Predicting the effect of livestock inputs of E. coli on microbiological compliance of bathing waters

Three alternative approaches to predicting delivery of faecal indicators from livestock sources to surface water in the catchment of the River Irvine, Ayrshire, Scotland, are described. These are a soil transport model which assumes all E. coli are transported through the soil, a regression model using observed E. coli concentrations in surface waters, and a distributed catchment model (PAMIMO). Each of these is linked to a simple group of equations describing inputs of E. coli from livestock to land, transport and inactivation in the river Irvine and mixing and inactivation in the sea. The models predict E. coli content of bathing water for Irvine beach. The regression model gives the best predictions of bathing water quality. The low values predicted by the soil transport model suggests that preventing surface runoff of faecal indicators from livestock would provide an adequate solution to the problem of bathing water contamination.     

Faecal bacteria yields in artificial flood events: quantifying in-stream stores

Stream sediments have been recognised as an in-channel store of faecal contamination that can be mobilised during floods or other sediment-disturbing events. We studied this store of faecal contamination by creating artificial floods during dry weather when, in the absence of overland flow from the catchment, the only source of faecal bacteria was stores within the channel. Artificial floods, created by releasing water from a supply reservoir, increased the E. coli concentration in the water column by two orders of magnitude, from a background level of 10(2) cfu per 100 mL to over 10(4) cfu per 100 mL. The bacterial peak concentrations and yields declined systematically through a triplicate flood series. The size of the total in-channel store, calculated as the sum of yields of an infinite series of artificial floods, was approximately 10(8) cfu m(-2) of streambed area. Direct measurements of sediment E. coli found few sites (only those associated with cattle crossings) with areal concentrations as high as 10(8) cfu m(-2), consistent with flood yields. Concentrations of E. coli in the biofilms on exposed rocks were orders of magnitude lower, indicating that exposed rocks were not a source of E. coli released by the artificial floods.     

The Bacterial Quality of an Upland Stream

Faecal coliform concentrations were determined at stream and water inflow sampling sites within an upland catchment in North Derbyshire containing rough, semiimproved and improved pastures. The results indicated that, during its passage through the catchment, the faecal indicators of the stream increased, and this may constitute a health risk to cavers using a popular recreational cave into which the stream drains. The decline in streamwater quality suggests that a semipermanent store of bacteria existed in catchment soils, and hydrological processes operated to transfer bacteria from the land to the stream channel. Contrary to expectations, intra-catchment differences in the bacterial quality of water at inflow and stream sampling sites could not be related to the intensity of agricultural land use in adjacent areas. This is explained with reference to changes in the nature of hydrological transport pathways associated with the improvement of pastures.     

Riparian zone influence on stream water chemistry at different spatial scales: a GIS-based modelling approach, an example for the Dee, NE Scotland

A geographical information system (GIS-ARC/INFO) was used to collate existing spatial data sets on catchment characteristics to predict stream water quality using simple empirical models. The study, based on the river Dee catchment in NE Scotland, found that geological maps and associated geochemical information provided a suitable framework for predicting chemical parameters associated with acidification sensitivity (including alkalinity and base cation concentrations). In particular, it was found that in relatively undisturbed catchments, the parent material and geochemistry of the riparian zone, when combined with a simple hydrological flow path model, could be used to accurately predict stream water chemistry at a range of flows (Q95 to > Q5) and spatial scales (1-1000 km2). This probably reflects the importance of the riparian zone as an area where hydrological inputs to stream systems occur via flow paths in the soil and groundwater zones. Thus, evolution of drainage water chemistry appears to retain the geochemical characteristics of the riparian area as it enters the channel network. In more intensively managed catchments, riparian land use is a further influential factor, which can be incorporated into models to improve predictions for certain base cations. The utility in providing simple hydrochemical models, based on readily available data sets, to assist environmental managers in planning land use in catchment systems is discussed.     

