Do Septic Tanks Pose a Eutrophication Risk in Rural Headwaters?

Home , Septic tank

Do septic tanks pose a eutrophication risk in rural headwaters?

P.J.A. Withers,1*, H.P. Jarvie2, C. Stoate3

1School of Environment, Natural Resources and Geography, Bangor University, Bangor,

Gwynedd, LL57 2UW, UK

2Centre for Ecology and Hydrology, Wallingford, Oxfordshire OX10 8BB, UK

3The Game & Wildlife Conservation Trust, Allerton Project, Loddington, Leicestershire LE7

9XE, UK

*Corresponding author

Tel: +44 (0) 1248 382631

Fax: +44 (0) 1248 354997

Email: [email protected]

Abstract

Septic tanks are a potential source of nutrient emissions to surface waters but few data exist in the UK to quantify their significance for eutrophication. We monitored the impact of septic tank systems on nutrient concentrations in a stream network around a typical English village over a 1-year period. Septic tank effluent discharging via a pipe directly into one stream was highly concentrated in soluble N (8-63 mg L-1) and P (<1-14 mg L-1) and other nutrients (Na, K, Cl, B and Mn) typical of detergent and household inputs. Ammonium-N

(NH4N) and soluble reactive P (SRP) fractions were dominant (70-85% of total) and average

-1 concentrations of nitrite-N (NO2N) were above levels considered harmful to fish (0.1 mg L ).

Lower nutrient concentrations were recorded at a ditch and a stream site, but range and average values downstream of rural habitation were still 4 to 10-fold greater than those in upstream sections. At the ditch site, where flow volumes were low, annual flow-weighted

-1 concentrations of NH4N and SRP increased from 0.04 and 0.07 mg L , respectively upstream to 0.55 and 0.21 mg L-1 downstream. At the stream site, flow volumes were twice as large

-1 and flow-weighted concentrations increased much less; from 0.04 to 0.21 mg L for NH4N and from 0.06 to 0.08 mg L-1 for SRP. At all sites, largest nutrient concentrations were recorded under low flow and stream discharge was the most important factor determining the eutrophication impact of septic tanks. The very high concentrations, intercorrelation and dilution patterns of SRP, NH4-N and the effluent markers Na and B suggested that septic tank soakaways in the heavy clay catchment soils were not retaining and treating the septic tank effluents efficiently, with profound implications for stream biodiversity. Rural communities therefore need to be made more aware of the potential impacts of septic tank systems on water quality so that their management can be optimised to reduce the risk of potential eutrophication and toxicity to aquatic ecosystems during summer low flow periods.

Key words: septic tanks;, eutrophication; phosphorus; nitrogen; rural populations

1. Introduction

Eutrophication of inland and coastal waters is a widespread environmental problem caused by the increased cycling and fluxes of phosphorus (P) and nitrogen (N) associated with urbanization, deforestation and intensification of agriculture. Eutrophication in rivers is most prevalent under low-flow conditions when residence times are greatest, and the role of major point sources in maintaining high nutrient concentrations, especially P, during summer low flows is now well documented (Jarvie et al., 2006; Neal et al., 2010). In rural catchments with no major sewage treatment works (STW), semi-continuous discharges of nutrients from

‘mini-point’ sources such as farmyards and septic tanks can also have a large impact on local stream chemistry and ecology during summer (Marshall and Winterbourn, 1979; Hooda et al., 2000; Steffy and Kilham, 2004; Palmer-Felgate et al., 2010). However, these ‘mini-point’ sources are usually ignored in catchment management programmes due to a lack of data on their patterns of nutrient emission and a lack of understanding regarding their relevance for eutrophication. In particular, septic tanks (sometimes referred to as ‘on-site wastewater systems’) are often assumed to contribute negligible amounts of nutrients on an annual basis relative to agricultural land, because of their sporadic and isolated distribution across catchments. In the UK, no one single authority has responsibility for septic tanks, many are not registered and they are often not sited or maintained adequately (Bulter and Payne, 1995;

May et al., 2010). Because they are not registered, and because many are very old, their design, operation and distribution within catchments are often not clear. Consequently there is confusion over their contribution to downstream eutrophication and the need to control their emissions.

Recent catchment-based studies indicate that septic tanks may be a much more important eutrophication risk than just considering their contribution to annual nutrient loads, especially in headweater streams where the capacity for dilution with water from the upstream

3 catchment areas is low. Using a continuous bankside analyser, Jordan et al. (2007) found that both chronic storm-driven and acute storm-independent point source P inputs under low flow conditions (i.e. not derived from agricultural land) were maintaining highly eutrophic conditions in a tributary of the Blackwater river in Northern Ireland. Arnscheidt et al. (2007) further identified these point sources as septic tanks due to (a) the presence of human faecal contamination in the river bed sediments, (b) simultaneous increases in stream soluble reactive P (SRP) and boron (B, a marker for detergents) concentrations under low flow, and

(c) a significant correlation between stream P status and the density of poorly maintained septic tanks in different areas of the catchment. Positive relationships between SRP and B, and negative relationships with flow indicative of point source inputs, have also been found in rural headwater streams in England (Neal et al., 2005; Jarvie et al., 2008, 2010), and in runoff entering these streams (Withers et al., 2009). In a recent review, May et al. (2010) found that septic tanks can collectively make up to 20% of catchment P loads in rural areas.

Septic tanks are considered to be an effective means of wastewater treatment in rural areas provided they are designed, located and maintained satisfactorily. According to Mara (2004), septic tanks are suitable for village populations up to 500. Settlement and anaerobic digestion of organic (largely faecal) matter occurs within the tank and the discharged effluent is treated by the soil in a soakaway area (often referred to as a ‘drainfield’ or ‘soil absorption field’).

The discharge from the tank is highly enriched in reduced inorganic nutrients (e.g. methane

(CH4), ammonium-N, (NH4N), SRP, sulphides (H2S) and manganese (Mn)) and pathogens

(Reneau et al., 1979; Wilhelm et al., 1994; Kay et al., 2008). When first discharged, tank effluent clogs the surface soil (5-15 cm) within the soakaway forming a ‘biomat’, which then acts as a site of further anaerobic digestion by microorganisms as infiltration rates and oxygen supply decreases within the biomat under continued loading (Beal et al., 2005). As pathogens and reduced nutrients percolate into the unsaturated zone beneath the biomat, the

4 pathogens die off and the nutrients become transformed via oxidation, sorption, precipitation and plant uptake, resulting in high concentrations of nitrate N (NO3N) but low concentrations of SRP and pathogens. Lack of C in the unsaturated zone inhibits denitrification of the NO3N produced, and seepage of nitrate into the groundwater is considered to be the most ubiquitous environmental problem associated with correctly working septic tanks (e.g. Weiskel and

Howes, 1992; Gill et al., 2009).

