Nitrate concentrations in river waters of the upper Thames and its tributaries

Science of the Total Environment 365 (2006) 15 – 32 www.elsevier.com/locate/scitotenv

Nitrate concentrations in river waters of the upper Thames and its tributaries Colin Neal ⁎, Helen P. Jarvie, Margaret Neal, Linda Hill, Heather Wickham Centre for Ecology and Hydrology, Maclean Building, Crowmarsh Gifford, Wallingford, OXON, OX10 8BB, UK Available online 17 April 2006

Abstract The spatial and temporal patterns of in-stream nitrate concentrations for the upper Thames and selected tributaries are described in relation to point and diffuse sources for these rural catchments. The rivers associated with catchments dominated by permeable (Cretaceous Chalk) bedrock show a smaller range in nitrate concentrations than those associated with clay and mixed sedimentary bedrock of lower permeability. The differences reflect the contrasting nature of water storage within the catchments and the influence of point and diffuse sources of nitrate. Nitrate concentrations often increase in a gradual way as a function of flow for the rivers draining the permeable catchments, although there is usually a minor dip in nitrate concentrations at low to intermediate flow due to (1) within-river uptake of nitrate during the spring and the summer when biological activity is particularly high and (2) a seasonal fall in the water table and a change in preferential flow-pathway in the Chalk. There is also a decrease in the average nitrate concentration downstream for the Kennet where average concentrations decrease from around 35 to 25 mg NO3 l− 1. For the lower permeability catchments, when point source inputs are not of major significance, nitrate concentrations in the rivers increase strongly with increasing flow and level off and in some cases then decline at higher flows. When point source inputs are important, the initial increase in nitrate concentrations do not always occur and there can even be an initial dilution, since the dilution of point sources of nitrate will be lowest under low-flow conditions. For the only two tributaries of the Thames which we have monitored for over 5 years (the Pang and the Kennet), nitrate concentrations have increased over time. For the main stem of the Thames, which was also monitored for over 5 years, there is no clear increase over time. As the Pang and the Kennet river water is mainly supplied from the Chalk, the increasing nitrate concentrations over time clearly reflect increasing nitrate concentrations within the groundwater. It primarily reflects long-term trends for agricultural fertilizer inputs and significant aquifer storage and long water residence times. The results are discussed in terms of hydrogeochemical processes and the Water Framework Directive and are compared with data from other eastern UK rivers. The importance of diffuse sources of nitrate contamination is highlighted. On a flow weighted basis, the average diffuse component of nitrate is around 95% for the Thames Basin rivers draining Chalk and for the corresponding rivers draining less permeable strata, there is a more significant but not major point source component (at least in terms of flux); the average diffuse component is 79% in this case. These data fit well with earlier assessments of agricultural sources to UK surface waters. Under baseflow conditions the diffuse sources remain dominant for the Chalk fed Thames Basin rivers, but point sources can be dominant for the low permeability cases. On a proportionate basis, the Thames Basin rivers are similar to the rural rivers of the Tweed and Humber Basins in terms of percentage diffuse components although the lower intensity agriculture occurring for the rivers monitored means that the average nitrate concentrations are lower for the rural rivers of central and northern England and the borders with Scotland: the Humber and Tweed. © 2006 Elsevier B.V. All rights reserved. Keywords: Nitrate; RELU; River; Chalk; Aquifer; Cherwell; Dun; Kennet; Lambourn; LOCAR; LOIS; Pang; Ray; Thame; Thames; Water Framework Directive ⁎ Corresponding author. E-mail address: [email protected] (C. Neal). 0048-9697/$ - see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.scitotenv.2006.02.031

16

C. Neal et al. / Science of the Total Environment 365 (2006) 15–32

1. Introduction The maintenance of good quality river water and the clean-up of polluted water are of major European concern. A major impetus for improving the riverine environment has been the introduction of the Water Framework Directive (WFD) that demands new approaches for managing and improving surface and groundwater quality across the European Union, with emphasis shifting from chemical towards ecological water quality standards (CEC, 2000). There is also a need to improve the water quality of groundwater under a proposed daughter directive (CEC, 2003). For rural and agricultural areas, there are major issues linked to water resources in terms of potability and amenity value. Major emphasis is now put on nutrient inputs to rivers in rural and agricultural areas in relation to both eutrophication, where nitrogen and phosphorus are an issue, and to potability, where waters high in nitrate are of concern (DoE, 1993; Dwyer et al., 2002; Defra, 2004a,b; EA, 2000; RPA, 2003; Withers and Lord, 2002). Indeed, there is an increasing impetus for environmental modelling and management within the context of the WFD that is generating considerable research interest and dynamics (Heathwaite et al., 2005; Wade et al., in press). In this paper, the variations in one of the key nutrient components associated with both agricultural and sewage sources, nitrate, is examined for a key rural/agricultural typology in the south east of England, the upper Thames Basin. Typically, around 80% of its area comprises arable, horticultural and grassland. This basin is important in terms of its size and economic and social relevance for the UK with regards to the WFD. This is because there are significant diffuse agricultural sources of nutrients and major pressures for the urban development. The region is essentially ‘green belt’ with increasing point-source nutrient inputs (Evans et al., 2003). The paper brings together a wide range of data from seven catchment studies for both permeable and less permeable catchments (Neal et al., 2000a,b,c, 2004a,b, 2005a,b, in press-a,b). Here the term ‘less permeable’ is used to denote that some of the catchments, while having low permeability geology (clays in particular), usually are of mixed geology with some permeable zones that can maintain baseflow. N.B. some of the permeable catchments also have a small component of low permeability strata and these can occasionally affect high-flow extremes. This study complements published material that relates specifically to the other main nutrient of concern, phosphorus (Jarvie et al., in press; Neal et al., 2005a). It adds to the wide ranging studies on the hydrology, water quality and

ecological functioning of the upper Thames and to Europe-wide research for rural and agriculturally impacted catchments (Neal, 2002; Neal and Whitehead, 2002; Heathwaite et al., 2005). The work provides an important empirical/descriptive base for modelling work both within the Thames region (e.g. Flynn et al., 2002; Whitehead et al., 2002b) as part of programmes such as LOCAR (Wade et al., in press) and European-wide initiatives within, for example the INCA modelling framework (Neal and Whitehead, 2002; Wade et al., 2005). The paper is also important in relation to the WFD and assessment of point versus diffuse sources: high biological activity occurs when flows are generally low and point source inputs are particularly important. It has been estimated that over 70% of nitrates in English waters originate from agricultural land (Defra, 2004b). The work links the Land Ocean Interaction Study of the 1990s (LOIS: Leeks and Jarvie, 1998), and its integration within a socioeconomic context (EUROCAT: Cave et al., 2003), to issues of sustainability of UK farming within the Rural Economy Land Use programme (RELU: www.relu.ac.uk). 2. Study area and background material This study concerns the main stem and four tributaries of the upper River Thames, the Cherwell, Kennet, Pang and Thame, as well as inputs to two of these tributaries: the Dun and Lambourn which join the Kennet and the Ray which joins the Cherwell. The locations of these tributaries and their relation to the Thames are shown in Fig. 1. For all the tributaries, their catchments are either largely associated with drainage from permeable (Cretaceous Chalk) aquifers (Dun, Kennet, Pang and Lambourn) or with low permeability sedimentary rocks mainly of Jurassic age (Cherwell, Ray and Thame). The upper Thames catchment comprises a mix of both high and low permeability strata. Details of the general water quality and the location of individual sites are outlined in earlier publications (Neal et al., 2000a,b,c, 2004a,b, 2005b, in press-a,b). Table 1 provides information on the hydrogeology and the land use for parts of each catchment studied. The salient features concerning sampling locations are: • The Thames. The main stem of the Thames has been monitored, beginning in 1997, at one site near Howberry Park, just upstream of Wallingford (Neal et al., 2000a). The catchment area at this location is approximately 3500 km2. • The Pang drains to the Thames at the town of Pangbourne and the catchment area is approximately

C. Neal et al. / Science of the Total Environment 365 (2006) 15–32

17

village of Islip. These sites together with the Ray have been monitored since 2000 (Neal et al., in press-a). ○ The Ray drains approximately 30% of the Cherwell catchment (catchment area approximately 284 km2) and it has been monitored at one point within the village of Islip (Neal et al., in press-a). • The Thame drains to the Thames near the village of Dorchester, upstream of the town of Wallingford and the Howberry Park monitoring site of the Thames. In the upper part of the catchment is the town of Aylesbury and the catchment area is approximately 800 km2. Three sites have been monitored on the Thame. Two of these sites are just upstream of Aylesbury (approximately 1.2 km upstream of Aylesbury STW) and 5 km downstream of Aylesbury near the village of Cuddington, while the other site is about half way along its length near the village of Wheatley (Neal et al., in press-b). Monitoring began in 1998 for the Thame at Wheatley and in 2002 for the other two sites. Fig. 1. Location map for the catchment studies in the upper Thames Basin. The letters denote the location of the major towns and cities for the catchments studied, as listed in the key to the base of this figure.

