5. Concept of the proposed networks

5.1. Representative Assessment of Water Resources

The proposed network is designed to give a representative view or assessment of water types within a Member State and also across the EEA area. It will ensure that similar types of water body are compared. For example at present comparisons are made on the basis of loose selection criteria such as important rivers, large rivers etc. The proposed design will ensure, for example, that small rivers are compared with small rivers, deep lakes with deep lakes, and so on. The need to compare like-with-like has led to a stratified design with the identified and defined strata containing similar water bodies. The use of the same criteria for selecting strata and water types across Member States will ensure that valid status comparisons will be obtained.


5.2. Overall Objective

The overall objective of the monitoring (information) network is:

"To obtain timely, quantitative and comparable information on the status of inland waters (groundwater, lakes/reservoirs, rivers and estuaries) from all EEA Member States so that valid temporal and spatial comparisons can be made, and so that key environmental problems associated with Europe’s inland waters can be defined, quantified and monitored".

To minimise cost implications, where possible the monitoring network will be based on existing national and international networks, use existing sources of monitoring information and create, only if necessary, an EEA database of aggregated data and information rather than of raw non-processed data.

It should be emphasised that the information provided by the network will not be for the assessment of compliance of Member States with the requirements of European Commission directives.


5.3. References Conditions and Stations

The process of quantifying the effect of human activities (to establish cause-effect relationships) on water resources requires the separation of natural factors and their effects from the impact due to human activities. This will be aided by the identification of reference rivers, lakes and aquifers. In these locations or catchments there will only be minimal or no human activities, there would be natural land cover and would be in most cases be representative of pristine conditions. The quantification of chemical, physical and biological conditions at these sites would give a measure of reference conditions or levels against which ‘impacted’ rivers etc. would be compared. In this way anthropogenic impacts can be separated from natural variability so that valid comparisons can be made. In some parts of Europe and for some water types reference conditions would not occur because landscapes and water bodies have a long history of human activities and changes. In others such as in the more remote Nordic regions there would be many examples of reference conditions. The statistical aim of the network is to detect significant differences between ‘similar’ areas (or strata) and the trends with time within areas (or strata). In this way improvements due to investment and the extent of problems will be detected and quantified.


5.4. Stratification Criteria

The natural criteria that might be appropriate for the selection of strata include bio-geographical, hydrological, catchment altitude, stream order (size), lake depth and aquifer geology. In this way similar water bodies would be identified, reference conditions quantified and valid comparisons made. The anthropogenic factors which must also be taken into account, particularly if the relationship between state and pressures are to be established, include catchment use and activities (for example agricultural forested catchments, urban/industrial catchments), population density, abstraction rates and discharges of pollutants from point sources (pipes) and from run-off from land or atmospheric deposition (diffuse sources).

Once appropriate strata have been defined, rivers, river reaches, lakes and aquifers would be allocated to the strata. Ideally there would be existing monitoring stations for each of the stratified water bodies. In reality in many cases there will not be. It is therefore proposed as an initial step that existing monitoring stations should be stratified. At some stage the stratification based on water body will be compared with that based on monitoring station. If there are large discrepancies between the two (for example if small rivers are not sampled representatively) then additional or re-assigned monitoring may have to be undertaken. In some Member States there will be large gaps in the proposed network because, for example, there are currently no national lake or groundwater monitoring programmes.

The allocation of rivers etc. into comparable strata also offers the possibility of reducing the amount of sampling or sites that need to be included in the network without the loss of information. This is because the amount of overall variability in indicators (for example a chemical or biological measure) might be potentially reduced. There is also potential for reducing sampling bias by selecting monitoring sites from similar water types/catchments etc., thus obtaining fairer comparisons between regions on by stratum basis. If, however, too many strata are produced there is always an opportunity of recombining data from different strata into larger strata perhaps using some form of stratum weighting (e.g. on the basis of frequency of occurrence of particular ‘water type’ in the whole water resource).


5.5. Example of Stratifying Rivers

An example of the possible stratification of rivers is given in Table 5.1 where the two stratifying criteria are size of river (perhaps defined by stream order), and altitude. Here monitoring stations have been divided into reference, representative and flux stations. Reference stations have been described. Representative stations will give a representative view of water resources in Member States. They will include impacted water bodies and also flux stations. Flux stations are required to calculate the loads of contaminants passing into and from water bodies (e.g. rivers discharging into seas) or where there is transboundary movement of water.

Table 5.1 Stratification of rivers into similar types for sampling and comparison

Type of monitoring site

Relative size

small rivers
(1st to 3rd order)

medium rivers
(4 to 5th order)

large rivers
(6th order and above)

  Relative European altitude (class)

















Altitude classes:
a = >800 m
b = 500 to 800 m
c = 200 to 500 m
d = 100 to 200 m
e = <100 m

The stratification framework illustrated in Table 5.1 has been applied to the 6939 river sites currently monitored in England and Wales by the former National Rivers Authority (now the Environment Agency). The allocation of sites to each strata is given in Table 5.2. Within this network most stations are on small rivers, there being relatively few large rivers (on a European scale) in England and Wales. The stratification would look different for other Member States. For example, countries like Austria are likely to have rivers in higher altitude strata which are not present in England and Wales. Once established data from the stations would be analysed in a consistent way by Member States and reported to the EEA. Comparisons of similar rivers etc. could then be made by the Agency or the ETC/IW.