Contamination of groundwater and surface waters with NH4N and/or SRP (e.g. Lapointe et al., 1990; Robertson, 2008), or pathogens (Ahmed et al., 2005), implies either direct discharge (i.e. no soakaway), existence of preferential flow pathways or shallow groundwater within the soakaway area, or a system failure due to poor design, soil type, location, age, or a lack of maintenance. There are a number of examples in the literature indicating reduced effectiveness. Ahmed et al. (2005) found that two-thirds of 48 septic tank systems surveyed in an area of Queensland, Australia needed cleaning out and half had ‘soggy’ ineffective soakaways. Arnscheidt et al. (2007) found that 35% of 113 systems surveyed in the Lough

Neagh catchment in Ireland were at high risk of nutrient leakage and over 70% were at medium risk of leakage. May et al. (2010) suggest that over 80% of septic tank systems in the

UK are probably working inefficiently, and consequently that are a potentially significant and underestimated source of phosphorus to nearby watercourses. Siting of systems on impermeable soils with a limited capacity for effluent infiltration appears a common cause of failure (Day, 2004; Beal et al., 2005).

As part of a wider study (PARIS: Phosphorus from Agriculture: Riverine Impacts Study) on the chemical and ecological impacts of various agricultural sources of P in England

(Palmer-Felgate et al., 2009; Jarvie et al., 2010), we investigated the potential eutrophication impact of rural septic tank systems at three sites located around the village of Loddington in

Leicestershire. The aim of the study was to characterize the nutrient composition of septic

5 tank discharged to a watercourse and to provide some data on their relative contribution to ambient nutrient concentrations in the lotic waters (ditches and streams) around Loddington.

The stream network at Loddington feeds into a headwater tributary (Eye Brook) of the River

Welland which flows through an undulating glacial till landscape dominated by mixed arable and livestock farming on heavy-textured, underdrained, chalky boulder clay soils

(Denchworth Association). Denchworth soils are characterized as having an impermeable clay horizon at about 25 cm depth and high proportion of rainfall generates surface runoff

(Soil Survey of England and Wales, 1983).

2. Methods

2.1 Study sites, field sampling and flow measurements

The three sites were monitored approximately weekly for just over 12 months. At one site

(Whitehorse Creek, WC), septic tank effluent from older houses in the main Loddington village (exact number not known) was discharged via a pipe directly into a stream. At a ditch site (Village East, VE), samples were taken upstream and downstream of a cluster of four houses (ca. 8 people) and a small but regularly used visitor centre (estimated visitor numbers of 600-1100 per year) with two offices at the eastern end of the village (Fig. 1). At a stream site (Loddington North, LN), samples were taken upstream and downstream of a cluster of eight houses (ca. 14 people) located in the Belton Bridge catchment monitored by Jarvie et al.

(2010). The septic tank systems serving the houses and visitor centre at VE and LN mainly discharge to the streams via soakaways and ditches.

The sampling period was from February 2007 until February 2008 at WC, and from

October 2006 – October 2007 at VE and LN, except during April and early May 2007 when there was too little flow to sample all sites during a very dry period (Fig. 2a). All monitoring sites were sampled on the same day. The samples were filtered in the field (using 0.45 µm

6 cellulose nitrate filter membranes) and sent in cool boxes by overnight courier for chemical analysis. Instantaneous flow was measured on 36 of the 37 sampling occasions at WC by timed collection of a known discharge volume from the pipe, and 12 out of the 46 sampling occasions at VE and LN by timing a floating object and measuring the volume of the stream channel. Daily rainfall over the monitoring period was measured on-site at Loddington to the nearest 0.2 mm and compared to the long-term average rainfall for the area. May, June and

July were particularly wet months with almost twice the long-term average rainfall (Fig. 2a).

2.2 Chemical analysis and load estimates

2.2.1 Analytical procedures

All pipe, ditch and stream samples were analysed for the following range of determinands: pH, Gran alkalinity (Alk), soluble reactive P (SRP), total dissolved P (TDP), total P (TP), ammonium-N (NH4-N), nitrate-N (NO3-N), nitrite-N (NO2-N), total dissolved N (TDN), sulphate-S (SO4-S), chloride (Cl), sodium (Na), potassium (K), calcium (Ca), magnesium

(Mg), boron (B), manganese (Mn), iron (Fe), and dissolved organic carbon (DOC). Samples were refrigerated before analysis and P fractions were determined within c. 24 hours of sampling to avoid significant sample deterioration on storage.

Sample pH was measured using a Radiometer GK2401c combination electrode and meter. Total alkalinity was determined by Gran titration with 0.05M HCl in the pH range 3-4.

Phosphorus fractions were determined colorimetrically (Neal et al., 2000) either directly for

SRP or after persulphate digestion (Eisenrich et al., 1975) for TDP and TP. A dissolved hydrolysable P (DHP) fraction was taken as the difference between SRP and TDP and a particulate P (PP) fraction was taken as the difference between TDP and TP. Nitrate, nitrite, sulphate and chloride were measured using ion chromatography, NH4-N was determined colorimetrically using an indophenol blue method (Leeks et al., 1997), and TDN was

7 measured by an automated thermal oxidation process, by which N in the sample is converted to its oxide (NO) and measured using a chemoluminescent detector. Dissolved organic nitrogen (DON) was calculated as the difference between TDN and the sum of NH4-N, NO3-

N and NO2-N. The major elements (Na, K, Ca, Mg) and micro elements (Mn, Fe, and B) were measured on samples filtered through 0.45 µm membranes using an Inductively

Coupled Plasma Optical Emission Spectrometer (ICP-OES). Dissolved Organic Carbon

(DOC) was measured by automated thermal oxidation to CO2 and non-dispersive infrared gas analysis.

