OR/16/033 The UK groundwater nitrate legacy
|Stuart, M E, Ward, R S, Ascott, M, and Hart A J1. 2016. Regulatory practice and transport modelling for nitrate pollution in groundwater. British Geological Survey Internal Report, OR/16/033.|
1 Environment Agency
| Nitrate units of measurement
Nitrate concentrations can either be quoted as nitrate or NO3 (mg NO3 l-1) or as the equivalent amount of N (mg NO3-N l-1). These units are converted using the factor molecular weight NO3/molecular weight N = 62/14 =4.43. For example the drinking water limit for nitrate is 50 mg NO3 l-1 or 11.3 mg NO3-N l-1. In this review mg NO3-N l-1 is used except where graphical material uses NO3.
The increase of nitrate in groundwater was first discussed by Greene and Walker (1970) who showed nitrate concentrations rising at the Friston and Cornish supply sources in the Chalk close to Eastbourne from 4 to 6 mg NO3-N l-1 and 5 to 8 mg NO3-N l-1 respectively between 1954 and 1966. Through the 1970s, awareness of the extent of high and rising nitrate in groundwater gradually increased, and it became clear that concentrations in public supply sources often exceeded the World Health Organization (WHO) drinking water guideline values used at this time in the absence of UK or EU drinking water standards. In 1973 the Water Act led to a reorganisation of the water industry and this resulted in:
- the establishment of regional water authorities which raised awareness of nitrate by bringing disparate groundwater quality data together. For the first time, the scale of the problem became apparent. The outcrop areas of all of the principal aquifers in the UK were affected, and a diffuse rather than point source nitrate contamination was implicated;
- the establishment of the Water Research Centre (WRc) and initiation of work on nitrate pollution of groundwater.
By the late 1970s the importance of storage of nitrate in unsaturated zone porewater was becoming recognised (Foster and Crease, 1974; Foster and Young, 1980; Oakes et al., 1981; Young et al., 1976b). Initially this was recognised in shallow pits in chalk (Foster and Crease, 1974) (Figure 2.1).
Pioneering work in understanding nitrate leaching to groundwater was carried out by drilling cored boreholes through the Chalk unsaturated zone to obtain profiles of the porewater nitrate concentration as a function of depth (Foster et al., 1982; Young et al., 1976a). At sites with good cropping records a relationship between historical land use and porewater nitrate concentration could be determined (Figure 2.2) (Young et al., 1976a).
Unsaturated zone travel time was studied using tritium as a conservative tracer (Geake and Foster, 1989; Young et al., 1976b). It was also estimated by a programme of redrilling and obtaining porewater profiles at sites where a well-defined peak could be identified (Figure 2.3). In this figure the main peak can be shown to have migrated from about 3 m depth to about 5 m over 2.5 years, a rate of downwards movement of about 0.8 m per year.
The rates of travel have been established for the three main aquifers, the Chalk, Sherwood Sandstone and the Lincolnshire Limestone (Table 2.1). These values suggest that intergranular velocity can be reasonably estimated from porosity and effective rainfall. Subsequently a large number of porewater profiles collected for the major aquifers of the United Kingdom have been collated by Stuart (2005) (Figure 2.4). This body of work showed how the loading of nitrate in the unsaturated zone had significantly increased due to post-1945 agricultural intensification.
|Porosity (%)|| Effective rainfall
| Unsaturated zone velocity|
|White Chalk Subgroup||25–45||33.1||150–350||250||0.3–1.4||0.76|
|Grey Chalk Subgroup||27.9||250||0.90|
|Lincolnshire Limestone Formation||10–25||18||150–250||200||0.6–2.5||1.11|
|Sherwood Sandstone Group||15–35||26||200–350||275||0.6–2.3||1.06|
Groundwater nitrate concentrations have continued to increase in many areas. Stuart et al. (2007) analysed UK groundwater nitrate using robust linear regression to define past trends and estimate future concentrations. Of the series analysed, 21% showed a significant improvement in the overall fit when a break was included. Half of these indicated an increase in trend with time. Significant seasonality was found in about one-third of the series, with the highest nitrate concentrations usually found during winter months. Inclusion of nearby water-level data as an additional explanatory variable successfully accounted for much of this seasonality. Based on 309 datasets from 191 distinct sites, nitrate concentrations were found to be rising at an average of 0.077 mg NO3-N l-1 year-1. In 2000, 34% of the sites analysed exceeded the EU drinking water standard. If the trends at that time continued, the authors predicted that 41% could exceed the standard by 2015. Rivett et al. (2007) concluded similarly that 60% of groundwater bodies may fail to reach good status by this time.