171 km2. Six sites on the Pang have been monitored at various times since 1997 together with a spring discharge that provides significant flow during baseflow to the lower Pang, the Blue Pool (Neal et al., 2000b, 2004b). • The Kennet drains to the Thames at Reading and the catchment area is approximately 1200 km2. The upper half of the Kennet has been monitored at 10 sites since 1997 (Neal et al., 2000c) ○ The Dun drains into the Kennet near Hungerford. The catchment area is approximately 110 km2. There have been a variety of reservoir, canal (Kennet and Avon Canal) and river sites monitored since 2000 (Neal et al., 2005b). For this study, only the river sites are considered, for which there are four, as well as two stream inputs to the Dun, the Froxfield and the Shalbourne streams. 2 ○ The Lambourn has an area of approximately 250 km down to its confluence with the Kennet at the town of Newbury. Three sites have been monitored since 2002 (Neal et al., 2004b). • The Cherwell drains to the Thames at Oxford and it has an area of approximately 943 km2. There are 3 monitoring points on the river; the uppermost is in the market town of Banbury, about a kilometre upstream of Banbury sewage treatment works (STW), the second site is about 5 km downstream of Banbury STW and the lowest site is just upstream of the confluence with its main tributary, the Ray, near the

Monitoring across all the study sites has taken place over a variety of sampling periods (Table 2) according to differing research needs and funding restrictions. The monitoring has largely been on a weekly basis, but monitoring ceased at several of the sites between the 28th February 2001 and the 13th June 2001 because of lack of access as a result of a Foot and Mouth outbreak. Water samples were filtered in the field through 0.45 μM membranes and the nitrate concentrations were then determined on return to the Wallingford laboratories by ion chromatography using a DIONEX system. Within these monitoring programmes, a wide variety of major (Na, K, Mg, Ca, Cl, SO4) and minor/ trace elements (B, Ba, Sr, first row transition metals), and nutrients (N and P species, Si) together with general water quality indices (pH, electrical conductivity and alkalinity) were analysed for: details of methodologies and general water quality can be found in Neal et al. (2000a,b,c, 2004a, b, 2005b, in press-a,b). Flow information used within this study was obtained for key monitoring sites based on a network of gauging sites (Marsh and Lees, 2003). For these sites, there is also a measure of degree of inputs from groundwater sources using a measure the baseflow index (BFI; Gustard et al., 1992). The BFI values typically range between 1 for highly permeable systems and 0.2 for low permeable catchments. For rivers predominantly supplied from aquifers such as the Chalk have BFIs typically in the range 0.8 to 1 and usually close to the upper limit of 1. Table 2 provides information on the BFI for the catchments studied in this paper. N.b. the baseflow index (BFI) given in Table 2 represents the value where available for some part of the catchment. The location of the gauging site on

18

C. Neal et al. / Science of the Total Environment 365 (2006) 15–32

Table 1 Catchment details Moderate High Moderate permeability permeability permeability (fiss) (fiss) (intergran)

Mixed Very low Wood Arable and Grassland Mount. permeability permeability land horticultural Heath and Bog

Hydrogeology Thames Days Lock

Built- Inland up water areas

Land use

17.7

0.8

0.4

0.0

53.4

10.8

46.5

34.1

0.8

7.2

0.6

Kennet Marlborough Hungerford Theale

0.0 0.0 0.0

0.0 0.0 0.0

0.0 8.4 2.3

0.0 9.9 12.2

0.0 0.0 13.9

6.8 24.9 15.0

55.8 48.9 45.3

31.9 21.9 31.5

3.5 1.7 3.1

2.0 2.6 5.0

0.0 0.0 0.2

Pang Frilsham Pangbourne

0.0 0.0

0.0 0.0

0.0 0.0

2.6 17.1

0.0 7

7.3 17.6

57.2 45.5

27.7 28.2

5.7 4.3

2.0 4.4

0.1 0.0

Lambourn East Sheff. Shaw

0.0 0.0

0.0 0.0

0.0 0.0

0.0 2.7

0.0 0.0

8.4 10.3

50.7 53.8

35.1 30.2

3.4 3.1

2.0 2.3

0.3 0.3

18.6 7.9

0.0 0.0

0.0 0.0

0.0 0.0

81.4 92.1

9.1 7.8

50.4 50.6

33.4 35.3

0.4 0.4

6.4 5.6

0.2 0.3

6.4

0.0

0.0

0.0

93.6

16.9

37.2

41.1

0.0

4.6

0.2

Thame Shabbington 16.7 Wheatley 21.4

1.4 1.7

3.1 2.6

0.0 0.0

62.2 60.6

8.1 9.7

37.5 36.2

45.3 45.5

0.1 0.1

8.9 8.3

0.2 0.2

Cherwell Enslow Mill Banbury Ray Gr. Und.

The hydrogeological information was derived from British Geological Survey Datasets, while the land use statistics derived from the CEH Land Cover Map for 2000. fiss = fissure; intergran = intergranular; Gr. Und. = Grendon Underwood.

which the BFI has been assessed this can be near the top, intermediate or lower parts of the catchment. The BFI data in Table 2 is given for the nearest location to the monitoring point. Table 2 also provides a detailed breakdown of agricultural land use. 3. Results 3.1. General comments Nitrate concentrations across the catchments are very similar (Table 2). Thus, for the permeable sub-catchments of the Thames (the Dun, Kennet, Lambourn and Pang), the mean nitrate concentration is around 28.5 mg NO3 l− 1 with a range in mean nitrate concentration across the sites of 21.9 to 35.6 mg NO3 l− 1; mean minima and maxima are 20 and 35.7 mg NO3 l− 1, respectively. Correspondingly, for the lower permeability catchments (Cherwell, Ray and Thame) the mean nitrate concentration is around 29.2 mg NO3 l− 1, with a range in mean nitrate concentration across the sites of 21.1 to 35.6 mg NO3 l− 1: mean

minima and maxima are 15.2 and 55.2 mg NO3 l− 1, respectively. There is some pattern to the degree of nitrate concentration variation for the different sites as examination of Table 2 reveals. Thus, for the less permeable catchments there is a larger maximum, range, standard deviation and standard deviation normalized to the mean values. This is illustrated using linear regression for all the sites (excluding the Blue Pool spring site) which reveals no statistically significant relationship with the BFI for the mean (p N 0.10) and minimum (p N 0.02) concentrations, but statistically significant relationships for the BFI with the maximum, standard deviation and standard deviation normalized to the mean values (p all less than 0.001). For example, Maximum nitrate concentration ¼ 90:8−59:2⁎BFI

ð1Þ

where N = 33 and r2 = 0.713. Further, in the case of the Blue Pool (BFI = 1), which is spring fed, the degree of variation in nitrate

C. Neal et al. / Science of the Total Environment 365 (2006) 15–32

19

Table 2 Nitrate concentrations (mg NO3 l− 1) in the Thames basin Mean

Minimum

Maximum

35.6 30.8 34.6 35.6 31.0 29.9 31.3

28.5 22.0 18.5 11.5 25.6 20.0 28.2

46.0 37.2 42.0 45.5 37.6 41.6 34.4

Std. Dev.

