Deterministic and stochastic modelling of water, heat and nitrogen dynamics on different scales with WHNSIM

With respect to groundwater quality objectives, it is difficult to understand the nitrogen balance and nitrogen dynamics of agricultural ecosystems without the support of deterministic and stochastic simulation tools. This paper focuses on the deterministic and stochastic modelling of water, heat and nitrogen dynamics in the soil compartments of agricultural ecosystems. The deterministic nitrogen simulation model for agricultural soils presented in this paper, WHNSIM, was originally designed for site conditions typical of southern Germany but can easily be applied to other regions due to its physically based design. Model results are compared with data originating from different experimental sites in Germany. Special consideration is given to the effects of spatial variability. Two different techniques of Monte Carlo simulation are used to study the impact of parameter variability on target variables of nitrogen dynamics on the field scale. Moreover, a procedure is presented that allows modelling nitrogen dynamics on the scale of small catchment areas, combining the geographical information system Arc Info with agricultural management data and model calculations.     

Effects of land use and land cover, stream discharge, and interannual climate on the magnitude and timing of nitrogen, phosphorus, and organic carbon concentrations in three coastal plain watersheds.

In-stream nitrogen, phosphorus, organic carbon, and suspended sediment concentrations were measured in 18 subbasins over 2 annual cycles to assess how land use and land cover (LULC) and stream discharge regulate water quality variables. The LULC was a primary driver of in-stream constituent concentrations and nutrient speciation owing to differences in dominant sources and input pathways associated with agricultural, urban, and forested land uses. Stream discharge was shown to be a major factor that dictated not only the magnitude of constituent concentrations, but also the chemical form. In high discharge agricultural subbasins, where nitrate was the dominant nitrogen form, there was a negative correlation between discharge and nitrate concentration indicating groundwater inputs as the dominant pathway. In urban settings, however, nitrate was positively correlated with discharge, and, in forested subwatersheds, where dissolved organic nitrogen (DON) was the dominant nitrogen form, there was a positive correlation between discharge and DON, indicating washoff from the watershed as the dominant input pathway. Similarly, phosphorus concentrations were strongly regulated by LULC, discharge, and seasonality. This comparative study highlights that different mechanisms regulate different forms of nitrogen, phosphorus, and carbon, and thus field programs or water quality models used for regulatory purposes must assess these nutrient forms to accurately apply management plans for nutrient reductions.     

Intrinsic and induced isoproturon catabolic activity in dissimilar soils and soils under dissimilar land use

The catabolic activity with respect to the systemic herbicide isoproturon was determined in soil samples by (14)C-radiorespirometry. The first experiment assessed levels of intrinsic catabolic activity in soil samples that represented three dissimilar soil series under arable cultivation. Results showed average extents of isoproturon mineralisation (after 240 h assay time) in the three soil series to be low. A second experiment assessed the impact of addition of isoproturon (0.05 microg kg(-1)) into these soils on the levels of catabolic activity following 28 days of incubation. Increased catabolic activity was observed in all three soils. A third experiment assessed levels of intrinsic catabolic activity in soil samples representing a single soil series managed under either conventional agricultural practice (including the use of isoproturon) or organic farming practice (with no use of isoproturon). Results showed higher (and more consistent) levels of isoproturon mineralisation in the soil samples collected from conventional land use. The final experiment assessed the impact of isoproturon addition on the levels of inducible catabolic activity in these soils. The results showed no significant difference in the case of the conventional farm soil samples while the induction of catabolic activity in the organic farm soil samples was significant.     

Ammonia emission factors for N fertilizers applied to two contrasting grassland soils

Ammonia volatilization from nitrogen (N) fertilizer applied throughout the year to two soil types was measured using a system of small wind tunnels. Losses from urea ranged from 12 to 46% of the applied N. Small losses, averaging <1%, were measured from ammonium nitrate (AN) and calcium nitrate applications. Factors influencing these losses are discussed. Using these results and those from other workers, emission factors for urea and AN applications to grassland in the UK were determined as 23.0 and 1.6% of the applied N, respectively. Emission factors for these fertilizers when applied to arable land were estimated as 11.8 and 0.8%, respectively. The emission factor for all other applied N (as straight and compound fertilizers) was assumed to be similar to that for AN. Calculations showed that fertilizer applications to agricultural land in the UK contributes 34 kt NH3-N per year, equivalent to 17% of the total annual NH3 emission.     