Statistical examination of over 8000 nitrate measurements in the Dorset and Hampshire Basin Chalk aquifers indicates that nitrate varies significantly with time, borehole depth and groundwater level, and between Chalk stratigraphic unit, land-use and groundwater body. Arable and urban land-uses are significantly more likely to be associated with higher groundwater nitrate concentrations than managed grassland (Roy et al., 2007).
Much effort has been focussed on understanding the processes associated with nitrate transport and degradation (Geake and Foster, 1989; Lawrence and Foster, 1986; Mathias et al., 2007; Rivett et al., 2007; Wellings and Bell, 1980), on mapping the spatial extent of nitrate contamination of groundwater (Rivett et al., 2007) and aquifer vulnerability to nitrate contamination (Foster, 1993; Lake et al., 2003).
Rivett et al. (2008) conclude that denitrification is the dominant nitrate attenuation process in groundwater. The critical limiting factors are oxygen concentration (denitrification is in the main microbially mediated and the enzyme systems responsible are inhibited by oxygen so anaerobic conditions are required for denitrification to proceed) and electron donor availability. Kinniburgh et al. (1999) concluded that denitrification beneath the soil zone in the unsaturated zone of UK aquifers was probably insignificant relative to the nitrate flux. Other available field studies suggest that denitrification in unconfined aquifers is relatively limited and have demonstrated only minor decreases in nitrate concentrations, estimated at just 1–2% of the nitrate load within infiltrating water in principal aquifers. Such decreases are unlikely to significantly influence regional groundwater quality. Within the saturated zones of the Chalk, Sherwood Sandstone and Jurassic Limestone aquifers, denitrification was only significant once these aquifers became confined and dissolved oxygen had depleted. However, evidence for denitrification is typically weak at the regional aquifer scale and low nitrate concentrations may sometimes be simply ascribed to dilution, lack of pollution or to slow transport of plumes.
Process-based models, typically at the source to catchment scale, have been used to provide estimates of future trends (Whitehead et al., 1998) but as a range of factors affect nitrate fate and transport these models tend to be specific to the study area (Smith et al., 2010). Consequently, it is difficult to generalise observations from these process-based predictive models and they do not enable systematic assessments of future trends in average nitrate concentration. Until recently the application of complex GIS models has only been practical at the catchment scale (Wang and Yang, 2008) and not at regional or national scales.
BGS nitrate legacy model (NTB)
A simple GIS model for Great Britain was developed on a 1 km × 1 km grid within BGS to predict nitrate arrival time at the water table (Wang et al., 2012). In this model the distribution of nitrate arriving at the water table depended on only three functions: the nitrate input at the land surface (the temporally varying but spatially uniform leaching of nitrate from the base of the soil); the rate of travel of nitrate through the unsaturated zone (spatially dependent on variations in hydrolithological characteristics); and the thickness of the unsaturated zone (Figure 2.5).
The unsaturated zone thickness and nitrate velocity are combined to estimate the spatial distribution of nitrate travel time in the unsaturated zone and from this the input year for nitrate reaching the water table at any defined time. A nitrate input function over time can then be used to estimate the concentration reaching the water table at any point and defined time, assuming that nitrate is conservative.
The presence of thick, low-permeability, superficial deposits limits the amount of nitrate which is able to enter the aquifer and this was accounted for by switching off the input function where such deposits are present. Spatio-temporal variations in recharge rate, nitrate degradation, and diffusive and dispersive processes in the soil and unsaturated zones will all influence the loading of nitrate at the water table, but here these factors are considered to be secondary and were not considered. Factors such as average saturated groundwater flow and groundwater discharge rates which will affect trends in nitrate concentration in the aquifer as a whole were also excluded.
The model is based on the following assumptions:
- nitrate input/loading is from the base of the soil;
- movement is through the matrix only in dual-porosity strata;
- the mass of nitrate in the unsaturated zone is preserved except where the bedrock is overlain by low-permeability superficial deposits;
- nitrate moves vertically from the land surface to the water table;
- nitrate moves at a constant velocity through the unsaturated zone; and
- there is no hydrodynamic dispersion of nitrate in the unsaturated zone.