Std/avg %

BFI

4.9 4.4 3.8 5.9 3.4 3.8 1.4

13.7 14.3 10.9 16.5 10.8 12.5 4.5

0.86

Pang Frilsham Upstream Blue Pool Downstream Blue Pool Bucklebury Bradfield Tidmarsh Blue Pool

10/04/2002 22/07/1997 10/04/2002 10/04/2002 22/07/1997 22/07/1997 22/07/1997

Lambourn E. Shefford Boxford Shaw

10/04/2002 to 5/11/2003 15/05/2002 to 5/11/2003 10/04/2002 to 5/11/2003

32.2 31.7 30.8

21.6 28.4 24.0

39.0 36.5 38.0

2.9 1.9 2.9

9.1 5.9 9.4

0.97

Kennet Clatford Glebe House Mildenhall Stitchcombe Axford Ramsbury Knighton Hungerford Kintbury Woolhampton

3/06/1997 to 9/03/2005 3/06/1997 to 28/02/2001 3/06/1997 to 9/03/2005 3/06/1997 to 28/02/2001 3/06/1997 to 28/02/2001 3/06/1997 to 28/02/2001 3/06/1997 to 28/02/2001 2/10/2000 to 23/07/2003 2/10/2000 to 2/01/2002 10/04/2002 to 13/08/2003

34.4 28.9 30.6 28.1 27.7 25.4 24.7 26.3 24.1 24.4

21.2 21.6 21.5 21.4 20.6 13.0 13.8 19.2 18.4 18.0

43.0 36.8 38.5 36.0 36.0 36.4 33.6 34.5 31.2 31.5

4.0 3.1 3.2 3.3 3.2 4.5 4.2 3.4 2.9 3.6

11.5 10.8 10.5 11.9 11.7 17.8 16.9 12.9 12.0 14.9

0.95

Dun Fore Bridge Great Bedwyn Dun Cottage Hungerford Froxfield stream Shalbourne stream

02/10/2000 02/10/2000 02/10/2000 02/10/2000 02/10/2000 02/10/2000

25.3 25.9 24.3 24.2 25.6 21.9

19.2 18.0 17.2 17.2 20.8 18.4

28.8 32.0 27.6 29.0 28.8 25.2

2.4 2.7 2.2 2.2 1.9 1.9

9.5 10.4 9.1 9.0 7.3 8.7

0.95

Pang/Lambourn/Dun/Kennet Pang/Lambourn/Dun/Kennet: combined data

28.5

11.5

46.5

3.3

11.5

0.93

Cherwell/Ray/Thame Cherwell Banbury Cherwell Kings Sutton Cherwell Islip Ray Islip Thame Aylesbury Thame Cuddington Thame Wheatley

26.2 35.1 31.1 32.0 21.1 29.4 32.9

9.8 17.8 23.4 12.4 6.0 17.6 19.2

51.0 69.6 50.4 54.6 57.0 50.8 53.0

10.7 9.4 6.5 8.2 11.3 7.4 7.0

40.9 26.7 20.9 25.5 53.5 25.1 21.2

0.65

Cherwell/Islip/Thame Cherwell/Islip/Thame: combined data

29.7

6.0

69.6

8.6

30.5

0.60

Thames Howberry Park

35.0

18.6

57.0

6.8

19.5

0.64

to 5/11/2003 to 20/09/2000 to 5/11/2003 to 25/06/2003 to 20/09/2000 to 5/11/2003 to 20/09/2000

to 2/01/2002 to 2/01/2002 to 2/01/2002 to 23/07/2003 to 2/01/2002 to 2/01/2002

26/07/2000 to 11/12/2001 26/07/2000 to 11/12/2001 26/07/2000 to 11/12/2001 26/07/2000 to 11/12/2001 26/07/2000 to 11/12/2001 26/07/2000 to 11/12/2001 3/06/1998 to 11/12/2001

9/04/1997 to 31/07/2002

1.00

0.92

0.54

The italics represent (a) the Blue Pool which is a spring discharge and (b) the Froxfield and Shalbourne, two tributaries of the Dun.

concentration is particularly small (e.g. range 28.2 to 34.4 mg NO3 l− 1). These results show that the more permeable the system the greater the degree of dampening and the lesser

the peak nitrate concentrations, even though the means are very similar across the sites. In the case of the longest record for two sites on the same river (the Kennet at Clatford and Mildenhall) there

20

C. Neal et al. / Science of the Total Environment 365 (2006) 15–32

is a strong linear relationship between the nitrate concentrations measured at both sites while the gradient is statistically significantly different from unity (p b 0.001 for both cases): NitrateMildenhall ¼ 0:63⁎NitrateClatford þ 8:9

ð2Þ

2

N = 383, r = 0.61. 4. Longer-term temporal changes For the Pang and Kennet sites mean nitrate concentrations are linked in part to the length and timing of the monitoring period, with the means for the data covering the earlier part of the monitoring programme often being lower than those for the later part of the programme. For example, in the case of the Pang mean nitrate concentrations at Frilsham and Bucklebury are 35.6 and 34.6 mg NO3 l− 1, respectively, for a monitoring period of 2002 to 2003, while the corresponding means are 30.8 and 31.0 mg NO3 l− 1 for the nearby sites upstream of the Blue Pool and at Bradfield respectively, that were monitored during 1997 to 2000; these two sets of meads are statistically significantly different (p b 0.01). To examine the temporal change in nitrate concentration for the Pang and Kennet, the time series for three sites having the longest period of data record (the Pang at Tidmarsh, the Kennet at Clatford and Mildenhall) were examined to avoid potential issues of systematic variations from location to location. For this exercise, the time series for the Thames at Howberry Park was also examined for comparative purposes, since this site has 6 years of data. The results of this exercise (Fig. 2) show distinct patterns of increasing nitrate concentrations for the Pang and Kennet sites over time, but no clear corresponding increase for the Thames. For many groundwater dominated systems such as the Pang and Lambourn there has been long-term increases in nitrate concentration due to a combination of increasing fertilizer use and long-term groundwater storage that ensures a lagged response for groundwater drainage to the river. In the case of the Thames, this feature may be obscured by the components of flow from the less permeable bedrock areas. Vinten and Smith (1993) showed that, in general, fertilizer use increased across the UK up to 1984 with a subsequent decline, while Limbrick (2003) showed continuing linear increases in spring discharge from the Chalk from 1975 up to 2000. The longer-term changes in nitrate concentration are discussed more fully for individual rivers later in this paper.

5. Within-year and flow related changes in nitrate concentrations • There are seasonal patterns in the measured nitrate concentrations, although the degree of seasonality varies in amplitude from catchment to catchment. This is illustrated in Fig. 2 where there are clearly higher nitrate concentrations during the winter to early spring months for the Kennet at Mildenhall, the Pang at Tidmarsh and the Thames at Howberry Park, but a much more damped response for the Kennet at Clatford (even though the Clatford site is only a few kilometres upstream of the Mildenhall site). This seasonal pattern is probably linked to factors such as: • Higher nitrate input concentrations from agricultural sources to the river during the winter period when the catchments have wetted up and there is the greatest runoff from the land. This represents the situation where there is insufficient rainfall volume to ‘dilute out’/deplete the agricultural sourced nitrate. Rainfall is generally low in nitrate compared to stream runoff (Neal et al., 2004a) and occasionally rainfall bypass mechanisms can sometimes be significant (Neal et al., 2004b; see also comment below). • A seasonal variation in the water-table, and a change in preferential flow-pathway in the chalk which drains water with a lower nitrate concentration. • Higher flushing of septic tanks discharging to the unsaturated zones near to the river, when groundwater levels are at their highest. • Greater within-river uptake of nitrate in the spring and summer months when flows, river water residence times and water column volumes are at their lowest, and biological activity is at its highest. • Superimposed on this may be occasional dilution of nitrate with increasing flow due to water from nearsurface pathways entering the stream which has a lower concentration than in the groundwater (i.e. a significant dilute rainfall component). However, this component may be very short lived and not picked up regularly with the weekly monitoring programme (cf. the complexity of the high resolution water quality data for the Lambourn indicated by detailed monitoring undertaken by Prior and Johnes, 2002). Fig. 3a–c shows the patterns of nitrate concentration variation with flow. The salient features are as follows. 5.1. The Pang The Pang shows some increase in nitrate concentration as flow increases for all three sites. For the upper

C. Neal et al. / Science of the Total Environment 365 (2006) 15–32

21

Fig. 2. Nitrate time series for the Pang at Tidmarsh, The Kennet at Clatford and Mildenhall and the Thames at Howberry Park.

