5. Representativeness of Current Monitoring Programmes

5. REPRESENTATIVENESS OF CURRENT MONITORING PROGRAMMES

5.1 Definition of representativeness

Representativeness can be expressed in a number of ways. For example, it can relate to how well represented or quantified are the water resources of a particular country in terms of the total national water resource or total EEA resource. Alternatively it could relate to how well a particular water problem had been quantified, for example acidification of small streams or eutrophication of lakes. There are procedures that can be applied to determine how statistically representative the ‘sampled’ population’ is of the ‘total population’. In this sense population refers to the total number of the water type being assessed (e.g. all small rivers, all lakes). Alternatively the representativeness may be expressed in relation to the international requirements and obligations for monitoring (MW1) or against the information requirements of the Agency.

Since much of the source information for the first type of assessment is generally lacking or difficult to obtain in the time scales available (e.g. numbers of small streams, temporal and spatial variability of determinands within and between water bodies) the emphasis in this section has been on the latter, more superficial, type of assessment of representativeness. The main source of information for the assessments was responses to the questionnaires circulated to each EEA State through the National Focal Points. The information and responses obtained from the questionnaires were variable, and any such shortcomings are indicated in the following text. The issue of statistical representativeness is discussed in more detail in Appendix A. Again, the section has been divided somewhat arbitrarily into sub-sections of surface waters quality and quantity, and groundwater quality and quantity. Where there are clear overlaps these are indicated in the text. The sub-sections summarise the main conclusions of work carried by ETC members which is reported in full detail in the Project Record (ETC 1995).

5.2 Surface water quality

This task drew heavily on the work on surface water quality monitoring networks initially funded by DGXI but then also supported by the EEA as part of the 1994 subvention. This section evaluates the representativeness of existing national river and lake monitoring programmes. Focus has especially been put on description of network design (number of sampling sites).

5.2.1 Rivers

Nearly all the countries in the EEA area have a national rivers monitoring programme generally based on chemical and physical indicators of quality. Additionally some international programmes, such as the EU Exchange of Information Decision (77/795/EEC), and the OECD and GEMS/WATER networks, focus on chemical and physical water quality primarily of large rivers. Thirty-one monitoring programmes of this type have been identified, these also include more specific monitoring programmes such as, for instance, monitoring of transboundary rivers and estimation of loadings into coastal areas.

The river monitoring networks can be divided into three categories according to their main purpose:

1. General characterisation of rivers and streams in a country.
2. Monitoring of water quality of rivers draining specific areas such as, for instance, reference sites in forested or uncultivated areas, or leaching of substances from agricultural watersheds.
3. Networks designed to estimate the riverine loading from land into coastal areas, or the loading of transboundary rivers from one country to a neighbouring country.

Many monitoring networks are multi-purpose and may be assigned to more than one of the above categories. The results from a network may, for instance, be used both to make a general characterisation of river water quality and to estimate the nutrient loading of coastal areas.

There are 20 monitoring programmes which have networks specifically designed to elaborate a general characterisation of rivers and streams in a country. Most of these networks are based on more than 100 sampling sites located on all major river systems and rivers in a country. According to most of these programmes, samples are taken annually with a sampling frequency ranging from 4 to 26 sample per year. The number of variables measured varies from 4 to 120, but all programmes generally include the determination of basic variables (e.g. pH, conductivity, water temperature), organic pollution indicators (e.g. dissolved oxygen, BOD), nutrients and suspended solids. Many programmes also include determination of specific ions (e.g. chloride, sulphate, calcium) and heavy metals. Additionally, the determination of more specific contaminants such as organic micropollutants and radionuclides is included in some monitoring programmes.

The areal density of sampling sites varies from one sampling site per 10,000 km2 to more than five sampling sites per 1,000 km2, with 1 to 2 sampling sites per 2,000 km2 generally being found. The density of sampling sites in relation to population varies from 2 to 500 sites per million inhabitants. Each sampling site also represents from 6 to 6,000 km of river. The river length used in this calculation was based on Morris and Kronvang (1994) and only includes rivers mapped at a 1:50,000 scale.

In the Nordic countries, Denmark, Finland, Norway and Sweden, there are monitoring networks with the purpose of monitoring water quality and loading from specific catchments. These monitoring networks generally consist of up to 20 relatively small stream catchments and involve detailed integrated studies of both river water quality and of the catchment (for example, land use, soil type). The main purposes of these networks are to monitor reference areas, loadings from agricultural land or the impact of acid precipitation.