Of these model functions, the unsaturated zone velocity and the depth to water are assumed to be constant over the modelled period and can be relatively well characterised from current hydrogeological data, whereas nitrate leaching will have changed over time and is based on a series of assumptions.
Unsaturated zone velocities
The model requires an effective vertical velocity of nitrate in the unsaturated zone for each 1 km by 1 km cell. The digital 1:625 000 hydrogeological mapping of Great Britain (BGS, 2010) was used as the basis for assigning the spatially dependent nitrate velocities. This was divided into three main classes of aquifer units: i) aquifers with significant intergranular flow, ii) aquifers in which flow is virtually all through fractures and other discontinuities and iii) rocks with essentially no groundwater (Figure 2.6). Within the first two classes aquifers were assessed as high, moderate or low productivity. Using a combination of these classes and other factors such as grain-size and age (as a surrogate for induration) each of the bedrock formations was attributed with a water movement rate.
Smith et al. (1970) used tritium profiles to measure rates of water movement through low permeability strata and obtained a value of 0.09 m year-1 for the Oxford Clay Formation. The latter value relates to autumn recharge through cracks (fractures) produced by a summer soil moisture deficit. A value of 0.1 m year-1 was therefore used for this and similar clay strata. For all other formations, the values were attributed heuristically using the criteria in Table 2.2. Where formations formed multi-layered aquifers and intergranular flow was significant in the permeable horizons, the prevalence of clay layers, as well as the predominant grain-size of the permeable horizons, was taken into account, to obtain the value.
|Type||Examples||Unsaturated zone flow rate (m year-1)|
|Aquifers with significant intergranular flow||Predominantly sands||Crag Group, Bracklesham and Barton Groups, Upper Greensand Formation, Lower Greensand Group, Bridport Sand Formation||3|
|Predominantly silts||Solent Group, Lambeth Group, Thanet Sand Formation||0.3|
|Fractured aquifers||Karstic||Zechstein Group dolomite, Dinantian limestone, Durness Group||10|
|Multi-layered Mesozoic aquifers||Corallian Group, Mercia Mudstone Group||1|
|All Palaeozoic (except Zechstein Group dolomites and Permian mudstones), igneous and metamorphic rocks||Old Red Sandstone Supergroup, Coal Measures Group, Millstone Grit Group, granite, Lewisian complex||1|
|Aquitards||Clays (Jurassic and younger)||Thames Group, Kimmeridge Clay Formation, Oxford Clay Formation, Lias Group||0.1|
The model does not take account of the wide variation in precipitation across Great Britain with over 2000 mm/year in upland areas of the north and west and less than 600 mm in parts of East Anglia. However, most of the important aquifers are located away from the north and west and it has been assumed that unsaturated zone annual travel time within aquifers is uniform at the national scale.
Depth to groundwater at the national scale
A representative depth to groundwater was estimated for each 1 km × 1 km cell across Great Britain based on:
- groundwater levels inferred from estimated river base levels (RBL);
- groundwater levels taken from contours on published hydrogeological maps (generally at 1:100 000 scale) and from other digitised contours;
- point measurements from national networks of observation wells and from well inventories.
Areas of low-permeability rocks are difficult to deal with by this approach so to avoid unrealistic estimations of groundwater levels in low permeability areas with pronounced topography the dataset was filtered so that the maximum thickness of the unsaturated zone was constrained to no more than 10 metres in areas underlain by low permeability rocks.
The RBL surface is an interpolated surface that assumes that rivers are hydraulically connected to aquifers, and approximate to the level of the water table in the aquifer (Figure 2.7). The river network used is derived from the NextMap Digital Surface Model (DSM), with drainage densities appropriate to different hydrolithological units. The depth to groundwater was obtained by subtracting the mean groundwater levels from the NextMap DSM mean topographic elevations for each 1 km by 1 km grid square.
The resulting dataset was compared to field measurements from 30 index boreholes in the National Ground Water Level network. The modelled water levels are within the observed ranges, where observation boreholes were unconfined. Where discrepancies were noted these were generally a result of observations being made close to valley floors, and hence where water tables are shallower than the average over a one kilometre square, which is the value used in the model. The model gives a realistic water table in permeable unconfined aquifers, and close to surface drainage.