Evaluation of groundwater contamination in a coastal area of south-eastern Sicily

This investigation was under taken to evaluate the groundwater resources contamination due to intensive agricultural practices (particularly greenhouses). The study-area is located in the coastal area of the Ragusa province (South-East Sicily), where numerous existing greenhouses may cause the contamination of groundwater systems (unconfined and confined aquifers) beneath the cropped land. The pollution risk is mainly related with the seepage process of macro-elements nitrogen (N), phosphorus (P), potassium (K), held in the irrigation water and the massive use of fertilizers and pesticides, that may pass through the unsaturated zone of the soil profile. Moreover, the area is characterized by the presence of several wells (about 15 wells/km2) for agricultural use that cause the aquifer overexploitation and the consequent risk of seawater intrusion. The agriculture practices adopted in the study area (irrigation volumes, fertilizer concentrations, use of pesticides…) were monitored since February 2009; moreover, the pollution risk of the aquifers was evaluated through the analysis of groundwater water samples collected (monthly) in the monitoring wells; in particular, nitrogen compounds, soluble phosphorous (PO?2?), potassium, as well as the main pesticides commonly used in the study area, were measured.The results show that electrical conductivity and chloride concentration values can cause reduction of production and leaf damage problems, respectively, for most of the monitored farm systems. The high nitrogen compounds concentrations observed in the monitored wells can cause health and environmental problems. Moreover high pesticide contamination of groundwater was found in two of the five monitored wells.     

Arsenic and other heavy metals in soils from an arsenic-affected area of West Bengal,India

Domkal is one of the 19, out of 26 blocks in Murshidabad district where groundwater contains arsenic above 0.05 mg/l. Many millions of cubic meters of groundwater along with arsenic and other heavy metals are coming out from both the hand tubewells, used by the villagers for their daily needs and shallow big diameter tubewells, installed for agricultural irrigation and depositing on soil throughout the year. So there is a possibility of soil contamination which can moreover affect the food chain, cultivated in this area. A somewhat detailed study was carried out, in both micro- and macrolevel, to get an idea about the magnitude of soil contamination in this area. The mean concentrations (mg/kg) of As (5.31), Fe (6740), Cu (18.3), Pb (10.4), Ni (18.8), Mn (342), Zn (44.3), Se (0.53), Mg (534), V (44.6), Cr (33.1), Cd (0.37), Sb (0.29) and Hg (0.54) in fallow land soils are within the normal range. The mean As (10.7), Fe (7860) and Mg (733) concentrations (mg/kg) are only in higher side whereas Hg (0.17 mg/kg) is in lower side in agricultural land soils, compared to the fallow land soils. Arsenic concentrations (11.5 and 28.0 mg/kg respectively) are high in those agricultural land soils where irrigated groundwater contains high arsenic (0.082 and 0.17 mg/l respectively). The total arsenic withdrawn and mean arsenic deposition per land by the 19 shallow tubewells per year are 43.9 kg (mean: 2.31 kg, range: 0.53-5.88 kg) and 8.04 kg ha(-1) (range: 1.66-16.8 kg ha(-1)) respectively. For the macrolevel study, soil arsenic concentration decreases with increase of distance from the source and higher the water arsenic concentration, higher the soil arsenic at any distance. A proper watershed management is urgently required to save the contamination.     

Fate of Triclosan and Methyltriclosan in soil from biosolids application

This study investigates the persistence of Triclosan (TCS), and its degradation product, Methyltriclosan (MeTCS), after land application of biosolids to an experimental agricultural plot under both till and no till. Surface soil samples (n = 40) were collected several times over a three years period and sieved to remove biosolids. Concentration of TCS in the soil gradually increased with maximum levels of 63.7 ± 14.1 ng g(-1) dry wt., far below the predicted maximum concentration of 307.5 ng g(-1) dry wt. TCS disappearance corresponded with MeTCS appearance, suggesting in situ formation. Our results suggest that soil incorporation and degradation processes are taking place simultaneously and that TCS background levels are achieved within two years. TCS half-life (t(0.5)) was determined as 104 d and MeTCS t(0.5), which was more persistent than TCS, was estimated at 443 d.     