two sites (Frilsham and Bucklebury) the patterns of increase are relatively uniform with the greatest nitrate concentration increases occurring at low to intermediate flows with a levelling off thereafter. At very low flows, there may be a small increase in nitrate concentrations as flows decreases. The lower Pang site at Tidmarsh shows a similar feature, but there is a much greater data scatter, especially at high flows that partially obscures the pattern observed at the upper two sites. For Tidmarsh, the data record is longer than for the other two sites and given that nitrate concentrations have been increasing over time, then a greater scatter would be expected at this site. Plotting the data for Tidmarsh just for the

period where all three sites have been monitored does reduce the scatter, but not to the level observed at the other two sites. A more detailed analysis of the Tidmarsh data does show a more complex pattern when comparing year to year patterns. For some years, when the flows are around 1 and 2 cumecs and above, there can be much greater scatter to the nitrate concentration data, with particularly low concentrations at such times. There may even be a clear threshold flow for this transition in a particular year. The Tidmarsh site drains both the Chalk and Eocene clays of the Bourne subcatchment and the greater scatter may well reflect some dilution at greater flows when the Eocene contribution

22 C. Neal et al. / Science of the Total Environment 365 (2006) 15–32 Fig. 3. (a) The relationship between nitrate concentration and flow for the Pang and Lambourn. The data plotted cover the full period of monitoring: the Pang at Tidmarsh data is most extensive as there was a much longer data run. (b) The relationship between nitrate concentration and flow for the Kennet and Thames. (c) The relationship between nitrate concentration and flow for the Cherwell, Ray and Thame.

C. Neal et al. / Science of the Total Environment 365 (2006) 15–32

23

Fig. 3 (continued).

will be at its highest owing to the greater near surface flow component. There seems to be some systematic change in nitrate concentration from the upper Pang sites to the Tidmarsh site, with a small lowering of nitrate concentration, across the flow range: the mean nitrate concentrations proceeding downstream were 35.6, 35.6 and 31.4 mg NO3 l− 1 for the period when all three sites were monitored. The trends for increased nitrate at higher flows are not confounded by the higher nitrate values over time.

and to the upper Pang: nitrate concentrations increase uniformly from low to intermediate flows and level off at higher flows; small increases in nitrate concentrations occur as flows decreases for very low flows during the autumn period. The patterns are more uniform and the magnitude of the changes lower for the Lambourn compared with the Pang. There also seems to be a very small reduction in nitrate concentration, across the flow range proceeding downstream, mean nitrate concentrations being 32.2, 31.7 and 30.8 mg NO3 l− 1, respectively.

5.2. The Lambourn

5.3. Kennet/Dun

The pattern of nitrate concentration across the three Lambourn monitoring sites is very similar to each other

Nitrate concentration changes with flow for the Kennet in a similar way to the Pang and, to a lesser

24

C. Neal et al. / Science of the Total Environment 365 (2006) 15–32

extent, the Lambourn. Thus, (a) there is larger data scatter where there are longer data runs (Clatford, Mildenhall and Hungerford), and (b) nitrate concentrations decrease slightly at very low flows, increase with increasing flow from low to intermediate flows, and level off at higher flows. Mean nitrate concentrations show a clear decrease in a downstream direction. For example, for equivalent period of record, mean nitrate concentrations are 37.7 mg NO3 l− 1 at Clatford, 32.8 mg NO3 l− 1 at Mildenhall, 25.0 mg NO3 l− 1 at Hungerford, and 23.1 mg NO3 l− 1 at Woolhampton. In the case of the Dun, there is no clear pattern of change with flow, although there is a notable range in nitrate concentration at very low flows and this may well result from high macrophyte biomass in the river producing variable loss of nitrate at very low flows (cf. Neal et al., 2005b). 5.4. Cherwell/Ray Nitrate concentrations vary with changing flows in different ways across the three monitoring sites on the Cherwell. At the uppermost site (Banbury), which represents the most ‘rural setting’, nitrate concentrations show strong increases with increasing flows at low to moderate flows and then start to decline at higher flows. For this ‘rural setting’, baseflow nitrate concentrations are much lower than storm flow nitrate concentrations. At the intermediate site (Kings Sutton), downstream of Banbury STW, nitrate concentrations show a marked decline as flow increases at very low flows. Above baseflow, nitrate concentrations increase and then subsequently decline as flow increases. For this site, there seems to be a composite pattern for the rural signal observed at the Banbury site and a dilution of point sources with increasing flow (nitrate concentrations in STW effluent are typically around 67 mg NO3 l− 1, Neal et al., 2005a). However, the very steep initial decline in nitrate concentrations with changing flow at the Kings Sutton site may well be linked to the contrasting patterns of point source inputs high in nitrate, low dilution potential at low flows and high biological activity and removal from the water column when water volumes and flows are low to intermediate. For the downstream Cherwell site, at Islip, the point source influence of Banbury STW on nitrate concentrations is in part diluted and there may be within-river loss of nitrate. Hence nitrate concentrations increase and then decrease as flow increases. However, unlike at the Banbury site, baseflow nitrate concentrations are similar to high flow nitrate concentrations. Mean nitrate concentrations are highest at the Kings Sutton site due to the additional point

source input. Mean nitrate concentrations change, in sequence upstream to downstream, 26.2, 35.5 and 31.1 mg NO3 l− 1. For the Ray, there is no clear pattern of change in nitrate concentration as a function of flow. This site is downstream of an extensive wetland area, Otmoor, which results in a marked damping of the nutrient response of the catchment. 5.5. Thame The Thame shows similar features to the Cherwell. Thus, there is a clear ‘rural’ signal at the uppermost site (increasing nitrate concentrations then slight decreasing nitrate concentrations as flow increases) and the influence of point sources is evident further downstream with a strong decrease in nitrate concentration at very low flows as flows increase. The mean nitrate concentration progressively increases downstream for the three sites (21.1, 29.4 and 32.9 mg NO3 l− 1, respectively). However, there are two differences for the Thame and the Cherwell. Firstly, extreme baseflow nitrate concentrations are not as high on the Thame. This may reflect a greater dilution of the effluent from Aylesbury STW or lower nitrate concentrations in the Aylesbury STW effluent. It may also represent greater within river loss due to the high biological productivity within the river. However, apart from a clear increase in phosphate levels downstream compared with upstream of the Aylesbury STW (phosphate is the limiting nutrient) there are no clear changes in biological indices of change other than in an increase in the percent of pollution tolerant benthic diatom species (Jarvie et al., 2002). Secondly, mean nitrate concentrations increase progressively downstream. This probably results from several point sources entering the Thame along its length. In particular, the town of Thame is located near to the river between the Cuddington and Wheatley monitoring points and there will also be inputs from Princes Risborough along this stretch. Thus, the Thame at Wheatley might well be expected to have the highest mean nitrate concentrations of the three sites. 5.6. Thames Nitrate concentrations show a clear pattern of change with flow. At low to intermediate flows, nitrate concentrations increase with increasing flow, while at moderate to high flow, nitrate concentrations decrease with increasing flow. There is a marked scatter around this pattern and the variation is not uniform.

C. Neal et al. / Science of the Total Environment 365 (2006) 15–32

25

6. Nitrate concentrations in the Thames basin: a comparison with other eastern UK rivers

average are provided to illustrate pertinent features for contrasting situations:

With regards to how the current data fit within the nitrate values observed across the UK, comparison is made here with the extensive data obtained for eastern UK Rivers as part of the LOIS (Neal and Robson, 2000). Table 3 provides a summary of average concentrations of boron and nitrate including averages for low flow, high flow and flow-weighted values. Here the boron data is provided as a surrogate for point source inputs from sewage (Table 3). Under low flow conditions, dilution potential is lowest, and typically boron concentrations are at their highest: the exception is for the Dun and Pang, where sewage effluent is not directly discharged to the river, but is drained to the unsaturated zone and flushed during periods of high flows when the water table is high (Neal et al., 2004a, 2005b). Boron has been used as a tracer for sewage sources as it is present at high levels in sewage effluent and it is chemically of low reactivity in the hydrosphere (Neal et al., 1998). For Table 3, four types of

The straight average This provides a measure of how the average concentrations vary across the rivers. Low flow average This provides a measure of the average nitrate concentration under baseflow conditions when point source inputs will be at their highest. High flow average This provides a measure of the average stormflow nitrate concentration, when diffuse sources are at their highest. Flow weighted average This provides a weighted nitrate average that can be linked to flux when combined with average flow values. Here it is used as a base for assessing the relative flux input from point and diffuse sources of nitrate contamination. In terms of the separation between sewage and other sources, it must be noted that there is not a straight forward divide between what is classically considered