Nitrate input function
The nitrate input function used, shown as a red line in Figure 2.8, was based on estimates of the time-varying nitrate content found in the unsaturated zone immediately beneath the soil layer. The curve was divided into six time slices or spans as follows:
Span 2, from 1940 to 1955, consists of a 1 kg N ha-1 year-1 rise in input from 25 kg N ha-1 in 1940 to 40 kg N ha-1 in 1955. This rise is the result of the gradual intensification of agriculture during and just after WWII (based on Foster et al., 1982).
Span 3, from 1955 to 1975, shows a more rapid rise of 1.5 kg N ha-1 year-1 from 40 kg N ha-1 in 1955 to 70 kg N ha-1 in 1975. This steeper rise is due to increases in the use of chemical based fertilisers (Addiscott et al., 1991; Foster et al., 1982).
Span 4, from 1975 to 1990, is a constant peak nitrate input value based on the average value obtained by Lord et al. (1999) beneath a range of land-uses.
Span 5, from 1991 to 2020, has a gradual decline of 1 kg N ha-1 year -1 from 70 kg N ha-1 in 1991 to 40 kg N ha-1 in 2020 due to restrictions on fertiliser application as a result of the implementation of nitrate sensitive areas and vulnerable zones (Lord et al., 1999) and also due to a general decrease in nitrate application of about 30% in fertiliser use between 1990 and 2000 (ADAS, 2003).
Span 6, from 2020 to 2050 (the end of the modelled input), is a constant 40 kg N ha-1 assuming a return to nitrate input levels similar to those associated with early intensified farming in the mid-1950s.
The nitrate input function was compared with nitrate concentration data from the porewaters of almost 300 cored boreholes from major aquifers (Stuart, 2005). The function was converted from kg N ha-1 to mg NO3-N l-1 by assuming a constant effective rainfall of 250 mm year-1. The porewater data were used to back estimate the nitrate in infiltration entering the unsaturated zone during the past 100 years, using the date at which the samples were taken, their depth below ground surface and an estimate of velocity in the unsaturated zone derived from tritium profiles (Table 2.1). Annual averages show an excellent agreement with the overall modelled input function. The apparent large applications between 1995 and 2000 may be an artefact of both the relatively small number of recent data points and a bias imposed by the focus of recent studies on areas with a nitrate problem.
The distribution of travel times for the unsaturated zone from the surface to the water table for nitrate, and indeed for any conservative tracer, is presented in Figure 2.9. The calculated nitrate travel time ranges between 1 and over 400 years. On the basis of the model, nitrate is projected to reach the water table of 88.1% of the areas of Great Britain within 20 years of input. It is predicted to take 1 year for nitrate to reach the water table in roughly 27% of areas.
The results can be summarised as:
- the NTB model gave the first modelled national assessment of legacy nitrate in the unsaturated zone;
- this showed that the 1980–90 peak nitrate applications are still in the unsaturated zone in aquifers with thick unsaturated zones, generally due to relief;
- this was based on measured values for travel time in the major aquifers, but used heuristic values for the others;
- it was based on hydrogeological mapping at the 652k scale;
- it used a literature-based nitrate input function applied uniformly to the whole land area;
- water levels were predominantly derived from hydrological maps and represented autumn minimum levels.
Integrating with the saturated zone at the catchment scale
Wang et al. (2013) developed an integrated modelling method to simulate the nitrate transport in both the unsaturated and the saturated zones at the catchment scale. Three BGS numerical models, the NTB model — described above, GISGroundwater — a groundwater flow model (Wang et al., 2010) and N–FM a nitrate transport model for the saturated zone developed for this work, were integrated to verify and support each other to provide information on nitrate lag time in the groundwater system at a catchment scale. The Eden Valley, which has thick Permo-Triassic Sandstone unsaturated zones and a nitrate groundwater pollution problem, was selected as a case study area.
Modelling water levels
The unsaturated zone (USZ) thicknesses used in the NTB model are too coarse for a catchment scale study. Therefore, a simple and easy-to-use groundwater flow model, GISGroundwater, was used to simulate the long-term average steady-state groundwater levels (GWLs) for the area to derive high spatial resolution of the thicknesses of the Permo-Triassic sandstone USZs in the Eden Valley. GISGroundwater is a seamless GIS 2-dimensional numerical finite difference groundwater flow model (Wang et al., 2010).