Sulfamethazine and flubendazole in seepage water after the sprinkling of manured areas

Pharmaceuticals are widely used in modern livestock production and can reach the environment via the application of manure containing excreted drugs. Limited information is available on the transport and fate of veterinary medicines applied to soils. Therefore, we assessed the potential for the sulphonamide antibiotic sulfamethazine (SMT) and the antiparasitic drug flubendazole (FLUB) including their metabolites to move from agricultural manure to drainage waters at 1m depth. A comparison was made of losses from sites under different land use (grassland versus arable cropping) as well as losses from neighbouring plots under the same land use. Liquid manure from pigs treated with SMT and FLUB was spread on 10 x 30 m2 plots (750l per plot). SMT concentration in slurry ranged from 600 to 1700 microg l(-1) (metabolite acetyl-SMT 280-1700 microg l(-1)) and FLUB concentration ranging from 25 to 56 microg l(-1) (metabolite amino-FLUB 32-110 microg l(-1), hydroxy-FLUB 19-38 microg l(-1)). About 1h after application heavy rainfall (50mm in 2.5h) was simulated by sprinkler irrigation. Drainage flow started within 1h after the commencement of sprinkling. SMT and FLUB concentrations in leachate reached values of up to 16 microg l(-1) and 0.3 microg l(-1), respectively. Loss rates (relative to the applied amounts) from the neighbouring sites under arable cropping ranged from 2.8% to 5.4% for SMT and 0.8% to 3.1% for FLUB (including metabolites). On the permanent grassland plot, due to its multitude of macropores, loss rates reached values up to 10% for SMT and 16% for FLUB (including metabolites). These results demonstrate that the variability in leaching of veterinary drugs may be high even between large neighbouring plots, depending on site heterogeneity and land use.     

A screening method for preliminary assessment of risk to groundwater from land-applied chemicals

A simple mathematical model for initial screening is presented that can aid in evaluating the relative risk to groundwater from applying nonpolar synthetic organic chemicals to soil. The basic premise is that the magnitude of the quotient of the chemical concentration of the water entering the aquifer and the maximum allowable concentration (as established by EPA or Health Departments) represents the health risk of a chemical. The chemical concentration of the soil water is estimated based on conservative, simplifying assumptions and requires only readily available data such as: basic soil properties (organic matter and saturated hydraulic conductivity), organic chemical properties (octanol-water partition coefficient and degradation rate) and environmental factors (recharge rate and depth to groundwater).The methodology was applied to assess the relative risk of organic chemicals in municipal sewage sludge and pesticides applied to agricultural land. The results are realistic.     

Assessment of Aquifer Vulnerability in an Agricultural Area in Spain Using the DRASTIC Model

The assessment of aquifer vulnerability is a very important task, especially in agricultural areas because the quality and availability of groundwater affects both the sustainability of agriculture and the quality of life. In this study, an integrated approach is considered, with the use of the generic and agricultural DRASTIC models as well as a geographic information system (GIS), to assess groundwater vulnerability in the agricultural area of Barrax, in the province of Albacete, in Spain. Seven parameters—depth to water, net recharge, aquifer media, soil media, topography, impact of vadose zone media, and hydraulic conductivity of the aquifer (DRASTIC)—have been considered as weighted layers to enable an accurate groundwater risk mapping. The results of the generic DRASTIC model indicated very low vulnerability to contamination for Barrax groundwater due to limited urban and industrial development in the wider area. However, agricultural activities impose pressure to groundwater resources and the results of the agricultural DRASTIC model show that 6.86% of the study area is characterized by very high, 2.29% by high, 47.28% by medium, 38.28% by low, and the remaining 5.29% by no vulnerability to groundwater contamination. The distribution of nitrate concentration in groundwater in the area under study is quite well correlated with the agricultural DRASTIC vulnerability index. Sensitivity analysis was also performed to acknowledge statistical uncertainty in the estimation of each parameter used, to assess its impact, and thus to identify the most critical parameters that require further investigation. Depth to water and impact of vadose zone are the parameters that had the most noticeable impact on the generic DRASTIC vulnerability index followed by the soil media and topography. In contrast, the agricultural DRASTIC method is more sensitive to the removal of the depth to water parameter followed by the topography and the soil media parameters.     