Table 3 Average boron and nitrate concentrations for eastern UK rivers River

Label

B

NO3

μg/l

mgNO3/l

Average Tweed1 Tweed2 Teviot Wear Swale1 Swale2 Nidd Ure Ouse1 Ouse2 Derwent Wharf Aire Calder Don Trent Great Ouse Cherwell/Ray Dun Kennet Lambourn Pang Thame Thames K&A Canal

Tw1 Tw2 Te We S1 S2 N U O1 O2 De Wf A C Do Tr GtO Ch Du K L P Th Ths K&A

23.1 14.8 20.6 66.0 44.0 66.0 131.0 38.0 74.0 117.0 62.0 43.6 470.0 408.0 611.0 600.0 386.6 147.8 20.6 33.0 15.9 22.9 203.7 163.6 27.2

6.8 4.7 6.4 11.7 7.0 15.0 18.7 12.6 13.7 15.4 16.3 10.4 32.9 24.5 35.6 39.2 37.3 31.2 24.7 28.8 31.6 32.1 29.2 34.8 21.7

B

NO3

μg/l

mgNO3/l

B

NO3

μg/l

mgNO3/l

B

NO3

μg/l

mgNO3/l

Average low flow

Average high flow

Average flow weight

46.2 22.5 36.2 274.0 55.3 83.6 182.3 44.7 102.8 163.9 76.1 79.8 575.6 502.0 725.6 706.0 541.7 269.8 15.0 61.9 15.2 17.9 378.2 293.1 25.8

10.6 6.1 8.9 53.1 10.9 18.5 33.2 10.2 16.7 22.6 30.6 13.1 93.9 82.6 173.1 175.4 141.2 60.6 29.4 25.6 15.8 25.6 91.0 73.9 28.9

21.4 13.6 18.3 99.9 24.1 43.4 94.4 22.4 43.9 58.4 50.1 38.9 341.6 304.5 439.4 448.4 366.4 91.9 23.0 27.8 15.9 24.1 132.3 108.4 27.9

5.2 2.8 2.0 9.3 6.5 11.7 17.8 10.1 10.9 12.5 12.0 11.0 40.1 25.5 39.5 38.0 27.6 28.8 23.3 23.1 30.7 29.6 27.1 28.7 18.8

6.5 4.1 7.7 8.3 8.8 17.8 19.2 9.2 17.1 15.9 28.4 8.5 16.2 13.6 27.2 28.1 53.3 31.7 26.6 31.7 36.0 36.6 26.7 34.5 27.4

7.0 4.3 7.4 9.0 13.5 15.2 17.1 9.6 14.9 16.0 24.7 8.2 20.2 17.6 30.4 33.0 44.6 33.3 25.3 30.3 32.8 33.3 29.6 36.7 23.8

26

C. Neal et al. / Science of the Total Environment 365 (2006) 15–32

Fig. 4. Nitrate–boron relationships for eastern UK Rivers. Plots for average, and averages for low flow, high flow and flow weighted values. The plotted labels correspond to the abbreviations listed against their full titles in Table 3.

point (i.e. sewage effluent) and diffuse (i.e. agricultural) sources. For example as discussed elsewhere in the paper, in areas such as the Chalk catchments of the upper Thames, sewage effluent sources often enter the unsaturated zone either via discharge from sewage treatment works or from septic tanks. Similarly, farm effluent can be stored in containing areas on farms and be discharged to the river during hydrological events. Neal et al. (2005b) distinguished four types of input— point, point-diffuse, diffuse-point and diffuse. Here, sewage effluent sources where sewage sources are being estimated using boron as a marker, the sewage effluent component is not simply that of the point source. Rather, it represents an integral value for the different types of input to the river. Here, the remaining component of the nitrate in the river is equated primarily with diffuse sources and is referred to in this way.

Analysis of the data in Table 3 indicates that the nitrate concentrations in the Thames basin are similar to those for urban and industrial rivers of the central and northern England (Aire, Calder, Don and Trent) as well as the agricultural river the Great Ouse. They are of higher concentrations than the rural rivers (Teviot, Tweed, Wharf, Swale, Nidd, Ure, Ouse and Derwent). In the case of the urban/industrial rivers, boron concentrations are elevated relative to the Thames Basin rivers (and the Great Ouse), and the urban/industrial rivers show clear reductions in boron and nitrate concentration with increasing flow. The urban/industrial rivers have an important point source component for both boron and nitrate. For the Thames Basin rivers, the high nitrate concentrations relative to the rural rivers and the low boron concentrations relative to the industrial rivers illustrates the relative importance of diffuse (non-

C. Neal et al. / Science of the Total Environment 365 (2006) 15–32 Table 4 Nitrate to boron ratios and estimates of non-sewage source (diffuse) inputs of nitrate for eastern UK rivers River

Tweed1 Tweed2 Teviot Wear Swale1 Swale2 Nidd Ure Ouse1 Ouse2 Derwent Wharf Aire Calder Don Trent Gt Ouse Ch./Ray Dun Kennet Lambourn Pang Thame Thames K&A

NO3/B (mgNO3/μgB)

% NO3 diffuse source

avg

lf

hf

fw

avg

lf

hf

fw

0.30 0.32 0.31 0.18 0.16 0.23 0.14 0.33 0.19 0.13 0.26 0.24 0.07 0.06 0.06 0.07 0.10 0.21 1.20 0.87 1.98 1.40 0.14 0.21 0.80

0.11 0.12 0.05 0.03 0.12 0.14 0.10 0.23 0.11 0.08 0.16 0.14 0.07 0.05 0.05 0.05 0.05 0.11 1.55 0.37 2.02 1.65 0.07 0.10 0.73

0.61 0.68 0.86 0.16 0.80 0.96 0.58 0.90 1.02 0.71 0.93 0.65 0.17 0.16 0.16 0.16 0.38 0.52 0.90 1.24 2.28 1.43 0.29 0.47 0.95

0.33 0.32 0.40 0.09 0.56 0.35 0.18 0.43 0.34 0.27 0.49 0.21 0.06 0.06 0.07 0.07 0.12 0.36 1.10 1.09 2.06 1.39 0.22 0.34 0.85

79 80 80 65 61 73 56 81 66 53 76 74 11 −4 −7 4 35 70 95 93 97 96 56 71 92

44 49 −16 −84 47 55 36 72 41 18 61 55 10 −23 −15 −16 −22 42 96 83 97 96 13 36 91

90 91 93 60 92 93 89 93 94 91 93 90 64 62 60 61 83 88 93 95 97 96 79 87 93

81 80 85 31 89 82 66 85 82 77 87 70 −5 −8 10 15 49 83 94 94 97 95 72 82 93

industrial areas there is also seepage from contaminated land associated with industrial waste and derelict mines (Neal et al., 1998). Within the present study, the estimates of the diffuse sources of boron and nitrate can only be approximate and detailed assessment must rely on analysis using computer based dynamic process based models such as INCA (Whitehead et al., 2002a, Wade et al., 2005, in press), QUASAR (Lewis et al., 1997) and QUESTOR (Boorman, 2003a,b) that can accommodate multiple sources of inputs and withinriver processing of nitrogen. For this study, it is assumed, following the study of Neal et al. (1998), that on average 60% of the boron is associated with point sources (i.e. 40% with sea-salt and industrial and other non-diffuse sources). Fig. 4 provides a plot of the average nitrate and boron concentrations for the eastern UK Rivers. On the diagram, the sewage effluent line is plotted together with the line for average baseflow for the industrial rivers, which represents the situation where there is maximal concentrations of boron and nitrate associated with

Table 5 Summary statistics for the percentage diffuse sources of nitrate in the LOIS and Thames rivers River

avg = average, lf = low flow, hf = high flow, fw = flow weighted.

point) sources of nitrate. In the case of the less permeable catchments of the Thames Basin (the main stem of the Thames and the Cherwell and Thame) there are elevated concentrations of boron and marked dilution in concentration with increasing flow and this indicates a clear point source influence. 7. The relative importance of point and diffuse sources of nitrate In order to assess the relative importance of diffuse sources, there is a need to factor out the point sources input. For this, the boron data is used here as a marker for the sewage effluent component and the nitrate component in the effluent is assessed using the ratio of nitrate to boron in sewage effluent, based on the sewage effluent chemistry data provided in the study of Neal et al. (2005a): the nitrate to boron ratio used for this assessment is 0.104 mgNO3/μgB. Not all of the boron in the river is associated with sewage effluent as there is also a component associated with atmospheric inputs from marine (sea-spray) and pollutant (e.g. fly ash) sources (Neal, 1997; Jahiruddin et al., 1998) and in the

27

Rural rivers Avg. Min. Max.