Many monitoring networks are established to estimate the riverine loading of contaminants from land to sea, or in transboundary rivers. Generally these networks consist of sampling sites located at downstream points in all major river systems. Those countries that have a relatively long coastline compared to their area, for example, the United Kingdom, Ireland, Norway, Sweden, Denmark and Greece, generally have a large number of relatively small river systems. Consequently, the number of sampling sites needed to estimate loads to coastal areas is high, whereas fewer sampling sites are required in countries dominated by a few large river systems. In Denmark, for instance, sampling is undertaken in 124 river systems which equates to the loading to sea from around 60 % of the land area, whilst sampling undertaken in the eight largest Spanish rivers equates to approximately 75 % of the loading from the Spanish land area. The sample analysis programmes generally include determination of nutrients and suspended matter. Additionally, loading by heavy metals and organic han natural lakes. In Spain, for instance, there are more than 1,000 large reservoirs.

Only a few countries in the EEA area have national monitoring programmes for the assessment of the chemical and physical water quality of lakes. Some countries, however, undertake local monitoring of lakes. The German Federal States (Länder), for instance, monitor the environmental state of lakes in their respective areas. Local lake monitoring activities are generally not standardised at a national level, and the variables and sampling frequency vary. During the last 10 to 15 years some countries have made national lake inventories and collected data and elaborated reports on the general environmental state of lakes based on locally gathered information. In the Nordic countries, in which there are many natural lakes, monitoring programmes cover a vast number of lakes. Some countries have a long tradition for monitoring large nationally important lakes, Austria has, for example, monitored Lake Constance and Neusiedler See since 1961 and 1972, respectively, and the Norwegian Lake Mjösa has been studied since 1971. Several countries, for example the Netherlands and Portugal, do not have specific lake monitoring programmes, but include their lakes in river or inland water programmes.

The number of determinants measured is generally in the order of 20 to 30. Most programmes include determination of basic variables (temperature, pH, conductivity, dissolved oxygen), organic pollution indicators (total organic carbon, biochemical oxygen demand, chemical oxygen demand), eutrophication indicators (nitrogen and phosphorus species, chlorophyll-a, Secchi depth) and major specific ions (Ca, Mg, Na, K, etc.). Some countries also include determination of heavy metals (Finland, Sweden).

The general national lake monitoring programmes can be divided into two categories: the ‘survey-type’ and ‘intensive’ programmes. The ‘survey-type’ programme typically covers a large number of lakes that are sampled at long intervals. Examples of this type are found in Norway and Sweden, and include around 1,000 lakes in each. In Ireland a national lake survey was performed in the period 1987 to 1990, and included a total of 170 large lakes and some representative smaller lakes. Additionally, a remote sensing survey of 360 Irish lakes was performed in 1989 and 1990. More intensive monitoring programmes with a sampling frequency of several times a year (e.g. in Denmark and Sweden) typically cover a smaller number of lakes. Survey-type lake monitoring programmes provide a general description of the environmental state of a wide range of lakes, whereas more frequent monitoring provides information on dynamics and seasonal variation that may be used to detect trends.

Biological variables are part of the sampling routine of many general lake monitoring programmes as well as programmes concerning specific localities. One Finnish programme includes only biological variables. Sampling and investigation of phytoplankton and zooplankton are components of several monitoring programmes. Apart from a general evaluation of the phytoplankton community, the objectives of some programmes are more specific such as assessment of the occurrence of potentially toxic blue-green algae in waterbodies used for bathing or drinking water supply (for example in the UK). Bottom fauna (invertebrates), macrophytes and fish are also studied in some of the lake monitoring programmes.

The monitoring of acidification effects is co-ordinated internationally in an ECE programme, the International Co-operative Programme on Assessment and Monitoring of Acidification of Rivers and Lakes, arising from the Convention on Long-range Transboundary Air Pollution. Twelve European countries, Canada and USA participate and report chemical and biological variables to the programme centre.

National acidification monitoring programmes are restricted to the countries affected by surface water acidification. Finland, Norway and Sweden have, for instance, a long tradition of assessment of surface water acidification. The acidification monitoring programmes can be divided into:

• nation-wide surveys to assess the extent of acidification; and,
• monitoring programmes involving detailed studies of a few catchments with the purpose of understanding the process of acidification and analysing trends.