Irrigation of treated wastewater in Braunschweig, Germany: an option to remove pharmaceuticals and musk fragrances

In this study the fate of pharmaceuticals and personal care products which are irrigated on arable land with treated municipal wastewater was investigated. In Braunschweig, Germany, wastewater has been irrigated continuously for more than 45 years. In the winter time only the effluent of the sewage treatment plant (STP) of Braunschweig is used for irrigation, while during summer digested sludge is mixed with the effluent. In the present case study six wells and four lysimeters located in one of the irrigated agricultural fields were monitored with regard to the occurrence of 52 pharmaceuticals and two personal care products (PPCPs; e.g. betablockers, antibiotics, antiphlogistics, carbamazepine, musk fragrances, iodinated contrast media (ICM) and estrogens). No differences in PPCP pollution of the groundwater were found due to irrigation of STP effluents with and without addition of digested sludge, because many polar compounds do not sorb to sludge and lipophilic compounds are not mobile in the soil-aquifer. Most of the selected PPCPs were never detected in any of the lysimeter or groundwater samples, although they were present in the treated wastewater irrigated onto the fields. In the groundwater and lysimeter samples primarily the ICM diatrizoate and iopamidol, the antiepileptic carbamazepine and the antibiotic sulfamethoxazole were detected up to several mugl(-1), while the acidic pharmaceuticals, musk fragrances, estrogens and betablockers were likely sorbed or transformed while passing the top soil layer. Potential estrogenic effects are likely to disappear after irrigation, since the most potent steroid estrogens were not measurable.     

Nutrient losses by surface run-off following the application of organic manures to arable land. 2. Phosphorus

Phosphorus (P) surface run-off losses were studied following organic manure applications to land, utilising a purpose-built facility on a sloping site in Herefordshire under arable tillage. Different rates and timing of cattle slurry, farm yard manure (FYM) and inorganic nitrogen (N) and P fertiliser were compared, over a 4-year period (1993-97). N losses from the same studies are reported in a separate paper. The application of cattle FYM and, especially slurry, to the silty clay loam soil increased both particulate and soluble P loss in surface water flow. Losses via subsurface flow (30 cm interflow) were consistently much lower than via surface water movement and were generally unaffected by treatment. Increased application of slurry solids increased all forms of P loss via surface run-off; the results suggested that a threshold for greatly increased risk of P losses via this route, as for N, occurred at ca. 2.5-3.0 t/ha solids loading. This approximates to the 50 m3/ha application rate limit suggested for slurry within UK 'good agricultural practice'. The studies also provided circumstantial evidence of the sealing of the soil surface by slurry solids as the major mechanism by which polluting surface run-off may occur following slurry application on susceptible soils. Losses of total and soluble P, recorded for each of the 4 years of experiments, reached a maximum of only up to 2 kg/ha total P (TP), even after slurry applications initiating run-off. Whilst these losses are insignificant in agronomic terms, peak concentrations of P (up to 30,000 micrograms/l TP) in surface water during a run-off event, could be of considerable concern in sensitive catchments. Losses of slurry P via surface run-off could make a significant contribution to accelerated eutrophication on entry to enclosed waters, particularly when combined with high concentrations of NO3(-)-N. Restricting slurry application rates to those consistent with good agronomic practice, and within the limits specified in existing guidelines on good agricultural practice, offers the simplest and most effective control measure against this potentially important source of diffuse pollution.     

Water quality dynamics and hydrology in nitrate loaded riparian zones in the Netherlands

Riparian zones are known to function as buffers, reducing non-point source pollution from agricultural land to streams. In the Netherlands, riparian zones are subject to high nitrogen inputs. We combined hydrological, chemical and soil profile data with groundwater modelling to evaluate whether chronically N loaded riparian zones were still mitigating diffuse nitrate fluxes. Hydraulic parameters and water quality were monitored over 2 years in 50 piezometres in a forested and grassland riparian zone. Average nitrate loadings were high in the forested zone with 87 g NO(3)(-)-N m(-2) y(-1) and significantly lower in the grassland zone with 15 g NO(3)(-)-N m(-2) y(-1). Groundwater from a second aquifer diluted the nitrate loaded agricultural runoff. Biological N removal however occurred in both riparian zones, the grassland zone removed about 63% of the incoming nitrate load, whereas in the forested zone clear symptoms of saturation were visible and only 38% of the nitrate load was removed.     