The Penrith and St Bees Sandstone formations were simplified as a single layer aquifer with a distribution of hydraulic conductivity values. The modelling extent was defined by a (100 m × 100 m) GIS layer. A GIS layer containing distributed K values was entered into the model; river nodes and river stages entered were derived from a Centre for Ecology and Hydrology (CEH) river system dataset and a digital elevation model) dataset from CEH; groundwater abstraction data were also entered into the model using a GIS layer.
An average groundwater recharge of 1 mm day-1 was used in the groundwater flow model which was calibrated by comparing the simulated long-term average GWLs with observed ones in 39 boreholes. Figure 2.10 shows that the modelled and observed GWLs correlate indicating that the steady-state groundwater flow model for the study area was well calibrated. The K values for modelling the groundwater flow in the Penrith and St Bees Sandstones were 3.5 m day-1 and 0.6 m day-1 respectively. The distributed Permo-Triassic Sandstone USZ thickness map for the area was then derived by subtracting the modelled long-term average GWLs from the DSM dataset.
In the study area, the modelled thickness of the Permo-Triassic sandstone USZs is greatest, 183 m in the northwest of the Eden Valley, and reduces to 0 m (i.e. GWLs are the same elevation as the river stages) along the River Eden and its tributaries. Notably SPZs generally have a thicker USZ than other parts of the study area. The nitrate travel time in the Permo-Triassic sandstone USZs correlating with the USZ thickness, ranges from 0 to 61 years with a mean value of 12 years; strip areas along streams have short travel times (0–1 year) due to thin USZs, whilst mountainous areas in the east and west of the Eden Valley have longer nitrate travel times.
Modelling nitrate dilution in the saturated zone
N–F M — a GIS nitrate transport model for the saturated zone was developed to simulate yearly nitrate concentration at a borehole by considering the process of nitrate leaching from the bottom of the soil zone, the nitrate movement in the USZ and dilution in the saturated zone. The simulated pumped nitrate concentration in boreholes was compared with observed values to validate the numerical modelling parameters, such as the nitrate transport velocity in the USZ, the thickness of the USZ, and the aquifer hydraulic conductivity values used for deriving the thickness of the USZ.
Figure 2.11 shows the conceptual model of N–FM. The dilution process was simplified by assuming that nitrate arriving at a borehole dilutes in water pumped out of the borehole, and the groundwater flow within a groundwater Source Protection Zone (SPZ), reaches a steady-state, i.e., the long-term recharge volume within a SPZ equals water pumped out of the borehole in the SPZ. Not all leached nitrate reaches the abstraction borehole due to attenuation in the USZ and the saturated zone. Nitrate concentration may be lowered due to denitrification and absorption in USZs; nitrate in the saturated zones will be absorbed by small pores or transported outside of SPZ due to the diffusion and dispersion processes. Therefore a nitrate attenuation coefficient (NAC) was introduced into this model. With this conceptual model, the depth of the saturated zone, and the thickness of active groundwater zone can be ignored, and the nitrate dispersion and diffusion processes can be simplified in simulating yearly nitrate concentration at a borehole in the SPZ.
The modelling showed that the peak nitrate loading around 1983 has affected most of the study area (Figure 2.12). For areas around the SPZs of Bowscar, Beacon Edge, Low Plains, Nord Vue, Dale Springs, Gamblesby, Bankwood Springs, and Cliburn, the peak nitrate loading will arrive at the water table in the next 34 years (Figure 2.12). Statistical analysis shows that 8.7% of the Penrith Sandstone and 7.3% of the St Bees Sandstone have not been affected by peak nitrate.
Distributed maps were produced for nitrate concentration at the water table for each year between 1925 and 2040. The results show that the average nitrate concentration at the water table across the study area has reached its peak and will decrease over the next 30 years (Figure 2.13). Some unaffected areas with thicker USZs around Beacon Edge, Fairhills, Bowscar, Nord Vue, Low Plains, Gamblesby, and Bankwood Springs, will be affected by peak nitrate loadings between 2020 and 2030, and then retain a high nitrate concentration (172 mg NO3 l-1) (before any groundwater dilution) around 2040.