2.2.2. Load estimation

An interpolation technique was employed to estimate annual stream load at each monitoring point as the product of the flow-weighted mean concentration (FWMC) and the mean daily flow over the monitoring period. This algorithm was chosen as the estimates produced have relatively small bias and lower variance than other comparable estimators using comparable routine weekly river water quality data (Webb et al., 1997):

River load =Kr( ∑ (Ci * Qi)/ ∑Qi) * (Qr)

Where:

Ci is the instantaneous concentration in the stream at the time of sampling

Qi is the instantaneous river flow at the time of sampling

Qr is the average long-term river flow record over the period of record

Kr is a conversion factor to take account of the units and period of record

At Whitehorse Creek, FWMCs were calculated from actual instantaneous flow measured on each sampling occasion (n = 36); FWMC = ∑ (Ci * Qi)/ ∑Qi. Daily flows over the whole monitoring period at WC were estimated from the linear relationship between instantaneous flow and the daily flow record from an Environment Agency gauging station on a nearby first order stream (GS31025; the River Gwash South Arm at Manton): r2 0.94, P <0.001. Daily load values were then summed to give total load over the whole monitoring period and adjusted pro-rata to a year. As the number of instantaneous flow measurements taken at VE and LN represented only a 3 month period, daily mean flows for days when samples were taken (for calculation of FWMCs), and over the whole monitoring period (for calculation of annual loads) were also estimated from the Gwash gauging station. The relationship between measured flow at the upstream LN station and Gwash (r2 0.47, P <0.001) was not as strong as at WC, but the estimated flows at LN were fully consistent with the flows at Belton Bridge reported by Jarvie et al. (2010) after taking into account differences in catchment area

(approximately half). Measured flows in the ditch at the upstream VE station were 27% of the measured stream flow at the upstream LN station (r2 0.92, P <0.001) and this ratio was used to estimate the daily flow record for the upstream station at VE. Daily flows at the downstream stations were based on the linear relationships between upstream and downstream measured flows; at LN, downstream flow was 28% greater than upstream (r2

0.99, P <0.001) and the corresponding increase at VE was 15% (r2 0.99, P <0.001).

3. Results

3.1 Whitehorse Creek

The composition of the effluent discharging from the pipe at WC was highly variable but considerably enriched in N, P, Cl, SO4S, Na, K, B and Mn and with high alkalinity (>6000

µEq L-1) relative to receiving waters (Table 1). Concentrations of TP ranged up to 21 mg L-1

(median 2.5 mg L-1) and were dominated by dissolved P forms (TDP was always >60% of

TP). Although SRP was on average over 80% of TDP, the DHP fraction (organic and polyphosphate P forms) was also significant, ranging up to 6 mg L-1. Concentrations of TDN ranged up to 60 mg L-1 (median >20 mg L-1) and were dominated by ammonium-N (ca. 70% of TDN), median concentrations of which were four-fold greater than those of nitrate. Nitrite-

N was also high (up to 3 mg L-1). Concentrations of other enriched nutrients, including the detergent markers Na and B, and DOC were 2 to 4-fold greater than the background values observed in the stream at Belton Bridge from 2004-2006 (Jarvie et al., 2010). A neutral to alkaline pH and high concentrations of Ca and Mg reflected the slightly calcareous lithology.

Temporal fluctuations in concentrations over the monitoring period were similar for all enriched nutrients, except nitrate which showed an opposite trend. The typical pattern is illustrated for SRP in Figure 2b. Largest concentrations were obtained during periods of low flow, particularly during September and October, and declined when flow rates increased.

For SRP, NH4N, Na, SO4S and Mn, flow accounted for over 50% of the variance in nutrient concentrations; the example for Mn is shown in Fig. 2c. For DHP, Cl, B and K, flow- concentration relationships were also significant (P <0.05-0.001) but the variance accounted for was low (<30%), particularly for B (12%). All enriched nutrients except B were significantly (P <0.05) correlated with one another being particularly strong between SRP,

NH4N and Na (Table 2, Fig. 3a). A correlation between B and Na in the pipe discharge might have been expected because these two elements are both constituents of detergents (Neal et al., 2010), but there was wide scatter suggesting complex sources (Fig. 3b).

3.2 Village East

Nutrient concentrations in the stream at VE were considerably lower than in the pipe discharge at WC, but downstream values at VE were considerably elevated compared to

10 upstream values, despite the greater flow rates downstream (Table 3). Highly significant (P

0.01 - <0.001) increases in the mean values all forms of P except the particulate P fraction, and in all forms of N except the DON fraction, occurred downstream of the houses and visitor centre. Median concentrations of SRP increased from 59 to 239 µg L-1, whilst median

NH4N and NO2N concentrations increased by up to an order of magnitude to 0.6 and 0.1 mg

-1 L , respectively. Downstream increases in the concentrations of NH4N and NO2N were positively correlated with one another (r2 0.44, P<0.001). The proportions of TP occurring in dissolved forms increased from 61% upstream to 85% downstream, while NH4-N as a proportion of TDN increased from 4% upstream to 13% downstream. The molar ratio of inorganic N to inorganic P ratio decreased from 190 upstream to 74 downstream reflecting the proportionally greater enrichment with P than with N downstream. Mean concentrations of Na, K, Cl, SO4S and B increased at least two-fold downstream (P <0.001) and alkalinity rose to over 5000 µEq L-1. The mean concentration for Mn upstream was heavily skewed by one extremely large value and median concentrations downstream were also double those obtained upstream. While most nutrients increased downstream, average DOC and Fe concentrations decreased.

Temporal trends in nutrient concentrations at VE showed that the downstream increases

(or decrease for DOC) were maintained over the whole monitoring period; data for SRP,

NH4N and the effluent marker Na are shown in Fig. 4. Other enriched nutrients showed the same trend. Downstream nutrient concentrations increased most during low flow periods, especially in September, October and early November in both 2006 and 2007. Smaller increases in SRP and NH4N compared to the effluent markers Na and B, and the conservative tracer Cl, during summer low flow periods suggest some additional P and N assimilation/oxidation within the ditch. Downstream concentrations of nitrate were

-1 consistently increased (3-4 mg L ) over the period January to August when NH4N

11 concentrations were low. While upstream concentrations showed little (Na, K, B and Mn), or no (SRP, NH4N), significant trend with flow, downstream concentrations became significantly (r2 up to 0.47, P<0.001) more diluted as flow rates increased. With the exception of K and Cl, upstream nutrient concentrations showed no significant inter-correlation with other nutrients, whereas downstream nutrient concentrations were usually well correlated with one another (e.g. Table 2). In contrast to WC, there was a significant correlation between B and Na at the downstream station (r2 0.56, P <0.001; Fig. 3b), but there was no correlation between B and Na at the upstream station.

3.3 Loddington North

Differences in nutrient concentrations between upstream and downstream monitoring stations at LN were generally less marked, and not apparent for as many determinands, compared to the VE site. However, mean values of alkalinity, SRP, DHP, NH4-N, Na, K, B and Mn were still significantly (P<0.05) greater downstream of the cluster of houses (Table 4). Trends in median concentrations followed a similar pattern to mean values. The increases in alkalinity,

Na, K and B were small (<10%), but median SRP, DHP, NH4-N and Mn values increased by

194%, 83%, 850% and 40%, respectively. Concentrations of Mg apparently fell very slightly downstream, but there were no other significant differences. Examination of temporal trends revealed that increases in downstream concentrations only occurred on a very limited number of occasions (Fig. 4). On 10 October 2006, 6 November 2006, 11 June 2007 and in

September 2007, there were coincident large increases in SRP, NH4N, Na, K, and Mn, and to a lesser extent B. The increase in nutrient concentrations observed on these four occasions was greater than was observed at VE.All these sampling occasions were characterized by low flow conditions and there was a consistent and significant trend for downstream concentrations to decline as flow increased (Fig. 5). This dilution was notably absent at the

12 upstream station, which tended either to show a concentration increase at higher flows (e.g.