Environmental fate studies of aldicarb and aldoxycarb applications to Wisconsin potatoes

Aldicarb and aldoxycarb pesticides were applied to potato fields in central Wisconsin to study the degradation and movement of their carbamate residues within the soil profile. Aldicarb and aldoxycarb residues degraded at similar rates with half-lives ranging from 0.9 to 1.4 months. Although unsaturated-zone residue measurements indicated relatively minor differences among the three plots, residues were detected in shallow groundwater beneath plots treated with aldicarb or aldoxycarb at planting. Residues beneath the plot treated with aldicarb at emergence were much lower and the timing of these residues suggests they might have resulted from migration of residues from the plot treated with aldicarb at planting. This study confirms the results of previous and concurrent research programs which suggest that emergence applications of aldicarb reduce the potential for residues to enter shallow groundwater without reducing insect control or potato yields.     

Influence of the nitrification inhibitor 3,4-dimethylpyrazole phosphate (DMPP) on ammonia-oxidizing bacteria and archaea in rhizosphere and bulk soil

In agricultural plant production nitrification inhibitors like 3,4-dimethylpyrazole phosphate (DMPP) are used to retard the microbial nitrification process of fertilized ammonium to enhance the nitrogen supply for cultivated crops and to reduce nitrogen losses from the production system. Besides the well-known ammonia-oxidizing bacteria (AOB) it is known for a few years that also ammonia-oxidizing archaea (AOA) are able to perform the first step in nitrification, hence being also a target for a nitrification inhibitor. However, so far no information are available concerning the effectiveness of DMPP and its extent towards AOB and AOA, neither in bulk soil nor in the root-rhizosphere complex. We investigated in a field experiment performed according to agricultural practice the effect of DMPP on the abundance of AOB and AOA two, four and eight weeks after fertilization. We observed impaired abundances of AOB but not of AOA in both soil compartments that were still visible eight weeks after application, possibly indicating a reduced effectiveness of the nitrification inhibitor in our study.     

Investigations of nitrogen fluxes and pools on a limestone site in the Alps

In the North Tyrolean Limestone Alps a site was investigated over a four-year period (1998–2001) in order to assess the nitrogen saturation status, the nitrogen budget (quantification of the net uptake of nitrogen by the canopy and of the nitrogen mineralization, nitrogen uptake from roots and N₂O emission rates, proof of the origin of nitrate in the soil water with stable isotope analyses), and the effects of the actual nitrogen input on ground water quality. The main goals were to quantify the nitrogen input rate, the nitrogen pools in above-ground and below-ground compartments, nitrogen turnover processes in the soil as well as the output into the groundwater and into the atmosphere. The findings are based on continuous and discontinuous field measurements as well as on model results.While nitrogen input exceeded the Critical Loads of the WHO (1995), nitrogen deficiency and nutrient imbalances were verified by needle analyses. The atmospheric input of inorganic nitrogen was higher than the nitrogen output in 50 cm soil depth. A tracer experiment with¹⁵N helped to prove that not more than half of the applied nitrate could be discharged. This allows the conclusion that nitrogen is stored in the system and that the site cannot yet be said to be saturated with nitrogen. The same result was also obtained by modelling. In addition, it was proved that the nitrogen discharge did not stem from deposition but from processes within the system.     

Groundwater pollution source apportionment using principal component analysis in a multiple land-use area in southwestern China

Identification of different pollution sources in groundwater is challenging, especially in areas with diverse land uses and receiving multiple inputs. In this study, principal component analysis (PCA) was combined with geographic information system (GIS) to explore the spatial and temporal variation of groundwater quality and to identify the sources of pollution and main factors governing the quality of groundwater in a multiple land-use area in southwestern China. Groundwater samples collected from 26 wells in 2012 and 38 wells in 2018 were analyzed for 13 water quality parameters. The PCA results showed that the hydro-geochemical process was the predominant factor determining groundwater quality, followed by agricultural activities, domestic sewage discharges, and industrial sewage discharges. Agriculture expansion from 2012 to 2018 resulted in increased apportionment of agricultural pollution. In contrast, economic restructure and infrastructure improvement reduced the contributions of domestic sewage and industrial pollution. Anthropogenic activities were found the major causes of elevated nitrogen concentrations (NO₃^?, NO₂^?, NH₄⁺) in groundwater, highlighting the necessity of controlling N sources through effective fertilizer managements in agricultural areas and reducing sewage discharges in urban areas. The applications of GIS and PCA successfully identified the sources of pollutants and major factors driving the variations of groundwater quality in tested years.