A model framework to assess the effect of dairy farms and wild fowl on microbial water quality during base-flow conditions

There is concern regarding microbial water quality in many pastoral catchments in New Zealand which are home to numerous livestock and wild animals. Information on microbial impacts on water quality from these animals is scarce. A framework is needed to summarise our current knowledge and identify gaps at the scale of an individual farm. We applied a Monte Carlo modelling approach to a hypothetical dairy farm based on the extensive data sets available for the Toenepi Catchment, Waikato, New Zealand. The model focused on quantifiable direct inputs to the stream from ducks, cows and farm dairy effluent (FDE) during base-flow conditions. Most of the inputs of Escherichia coli from dairy farms occur sporadically and, therefore, have little effect on the expected median stream concentrations. These sporadic inputs do however, have a strong influence on extrema such as 95th percentile values. Current farm mitigations of fencing streams and using improved management practices for applying FDE to land, such as low application rate deferred FDE irrigation systems, would appreciably reduce faecal microbial inputs to the stream. However, the concentrations of E. coli in rural streams may not reduce as much as expected as wild fowl living in streams would have a larger effect on water quality than a farm in which environmental mitigations are widely implemented.     

Comparison of the efficacy of an existing versus a locally developed metabolic fingerprint database to identify non-point sources of faecal contamination in a coastal lake

A comparison of the efficacy of an existing large metabolic fingerprint database of enterococci and Escherichia coli with a locally developed database was undertaken to identify the sources of faecal contamination in a coastal lake, in southeast Qld., Australia. The local database comprised of 776 enterococci and 780 E. coli isolates from six host groups. In all, 189 enterococci and 245 E. coli biochemical phenotypes (BPTs) were found, of which 118 and 137 BPTs were unique (UQ) to host groups. The existing database comprised of 295 enterococci UQ-BPTs and 273 E. coli UQ-BPTs from 10 host groups. The representativeness and the stability of the existing database were assessed by comparing with isolates that were external to the database. In all, 197 enterococci BPTs and 179 E. coli BPTs were found in water samples. The existing database was able to identify 62.4% of enterococci BPTs and 64.8% of E. coli BPTs as human and animal sources. The results indicated that a representative database developed from a catchment can be used to predict the sources of faecal contamination in another catchment with similar landuse features within the same geographical area. However, the representativeness and the stability of the database should be evaluated prior to its application in such investigation.     

Sunlight, season, snowmelt, storm, and source affect E. coli populations in an artificially ponded stream

Reducing fecal indicator bacteria, such as Escherichia coli (E. coli), in streams is important for many downstream areas. E. coli concentrations within streams may be reduced by intervening ponds or wetlands through a number of physical and biological means. A section of Dunes Creek, a small coastal stream of southern Lake Michigan, was impounded and studied for 30 months from pre-through post-construction of the experimental pond. E. coli reduction became more predictable and effective with pond age. E. coli followed the hydrograph and increased several-fold during both rainfall and snowmelt events. Seasonally, the pond was more effective at reducing E. coli during summer than winter. Late summer, non-solar reduction or inactivation of E. coli in the pond was estimated at 72% and solar inactivation at 26%. E. coli DNA fingerprinting demonstrated that the winter population was genetically more homogeneous than the summer population. Detection of FRNA coliphages suggests that there was fecal contamination during heavy rain events. An understanding of how environmental factors interact with E. coli populations is important for assessing anticipated contaminant loading and the reduction of indicator bacteria in downstream reaches.     

Defining the sources of low-flow phosphorus transfers in complex catchments

Nutrient transfers from the land to rivers have the potential to cause persistent eutrophic impacts at low flows even though the transfers may constitute a minor percentage of total annual fluxes. In rural catchments, the contribution from agricultural soils during storm events can be particularly large and untangling the relative contributions from multiple sources that vary in time and space is especially problematic. In this study, the potential for domestic septic tank system pollution during low flows was investigated in 3 small catchments (3 to 5 km(2)) using an integrated series of methods. These included septic system surveys, continuous (10 min) total phosphorus (TP) monitoring at the outlet of each catchment, repeated low-flow water quality surveys in sub-catchments upstream of the catchment outlets and single day river-walk water quality surveys. A series of faecal matter and grey-water fingerprinting techniques were also employed. These included determining sterol ratios in stream sediments, monitoring the presence of proteins, E. coli and enterococci bacterial signatures and boron. The total density and density of poorly maintained septic systems mirrored the magnitude of frequent TP concentrations in the catchments although this relationship was less apparent in the nested sub-catchments. The exception was possibly related to the simple hydraulics in one particular catchment and indicated temporary effluent attenuation in the other catchments. Repeated low-flow and river-walk water quality surveys highlighted discrete areas and reaches where stepped changes in nutrient concentration occurred. Bio-chemical fingerprinting showed that between 7% and 27% of sediments were contaminated with human faecal material and correlation matrices indicated that, at least during low flows, P fractions were positively correlated with some markers of faecal and grey-water contamination.     