Norway, Finland and Sweden each have nation-wide surveys with the purpose of assessing the extent of acidification. The surveys include national sampling of more than 1,000 lakes and are generally performed at intervals of five to ten years. In 1995 a co-ordinated lake acidification survey will be performed in each of the three countries. The countries also take annual samples in fewer lakes; 176 lakes in Finland, 100 in Norway and 85 lakes in Sweden. In addition, a number of small streams are sampled. The annual programmes are used for analyses of acidification trends. One sample is taken from each waterbody both in the survey programme and in the annual programme, and it is then analysed for general acidification variables. The most common variables for all the monitoring programmes are pH, conductivity, alkalinity, total organic carbon, nitrate, four major cations (potassium, calcium, magnesium, and sodium) and anions (sulphate and chloride), and various aluminium fractions. Some monitoring programmes also include measurements of total phosphorus, total nitrogen and ammoniacal nitrogen. The extent of acidification is also assessed using of various biological indicators such as zoobenthos, phytoplankton, and fish.

Seven integrated acidification monitoring programmes are in operation in five countries: Norway, Sweden, Finland, the UK, and Ireland. Generally the programmes include extensive investigations of a limited number of waterbodies or catchments and involve frequent sampling and determination of many variables. Chemical analyses of surface water samples are made and in some cases also of precipitation and groundwater. Water samples are analysed for all the previously mentioned acidification variables. In some of the programmes detailed studies of the biological communities are also performed, examples being studies of macroinvertebrates in streams and the littoral zone of lakes, as well as studies of phytoplankton, macrophytes and fish. In some lakes the record of acidification is reconstructed by use of palaeolimnological indicators (primarily diatoms).

5.3 Surface water quantity

There a lack of sufficiently detailed responses to the MW2 questionnaires to enable a thorough assessment of representativeness of flow gauging stations to be made. Only nine countries had given reasonably detailed information on specific gauging stations by the end of September. From these only five provided information about the altitude of the flow gauges and only one answered all questions giving information on the maximum altitude of each basin draining to a flow gauge, average catchment precipitation and average flow. Some of the countries gave only geographical information (names and co-ordinates) about certain gauges with no information as to the period of activity. However some weak points can be detected, the biggest being the relatively high concentration of gauges in the lowlands. Although also incomplete, the information on monitoring frequency and geographical spread of gauges is currently being analysed.

5.4 Groundwater quality

5.4.1 Summary of information

This evaluation was based on the MW2 returned questionnaires and showed that the objectives ‘general surveillance purpose’ and ‘water quality trend identification’ are part of all monitoring systems. Other objectives include assessment of compliance with national or EC legislation for example the control of drinking water quality (80/778/ECC) or monitoring for compliance with the EC Nitrates Directive (91/676/ECC). Monitoring systems in Italy and France are based on requirements for monitoring drinking water quality. Some other important purposes for groundwater quality monitoring in the EEA area include detection of sea water intrusion and evaluation of impacts caused by airborne pollutants. A sea water intrusion monitoring network has been installed in Spain and this problem is also the subject of investigations in UK and Portugal. The monitoring of impact from airborne pollutants in relation to acidification problems is mostly limited to the northern part of the EEA area (e.g. Norway).

Most monitoring networks include sampling sites which are distributed evenly within the whole groundwater area and/or are concentrated around drinking water wells. The objectives of these two network types may well differ but certainly both differ from the networks which are based on sampling sites concentrated around impacted areas.

Large differences exist between the monitoring networks as far as the number of sites and investigated areas are concerned. This is not only due to the different hydrogeological situations in the EEA area but also due to the different objectives (for example, impact or baseline station network, identification of effects from airborne pollution or effects from agricultural land use). The total national observation area, for example, for groundwater in porous media varies from 35 to 79,258 km² over the EEA countries, and sampling site density varies from 0.004 to 0.57 sites/km². The variation in the equivalent figures for karst groundwater and other groundwater is also very high.

Groundwater quality parameters can be divided into following groups:

1. Descriptive parameters (e.g. conductivity, pH, turbidity, odour);
2. Major ions (e.g., Ca, Mg, Na, K, N0₃, N0₂, NH₄, Cl, S0₄, HCO₃);
3. Additional parameters (e.g. DOC, boron, fluorine, cyanide, hydrocarbon benzene);
4. Heavy metals (e.g. Cr, Pb, Cd, Hg, Ni);
5. Organic substances including chlorinated solvents (e.g. trichloroethene; tetrachloroethene; 1,1,1 trichloroethane; 1,1-dichloroethene);
6. Pesticides (herbicides, insecticides).