     

Biodegradation of wastewater nitrogen compounds in fractures: laboratory tests and field observations

Throughout several coastal regions in the Mediterranean where rainfalls rarely exceed 650 mm per year municipal treated wastewater can be conveniently reused for soil irrigation. Where the coastal aquifer supplies large populations with freshwater in such area, an assessment of ground water quality around spreading sites is needed. In this study, the efficacy of natural filtration on nitrogen degradation in wastewater spreads on the soil covering the Salento (Southern Italy) fractured limestone was quantified by using laboratory tests and field measurements. In the laboratory, effluent from municipal wastewater treatment plants was filtered through a package of fractures made by several slabs of limestone. An analysis of wastewater constituent concentrations over time allowed the decay rates and constants for nitrogen transformation during natural filtration to be estimated in both aerated and non-aerated (i.e., saturated) soil fractures. A simulation code, based on biodegradation decay constants defined in the laboratory experiments, was then used to quantify the total inorganic nitrogen removal from wastewater injected in an aquifer in the Salento region (Nardò). Here the water sampled in two monitoring wells at 320 m and 500 m from the wastewater injection site and downgradient with respect to groundwater flow was used to verify the laboratory nitrification and denitrification rates.     

Past,present, and future use of phosphorus in Chinese agriculture and its influence on phosphorus losses

Large inputs of phosphorus (P) in chemical fertilizers and feed supplements since 1978 have improved soil P status in arable land in China, but have also created challenges by increasing P concentrations in manure and exacerbating water quality degradation. Arable land in China can be divided into five management zones based on soil P chemistry, with 15–92 % of arable land having lower P status than the agronomic optimum and 0.3–7.2 % having severe risks of P leaching losses. A scenario analysis of soil P budget and agronomic P demand during 2011–2030 highlighted the great pressure China faces in sustainable P management and the need for drastic changes in current practices. This includes new policies to reduce P supplementation of feed and improved P use efficiency by livestock and programs to expand the adoption of appropriate fertilization, soil conservation, and drainage management practices to minimize P losses.     

Predicting groundwater nitrate concentrations in a region of mixed agricultural land use: a comparison of three approaches.

We investigated whether nitrate-N (NO3(-)-N) concentrations of shallow groundwater (< 30 m from the land surface) in a region of intensive agriculture could be predicted on the basis of land use information, topsoil properties that affect the ability of topsoil to generate nitrate at a site, or the 'leaching risk' at different sites. Groundwater NO3(-)-N concentrations were collected biannually for 3 years at 88 sites within the Waikato Region of New Zealand. The land use was classed as either the predominant land use of the farm where the well or bore was located, or the dominant land use within a 500 m radius of the well or bore. Topsoil properties that affect the ability of soil to generate nitrate were also measured at all the sites, and a leaching risk assessment model 'DRASTIC' was used to assess the risk of NO3(-)-N leaching to groundwater at each site. The concentration of NO3(-)-N in shallow groundwater in the Waikato Region varied considerably, both temporally and spatially. Nine percent of sites surveyed had groundwater NO3(-)-N concentrations exceeding maximum allowable concentrations of 11.3 ppm recommended by the World Health Organisation for potable drinking water which is accepted as a public health standard in New Zealand. Over half (56%) of the sites had concentrations that exceeded 3 ppm, indicating effects of human activities (commonly referred to as a human activity value). Very few trends in NO3(-)-N concentration that could be attributed to land use were identified, although market garden sites had higher concentrations of NO3(-)-N in underlying groundwater than drystock/sheep sites when the land use within 500 m radius of a sampling site was used to define the land use. There was also some evidence that within a district, NO3(-)-N concentrations in groundwater increased as the proportion of area used for dairy farming increased. Compared to pastoral land, market gardens had lower total C and N, potentially mineralisable N and denitrifying enzyme assay. However, none of these soil properties were directly related to groundwater NO3(-)-N concentrations. Instead, the DRASTIC index (which ranks sites according to their risk of solute leaching) gave the best correlation with groundwater NO3(-)-N concentrations. The permeability of the vadose zone was the most important parameter. The three approaches used were all considered unsuitable for assessing nitrate concentrations of groundwater, although a best-fit combination of parameters measured was able to account for nearly half the variance in groundwater NO3(-)-N concentrations. We suggest that non-point source groundwater NO3(-)-N contamination in the region reflects the intensive agricultural practices, and that localised, site-specific, factors may affect NO3(-)-N concentrations in shallow groundwaters as much as the general land use in the surrounding area.