Impact of urbanization and agriculture on the occurrence of bacterial pathogens and stx genes in coastal waterbodies of central California

Fecal pollution enters coastal waters through multiple routes, many of which originate from land-based activities. Runoff from pervious and impervious land surfaces transports pollutants from land to sea and can cause impairment of coastal ocean waters. To understand how land use practices and water characteristics influence concentrations of fecal indicator bacteria (FIB) and pathogens in natural waters, fourteen coastal streams, rivers, and tidal lagoons, surrounded by variable land use and animal densities, were sampled every six weeks over two years (2008 & 2009). Fecal indicator bacteria (FIB; Escherichia coli and Enterococci) and Salmonella concentrations, the occurrence of Bacteroidales human, ruminant, and pig-specific fecal markers, E. coli O157:H7, and Shiga toxin (stx) genes present in E. coli, were measured. In addition, environmental and climatic variables (e.g., temperature, salinity, rainfall), as well as human and livestock population densities and land cover were quantified. Concentrations of FIB and Salmonella were correlated with each other, but the occurrence of host-specific Bacteroidales markers did not correlate with FIB or pathogens. FIB and Salmonella concentrations, as well as the occurrence of E. coli harboring stx genes, were positively associated with the fraction of the surrounding subwatershed that was urban, while the occurrence of E. coli O157:H7 was positively associated with the agricultural fraction. FIB and Salmonella concentrations were negatively correlated to salinity and temperature, and positively correlated to rainfall. Areal loading rates of FIB, Salmonella and E. coli O157:H7 to the coastal ocean were calculated for stream and river sites and varied with land cover, salinity, temperature, and rainfall. Results suggest that FIB and pathogen concentrations are influenced, in part, by their flux from the land, which is exacerbated during rainfall; once waterborne, bacterial persistence is affected by water temperature and salinity.     

Can sediment data be used to predict alkalinity and base cation chemistry of surface waters?

We hypothesise that stream sediment elemental composition can predict mean and minimum concentrations of alkalinity, Ca and Mg in the river water throughout a river network. We tested this hypothesis for the River Derwent catchment in North Yorkshire, England, by using 6 years of water chemistry data from the Environment Agency and a digital elevation model to flow path-weight British Geological Survey (BGS) sediment element concentration data. The predictive models for mean concentrations were excellent for Ca and alkalinity, but less good for Mg, and did not require land use data inputs as stream water sediment composition seems to reflect all aspects of the riparian zone soil system. Predictive model forms were linear. Attempts to predict minimum values for Ca and alkalinity also were less satisfactory. This probably is due to variations in hydrological response times to individual precipitation events across the catchment.     

Escherichia coli O157:H7 in drinking water from private water supplies in the Netherlands

The microbiological quality of drinking water from 144 private water supplies in the Netherlands was tested and additionally the occurrence of Escherichia coli O157 was examined. Faecal indicators were enumerated by using standard membrane filtration methods. The presence of E. coli O157 was determined using a specific enrichment method. Eleven percent of the samples contained faecal indicators whereas E. coli O157:H7 was isolated from 2.7% of the samples that otherwise met the drinking water standards. The E. coli O157 positive water supplies were located on camp-sites in agricultural areas with large grazer densities. Pulsed field gel electrophoresis (PFGE) analysis suggested that cattle might have been the cause of contamination. Our results indicate that compliance with microbiological quality standards obtained in routine monitoring does not always guarantee the absence of pathogens. The presence of pathogens such as E. coli O157 may suggest possible health consequences; however, a risk assessment process should be performed as the monitoring of both faecal indicator parameters and pathogens do not predict the effect of microbial contamination of drinking water on a population.     

