During the last 10 years the Dutch national government has, in co-operation with regional and local authorities, implemented an integrated regional policy concept in which region-specific measures could be taken to enhance invironmental quality or the socio-economic structure. However, the effects of measures taken within regional projects to improve environmental quality were largely unknown. The aims were, in general, not defined specifically enough and an ex-ante evaluation of measures to be taken was not carried out. From the year 2000 onward, policy will demand more specific aims and an analysis of the effectiveness of the measures. In this study, applied to the regional project, NUBL (a region in the south of the Netherlands, including large parts of the provinces Brabant and Limburg), the proposed methodology for ex-ante evaluation can only be carried out if the aims of the measures are specified in terms of environmental quality parameters. For example, the aims of the measures to stimulate environmental-friendly production were only specified in economic terms. Furthermore, it is necessary to evaluate the set of measures as a whole, since measures may have synergistic or weakening effects on each other. For example, the reduced application of fertilizers will lower the nitrate concentration in the groundwater, but if the extraction of groundwater is also reduced this effect is strengthened.<br>

Government legislation is directed to decreasing nitrogen loads on agricultural land in the future so as to decrease nitrate in shallow groundwater. We wanted to determine if this nitrate decrease could be measured. Groundwater was sampled annually on 99 farms in the sandy parts of the Netherlands during the period 1992-1995. Mean nitrate concentration in 1995 was half the 1992 concentration, although new laws were not yet effective. This was caused by the increasing amount of annual rain during this period. Correction for such climatological influences is neccassary to establish the effect of new laws. The dilution of a conservative compound was calculated using a soil model for eight sandy parts and real 10-day mean climatological data for the same period. After correction with the calculated dilution, the mean nitrate concentration no longer varied per year. A monitoring network to detect the effects of new laws has been proposed using these results. Statistical and practical considerations, as wel as boundary conditions, stipulated by policy, have been integrated in the proposal.<br>

The National Groundwater Quality Monitoring Network was set up by the RIVM in 1979. Since 1984 groundwater has been sampled every year at about 400 locations at depths of about 10 and 25 metres below land surface. After 1989 the number of locations increased because several Provincial Groundwater Quality Networks joined the National Network. In 1992 groundwater was sampled at about 600 locations. About 25 chemical components and some physical properties were measured. Information on sampling depth, soil- and land use and region in the Netherlands was used to classify the data into groups. This study resulted in estimates of mean concentrations and areal percentages of target value exceedance for the year 1992 and target value exceedance per region presented in the form of maps. The change in groundwater quality per group and region in the period 1984-1993 is also investigated. Correlation analysis was used to examine the relationship between concentrations and time.<br>

The National Groundwater Quality Monitoring Network (RIVM) was used to evaluate groundwater quality in the Netherlands. A cluster is a group of groundwater samples having the same characteristics with respect to sampling depth, land use, soil type, class of groundwater depth and location or region. The soil map, the national landuse database and several regional subdivisions were used to classify the data. The mean and the variance of the nitrate concentrations as well as the exceedance of the nitrate target value in groundwater were predicted for the year 1992. The exceedance of the target value, expressed as areal fractions was predicted for nine regions and presented in the form of maps. The target value exceedance was also predicted for 13 clusters for each year in the period 1984-1992. Finally the change in the concentration of nitrate in groundwater was predicted using regression between concentrations and time over the same period.<br>

Models relating nitrogen supply and nitrate leaching are based upon scientific knowledge and experiments. These deterministic models are used for scenario analysis to reduce nitrate leaching. The models also give a geographical view of nitrate in the uppermost groundwater. Because of the difference in scale between experimental fields and a geographical view, systematic differences between reality (measurements) and model calculations can occur. Our goal is to make a geographical view of the Netherlands for the occurence of nitrate in the uppermost groundwater of agricultural land and of forests, by statistical interpolation of measurements. Indications will be given of magnitude and origin of the differences between deterministic and statistical models.<br>