SRP) or show no effect (Na, B). As found at VE, nutrient concentrations were generally better inter-correlated with each other downstream than upstream (Table 2). Concentrations of Na and B were also signficantly correlated downstream (r2 0.46, P <0.001), but only because of the simultaneous increase in these nutrients on the 4 occasions when other nutrient concentrations increased (Fig. 3b).

3.4 Annual nutrient fluxes

Estimated mean daily discharges over the respective monitoring periods were 26 m3 at WC,

159 and 182 m3 upstream and downstream at VE, and 555 and 710 m3 upstream and downstream at LN. Annual nutrient exports were therefore much greater at LN due to the much larger volumes of flow (Table 5). Annual exports of TP were 15 kg at WC, 20 and 29 kg at VE and 69 and 73 kg at LN with 69%, 21 and 47% and 17 and 27%, respectively in

SRP form. Corresponding values for TDN export were 141 kg at WC, 389 and 599 kg at VE and 1055 and 1294 kg at LN of which 35, 0.5 and 6 and 0.7 and 4%, respectively were in

NH4N form. Hence there was a notable increase in the proportion of the N and P loads in

SRP and NH4N forms at the downstream stations.

Differences in annual export of dissolved P (11-12 kg) and dissolved N (210-238 kg) between upstream and downstream stations were of a similar magnitude, despite the large difference in flow between the two sites. Taking out the effect of flow, downstream flow- weighted concentrations of dissolved P and N fractions increased between 3 (SRP) to 14-fold

(NH4N) compared to upstream values at VE, but only SRP, NH4N and NO2N showed a small increase at LN. The effect of septic tank discharge on stream nutrient status was therefore much greater at VE than at LN because of the lower flow rates at VE. Flow-weighted

13 concentrations of nitrate decreased slightly at LN, while there was apparent retention of particulate P between monitoring stations at both sites, being slightly greater at LN than VE.

4. Discussion

In the UK, the design and location of many septic tanks reflects the historical legacy of the wastewater disposal infrastructure prevalent in rural areas during periods when regulations and environmental awareness were not so rigorously defined or implemented. Invariably septic tanks were located close to watercourses, and it is quite apparent that in many areas there is still direct discharge of septic tank effluent to adjacent streams, often together with discharges from other sources such as farmyards, roads and field drains (e.g. Withers et al.,

2009). There is no reason to suppose that Loddington is any different to other rural villages in the English landscape that rely on septic tank systems for waste disposal. There is no mains sewerage network at Loddington and the mixture of historic and modern housing delivers the effluent either directly to the stream channel (older systems) or via soakaways and ditches

(modern systems). The three sites that were monitored at Loddington therefore represent the range of inputs likely to be encountered in the UK rural landscape.

The effluent discharge from the pipe at Whitehorse Creek was highly concentrated in the range of nutrients that are typically associated with the inorganic products of anaerobic digestion of organic material in septic tanks and associated detergent ingredients; SRP,

NH4N, Na, K, Cl, B and Mn. The range in values, predominance of reduced forms (e.g NH4N

SRP and Mn) and high inter-correlation between enriched nutrients is similar to that obtained for neat septic tank effluent found in other studies (Beal et al., 2005). Whelan and Titamnis

(1982) found effluent was up to 20-fold enriched with nutrients compared to tap water but that the composition varied between households depending on water and detergent use. The median molar ratio of inorganic P to inorganic N in the pipe discharge at WC (18) is the same

14 as that found in septic tank effluent by Weiskel and Howes (1992) and the nutrient ratio required for algal growth (Redfield, 1958). SRP and NH4N were particularly well correlated to the detergent marker Na, but less well correlated to the detergent marker B. This may be due to the confounding release of B from the underlying soil parent material (Harder, 1970) since B tended to show some correlation with Ca and Mg. Varying Na:B ratios at the different sites may indicate differences in domestic practices.

The same cations and anions observed in high concentrations in the pipe discharge were also clearly elevated, and intercorrelated, at the monitoring stations downstream of the septic tank systems at both the Village East and Loddington North sites. Range and average concentrations were generally lower than at WC due to the greater opportunity for nutrient attenuation within the soakaway and/or ditch system, and dilution on entering the stream network, especially at Loddington North where flow volumes were greater. Attenuation of

SRP and NH4N (an order of magnitude lower) was very much greater than for the other more conserved nutriuents, Na, K, Cl, B and Mn (2 to 5-fold lower) as might be expected in the more aerobic environment of ditches and streams (Stutter et al., 2010). A consistent pattern of exponential nutrient dilution as flow increased was observed at WC and both downstream stations at VE and LN. Dilution curves for nutrient elements in water are very characteristic of point source inputs (Jarvie et al., 2006) and there was no indication of a storm-dependent discharge of septic tank effluent at either VE or LN sites.

The strong dilution patterns for SRP and NH4-N with increasing flow, as well as for conservative effluent tracers such as Na and B, combined with high concentrations of NH4-N and SRP, and small or no increases in concentrations of NO3-N, at downstream sites provide compelling evidence that septic tank effluent was not being effectively treated at Loddington.

Poor functioning of soakaways may well be a result of the low infiltration capacity of the impermeable boulder clay soils at Loddington, resulting in more rapid runoff of septic tank

15 effluent and greater hydrological connectivity between the septic tank and the ditch/stream channel network. In contrast, previous work in the Chalk stream of the Hampshire Avon

(Jarvie et al., 2006, 2008) showed that, for well-drained Chalk soils, septic tank soakaways were highly efficient at removing SRP and leakage of effluent into the stream channel occurred only under high flows, when rising groundwater intercepted the soakaways during storm events. This marked difference in the nature and timing of septic tank wastewater inputs to the streams at Loddington compared with the Hampshire Avon reflects a fundamental difference in the effectiveness of permeable and impermeable soils to treat septic tank effluent effectively.

