-stock) and dairy farms to quantify direct deposition associated with sporadic incursions. The hill-country study (Bagshaw 2002) suggested that beef cattle defecated in the stream or riparian zone (defined as within 2 m of the stream bank) at an average rate of 0.2 faeces (i.e., 0.2 of a pat) per cow per day. This equated to about 4% of the total number of defecations per day. Half this amount was deposited directly into the stream channel and the other half upon the riparian zone.
Dairy cattle, studied over two summers and one spring, spent on average 99.1% of their time in the paddock, 0.7% on the bank, and 0.1% in the stream. On average, 98.8%, 0.7%, and 0.5% of defecations were in the paddock, on the bank, and in the stream, respectively. However, considerable variation was found, potentially attributable to a range of factors e.g., stream size, ease of access, and the characteristics of the stream bed. The presence of dairy cows within a stream was also shown to increase streamwater E. coli concentrations, often by an order of magnitude or more, compared to background levels. This increase is attributed to a combination of both the direct deposition of faecal material and the stirring up, by the cattle, of bacteria previously settled on the stream bed.
A rigorous quantification of the importance of sporadic incursions relative to indirect pathways is not possible from information provided by the reported studies. However, a crude analysis indicates that sporadic incursions appear to be of broadly comparable importance as some indirect pathways (section 3.2). For example, the studies of dairy cattle behaviour found, on average, that 0.5% of defecations were deposited directly to waterways. Assuming a herd size of 175 animals, each excreting 12 pats per day, derives an approximate herd input (on average) of 10 pats per day, directly to waterways. Assuming (conservatively) that at least 109 E. coli are found within a fresh pat, indicates that about 1010 E. coli are deposited directly, per day, by the herd. This figure lies, therefore, within the upper range of storm period E. coli yields (flushed via wetlands) per hectare of grazed pasture (Section 3.2). In addition, it appears to be comparable to the larger yields reported for artificial drains (Section 3.2).
In addition to cattle, deer are also attracted to water. A study conducted by Environment Southland found that concentrations of suspended sediment, ammoniacal nitrogen, and faecal bacteria were 20 to 30 times higher downstream of a deer wallowing site, than upstream (Environment Southland 2000). Similarly, Davies-Colley and Nagels (2002) measured concentrations of faecal bacteria that were 2-10 times greater downstream of two large deer farms than upstream. Wallowing by deer is likely to enhance erosion of the stream bed and bank, and near-channel areas.
Aside from the deposition of faecal material directly into waterways, a number of other transmission routes exist that can be considered ‘indirect’, whereby faecal microbes are ultimately transferred to waterways via the flow of water over the surface of the land (surface runoff) or down through the soil horizons (subsurface flow). The nature and relative importance of each indirect pathway varies with a range of factors including the type of farming, management practices, the magnitude of a rain event, soil type, slope angle and distance to waterways.
On hill-country dry-stock farm land, the generally steep topography promotes the generation of a significant volume of surface runoff under heavy and/or prolonged rainfall, providing an efficient mechanism by which faecal microbes are delivered to waterways. Field experiments have shown that between 2 x 109 and 5 x 1012 E. coli, per hectare of steep hillside (grazed by sheep), are washed to a headwater stream during the course of very heavy rainfall (return period of 8 years) events (Collins et al. 2005a). Furthermore, the yield of E. coli strongly correlates (negatively) with the time elapsed since the hillside was last grazed.
Farm tracks in hill-country readily generate surface runoff due to soil compaction, and can be expected to deliver microbes and other pollutants to waterways. However, we are not aware of any quantitative information in the literature with which to assess the relative importance of this process.
The hill-country topography also promotes the convergence of surface and subsurface flows causing near-channel saturated areas or wetlands to form. These wetlands drain into waterways, transporting high numbers of faecal microbes, particularly during heavy rainfall. For example, Collins (2004) measured storm period E. coli yields (across a range of rain events) of between 2 x 106 and 4 x 1010 most probable number (MPN) per hectare of grazed hillside, at the outflow of a hill-country wetland.
On dairy farms, effluent derived from the milking shed accounts for about 10% of the total daily load of faeces excreted by dairy cattle (Cameron and Trenouth 1999). The standard form of treatment for dairy shed effluent has been by a two-pond system combining both an anaerobic and facultative pond (Sukias et al. 2001). Pond treatment is efficient at removing sediment and biochemical oxygen demand, but high concentrations of faecal indicator bacteria and nutrients can remain (Hickey et al. 1989), discharging to a waterway.
Following the introduction of the Resource Management Act in 1991, land disposal of dairy effluent is now favoured by most regional councils. This approach, relative to the two-pond system alone, has the potential to markedly reduce the yield of faecal microbes and nutrients to waterways, attributable to the filtration and adsorption properties of soil. In addition, some treatment may be provided during the storage process, whereby the effluent is held in an existing two-pond system or sump, prior to land application.