% NO3 diffuse source avg

lf

hf

fw

71 53 81

42 − 16 72

92 89 94

80 66 89

− 11 − 23 10

62 60 64

3 −8 15

93 83 97

95 93 97

95 94 97

84 79 88

79 72 83

Urban/industrial rivers Avg. 1 Min. −7 Max. 11 Thames rivers draining Chalk Avg. 95 Min. 93 Max. 97

Thames rivers draining low permeability strata Avg. 66 30 Min. 56 13 Max. 71 42

The rivers are subdivided into four groups (1) the rural rivers of the Tweed, Teviot and northern Humber (Derwent, Ouse, Swale, Nidd, Ure, Wharf), (2) the urban/industrial rivers of the southern Humber (Aire, Calder, Don and Trent), (3) the Thanes Basin draining the Chalk (Dun, Kennet, Pang and Lambourn) and (4) the Thames Rivers draining low permeability strata (Cherwell/Ray, Thame and main stem of the Thames). The Wear and Great Ouse are not included in the list as they represent mining and agriculturally polluted systems, respectively.

28

C. Neal et al. / Science of the Total Environment 365 (2006) 15–32

sewage, contaminated land and derelict mine drainage sources. Fig. 4 shows five important features. 1. In all cases of averaging, the urban/industrial rivers plot to the high boron side of the sewage effluent line. This indicates the earlier point that there is a nonsewage source for boron in these rivers. 2. The average ratio for nitrate to boron under baseflow for the urban/industrial rivers is around 0.05 mgNO3/ μgB, which is almost a half (times 0.48) of the sewage line. This value is slightly lower than that which would be estimated based on the information given by Neal et al. (1998) where 60% of the boron is associated with sewage sources (the resultant nitrate to boron ratio would be 0.104 × 0.6 = 0.062 mgNO3/ μgB. This data is reasonably close to that for the urban/industrial baseflow situation. The relatively small difference might reflect errors in the estimation process and/or losses of nitrate by denitrification processes in the river that would be maximal under baseflow conditions, and which is known to occur (Lewis et al., 1997; Neal et al., 2000d). 3. Under high flow conditions, the data plot to the high nitrate side of the sewage effluent line and this indicates the much higher influence of diffuse sources of nitrate. 4. For the Thames Basin rivers, the catchments draining primarily the Chalk have data which plot to the high nitrate side of the sewage effluent line (except in some cases under low flow conditions). This shows the relatively high influence of agricultural sources of nitrate. 5. For the Thames Basin rivers where the catchments drain a high proportion of impermeable strata, data which plot to the higher nitrate side of the sewage effluent line, but not as high as for the rivers draining the Chalk. This shows the combined influence of sewage and agricultural sources of nitrate. Table 4 provides information on the ratios of nitrate to boron in the river waters for the various averages presented in Table 3 together with an estimate of the percentage of the diffuse (non-sewage source) component. The percentage diffuse component is calculated as the term 100 ⁎ ([NO3] − (R ⁎ C ⁎ [B])) / [NO3] where [NO3] and [B] refers to the average values of nitrate and boron, respectively, R the ratio on nitrate to boron in the effluent (0.104 mgNO3/μgB) and C is a correction factor for non-sewage sources of B (0.6). The results of this exercise are presented in Table 4 and summarized in Table 5. The percentage diffuse component increases in the order low flow, average,

flow weighted and high flow, respectively and this marks the progressive increase in the diffuse source as flow increases. There are four separate catchment typologies with distinct relationships to nitrate: • The urban/industrial rivers. These are dominated by point source inputs. In some cases the point source inputs are computed to be greater than 100% and this represents estimation errors. Under high flow conditions, the diffuse component averages around 61%, but the flow weighted average is just a few percent. • The rural rivers. These are typically dominated by diffuse sources (typically around 80% for flow weighted and around 92% for high flow values) except for under baseflow conditions where sewage effluent source inputs are much more important. • The Thames Basin rivers draining Chalk. For these rivers, the diffuse source dominates across the flow range (average 93% to 95%). • The Thames Basin rivers draining low permeability strata. In this case there is a high proportion of a sewage effluent source under low flow conditions (average 70%—i.e. 30% diffuse source). However, under average, flow weighted and high flow conditions, the diffuse source dominates averaging 66%, 79% and 84% respectively. 8. Discussion This study shows that nitrate concentrations are similar across the rivers monitored in the upper Thames basin, and that there are some clear similarities and differences related to the permeability of the catchment and the extent of point and diffuse sources of nitrate. For the tributaries of the Thames associated with permeable catchments, the dominance of groundwater inputs from the Chalk to the river ensures that the nitrate concentration variations over time are relatively damped. However, the longer-term data record shows some marked scatter and sub-weekly monitoring from other studies reveals significant short-term variability (Prior and Johnes, 2002). On a longer timescale, nitrate concentrations for the Chalk tributaries have been increasing over time through the study. This type of trend is common to many rivers in the UK (Johnes and Burt, 1993) and for the Chalk and other aquifers both in terms of surface and groundwaters (Burt and Trudgill, 1993; Limbrick, 2003). This increasing trend is strongly associated with increasing usage of fertilizers from the 1950s to 80s: a problem of national and pan-national concern (Heathwaite et al., 1993; Johnes and Burt, 1993).

C. Neal et al. / Science of the Total Environment 365 (2006) 15–32

With regards to the main stem of the Thames at Howberry Park, the long-term nitrate increase has been observed for other sites on the Thames through the 1960s to 1980s (Johnes and Burt, 1993). For our study, there is no clear trend apart from perhaps a modest decline between 1998 and 2002. This decline (Fig. 2) may just represent unusually high nitrate concentrations in 1997–1998 at the beginning of our study. However, may link to a number of factors such as improved regulation of nutrient losses from catchments, reduced nitrate fertilizer use (coupled with a delay in the groundwater response), the introduction of nitrate vulnerable zones, climate variability and the influence of less permeable catchments in the area both in terms of lack of longer term storage and high runoff from the agricultural land. However, the degree to which each of these factors has influenced the longer-term patterns of change for this site on the Thames cannot be judged in this study and other sites with much longer and continuing data runs need further examination. For the less permeable catchments, two types of behaviour are observed. Firstly, where point sources are not important, nitrate concentrations initially increase with increasing flow as nitrate is flushed from the land surface, but as flows get higher, there is a dilution due to factors that are probably linked to the kinetics of nitrate release, the extent of water mixing and partial exhaustion of available nitrate. However, the transport of nitrate and its biogeochemical attenuation through catchments is complex (Armstrong and Burt, 1993). Withers and Lord (2002) note that for low permeability systems, flow during storm events is either through fissures or over the surface and that there is limited opportunity for equilibrium with the matrix pore water. They point out that under such circumstances that nitrate concentrations in drain and surface water flow tend to fall during these periods of high flow. Secondly, where point sources are important, river-water nitrate concentrations may be high during baseflow conditions, or as high as the nitrate concentrations under high flows when diffuse agricultural inputs are maximal. For both the permeable and less permeable catchments, there may well be significant within-river losses of nitrate (Heathwaite, 2003). This has been shown in other studies (including work on the Lambourn at Boxford, a site which is referred to in our study: Prior and Johnes, 2002) and this feature probably links to uptake by aquatic plants (algae, macrophytes, etc.) within the river and to wetland areas near to the river (Prior and Johnes, 2002; Howard-Williams and Downes, 2003). However, point source inputs are important at some of the sites in terms of concentrations at low flows. At some