Septic tanks had largest impact on the fluxes of soluble inorganic P and N. We consider the visitor centre probably contributes the equivalent of 6 people, which together with the estimated 8 people living in nearby houses gives 14 people at VE, the same as at LN. The annual per capita fluxes for both sites can therefore be calculated as ca. 0.64 kg SRP and 14 kg TDN, although there is clearly some uncertainty in these values. Although the annual nutrient fluxes of P and N from the septic tanks were very similar at VE and LN, there was a large difference in flow–weighted concentrations between the two sites. At LN, the larger flow rates were better able to accommodate the increase in nutrient loading such that annual flow-weighted concentrations of the different fractions remained largely unchanged. The number of occasions when instantaneous concentrations of P and N were increased was also much lower at LN than at VE. Flow rates are clearly critically important for understanding and quantifying the potential ecological impacts of septic tank inputs in headwater streams.

Under low flow conditions, effluent inputs will lead to much greater ambient nutrient concentrations (e.g. Fig. 5) and therefore have much greater ecological impact. The high ammonium concentrations are indicative of very high levels of organic pollution, which can have severe consequences for aquatic organisms through the toxic effects of unionized

16 ammonia and oxygen depletion; invertebrates are particularly sensitive to oxygen depletion whereas fish are particularly sensitive to ammonia toxicity and oxygen stress (Jones et al.,

2008). Sewage fungus indicative of gross pollution by anthropogenic effluents that have not received secondary treatment (Curtis and Harrington, 1971) is regularly observed in the ditch at VE. Concentrations of nitrite-N were also an order of magnitude greater than is normally observed in streams (0.01 mg L-1), and are considered toxic to fish at these elevated concentrations (Stanley and Maxted, 2008). Discharges of SRP in concentrated effluent are rapidly adsorbed by bottom sediments in streams where they may become a source of nutrient for submerged macrophytes, and/or re-released under anoxic conditions to provide a nutrient source for stream benthic communities (Zanini et al., 1998; Palmer-Felgate et al., 2010;

Stutter et al., 2010). Further population growth in rural areas and climate change are likely to greatly exacerbate the ecological impacts of septic tank discharges in the future. Recent climate change scenarios and forecasts predict significantly lower river flows during spring, summer and autumn allowing longer residence times for nutrient uptake by aquatic biota and greater eutrophication impact (Johnson et al., 2009).

Conclusions

Our results provide new quantitative evidence that septic tanks are a real concern for water quality in rural areas during ecologically sensitive periods, especially where the effluent is discharged directly into the stream network. Monitoring of the nutrient composition of pipe discharge and lotic waters downstream of rural habitation in a typical English village has shown that septic tanks systems deliver variably high concentrations of highly bioavailable and/or potentially toxic nutrients to the stream network. Flow-concentration relationships in downstream sections and the high concentrations of SRP and NH4-N suggest that soakaways on the heavy clay soils at Loddington are not adequately retaining and treating septic tank

17 effluents. Consequently, the septic tanks are simply acting as direct point source inputs into headwater streams with very low dilution capacity causing degradation in water quality.

Failure to take account of nutrient emissions from septic tanks may undermine current attempts to improve the ecological status of freshwaters through controls over major sewage treatment works and agriculture. Although it was not possible in this study to ascertain the causes of system failure, or how ‘well-managed’ the septic tank systems were, the widespread distribution of these (mini) point sources suggest there is a need to raise awareness of the potential environmental problems associated with septic tanks amongst rural communities in the UK.

Acknowledgments

This study was funded by the Department for Environment, Food and Rural Affairs. John

Szczur of The Allerton Trust collected samples and monitored flow.

References

Ahmed, W., Neller, R., Katouli, M., 2005. Evidence of septic system failure determined by a

bacterial fingerprinting method. J. Appl. Microbiol. 98, 910-920.

Arnscheidt, H., Jordan, P., Li, S., McCormick, S., McFaul, R., McGrogan, H.J., Neal, M.,

Sims. J.T., 2007. Defining the sources of low-flow phosphorus transfers in complex

catchments. Sci. Total Environ. 382, 1–13.

Beal, C.D., Gardner, E.A., Menzies, N.W., 2005. Process, performance, and pollution

potential: A review of septic tank-soil absorption systems. Aust. J. Soil Res. 43,781-802.

Curtis, E.J.C., Harrington, D.W., 1971. The occurrence of sewage fungus in rivers in the

United Kingdom. Water Res. 5, 281-290.

Day, L., 2004. Septic systems as potential pollution sources in the Cannonsville Reservoir

Watershed, New York. J. Environ. Qual. 33, 1989–1996.

Gill, L.W., O’Luanaigh, N., Johnston, P.M., Misstear, B.D.R., O’ Súlleabháin, C., 2009.

Nutrient loading on subsoils from on-site wastewater effluent, comparing septic tank and

secondary treatment systems. Water Res. 43, 2739-2749.

Harder, H., 1970. Boron content of sediments as a tool in facies analysis. Sedimentary Geol.

4(1-2), 153-175.

Hooda, P.S., Moynagh, M., Svoboda, I., Miller, A., 2000. Macroinvertebrate as bioindicators

of water pollution in streams draining dairy farm catchments. Chem. Ecol.17, 17-30.

Jarvie, H.P., Neal, C., Withers, P.J.A., 2006. Sewage effluent phosphorus: a greater risk to

river eutrophication than agricultural phosphorus? Sci. Total Environ. 360, 246-253.

Jarvie, H.P., Withers, P.J.A., Bowes, M.J., Palmer-Felgate, E.J., Harper, D., Wasiak, K.,

Wasiak, P., Hodgkinson, R.A., Bates, A., Stoate, C., Neal, M., Wickham, H.D., Harman,

S.A., Armstrong, L.K., 2010. Streamwater phosphorus and nitrogen across a gradient in

rural-agricultural land use intensity. Agric., Ecosyst.Environ. 135, 238–252.

Jarvie, H.P., Withers, P.J.A., Hodgkinson, R., Bates, A., Neal, M., Wickham, H.D., Harman,

S.A., Armstrong. L., 2008. Influence of rural land use on streamwater nutrients and their

ecological significance. J. Hydrol. 350, 166–186.

Johnson, A.C., Acreman, M.C., Dunbar, M.J. et al., 2009. The British river of the future:

How climate change and human activity might affect two contrasting river ecosystems in

England. Sci. Total Environ. 407, 4787-4798.

Jones, J.I., Davy-Bowker, J., Murphy, J., Keller, V., Williams, R., Davies, C., 2008. Review

of the evidence for organic pollution thresholds to protect rivers with special designations

for wildlife. NERC/Centre for Ecology & Hydrology, 213pp. (Rep.to Natural England;

CEH Rep.No.: C03400/2008/1).

Jordan, P., Arnscheidt, A., McGrogan, H., McCormick, S., 2007. Characterising phosphorus

transfers in rural catchments using a bank-side analyser. Hydrol. Earth Syst. Sci. 11, 372-

381.