In most monitoring programmes descriptive parameters and major ions are analysed, heavy metals are also often measured. However, there is a difference between the programmes in the case of chlorinated solvents and pesticides. The total number of organic substance determinands varies from 15 to 106 compared to 1 to 64 for pesticides. A further figure for comparison is the number of determinands which are included in ‘basic programmes’, these vary from 14 to 51.

The lowest sampling frequency for basic programmes is once every 2 years compared to the highest frequency of 12 times per year. These differences are due to differences in monitoring purpose or objectives (e.g. specific networks with small number of sampling sites, high sampling frequency and small number of variables).

Little useful information was received on the limits of detection achieved for the determinands. In some cases values above detection limit are only given as an order of magnitude value (mainly in monitoring systems for drinking water), If there are no exact values, statistical evaluations for example analysis of time series, are difficult.

5.4.2 Conclusions

The evaluation of information shows that national or regional groundwater quality monitoring networks in the EEA area have very different purposes and objectives. Because of this the structure and design of the networks are different. This means that in the EEA area existing individual groundwater quality monitoring networks can only be regarded as representative at their national level.

Evaluation of data arising from the existing, highly different monitoring networks in the EEA area (different objectives and different criteria for selection of sampling sites) will not give the reliable results which are needed by the EEA. With regard to the information needs of the EEA, it should be noted that the data are not entirely comparable and would certainly lead to wrong conclusions.

A three step approach could bring together existing, monitoring networks and data needs of EEA.

1. Lay down strategies for EEA groundwater quality monitoring.
2. Analyse which parts of existing national monitoring networks can be used.
3. Establishment of additional elements either in existing or new monitoring networks.

Point 2 and 3 can only be carried out in close co-operation with national institutions. This should be undertaken as part of the pilot implementation proposed for 1996.

5.5 Ground water quantity

5.5.1 Introduction

The information obtained on groundwater quantity monitoring by the MW2 questionnaire was somewhat limited and patchy for some items and for most countries, which has restricted the scope of this task. For example information that would help to understand and quantify links between the state and pressures arising from human activities, or to assess the state of the aquatic environment was not reported. Therefore, because of this lack of information and data, it is only possible to make a very broad assessment of representativeness based on very simple indicators. It should be noted that no information on monitoring programmes for Greece, Belgium and Luxembourg have been received. Furthermore, Italy did not provide information about sampling density and frequency of observations. Also due to organisational responsibilities 2 countries (France and Germany) have responded though regional authorities.

5.5.2 Objectives of monitoring

It should be noted that no current EC directive addresses specific requirements for groundwater quantity monitoring. Many monitoring networks are multi-purpose: About 90% are oriented to collect basic data, 59% for management purposes and 41% for scientific research purposes. In the context of this last objective the Alsace region of France uses its piezometric network to supply data in order to maintain the mathematical model of the Alsace aquifer. Two countries, Portugal and Spain, have specific networks designed to monitor saltwater intrusion in coastal aquifers. Within the ‘management’ objectives, 31% of countries/regions use a piezometric network to help the assessment of compliance with national legislation and to a lesser extent for EC legislation and transboundary obligations (Austria, Thüringen - Germany and Portugal). The Nord Pyreneen region of France collects piezometric information from programmes that have other purposes, for example, monitoring underground gas reservoirs.

5.5.3 Sample site density

In porous aquifers the areal density of sampling sites varies from 30 to 40 per 100 km² (Austria and the North Rhine Westfalia) to a minimum of 0.1 per 100 km² in Ireland. Almost all countries have observation sites evenly distributed within the whole groundwater area and very few concentrated around drinking water wells (Portugal and North Rhine Westfalia, Germany) or impact areas. Ireland and UK have some observation sites non-specifically located.

In karstic groundwater the areal density of sampling sites varies from 33 per 100 km² (North Rhine Westfalia, Germany) to a minimum of 0.066 per 100 km² in Austria. Almost all the countries have the observation sites evenly distributed over the whole groundwater area and only Portugal has observation sites concentrated around drinking water wells. It should be noted that as most karstic regions are located in mountainous areas, it is not possible to achieve an even site distribution, therefore, sampling stations tend to be situated in flat areas. Again Ireland and UK have some observations sites non-specifically located. In other aquifer systems the areal density of sampling sites varies from 1,000 per 100 km² (Finland) to a minimum of 0.6 per 100 km² in North Rhine Westfalia, Germany. Almost all the countries have the observations sites evenly distributed over the whole groundwater area. Germany, Spain, Great Britain and Ireland have some observations sites non- specifically located.