The new era of the livestock production in Mongolia: Consequences on streams of the Great Lakes Depression

Mongolia, a landlocked country of the Central Asian plateau, is experiencing a significant modification of herding practices coupled with an increase in livestock numbers. These modifications lead to increasing impacts of grazing on the Mongolian steppes with major consequences on the waterbodies. We researched the impacts of grazing intensity on the streams of the Great Lakes Depression in northwestern Mongolia. We assessed the level of watershed and stream bank erosion and the type of vegetation structure. We calculated the livestock densities per watershed and linked them to the stream water discharge through a new metric (I_CU). I_CU was created as a function of cattle unit density and water discharge, having water discharge at a stream section reflecting its location in the drainage and therefore accounting for the surface area drained upstream. We measured also the major nutrients in the stream water and researched the causalities between the grazing and the impairment of watersheds and streams. Our results suggest that the increase of livestock numbers is reaching beyond the grassland and affecting the stream ecosystem. Two major impacts were highlighted by this study, 1) the extensive watershed and stream bank erosion and 2) the increase in concentration of suspended particles and orthophosphate in stream systems. When compared with past values from literature, our results show recent eutrophication of the streams compared to the pre-liberalization of the herding activity in Mongolia (before 1991). Consequently the continued uncontrolled increase of livestock numbers could threaten the conservation of the Mongolian waterbodies, with notable consequences on the life of the nomadic population of the Central Asian Plateau.     

Faecal pollution source identification in an urbanising catchment using antibiotic resistance profiling, discriminant analysis and partial least squares regression

Increasing urbanisation and changes in land use lead to adverse impacts on the quality of natural water resources. The specific sources of contamination are often difficult to identify using conventional water quality monitoring techniques. This acts as a significant constraint to the development of appropriate management techniques to protect natural water resources. Consequently, alternative means of identifying pollutant sources and their locality are necessary. In this study, Antibiotic Resistance Patterns (ARP) were established for a library of 1005 known Escherichia coli source isolates obtained from human and non-human (domesticated animals, livestock and wild) sources in an urbanising catchment in Queensland State, Australia. Discriminant Analysis (DA) was used to differentiate between the ARP of source isolates and to identify the sources of faecal contamination. Partial Least Square (PLS) regression was then utilised on identified human source isolates to correlate their locality with specified sampling locations within the catchment. The resulting ARP DA indicated that a majority of the faecal contamination in the rural areas was non-human. However, the percentage of human isolates increased significantly in urbanised areas using onsite systems for wastewater treatment. The PLS regression was able to develop predictive models which indicated a high correlation of human source isolates from the urban area. The study results confirm the feasibility of using ARP for source tracking faecal contamination in surface waters, as well as predicting their point of origin.     

Integrated Lake and Catchment Phosphorus Model– A Eutrophication Management Tool. I: Model Theory

The paper describes an integrated lake and catchment model which was developed and tested against data from the Rutland Water catchment (River Nene and Welland) in Eastern England. The model uses export coefficients to predict annual diffuse losses of phosphorus from catchment land use, which are temporally disaggregated using empirical relationships between phosphorus load and discharge in headwater streams. Dynamic inputs of phosphorus are incorporated into a hydrologically based, multi-reach distributed catchment model which also simulates sewage discharges, abstractions and reservoirs/ lakes. The model allows a comparison of simulated output with observed discharge and water-quality data from various points within the river system.     

Genotype diversity of Escherichia coli isolates in natural waters determined by PFGE and ERIC-PCR

Most library-dependent bacterial source tracking studies using Escherichia coli (E. coli) have focused on strain diversity of isolates obtained from known human and animal faecal sources for library development. In contrast, this study evaluated the genotype variation of E. coli isolated from natural surface water using pulsed field gel electrophoresis (PFGE) and enterobacterial repetitive intergenic consensus sequence polymerase chain reaction (ERIC-PCR) to better understand these naturally occurring populations. A total of 650 water samples were collected over a nine month period from eleven sampling stations from Lake Waco and Belton Lake in Central Texas. Of the 650 water samples collected, 412 were positive for E. coli, yielding a total of 631 E. coli isolates (1-12 isolates collected per sample). PFGE and ERIC-PCR patterns were successfully generated for 555 isolates and were compared using the curve-based Pearson's product-moment correlation coefficient. The 555 E. coli isolates represented 461 PFGE genotypes, with 84% (386/461) of the genotypes being represented by individual isolates. The remaining 75 genotypes were represented by 2-5 isolates each. Using ERIC-PCR, the 555 E. coli isolates represented 175 genotypes, with 63% (109/175) of the genotypes being represented by individual isolates. In contrast to the PFGE results, two ERIC-PCR genotypes represented 37% of the E. coli isolates, (83 and 124 isolates, respectively), and were found throughout the watersheds both spatially and temporally. Based on the PFGE genotype diversity of water isolates, there is little evidence that a small number of environmentally-adapted E. coli represent dominant populations in the studied waterbodies. However, with the lower discriminatory power technique ERIC-PCR, an opposing conclusion might have been drawn. These results emphasize the importance of considering the resolving power of the source tracking technique being used when assessing strain diversity and geographical stability.     