Nitrate is measured in ground water under forests and heather of the sandy regions in the Netherlands. Possible explanations for higher concentrations are expressed in variables being available in national geographic databases. The maesured concentrations and variables are statistically related. It is plausibel for the concentrations to be influenced by soiltype (an increase in the order ; wet soils, poor soils, and rich soils), atmospheric ammonia deposition and vegetation (an increase by more potential ammonia deposition in the presence of higher vegetation, with more coniferous trees as part of the vegetation) and surface area of forest and heather in respect to other landuse in the surrounding (an increase as agriculture or other landuse has more influence). It is estimated for nitrate to exceed the reference value at least in 46% and at most in 56% of the 287000 hectares. It is also estimated for nitrate to exceed the EEG directive for drinking water at least in 23% and at most in 30% of the 287000 hectares. (Coastal dunes are not considered).<br>

The pH was measured in samples from the National Monitoring Network on Groundwater Quality at the laboratory up to 1989. The pH has been measured in situ and in the laboratory from 1989. To quantify differences results of the in situ pH measurements and the pH measurements in the laboratory are compared. Results of the comparison are presented for the period from 1989 up to 1992. For the accuracy of the pH-measurements in the Monitoring Network on Groundwater Quality a limit of +-0.2 pH units is specified. In situ pH measurements and the pH measurements at the laboratory do not compare well since 35 to 40% of the differences between them are larger than the specification. It is recommended to use the results of the in situ pH measurements from 1992 as primary item in studies in which the chemical composition of groundwater is of interest.<br>

Nitrate concentrations were measured in the uppermost groundwater in the presence of forest and heather in 1989-1990 and in the presence of agriculture in 1992-1993. The measurements were done in the sandy regions of the Netherlands. The measured concentrations were statistically related to variables representing nitrogen load and the soil type. This resulted in a regression model for agriculture and one for forest and heathland. Both models were combined for geographical estimation of nitrate concentrations. Nitrate concentrations in the presence of agriculture are about six times higher then in the presence of forest and heather. The surface groundwater surpassing the groundwater value is more than 94% of the total surface in the presence of agriculture. The surface groundwater surpassing the groundwater value is more than 23% of the total surface in the presence of forest and heathland.<br>

This report contains an integrated survey of the role of nitrogen in the environment. Emissions of several nitrogen components are given, and the fate and transport of nitrogen through the environment (air, soil, groundwater and surface water) is described. The ecological effects in both terrestric and aquatic ecosystems as well as risks for the public health are evaluated. The implications for the public water supply are discussed.<br>

The report documents the results of the four years of a monitoring programme started in 1992 to assess the quality of the upper groundwater in the sandy regions of the Netherlands affected by fertiliser and manure use in agriculture. This programme is a cooperative effort of the National Institute of Public Health and the Environment (RIVM) and the Agricultural Economics Research Institute (LEI-DLO). The upper 100 cm of groundwater occurring within five metres of the surface was sampled at 99 farms (80 cattle and 19 arable) situated in the central, northern, eastern and southern sandy regions of the Netherlands. The samples were analysed for chloride, nitrate, ammonium, potassium, dissolved organic carbon and phosphate (ortho and total). Taken on average, the nutrient load did not change in the study period. The groundwater characteristics, on the contrary, did change as a consequence of the increase in precipitation in the periode 1992-1995. Nitrate and chloride concentrations in the upper metre of the groundwater were halved in this period, while the phosphorus concentration doubled. The potassium and ammonium concentrations showed a slight decrease. The average nitrate concentration in the upper metre of the groundwater under farms in the sandy regions of the Netherlands in the period 1992 - 1995, corrected for precipitation deviation from average, was found to be 158 mg.l-1. This is over three times the limit value of 50 mg.l-1. More than 95% of the farms in the sandy regions have a (precipitation corrected) nitrate concentration higher than the limit value. A relationship was derived between groundwater quality on cattle farms (nitrate and potassium) and nutrient balance surplus, as well as between the percentage of silage maize and groundwater level. On arable farms no relationship could be derived because the number of farms was too small.<br>