Table 5.2 Stratification of river stations in England and Wales



200 to 500 m

100 to 200 m

<100 m

Stream order ¯

1 to 3





4 and 5









All sites






5.6. Statistical Implications

The reason for designing a monitoring network and sampling programme in the first place is because it is probably impractical to sample the whole water resource of a country. Therefore, the values obtained for the statistical objectives are estimated from a (usually) much smaller sub-population of samples and are, consequently, subject to a certain amount of error or uncertainty. Choosing the precision and confidence sets limits on how much of this uncertainty can be tolerated in the results of the programme.

Consider some quantity that has been estimated from the sampled data. This estimate will almost always differ from the true value (i.e. the quantity which would be calculated if all water bodies were sampled). Answering the following two questions will define the precision and confidence.

  • What is the largest discrepancy that can be tolerated between the answer given by the sampling programme and the true value? This is the desired precision.

  • What degree of confidence should there be that the answer obtained does in fact lie within the desired precision? This is the desired confidence.

Confidence is expressed as a percentage, so for example, a confidence of 99% means that if the sampling programme could be repeated 100 times, the answer would be within the precision tolerance on 99 occasions.

Figure 5.1 provides an example of the statistical power that can be achieved through stratification. The data from the stratified stations in Table 5.2 have been analysed to calculate the number of stations required to detect 50% and 25% changes in total inorganic nitrogen (TIN) and soluble reactive phosphorus (SRP) concentrations in rivers in England and Wales over 10 years with a 90% level of confidence and assuming that a sample is taken from each site each month over the 10 year period. As well as the four strata with enough data to test, all (general) unstratified stations and those submitted for the Exchange of Information Decision for England and Wales (14 sites on 11 rivers) are given for comparison.

Figure 5.1 illustrates that the number of sites required to obtain the same level of confidence differs for 2 water quality indicators with TIN being more variable (and hence requiring information from more sites) than SRP in small upland rivers but less variable in medium sized lowland rivers. Decisions will thus have to be made as to whether to sample at a different number of stations and frequency for each indicator or perhaps select the most important indicator and accept that the level of confidence associated with other indicators may be lower (or higher). The former option would lead to a more logistically difficult and expensive sampling programme.

It is also apparent that around 4 times more stations would be required to detect a 25% change compared to a 50% change in TIN concentrations. For SRP approximately 6 times more stations would be required for most strata. It is also interesting to note that it would not be possible to detect a 50% change in SRP concentrations over 10 years at the Exchange of Information Decision stations as more stations would be required than are actually reported.

a. 50% change


b. 25% change

Figure 5.1 Number of stations required in a stratified river monitoring network in England and Wales to detect significant changes in 10 years with a 90% confidence and with 12 samples taken at each site each year.


5.7. What is the Benefit to Member States in Providing this Information?

At present there is not enough comparable information to obtain a quantitative assessment of water resources across Europe. This can lead to unfair or incomplete comparisons being made and wrong conclusions drawn. By submitting information within this proposed framework a ‘level playing field’ will be obtained so that Member States will have confidence in the conclusions being drawn. In addition the information will enable environmental policies to targeted correctly and cost-effectively.

As an example the information on soluble reactive phosphorus from 55 stations on 40 rivers in England and Wales (as presented in the Dobris report) is compared to that obtained from a national database of 7,000 sites used to assess the general quality of rivers in England and Wales. The distribution of concentrations obtained from each data set is different, with the Dobris dataset indicating a higher proportion of poor quality and lower portion of better quality waters than that based on the larger dataset. It could be argued that the latter gives a more representative view of river quality in terms of phosphorus in England and Wales, and would indicate more accurately where improvements had to be made.

Figure 5.2 Comparison of distribution of soluble reactive phosphorus in rivers of England and Wales


5.8. Summary

In summary the EEA network will:

  1. Be representative of the size/numbers/types of water bodies in the EEA area (e.g. small rivers), variation in human pressures (e.g. population density and land use), and, should include a number of reference and flux stations.

  2. For rivers, have reference, representative, impact (part of representative network) stations, and flux monitoring stations at discharge into sea, or at international boundaries.

  3. For lakes, have a general surveillance network comprising reference and representative lakes, and if necessary, (in the light of experience) an impact network with lakes selected on the basis of population density. In addition the largest and most important lakes (nationally) will be included and possibly a specific cause/effect network of lakes.

  4. For groundwater, have a general surveillance network comprising representative stations on all nationally important aquifers (groundwater in porous media, karst groundwater and others) ideally at a density of 1 station per 20 to 25 km2 of aquifer. In addition the feasibility of establishing reference stations in aquifers not affected by human activities will be assessed.

lijn.gif (900 bytes)

up.gif (859 bytes)
left.gif (869 bytes) right.gif (869 bytes)


Document Actions