Kay, D., Crowther, J., Stapleton, C.M., Wyer, M.D., Fewtrell, L., Edwards, A., Francis, C.A.,

McDonald, A.T., Watkins, J., Wilkinson, J., 2008. Faecal indicator organism

concentrations in sewage and treated effluents. Water Res.42, 442-454.

Lapointe, B.E., O’Connell, J.D., Garrett, G.S., 1990. Nutrient couplings between on-site

sewage disposal systems, groundwaters, and nearshore surface waters of the Florida Keys.

Biogeochemistry 10, 289-307.

Leeks, G.J.L., Neal, C., Jarvie, H.P., Casey, H., 1997. The LOIS river monitoring network:

strategy and implementation. Sci. Total Environ. 194, 101–109.

Mara, D.D., 2004. Natural sewage treatment in the UK: Selection guidelines. Water Env. J.

18, 230-234.

Marshall, J.W., Winterbourn, M.J., 1979. An ecological study of a small New Zealand stream

with particular reference to the oligochaeta. Hydrobiol.65, 199-208.

May, L., Place, C., O’Malley, M.O., Spears, B., 2010. The impact of phosphorus inputs from

small discharges on designated freshwater sites. A Report to English Nature and Broads

Authority. 130pp.

Neal, C., Neal, M., Wickham, H., 2000. Phosphate measurement in natural waters: two

examples of analytical problems associated with silica interference using

phosphomolybdic acid methodologies. Science of the Total Environment 251, 511–522.

Neal, C., Jarvie, H.P., Withers, P.J.A., Whitton, B.A., Neal, M., 2010. The strategic

significance of wastewater sources to pollutant phosphorus levels in English rivers and to

environmental management for rural, agricultural and urban catchments. Sci. Total

Environ. 408, 1485-1500.

Palmer-Felgate, E.J., Jarvie, H.P., Withers, P.J.A., Mortimer, R.J.G., Krom, M.D., 2009.

Stream-bed phosphorus in paired catchments with different agricultural land use intensity.

Agric. Ecosyst. Environ. 134 (1-2), 53-66.

Palmer-Felgate, E.J., Mortimer, R.J.G., Krom, M.D., Jarvie, H.P., 2010. Impact of point-

source pollution on phosphorus and nitrogen cycling in stream-bed sediments. Environ.

Sci. Technol. 44, 908-914.

Reay, W.G., 2004. Septic tank impacts on ground water quality and nearshore sediment

nutrient flux. Ground Water 42, 1079-1089.

Redfield, A.C., 1958. The biological control of chemical factors in the environment.

American Scientist 46, 205–222.

Reneau, R.B. Jr., Hagedorn, C., Degen, M.J., 1989. Fate and transport of biological and

inorganic contaminants from on-site disposal of domestic wastewater. J. Environ. Qual.

18, 135-144.

Robertson, W.D., 2008. Irreversible phosphorus sorption in septic system plumes? Ground

Water 46, 51-60.

Soil Survey of England and Wales, 1983. Soil Map of England and Wales, Scale 1:250,000.

Soil Survey of England and Wales, Harpenden, UK.

Stanley, E.H., Maxted, J.T., 2008. Changes in the dissolved nitrogen pool across land cover

gradients in Wisconsin streams. Ecol. Appl.18, 1579-1590.

Steffy, L.Y., Kilham, S.S., 2004. Elevated δ15N in stream biota in areas with septic tank

systems in an urban watershed. Ecol. Appl.14, 637-641.

Stutter, M.I., Demars, B.O.L., Langan, S.J., 2010. River phosphorus cycling: separating

biotic and abiotic uptake during short-term changes in sewage effluent loading. Water Res.

44, 4425-4436.

Webb, B.W., Phillips J.M., Walling D.E., Littlewood I.G., Watts C.D., Leeks G.J.L., 1997.

Load estimation methodologies for British rivers and their relevance to the LOIS

RACS(R) programme. Sc. Total Environ. 194/195, 379-389.

Weiskel, P.K., Howes, B.L., 1992. Differential transport of sewage-derived nitrogen and

phosphorus through a coastal watershed. Environ. Sci. Tech. 26, 352-360.

Whelan, B.R., Titamnis, Z.V.,1982. Daily chemical variability of domestic septic tank

effluent. Water Air Soil Pollut.17, 131-139.

Wilhelm, S.R., Schiff, S.L., Cherry, J.A., 1994. Biogeochemical evolution of domestic waste-

water in septic systems. 1. Conceptual-Model. Ground Water 32, 905-916.

Withers, P.J.A., Jarvie, H.P., Hodgkinson, R.A., Palmer-Felgate, E.J., Bates, A., Neal, M.,

Howells, R., Withers, C.M., Whickham, H., 2009. Characterization of phosphorus sources

in rural watersheds. J. Environ. Qual. 38, 1998-2011.

Zanini, L., Robertson, W.D., Ptacek, S C.J., Schiff, L., Mayer, T., 1998. Phosphorus

characterization in sediments impacted by septic tank effluents at four sites in central

Canada. J. Contaminant Hydrol. 33, 405–429.

Legends to Tables

Table 1. Distribution statistics for nutrient concentrations in the septic tank effluent discharge at Whitehorse Creek (n = 37).

Table 2. Correlation coefficients (r) for the relationships between soluble reactive P (SRP) and enriched nutrients in the pipe discharge at Whitehorse Creek, and upstream and downstream monitoring points at Village East and Loddington North.

Table 3. Distribution statistics for nutrient concentrations in ditch water upstream and downstream of rural habitation at the Village East site (n = 46).

Table 4. Distribution statistics for nutrient concentrations in stream water upstream and downstream of rural habitation at the Loddington North site (n = 46).

Table 5. Annual flux (kg) of N and P fractions at Whitehorse Creek, Village East and

Loddington North monitoring sites. Flow-weighted concentrations (FWC, mg L-1) are given in parenthesis.