Results of field investigations into the impact of intermittent sewage discharges on the microbiological quality of wild mussels (Mytilus edulis) in a tidal estuary

Field surveys were designed to examine the effects of sewage contamination from storm overflow effluent on faecal coliform and Escherichia coli concentrations in the flesh of wild mussels (Mytilus edulis). Bags containing 30 mussels each were fixed at known inter-tidal locations and retrieved at intervals following discharge from a nearby combined sewer overflow (CSO). Concentrations of faecal coliform bacteria and E. coli were measured in the shellfish flesh and in samples of overlying water prior to collection of the mussel samples. Faecal coliform and E. coli concentrations in shellfish increased rapidly after CSO discharge. E. coli concentrations exceeded the European shellfish hygiene class C limit of 46,000 100g(-1), and decayed during subsequent CSO discharge-free periods. The concentration and depuration response was independent of the magnitude of CSO spill volume. First-order exponential decay functions were fitted to the data. Decay rates were lower than those found in corresponding microcosm experiments. This relates to the repeated pattern of inundation and exposure associated with the tidal cycles in the estuary. Relationships between E. coli and faecal coliform concentrations in the shellfish and overlying water samples were relatively weak (r<0.60), a pattern often seen with data from uncontrolled environmental experiments.     

In situ measurement and statistical modelling of Escherichia coli decay in small rivers

Numerous studies have been carried out on the environmental factors associated with the decay of faecal bacteria in open (fresh or marine) waters. The present study aimed at understanding the fate of bacteria in small streams (flow <20 m3 s(-1)) for which there is a lack of knowledge. An original in situ protocol was developed for measuring the die-off of Escherichia coli (E. coli) from wastewater treatment plants. Based upon 80 values of the decay first-order parameter (K or its inverse T90), collected from five rivers in Normandy (France), a median T90 of 10 h and a minimal T90 of 1.3 h were obtained. K was then modelled as a linear function of variables made up from flow, water temperature and suspended particulate matter (SPM). The set of significant co-variables did not include light indicators. E. coli decay is inversely related to the river flow and it becomes highly significant below 0.3 m3 s(-1). The positive effect of small flows on die-off is increased by water temperature over 15 degrees C, whereas it could be reduced by SPM. The major co-variable of the model (p < 10(-9)) is an empiric composite variable integrating the effect of flow and temperature that explains more than 40% of the variance of K. We interpreted this as an expression of predation by benthic micro-grazers which could be the main cause of E. coli die-off in small streams in temperate countries.     

The impact of cattle farming best management practices on surface water nutrient concentrations,faecal bacteria and algal dominance in the Lake Oconee watershed

The objective of this study was to assess water quality in the Lake Oconee watershed and evaluate the best management practices used by cattle farms to reduce water contamination. Inorganic nutrient concentrations, algal abundance and faecal bacteria were highest in the cattle farming areas. The diatom community where cattle had no access was dominated by Achnanthidium minutissimum (Kützing) Czarnecki and Fragilaria crotonensis Kitton, and in sites where cattle were allowed direct access to the lake, Asterionella formosa Hassal, Nitzschia palea (Kützing) Smith and Navicula rostellata Kützing dominated. The latter three taxa are well‐known high‐nutrient diatoms. High populations of green algae (coccoid Desmidiaceae) were found where cattle had access. Sources of faecal pollution were identified using polymerase chain reaction detection, with Bifidobacterium adolescentis as a marker of human faecal pollution and Bacteroides (BoBac) indicating cattle faecal pollution. Overall, riparian buffers were most effective at reducing pollution from cattle operations.     