The results show that:
- The model provides predictions of nitrate concentrations at public supply boreholes;
- catchment scale groundwater modelling simulated water levels and derived unsaturated zone thickness;
- the NTB nitrate application function and unsaturated zone velocities were applicable;
- measured water levels are needed to calibrate the flow model;
- SPZs are needed to calculate nitrate dilution in the saturated zone.
- GREENE, L A, and WALKER, P. 1970. Nitrate pollution of Chalk waters. Water Treatment and Examination, Vol. 19, 169–182.
- FOSTER, S S D, and CREASE, R I. 1974. Nitrate pollution of chalk ground water in east Yorkshire — a hydrogeological appraisal. Journal of the Institute of Water Engineers, Vol. 28, 178–194.
- FOSTER, S S D, and YOUNG, C P. 1980. Groundwater contamination due to agricultural land-use practices in the United Kingdom. 262–282 in Aquifer Contamination & Protection. Studies and Reports in Hydrogeology Series, 30. (UNESCO-IHP.)
- OAKES, D B, YOUNG, C P, and FOSTER, S S D. 1981. The effects of farming practices on groundwater quality in the United Kingdom. Science of the Total Environment, Vol. 21, 17–30.
- YOUNG, C P, OAKES, D B, and WILKINSON, W B. 1976b. Prediction of future nitrate concentrations in groundwater. Ground Water, Vol. 14, 426–438.
- FOSTER, S S D, CRIPPS, A C, and SMITH-CARINGTON, A K. 1982. Nitrate leaching to groundwater. Journal of Hydrology, Vol. 46, 343–364.
- YOUNG, C P, HALL, E S, and OAKES, D B. 1976a. Nitrate in groundwater — studies on the Chalk near Winchester, Hampshire. WRc Technical Report TR 31 (Medmenham).
- GEAKE, A K, and FOSTER, S S D. 1989. Sequential isotope and solute profiling in the unsaturated zone of British Chalk. Hydrological Sciences Journal, Vol. 34, 79–95.
- FOSTER, S S D, BRIDGE, L R, GEAKE, A K, LAWRENCE, A R, and PARKER, J M. 1986. The groundwater nitrate problem: a summary of research on the impact of agricultural land-use practices on groundwater quality between 1976 and 1985. British Geological Survey Research Report, 86/2.
- STUART, M E. 2005. Development of nitrate profiles database. British Geological Survey Internal Report, IR/05/058.
- BLOOMFIELD, J, BREWERTON, L, and ALLEN, D. 1995. Regional trends in matrix porosity and dry density of the Chalk of England. Quarterly Journal of Engineering Geology and Hydrogeology, Vol. 28, S131–S142.
- ALLEN, D J, BREWERTON, L J, COLEBY, L M, GIBBS, B R, LEWIS, M A, MACDONALD, A M, WAGSTAFF, S J, and WILLIAMS, A T. 1997. The physical properties of major aquifers in England and Wales. British Geological Survey Technical Report WD/97/34, Environment Agency R&D Publication 8. British Geological Survey Technical Report WD/97/34, Environment Agency R&D Publication 8.
- STUART, M E, CHILTON, P J, KINNIBURGH, D G, and COOPER, D M. 2007. Screening for long-term trends in groundwater nitrate monitoring data. Quarterly Journal of Engineering Geology and Hydrogeology, Vol. 40, 361–376.
- RIVETT, M O, SMITH, J W N, BUSS, S R, and MORGAN, P. 2007. Nitrate occurrence and attenuation in the major aquifers of England and Wales. Quarterly Journal of Engineering Geology and Hydrogeology, Vol. 40, 335–352.
- ROY, S, SPEED, C, BENNIE, J, SWIFT, R, and WALLACE, P. 2007. Identifying the significant factors that influence temporal and spatial trends in nitrate concentrations in the Dorset and Hampshire Basin Chalk aquifer of Southern England. Quarterly Journal of Engineering Geology and Hydrogeology, Vol. 40, 377–392.
- LAWRENCE, A R, and FOSTER, S S D. 1986. Denitrification in a limestone aquifer in relation to the security of low-nitrate groundwater supplies. Journal of the Institution of Water Engineers and Scientists, Vol. 40.