In order to know the history of the groundwater system, in particular the identification of short- and long-term changes in groundwater quantity, an appropriate sampling frequency must be used which in itself depends on the hydrogeological system being monitored and its interaction with other systems. Sampling frequencies (e.g. for water levels, temperature) vary a great deal from country to country, from a daily basis in Thüringen, Germany to 2 to 6 times per year in Spain, for basic programmes, and from continuous sampling (Bavaria, Germany) to a 2 to 3 times per year (Portugal), in special programmes. The earliest quantitative groundwater records are from the UK (1845). Only the Netherlands has a complete sampling frequency programme with observation in levels, temperature and other variables near to the proposed requirements. There is less information on the frequency of sampling springs to monitor the water level, the temperature, the discharge, the conductivity and other variables. Springs are also sampled in basic and special programmes. There is also no information with regard to the control of the seasonal changes in frequency of observations in areas with marked changes in groundwater levels during the year (e.g. in water supply areas of tourist centres, areas with important water abstraction).

5.5.4 Evaluation of representativeness

This can be analysed at two different scales: regionally, related to the control of global gradients and trends of the whole aquifer, and locally, more related to impacted areas, for example for the control of the over-exploitation of aquifers in industrial areas or in zones where there is intensive pumping for municipal water supply purposes. It is obvious that groundwater quantity monitoring networks are directly related with quality problems and it must be taken into account in the evaluation of representativeness. Two examples illustrate the point: there is a risk of groundwater contamination produced by saltwater intrusion from downward leakage induced by pumping in areas where the cones of depression extend beneath estuaries or the oceans; and, the risk of horizontal saltwater intrusion in coastal aquifers in areas where there is an over-pumping situation.

The representativeness of a national (reference) piezometric network can be evaluated by calculating the contribution of each existing monitoring station in the geographic aquifer coverage. This can be made by building an indicator based on the spatial correlation between the stations. For instance a method such as kriging provides the estimation of an average value of a specific parameter (i.e. piezometry, temperature) over the whole aquifer, and also an associated estimation error, highlighting the areas with poor monitoring.

The indicator can be improved by incorporating information related to sampling frequency (by simple trend analysis methods) and to the particular features of piezometric surface on the over-exploited areas of the aquifer system (in relation to pumping wells). In order to analyse the spatial and temporal variability of the groundwater quantity parameters a synthetic seasonal piezometric index could be built, based on available information. The index could characterise different trend patterns of piezometric evolution.

With the present information provided by the EEA Members it is not possible to achieve the ultimate objective of this Task as described in the technical work programme. Therefore, the evaluation of the representativeness of the existing monitoring networks was carried out only as a very broad assessment. This is particularly evident for sampling density. In fact no information was reported on the probable clustering of observation stations, the type of impact areas (heavily exploited areas or areas particularly subject to interactions with other systems: rivers, sea, lakes, estuaries) and what groundwater regions are monitored.

5.6 International databases

The Agency’s policy on reporting on the state of the environment is to rely on existing data as much as possible in order to reduce the burden of additional sampling on Member States. A key part of designing a monitoring network is therefore to assess available data held for international monitoring programmes that already exist. This section summarises the findings of a review of international monitoring databases within the EEA (France et al. 1996). The review identified a total of 19 databases involved with inland water quality monitoring. These are given in Table 5.1 below.

Table 5.1 Summary of databases holding monitoring data from inland freshwaters of potential interest to the EEA.

Table 5.1 (contd)

Table 5.1 (contd)

Four different sources of information were used in order to identify international databases and combine them to produce a detailed database of data sources and information:

• Data relating to International Conventions agreed between the EEA countries were drawn from the database developed in the project MW1.
• Published reports and other literature, e.g. PARCOM and Helsinki Commission reports, the Dobríš Assessment and UNEP publications.
• Information available on Internet.
• These data were supplemented by circulation of a questionnaire to key organisations.

The constructed database forms a partial catalogue of data sources pertinent to inland waters. It holds the history of monitoring for each country per database, the number of sites, monitored, water types sampled and the number of years of sampling. Additionally it details for each database; areas, regions specific water bodies and other related information that is available such as land use. It is envisaged that this database will provide essential information in selecting sites for the monitoring program.

Liaison with the ETC-CDS is now a priority to inform of the data sources that ETC-IW are aware of and to make them available to the EIONET.