29

sites, the within-river processing may mask expected elevated nitrate concentrations from point source inputs at low flows, due to elevated biological removal of nitrate from the water column. With regards to the WFD, the cut-off point for nitrate regulation of 50 mg NO3 l− 1 means that several of the Thames tributaries cannot be deemed problem cases from this standpoint, as yet. However, it is not clear where threshold nitrate concentrations should be at least in terms of eutrophication, as the biological activity will be much more strongly linked to phosphorus, the limiting nutrient at least in the rivers—cf. Jarvie et al. (1998) for a discussion for eastern UK rivers. This issue is also related to aquatic ecosystem functioning and explanations in terms of types of nutrient source are simplistic. Further, for the tributaries with permeable catchments, the nitrate concentrations are continuing to increase and even the 50 mg NO3 l− 1 cut-off may be crossed. For the Thames and its tributaries with less permeable catchments, the cut-off has already been crossed in some cases. Note however that (a) the monitoring has only been undertaken on a weekly basis and it probably will not have picked up the extremes and (b) the high nitrate concentrations may not correspond to the periods of high biological activity and eutrophication risk. The results show that there is a dominance of diffuse sources of nitrate (typically of the order of 80%), that is consistent with estimates from other sources. 9. Wider comments To study the changes in nitrate concentration within these catchments there is a need for long-term measurements of the order of decades to identify the changes associated with storage, fertilizer application rates and climate. There is also a need for intensive measurements as weekly monitoring programmes cannot easily identify the short term dynamics (Prior and Johnes, 2002). This feature is becoming more recognised within issues such as fractal processing (Kirchner et al., 2004), even for agricultural systems and nutrient dynamics (Harris and Heathwaite, 2005), and biological feedback mechanisms (Jarvie et al., 2004). For both the long and short term changes, the measurements in nitrate must be coupled to information on the other key nutrients (phosphorus in particular), other pollutants (e.g. pesticides and herbicides that enter the aquatic environment) and aquatic biology (e.g. phytoplankton, macrophyte, epiphyte abundance and diversity and indices such as mean trophic rank: Jarvie et al., 2002). This is needed to provide a balanced judgement over what the biological

30

C. Neal et al. / Science of the Total Environment 365 (2006) 15–32

responses are to changes in nitrogen flux to the aquatic environment in the context of other factors causing biological change, such as climate, land-use and rural/ urban development. In this paper, the need for evaluating both point and diffuse sources of nitrate has been identified. These findings agree with many other studies within the UK and elsewhere (Flynn et al., 2002; Grizzetti et al., 2005; Oenema et al., 2005; Wade 2006-this volume; Wade et al., 2005; Whitehead et al., 2002a,b). Indeed, a major element and advantage to dynamic modelling approaches (e.g. INCA, Wade et al., 2005; QUASAR, Lewis et al., 1997 and QUESTOR, Boorman, 2003a,b) is that they can not only make allowance for point and diffuse sources, but they can also apportion their relative contribution to nutrient concentrations and loads through the seasons as well as on an annual basis. This has been done in this paper using a simple two component mixing model with boron as a marker for sewage sources and at a semi-quantitative level the importance of diffuse sources comes to the fore except for the situation where point sources are of particular importance, i.e. the urban/industrial rivers and the low permeability catchments with market and major conurbations near to the river. Within this paper, focus has been given to catchment sources of nitrate applied directly onto the land and to the river via background, sewage effluent and agricultural sources. However, atmospheric sources of ammonium are often not considered in detail in relation to flux inputs, modifications and transfers even though the total nitrogen deposition flux might equal or exceed that in runoff (Neal et al., 2004a, in press-b). In order to understand the interplay between the atmosphere, land, groundwater and aquatic environment, there is a need to integrate the work in a more holistic way into water quality and environmental impact models for nitrate in the UK environment. The critical point with regards to eutrophication is that risk assessment must take into account that the risk is highest during the spring and summer times when flows are generally low and biological activity is high. This makes the need for point source input assessment especially important and this should also be highlighted for the other major nutrient, phosphorus, which is critically related to point source inputs (Jarvie et al., in press; Neal et al., 2005a). To understand the nitrogen as well as phosphorus dynamics and their combined influence on river ecology (including eutrophication), there is a major need to integrate the measurement, Geographical Information System (GIS) and modelling approaches so that issues of

sources, flux inventories, nitrate transfers and attenuation through catchments and within river processing can be addressed. This integration must also be linked to high-quality and high-resolution measurements that are sustained and linked to ecological response to changing policy drivers such as reform of the Common Agricultural Policy, land use change, climate change/ instability and feedback mechanisms. While there have been major strides forward over the past decade or more in these areas, there remains the need for clear and progressive research that underpins the technology and environmental management advances needed to make the Water Framework Directive work (Neal and Heathwaite, 2005; Neal and Jarvie, 2005). Acknowledgements The study is the result of a combination of research drives including the research programmes LOIS, LOCAR and RELU. The work was underpinned by funds from CEH, NERC, the Environment Agency and contributes to the Catchment Hydrology, Resources, Economics and Management (ChREAM) project, funded under the joint ESRC, BBSRC and NERC Rural Economy and Land Use (RELU) programme. References Armstrong AC, Burt TP. Nitrate losses from agricultural land. In: Burt TP, Heathwaite AL, Trudgill ST, editors. Nitrate: processes, patterns and management. Chichester: John Wiley and Sons; 1993. p. 239–67. Boorman DB. LOIS in stream water quality modelling. Part 1. Catchments and methods. Sci Total Environ 2003a;314–316:379–96. Boorman DB. LOIS in stream water quality modelling. Part 2. Results and scenarios. Sci Total Environ 2003b;314–316:397–410. Burt TP, Trudgill ST. Nitrate in groundwater. In: Burt TP, Heathwaite AL, Trudgill ST, editors. Nitrate: processes, patterns and management. Chichester: John Wiley and Sons; 1993. p. 213–38. Cave RR, Ledoux L, Turner K, Jickells T, Andrews JE, Davies H. The humber catchment and its coastal area: from UK to European perspectives. Sci Total Environ 2003;314–316:31–52. CEC. Directive 2000/60/EC of the European Parliament and of the Council establishing a framework for the Community action in the field of water policy; 2000. CEC, L 327, P. 0001–0073. http://europa. eu.int/comm/environment/water/water-framework/index_en.html. CEC. Proposal for a Directive of the European Parliament and of the Council on the protection of groundwater against pollution. Brussels; 2003. 19.9.2003, COM(2003) 550 final. Defra. Developing measures to promote catchment-sensitive farming. A joint Defra-HM Treasury Consultation. http://www.defra.gov. uk/environment/water/dwpa/index.htm2004. Defra. Nitrates-reducing water pollution from agriculture. http://www. defra.gov.uk/environment/water/quality/nitrate/default.htm2004. DoE, 1993. Methodology for identifying sensitive areas (Urban Wastewater Directive) and designating vulnerable zones (Nitrates Directive) in England and Wales. Department of the Environment,