Modelling denitrification at the catchment scale

The objective of this work was to evaluate the importance of heterotrophic denitrification in the fate of nitrogen surpluses at the catchment scale. For that purpose we modified the denitrification module of TNT2 model and calibrated the model on a small catchment where denitrification measurements had been performed in different locations. The main interest of the TNT2 model is its ability to simulate the dynamics of the zones where soil and shallow water table interact, making it possible to spatialize the denitrification process. Daily water and nitrogen flux at the outlet were relatively well simulated (Nash of 0.85 and 0.77). In average, the model correctly simulates the denitrification measurements (R = 0.68). Nitrogen flux towards the atmosphere, at the catchment scale (4.70 g N m^− 2 year^− 1), is of the same order of magnitude as the soluble N flux in the stream. The model was able to reproduce the distribution of denitrification in the riparian (mean of 9.26 g N m^− 2 year^− 1) and hillslope (mean of 3.45 g N m^− 2 year^− 1) domains of the catchment. The results confirm the importance of riparian denitrification, but show also that hillslope soils contribute significantly (60%) to the whole catchment denitrification. The variations of denitrification rates, and also of nitrate concentrations in stream were not very well simulated by the model, highlighting the complexity of the spatial and temporal controls of nitrogen dynamics in areas with high inputs of nitrogen fertilizers, especially under organic forms.     

Associations among pathogenic bacteria, parasites, and environmental and land use factors in multiple mixed-use watersheds

Over a five year period (2004-08), 1171 surface water samples were collected from up to 24 sampling locations representing a wide range of stream orders, in a river basin in eastern Ontario, Canada. Water was analyzed for Cryptosporidium oocysts and Giardia cyst densities, the presence of Salmonella enterica subspecies enterica, Campylobacter spp., Listeria monocytogenes, and Escherichia coli O157:H7. The study objective was to explore associations among pathogen densities/occurrence and objectively defined land use, weather, hydrologic, and water quality variables using CART (Classification and Regression Tree) and binary logistical regression techniques. E. coli O157:H7 detections were infrequent, but detections were related to upstream livestock pasture density; 20% of the detections were located where cattle have access to the watercourses. The ratio of detections:non-detections for Campylobacter spp. was relatively higher (>1) when mean air temperatures were 6% below mean study period temperature values (relatively cooler periods). Cooler water temperatures, which can promote bacteria survival and represent times when land applications of manure typically occur (spring and fall), may have promoted increased frequency of Campylobacter spp. Fifty-nine percent of all Salmonella spp. detections occurred when river discharge on a branch of the river system of Shreve stream order = 9550 was >83 percentile. Hydrological events that promote off farm/off field/in stream transport must manifest themselves in order for detection of Salmonella spp. to occur in surface water in this region. Fifty seven percent of L. monocytogenes detections occurred in spring, relative to other seasons. It was speculated that a combination of winter livestock housing, silage feeding during winter, and spring application of manure that accrued during winter, contributed to elevated occurrences of this pathogen in spring. Cryptosporidium and Giardia oocyst and cyst densities were, overall, positively associated with surface water discharge, and negatively associated with air/water temperature during spring-summer-fall. Yet, some of the highest Cryptosporidium oocyst densities were associated with low discharge conditions on smaller order streams, suggesting wildlife as a contributing fecal source. Fifty six percent of all detections of ≥2 bacteria pathogens (including Campylobacter spp., Salmonella spp., and E. coli O157:H7) in water was associated with lower water temperatures (<∼14 °C; primarily spring and fall) and when total rainfall the week prior to sampling was >∼27 mm (62 percentile). During higher water temperatures (>∼14 °C), a higher amount of weekly rainfall was necessary to promote detection of ≥2 pathogens (primarily summer; weekly rainfall ∼>42 mm (>77 percentile); 15% of all ≥2 detections). Less rainfall may have been necessary to mobilize pathogens from adjacent land, and/or in stream sediments, during cooler water conditions; as these are times when manures are applied to fields in the area, and soil water contents and water table depths are relatively higher. Season, stream order, turbidity, mean daily temperature, surface water discharge, cropland coverage, and nearest upstream distance to a barn and pasture were variables that were relatively strong and recurrent with regard to discriminating pathogen presence and absence, and parasite densities in surface water in the region.