Table 1

Determinand Range Median Mean (s.d.) pH (units) 6.9 – 7.9 7.3 7.3 (0.20) Alkalinity (µEq. L-1) 2205 - 10666 6085 6199 (2064) DOC (mg L-1) 3.8 – 28.5 10.9 11.1 (5.0)

SRP (µg L-1) 68 - 8920 2470 3206 (2614) DHP (µg L-1) 0 - 5780 290 693 (1218) PP (µg L-1) 55 - 6760 280 708 (1190) TP (µg L-1) 195 - 20560 3100 4606 (4279)

-1 NH4N (mg L ) 0.03 – 72.4 15.6 17.5 (15.5) -1 NO3N (mg L ) 0.06 – 10.4 4.1 4.2 (3.2) -1 NO2N (mg L ) 0.02 – 3.1 0.15 0.34(0.55) DON (mg L-1) 0 – 13.9 2.5 3.4 (3.4) TDN (mg L-1) 8.1-63 23.1 24.6 (13.0)

-1 SO4S (mg L ) 11 - 42 31 29 (9) Cl (mg L-1) 27 - 101 54 54 (16) Na (mg L-1) 13 - 119 43 49 (25) K (mg L-1) 9 - 40 27 26 (8) Ca (mg L-1) 45 - 139 96 96 (22) Mg (mg L-1) 5 - 18 14 12 (4) B (µg L-1) 54 - 213 103 109 (33) Mn (µg L-1) 11 - 169 52 68 (46) Fe (µg L-1) 36 - 267 101 123 (64)

Table 2

Site NH4N Na K Cl SO4-S B Mn DOC

WC - pipe 0.81 0.84 NS 0.68 0.50 NS 0.44 0.63

V - upstream NS NS 0.67 0.30 NS NS NS NS - 0.89 0.62 NS 0.51 0.40 0.33 0.63 NS downstream

LN - upstream NS NS 0.58 0.42 NS NS NS 0.41 - 0.94 0.88 0.82 0.64 0.33 0.67 0.85 NS downstream

NS – not significant (P >0.05)

Table 3 Determinand Upstream Downstream Difference Range Median Mean (s.d.) Range Median Mean (s.d.) (Down – up) pH (units) 7.0 – 7.9 7.6 7.5 (0.21) 7.3 – 7.8 7.5 7.5 (0.11) NS Alkalinity (µEq. L-1) 1428 - 5137 2781 2792 (880) 2257 - 9847 5086 4895 (1473) *** DOC (mg L-1) 5.9 – 19.1 9.0 9.4 (2.6) 2.3 – 14.1 5.1 5.9 ***

SRP (µg L-1) 12 - 494 59 78 (76) 83 - 895 239 321 (205) *** DHP (µg L-1) 10 - 157 49 57 (37) 0 - 445 72 81 (68) ** PP (µg L-1) 2 - 1135 83 117 (177) 4 - 1208 39 100 (209) NS TP (µg L-1) 66 - 1516 185 252 (237) 139 – 2028 400 502 (348) ***

-1 NH4N (mg L ) >0.01 - 0.31 0.03 0.05 (0.05) 0.06 -4.82 0.57 0.99 (1.14) *** -1 NO3N (mg L ) <0.1 - 13.6 3.3 3.9 (2.8) 2.3 – 12.8 6.7 6.7 (2.3) *** -1 NO2N (mg L ) 0.02 - 0.09 0.02 0.02 (0.015) 0.02 – 1.06 0.09 0.16 (0.20) *** DON (mg L-1) 0.15 - 15.7 0.90 1.37 (2.30) 0 – 11.8 1.17 1.50 (1.85) NS TDN (mg L-1) 0.6 - 20.4 4.3 5.3 (3.9) 4.5 – 21.6 8.8 9.3 (3.0) ***

-1 SO4S (mg L ) 4 - 22 13 14 (3.8) 8 - 34 25 23 (6.9) *** Cl (mg L-1) 4 - 18 12 11 (3.1) 9 - 47 24 23 (6.6) *** Na (mg L-1) 4.8 – 11.3 7.5 7.5 (1.5) 7.9 – 30.4 16.7 16.6 (4.8) *** K (mg L-1) 2.0 – 8.3 3.5 3.9 (1.5) 8.0 – 26.8 19.8 18.5 (5.5) *** Ca (mg L-1) 41 - 107 71 71 (16) 48 - 139 114 108 (24) *** Mg (mg L-1) 3.5 – 7.6 5.9 5.8 (1.0) 4.4 – 12.5 10.0 9.4 (2.3) *** B (µg L-1) 35 - 72 58 56 (10) 53 - 159 109 109 (28) *** Mn (µg L-1) 3 - 317 21 57 (74) 19 - 163 48 58 (33) NS Fe (µg L-1) 16 - 538 75 130 (125) 7 - 644 35 90 (131) * 1Statistical significance of the difference in upstream and downstream mean values: *** 0.001, ** 0.01, * 0.05, NS not significant.

Table 4 Determinand Upstream Downstream Difference1 Range Median Mean (s.d.) Range Median Mean (s.d.) (Down – up) pH (units) 7.3 – 8.1 7.8 7.8 (0.17) 7.4 – 8.0 7.8 7.8 (0.14) NS Alkalinity (µEq. L-1) 1885 - 5341 3241 3348 (902) 2207 - 6139 3320 3621 (969) ** DOC (mg L-1) 1.4 – 21.6 8.1 8.1 (3.5) 4.1 – 14.8 7.3 7.5 (2.4) NS

SRP (µg L-1) 4 - 226 31 39 (39) 8 - 1925 91 213 (342) *** DHP (µg L-1) 3 - 120 30 41 (30) 1 - 710 55 73 (105) * PP (µg L-1) 5 - 1830 35 106 (286) 10 - 898 70 124 (176) NS TP (µg L-1) 32 - 2176 102 186 (334) 52 - 2620 212 410 (543) **

-1 NH4N (mg L ) <0.01 – 0.17 0.02 0.03 (0.03) <0.01 – 15.6 0.19 1.32 (2.97) ** -1 NO3N (mg L ) 0.11 – 16.7 4.1 4.8 (3.5) 0.7 – 14.7 4.0 4.1 (2.7) NS -1 NO2N (mg L ) 0.015 – 0.03 0.015 0.017 (0.005) 0.015 – 0.58 0.015 0.068 (0.11 0 ** DON (mg L-1) 0.09 – 11.1 0.86 1.08 (1.57) 0 – 5.5 0.89 1.20 (1.17) NS TDN (mg L-1) 0.7 -19.0 4.7 5.9 (4.1) 1.7 – 21.6 5.7 6.7 (4.2) NS

-1 SO4S (mg L ) 14 - 47 25 26 (8.5) 13 - 44 25 26 (6.9) NS Cl (mg L-1) 11 - 55 24 22 (7.6) 13 - 67 23 24 (9.5) NS Na (mg L-1) 6.7 – 18.6 13.4 12.7 (3.0) 8.6 - 57 12.9 15.6 (8.9) * K (mg L-1) 1.4 – 12.9 3.1 3.4 (1.8) 1.9 – 14.0 3.4 4.3 (2.3) * Ca (mg L-1) 49 - 139 88 91 (23) 49 -126 94 92 (18) NS Mg (mg L-1) 3.8 – 16.1 9.5 9.8 (2.9) 5.3 – 13.5 9.2 9.0 (1.9) ** B (µg L-1) 29 - 77 54 52 (14) 30 - 108 57 56 (16) * Mn (µg L-1) 2 - 89 10 12 (13) 6 -179 14 29 (36) ** Fe (µg L-1) 1 - 999 44 106 (168) 8 - 455 49 90 (97) NS 1Statistical significance of the difference in upstream and downstream mean values: *** 0.001, ** 0.01, * 0.05, NS not significant