- MATHIAS, S A, BUTLER, A P, IRESON, A M, JACKSON, B M, MCINTYRE, N, and WHEATER, H S. 2007. Recent advances in modelling nitrate transport in the Chalk unsaturated zone. Quarterly Journal of Engineering Geology and Hydrogeology, Vol. 40, 353–359.
- WELLINGS, S R, and BELL, J P. 1980. Movement of water and nitrate in the unsaturated zone of Upper Chalk near Winchester, Hants., England. Journal of Hydrology, Vol. 48, 119–136.
- FOSTER, S S D. 1993. The Chalk aquifer — its vulnerability to pollution. 1–13 in The Hydrogeology of the Chalk of North-West Europe. DOWNING, R, PRICE, M, and JONES, G (editors). (Oxford: Clarendon Press.) ISBN 0198542852
- LAKE, I R, LOVETT, A A, HISCOCK, K M, BETSON, M, FOLEY, A, SÜNNENBERG, G, EVERS, S, and FLETCHER, S. 2003. Evaluating factors influencing groundwater vulnerability to nitrate pollution: developing the potential of GIS. Journal of Environmental Management, Vol. 68, 315-328.
- RIVETT, M O, BUSS, S R, MORGAN, P, SMITH, J W, and BEMMENT, C D. 2008. Nitrate attenuation in groundwater: a review of biogeochemical controlling processes. Water Research, Vol. 42, 4215–4232.
- KINNIBURGH, D G, GALE, I N, GOODDY, D C, DARLING, W G, MARKS, R J, GIBBS, B R, COLEBY, L M, BIRD, M J, and WEST, J M. 1999. Denitrification in the unsaturated zones of the British Chalk and Sherwood Sandstone aquifers. British Geological Survey Technical Report WD/99/2.
- WHITEHEAD, P G, WILSON, E J, and BUTTERFIELD, D. 1998. A semi-distributed Integrated nitrogen model for multiple source assessment in catchments (INCA): Part I—model structure and process equations. Science of the Total Environment, Vol. 210, 547–558.
- SMITH, J T, CLARKE, R T, and BOWES, M J. 2010. Are groundwater nitrate concentrations reaching a turning point in some chalk aquifers? Science of the Total Environment, Vol. 408, 4722–4732.
- WANG, J L, and YANG, Y S. 2008. An approach to catchment-scale groundwater nitrate risk assessment from diffuse agricultural sources: a case study in the Upper Bann, Northern Ireland. Hydrological Processes, Vol. 22, 4274–4286.
- WANG, L, STUART, M E, BLOOMFIELD, J P, BUTCHER, A S, GOODDY, D C, MCKENZIE, A A, LEWIS, M A, and WILLIAMS, A T. 2012. Prediction of the arrival of peak nitrate concentrations at the water table at the regional scale in Great Britain. Hydrological Processes, Vol. 26, 226–239.
- SMITH, D B, WEARN, P L, HICHARDS, H J, and ROWE, P C. 1970. Water movement in the unsaturated zone of high and low permeability strata by measuring natural tritium. IAEA Symposium on Isotope Hydrology, Vienna, 73–87.
- ADDISCOTT, T M. 2005. Nitrate, agriculture and the environment. (Wallingford: CAB International.)
- ADDISCOTT, T M, WHITMORE, A P, and POWLSON, D S. 1991. Farming, fertilizers and the nitrate problem. (Wallingford: CAB International.)
- LORD, E I, JOHNSON, P A, and ARCHER, J R. 1999. Nitrate sensitive areas: a study in large scale control of nitrate loss in England. Soil Use and Management, Vol. 15, 1–27.
- ADAS. 2003. Production and use of nitrogen fertilisers. ADAS UK Ltd Report for Defra Project NT2601 http://randd.defra.gov.uk/Default.aspx?Menu=Menu&Module=More&Location=None&Completed=0&ProjectID=11771#Relate Documents.
- WANG, L, BUTCHER, A S, STUART, M E, GOODDY, D C, and BLOOMFIELD, J P. 2013. The nitrate time bomb: a numerical way to investigate nitrate storage and lag time in the unsaturated zone. Environmental Geochemistry and Health, Vol. 35, 667–681.
- WANG, L, MANSOUR, M, and A H. 2010. Description of the GIS based regional groundwater flow model. British Geological Survey Internal Report, IR/10/070.