Subsurface artificial drains commonly underlie dairy pastures where soils have some form of drainage restriction. The presence of subsurface drains reduces saturation of the soil and the propensity for surface runoff, a process that can rapidly transfer microbes to waterways. For example, once un-drained Pallic Soils have rewet in late autumn or early winter, their hydraulic conductivity falls to as low as 0.01 mm per day (Horne 1985). In these locations, all winter rainfall in excess of evapotranspiration is discharged as surface runoff and shallow seepage flow. Following the installation of subsurface drains, surface runoff on Pallic Soils has been shown to reduce by 60-80% (Hedley et al. 2005).
Despite the benefits of subsurface drains in reducing soil saturation and the generation of surface runoff, on soils with poor natural drainage, they are known to rapidly transfer both pathogens and nutrients to waterbodies via open ditches. This transfer has been shown to occur in response to the irrigation of effluent when the maximum depth of effluent application exceeds the soil water deficit e.g., Houlbrooke et al. (2004a). In many soils, this causes bypass or preferential flow of water and microbes to occur, whereby their movement is through soil cracks, large pores and worm channels rather than through the fine pores of the soil matrix. Under bypass flow, soil-water contact is minimal providing little opportunity for the filtration or adsorption of microbes. Water, entrained microbes and nutrients can therefore be readily transported to subsurface drains (and hence to waterbodies via open ditches) and groundwater. In addition to the application of effluent, both rainfall and the irrigation of water alone can also generate bypass flow, on drained land.
Monaghan and Smith (2004) studied the impact of effluent irrigation to land upon the generation and contamination of subsurface drain flows. They found that when the soil was wet (small soil moisture deficit) prior to application, E. coli concentrations in the resulting drain flows - 105 to 107 coliform forming units (cfu) 100 mL-1 - reflected those of the applied effluent. Smaller soil moisture deficits, however, led to lower E. coli concentrations in drain flows (≈104 cfu 100 mL-1), indicating that some mixing of the effluent with older resident soil water had occurred. The yield of E. coli across these reported drainage events ranged between 108 and 5 x 1010 cfu from the 1080 m2 study plot. Monaghan and Smith (2004) also found non-uniform patterns of effluent application with the outside of a small rotating irrigator receiving double the average application depth, promoting ponding and bypass flow. Hedley et al. (2005) report peak drain-flow Campylobacter concentrations of > 103 MPN 100 mL-1, following the application of effluent at a time of negligible soil moisture deficit. Similarly, Ross and Donnison (2003) found that when preferential flow occurred, Campylobacter concentrations in drainage water were similar to those in the applied effluent.
Contaminated drain flows also occur in response to rainfall during or following grazing episodes. Hedley et al. (2005) report drain flow Campylobacter concentrations of ≈ 102 MPN 100 mL-1 following overnight grazing (80-100 cows per ha) and ≈ 101 MPN 100 mL-1 following grazing 19 days earlier.
On some soil types, appreciable surface runoff, contaminated by faecal microbes, can still be generated on dairy land underlain by artificial drainage. For example, Hedley et al. (2005) report that 46 mm and 179 mm of surface runoff were generated upon a study plot underlain by a Tokomaru silt loam soil, during 2003 and 2004, respectively. This compared with 258 mm and 388 mm of subsurface drainage from the same plots over the same periods. Furthermore, the surface runoff generated was contaminated by faecal microbes, with concentrations of E. coli and Campylobacter peaking at > 105 MPN 100 mL-1 and > 103 MPN 100 mL-1, respectively, immediately following grazing. Peak Campylobacter concentrations in surface runoff, generated following the application of effluent, were also > 103 MPN 100 mL-1.
The irrigation of water to encourage pasture growth can promote the flushing of faecal microbes, from faeces deposited on pasture by grazing livestock, down through the soil horizons (particularly via bypass flow) with the potential to cause contamination of groundwater. Border-strip irrigation, in particular, has been shown to lead to the faecal contamination of wells up to 11 m below ground level (Close et al. 2005). Campylobacter were identified in 12% of groundwater samples with concentrations ranging between < 0.6 and > 3.1 MPN per L, raising implications for public health associated with the use of this groundwater for household drinking. Generally, the highest Campylobacter and E. coli concentrations found in the wells occurred approximately 20-30 days after a period of grazing had coincided with a border strip irrigation event or a large rainfall event.
Fine-grained aquifers (e.g., sand, sandstone and pumice) are efficient at retarding microbes, through filtration. However, some large-pore heterogeneous aquifers (e.g., gravel, karst, fractured rocks) are susceptible to microbial contamination because of their high permeability, low filtration capacity, and presence of preferential flow paths (Davies-Colley et al. 2003). Protozoa are assumed to be more efficiently filtered by aquifer media compared to bacteria and viruses, due to a generally larger particle size.
Contact for Enquiries
North Island Regions
Sector Performance Policy
Private Bag 3123 Hamilton
Phone: +64 7 957 8313
Fax: +64 7 957 8315
Contact this person