Table 5

Determinand Whitehorse Village Loddington North Creek (pipe) Upstream Downstream Down-Up Upstream Downstream Down-Up

SRP 10.2 (1.07) 4.1 (0.07) 13.7 (0.21) 9.6 (0.14) 11.7 (0.06) 19.4 (0.08) 7.7 (0.02) DHP 1.8 (0.18) 3.6 (0.06) 5.8 (0.09) 2.2 (0.03) 11.1 (0.06) 14.0 (0.05) 2.9 (-0.01) PP 2.8 (0.29) 12.0 (0.21) 9.7 (0.15) -2.3 (-0.06) 45.8 (0.23) 39.1 (0.15) -6.6 (-0.08) TP 14.8 (1.54) 19.7 (0.34) 29.2 (0.44) 9.5 (0.10) 68.6 (0.34) 72.8 (0.28) 4.2 (-0.06)

NH4N 49 (5.08) 2 (0.04) 37 (0.55) 35 (0.51) 7 (0.04) 54 (0.21) 47 (0.17) NO3N 60 (6.25) 301 (5.20) 458 (6.88) 157 (1.68) 812 (4.0) 959 (3.70) 147 (-0.30) NO2N 2 (0.19) 1 (0.02) 6 (0.08) 5 (0.06) 3 (0.02) 7 (0.03) 3 (0.01) DON 20 (2.11) 85 (1.48) 99 (1.49) 14 (0.01) 235 (1.16) 272 (1.05) 37 (-0.11) TDN 131 (13.52) 389 (6.73) 599 (9.00) 210 (2.27) 1055 (5.21) 1294 (4.99) 238 (-0.22)

Legends to Figures

Figure 1. Loddington village in Leicestershire showing monitoring points in relation to rural habitation. The monitoring station at the outlet of the Belton Bridge catchment monitored by

Jarvie et al. (2010) is also shown. Arrows indicate the direction of flow.

Figure 2. Temporal variability in (a) monthly rainfall in relation to the long-term average

(1961-2007) at Loddington and (b) measured flow and concentrations of soluble reactive P

(SRP) in the septic tank effluent discharge at Whitehorse Creek over the monitoring period.

The typical relationship between flow and nutrient concentrations at Whitehorse Creek is illustrated for Manganese (Mn) in (c).

Figure 3. Relationships between (a) ammonium N (NH4N) and soluble reactive P (SRP), and

(b) sodium (Na) and boron (B) concentrations for Whitehorse Creek (WC), and downstream stations at Village East (VE) and Loddington North (LN). Data points for Na and B corresponding to the four dates when nutrient concentrations increased at the Loddington

North downstream station are highlighted.

Figure 4. Temporal variability in the concentrations of soluble reactive P (SRP), ammonium

N (NH4N) and sodium (Na) at upstream and downstream monitoring stations at the Village

East and Lodington North sites. Note the difference in the y axis scales.

Figure 5. Effects of flow on the concentrations of soluble reactive P (SRP), ammonium N

(NH4N) and sodium (Na) at upstream and downstream stations at the Village East and

Loddington North sites. Note the difference in the y axis scales.

Fig. 1.

Loddington North

Visitor centre

Loddington

village Village East

Monitoring station

Houses

Belton Bridge catchment outlet Whitehorse Creek

Fig. 2

140 (a)

120

100

80 60

Rainfall(mm) 40

2006 2007 2008

(b) 10 2.5 SRP 9

Flow

8 2

)

1 -

7 s

6 1.5 L (mg L (mg

5 Flow ( Flow

SRP 4 1 3

2 0.5

1 0 0 Feb-07 Apr-07 Jun-07 Aug-07 Oct-07 Dec-07 Feb-08

160

)

1 140

- 120

100

L (µg Mn 80 y = 21.43x-0.548 60 r² = 0.76 40 20

0 0 0.5 1 1.5 2 2.5 -1 Flow30 (L s )

Fig. 3

80 (a)

)

1 - 60

N (mg L (mg N 4 40

NH 30

0 0 2 4 6 8 10

SRP (mg L-1)

140 (b)

120 WC

) 10/10/06

1 - 100 VE

80 11/06/07 LN

L (mg Na 60 17/09/07

40 06/11/06

0 50 100 150 200 250

B (µg L-1)

Fig. 4

Village East Loddington North 1.0 2.0

0.9 1.8

1 0.8 1.6 - L 0.7 1.4 0.6 1.2 SRP 0.5 1.0 mg mg 0.4 0.8 0.3 0.6 0.2 0.4

0.1 0.2 0.0 0.0 Oct Dec Feb Apr Jun Aug Oct Oct Dec Feb Apr Jun Aug Oct

6 16

5 1 - Upstream

L 12

4 4 Downstream 10

g 3 8 m 6 2 4 1 2

0 0 Oct Dec Feb Apr Jun Aug Oct Oct Dec Feb Apr Jun Aug Oct

35 60

1 - 25 L 40

Na Na 20 g g 30

m 15 20 10

5 10

0 0 Oct Dec Feb Apr Jun Aug Oct Oct Dec Feb Apr Jun Aug Oct

Fig. 5 Village East Loddington North

0.6 11 0.250.25 2.5

)

- 0.5 ) 0.80.8 0.20.2 2 1 Upstream Downstream Upstream Downstream - 0.4 0.150.15 mg L mg 0.60.6 1.5

mg L mg 0.3 0.40.4 0.10.1 1 0.2 SRP(

SRP( 0.050.05 0.1 0.20.2 0.5 0 00 00 0 0 0.01 0.02 0.03 00 0.010.01 0.020.02 0.030.03 00 0.020.020.040.040.060.060.080.080.10.1 0 0.02 0.04 0.06 0.08 0.1

0.25

) 0.35 6 20 1 - L 0.3 5 0.2 0.25 15

mg 4 0.2 0.15

N ( N 3 10

4 0.15 0.1 2 NH 0.1 5 1 0.05 0.05 0 0 0 0 0 0.01 0.02 0.03 0 0.01 0.02 0.03 0 0.02 0.04 0.06 0.08 0.1 0 0.02 0.04 0.06 0.08 0.1

12 35 20 60

) 1 - 10 30 50 15 8 25 40 mg L mg 20 6 10 30 15 Na ( Na 4 20 10 5 2 5 10

0 0 0 0

0 0.01 0.02 0.03 0 0.01 0.02 0.03 0 0.02 0.04 0.06 0.08 0.1 0 0.02 0.04 0.06 0.08 0.1 3 -1 Flow (m33 s )