C. Neal et al. / Science of the Total Environment 365 (2006) 15–32 Ministry of Agriculture Fisheries and Food, Welsh Office. Consultation Document. Dwyer J, Eaton R, Farmer A, Balsock D, Withers P, Silkock P. Policy mechanisms for the control of diffuse agricultural pollution, with particular reference to grant aid. English nature report number 455; 2002. 135 pp. http://www.englishnature.org. uk/pubs/publication/. EA. Aquatic eutrophication of England and Wales: a management strategy. Bristol: Environment Agency; 2000. 32 pp. Evans SG, Spillett PB, Colquhoun K. South-east housing development —the quest for sustainability: water and sewerage needs. J Water Environ Manag 2003;17:257–61. Flynn NJ, Paddison T, Whitehead PG. INCA modelling of the Lee system: strategies for the reduction of nitrogen loads. Hydrol Earth Syst Sci 2002;6:467–83. Grizzetti B, Bouraoui F, de Marsily G, Bidoglio G. A statistical method for source apportionment of riverine nitrogen loads. J Hydrol 2005;304:302–15. Gustard A, Bullock A, Dixon JM. Low flow estimation in the United Kingdom. Institute of hydrology report, vol. 108. Wallingford: Institute of Hydrology; 1992. OXON, OX10 8BB, UK. Harris G, Heathwaite AL. Inadmissible evidence: knowledge and prediction in land and riverscapes. J Hydrol 2005;304:3-19. Heathwaite AL. Nitrogen cycling in surface waters and lakes. In: Burt TP, Heathwaite AL, Trudgill ST, editors. Nitrate: processes, patterns and management. Chichester: John Wiley and Sons; 2003. p. 99-140. Heathwaite AL, Burt TP, Trudgill ST. Overview—the nitrate issue. In: Burt TP, Heathwaite AL, Trudgill ST, editors. Nitrate: processes, patterns and management. Chichester: John Wiley and Sons; 1993. p. 3-21. Heathwaite AL, Billen G, Gibson C, Neal C, Steenvooden J, Withers P, editors. Nutrient mobility within river basins, vol. 304. J Hydrol; 2005. p. 1-492. Howard-Williams C, Downes MT. In: Burt TP, Heathwaite AL, Trudgill ST, editors. Nitrate: processes, patterns and management. Chichester: John Wiley and Sons; 2003. p. 141–67. Jarvie HP, Whitton BA, Neal C. Nitrogen and phosphorus in east coast British rivers: speciation, sources and biological significance. Sci Total Environ 1998;210/211:79-109. Jarvie HP, Lycett E, Neal C, Love A. Patterns in nutrient concentrations and biological quality indices across the upper Thames basin, UK. Sci Total Environ 2002;282–283:263–94. Jarvie HP, Williams RJ, Neal C. Assessing changes in phosphorus concentrations in relation to in-stream plant ecology in lowland permeable catchments: bringing ecosystem functioning into water quality monitoring. Water Air Soil Pollut Focus 2004;4:641–55. Jarvie HP, Neal C, Withers PJA, in press. Sewage-effluent phosphorus a greater risk to river eutrophication than agricultural phosphorus. Sci Tot Environ. Jahiruddin M, Smart R, Wade AJ, Neal C, Cresser MS. Factors regulating the distribution of boron in water in the River Dee catchment in N.E. Scotland. Sci Total Environ 1998;210/ 211:53–62. Johnes PJ, Burt TP. Nitrate in surface waters. In: Burt TP, Heathwaite AL, Trudgill ST, editors. Nitrate: processes, patterns and management. Chichester: John Wiley and Sons; 1993. p. 268–317. Kirchner JW, Feng X, Neal C, Robson AJ. The fine structure of water-quality dynamics: the (high-frequency) wave of the future. Hydrol Proc 2004;18:1353–9.

31

Leeks GJL, Jarvie HP. Introduction to the Land–Ocean Interaction Study (LOIS): rationale and international context. Sci Total Environ 1998;210/211:5-20. Lewis DR, Williams RJ, Whitehead PG. Quality simulation along rivers (QUASAR): an application to the Yorkshire Ouse. Sci Total Environ 1997;194/195:399–418. Limbrick KJ. Baseline nitrate concentration in groundwater of the Chalk in south Dorset, UK. Sci Total Environ 2003;314– 316:89–98. Marsh TJ, Lees ML, 2003. Hydrological data UK: Hydrometric Register and Statistics, 1991–95. Institute of Hydrology and British Geological Survey report, Institute of Hydrology, Maclean Building, Crowmarsh Gifford, Wallingford, OXON, OX10 8BB, UK. Neal C. Boron water quality of the Plynlimon catchments. Hydrol Earth Syst Sci 1997;1:619–27. Neal C, editor. Assessing nitrogen dynamics in catchments across Europe within an INCA modelling framework, vol. 6. Hydrol Earth Sys Sci; 2002. p. 297–615. Neal C, Heathwaite AL. Nutrient mobility within river basins: a European perspective. J Hydrol 2005;304:477–90. Neal C, Jarvie HP. Agriculture, community, river eutrophication and the Water Framework Directive. Hydrol Proc 2005;19:1895–901. Neal C, Robson AJ. A summary of river water quality data collected within the Land Ocean Interaction Study: core data for Eastern UK rivers draining to the North Sea. Sci Total Environ 2000;251/ 252:585–665. Neal C, Whitehead PG, editors. Water quality functioning of lowland permeable catchments: inferences from an intensive study of the River Kennet and upper River Thames, vol. 282/283. Sci Tot Environ; 2002. 500 pp. Neal C, Fox KK, Harrow ML, Neal M. Boron in the major UK rivers entering the North Sea. Sci Total Environ 1998;210/211:41–52. Neal C, Williams RJ, Neal M, Bhardwaj LC, Wickham H, Harrow M, et al. The water quality of the River Thames at a rural site downstream of Oxford. Sci Total Environ 2000a;251/252:441–58. Neal C, Neal M, Wickham H, Harrow M. The water quality of a tributary of the Thames, the Pang, southern England. Sci Total Environ 2000b;251/252:459–76. Neal C, Jarvie HP, Howarth SM, Whitehead PG, Williams RJ, Neal M, et al. The water quality of the River Kennet: initial observations on a lowland chalk stream impacted by sewage inputs and phosphorus remediation. Sci Total Environ 2000c;251/252:477–96. Neal C, House WA, Leeks GJL, Whitton BA, Williams RJ. Conclusions of the special issue of Science of the Total Environment concerning ‘The water quality of UK rivers entering the North Sea. Sci Total Environ 2000d;251/252:557–73. Neal C, Skeffington R, Neal M, Wyatt R, Wickham H, Hill L, et al. Rainfall and runoff water quality of the Pang and Lambourn, tributaries of the River Thames, south eastern England. Hydrol Earth Syst Sci 2004a;8:601–13. Neal C, Jarvie HP, Wade AJ, Neal M, Wyatt R, Wickham H, et al. The water quality of the LOCAR Pang and Lambourn catchments. Hydrol Earth Syst Sci 2004b;8:614–35. Neal C, Jarvie HP, Neal M, Love AJ, Hill L, Wickham H. Water quality of treated sewage effluent in a rural area of the upper Thames Basin, southern England, and the impacts of such effluents on riverine phosphorus concentrations. J Hydrol 2005a;304:103–17. Neal C, House WA, Jarvie HP, Neal M, Hill L, Wickham H. Phosphorus concentrations in the River Dun, the Kennet and Avon Canal and the River Kennet, southern England. Sci Total Environ 2005b;344:107–28.

32

C. Neal et al. / Science of the Total Environment 365 (2006) 15–32

Neal C, Neal M, Hill L, Wickham H, in press-a. The water quality of the River Thame in the Thames Basin of south/south-eastern England. Sci Tot Environ. Neal C, Neal M, Hill L, Wickham H, in press-b. River water quality of the River Cherwell: an agricultural clay-dominated catchment in the upper Thames Basin, southeastern England. Sci Tot Environ. Oenema O, van Liere L, Schoumans O. Effects of lowering nitrogen and phosphorus surpluses in agriculture on the quality of groundwater and surface water in the Netherlands. J Hydrol 2005;304:289–301. Prior H, Johnes PJ. Regulation of surface water quality in a Cretaceous Chalk catchment, UK: an assessment of the relative importance of instream and wetland processes. Sci Total Environ 2002;282– 283:159–74. RPA, 2003. Water Framework Directive—Indicative costs of agricultural measures. Final report prepared for the Department for the Environment, Food and Rural Affairs, by Risk and Policy Analysts Limited in association with ADAS and CEH, 48pp. (www.defra.gov.uk/environment/water/wfd/index.htm). Vinten AJA, Smith KA. Nitrogen cycling in agricultural soils. In: Burt TP, Heathwaite AL, Trudgill ST, editors. Nitrate: processes, patterns and management. Chichester: John Wiley and Sons; 1993. p. 39–73.

Wade AJ. Monitoring and modelling the impacts of global change on European freshwater ecosystems—Introduction. Sci Tot Environ 2006;365:3-14. doi:10.1016/j.scitotenv.2006.02.054. Wade AJ, Neal C, Whitehead PG, Flynn NJ. Modelling nitrogen fluxes from the land to the coastal zone in European systems: a perspective from the INCA project. J Hydrol 2005;304:413–29. Wade AJ, Butterfield D, Whitehead PG, in press. Towards an improved understanding of the nitrate dynamics in lowland, permeable river-systems: applications of INCA-N. J Hydrol. Whitehead PG, Johnes PJ, Butterfield D. Steady state and dynamic modelling of nitrogen in the River Kennet: impacts of land use change since the 1930s. Sci Total Environ 2002a;282– 283:417–34. Whitehead PG, Lapworth DJ, Skeffington A, Wade AJ. Excess nitrogen leaching and C/N decline in the Tillingbourne catchment, southern England. Hydrol Earth Syst Sci 2002b;6:455–66. Withers PJA, Lord EI. Agricultural nutrient inputs to rivers and groundwaters in the UK: policy, environmental management and research needs. Sci Total Environ 2002;282–283:9-24.