Proposal for a European coding system for hydrological features2nd Draft

1.  Introduction

Within the GIS guidance document on implementing GIS elements within the WFD the need for assigning identifiers to geographic features was recognised. The recommendations give precedence to unique coding rather than to structured hydrological coding. At that time, no agreement could be reached on a common recommendation for structured hydrological codes. It was, therefore, recommended to investigate the coding issue further in order to develop final recommendations on a European structured hydrological feature coding system.

2.  Purpose of coding system

One purpose of a European coding system for hydrographical features is to create an identifier for geographic features that are reported by the member states to the European Commission. The code will serve as a common identifier and provide a link between the European and the national information systems.

In addition the coding should enable topological analysis of the geographic features without having the need for utilising a geographic information system. For example, it should be possible to analyze upstream contributaries of a river system but as well follow a flow from a defined source to its sea outlet.

The code should support compliance checking, i.e. it should be possible to verify topological relationships and characteristics of reported features in a spatial and topological context. In this respect, redundant information is provided for verification purposes.

The code should identify hydrological reference features which are not man-made and are more stable in time. The code should not follow “administrative” criteria for discriminating different features.

More detailed objectives of the working group are to define a coding system that satisfies the following characteristics:

·  Uniqueness
Each hydrological reference feature should be uniquely identifiable via its code.

·  Topological code
Geographic features should be assigned with a structured hydrological code, i.e. coding should be done in relation to the hierarchy and topology of the river network and of its related catchments. Structured coding will enable analysis of topological relationships between water bodies avoiding time consuming geographical analysis of feature connectivity.

·  International
Transboundary features are identified with one unique code avoiding management of several national codes and assuring proper connectivity between national features.

·  Comprehensive coding system
The coding system has to cover all hydrological reference features that make part of the Water Framework Directive and should establish proper relationships to all non-reference features.

·  Functionality
Within the general principles, codes must be functional and easy to process. They must be as short as possible and nevertheless informative enough to ease error checking.

·  Top down approach
Coding should be done from the largest to the smallest feature. With this approach, coding could always be extended to higher resolutions, ensuring flexibility for code extension. The bottom-up approach would require the existence of a European dataset of rivers and catchment areas at the highest possible resolution, which is not feasible.

·  Management of codes
The codes should be managed, when relevant, on the lowest possible level, i.e. by international organisations for international river basins, by national organisations for features that are located entirely within a country. This principle should also be valid for sub-features, e.g. national tributaries to international river systems.

The application of a common European coding system should establish a stable link between the national and the European level and enable adequate reporting and analysis of the data on European level.

3.  Review of GIS guidance

The GIS guidance document gives precedence to a unique code rather than to a structured hydrological coding system. It proposed to generate a European code by preceding the national code with a 2 character member state identifier, corresponding to the ISO 3166-1-Alpha 2 country codes. The proposed code has the following format:

“MS#1#2…#22 where

MS = a 2 character Member State identifier,

in accordance with ISO 3166-1-Alpha-2 country codes, and

#1#2…#22 = an up to 22 character feature code that is unique within the Member State.

Special advice given is that

·  Alphabetical characters should always be in upper case, as systems will be case sensitive.

·  Special characters must be avoided, such as ‘$’, ‘!’, ‘&’, ‘ë’, ‘á’, etc.

·  Digits should be used where practical to help avoid the above problems.”

The usage of this kind of code is the only requirement when reporting hydrological features for the purpose of the WFD.

The codes are entirely managed by the member states. Although the GIS guidance does not force any structure of the code, it gives recommendations on establishing the codes in order to facilitate code management in the member states. There might be a need for setting up a code management system that is able to monitor code changes over time. Such a system would include a start and an end date of validity of the coded feature and links to its successors or predecessors.

4.  Application of the code

In general the code should be applied to all hydrological reference features reported within the WFD. In addition, it should be flexible enough to allow for extensions to other additional features in future or in the context of other applications. The following geographic features will be identified using a unique feature code:

  1. River (sub) basins
  2. (Main) Rivers
  3. River Segment (reaching from confluence to confluence)
  4. Lakes

In general, all hydrological reference features should be assigned with a primary feature code. Non-reference features, such as water bodies, monitoring stations, discharge points, point sources, non-point sources, etc., should be assigned with a non-hydrological identifier but should carry the code of the hydrological feature to which they are related as foreign key. This principle assures the identification of each feature and establishes necessary relationships to the hydrological reference data. In addition, the code of non-hydrological features could be completely managed by the member states and the hydrological code could be kept as short as possible. These features should be assigned with a country code and a unique string as defined in the GIS guidance document.

The temporal aspect of the non-reference features entails the follow-up of changes in time. It will be necessary to establish a system that manages changes of water bodies, i.e. lifetime periods and identification of predecessors and successors.

Features with hydrological code as foreign key are:

  1. River water bodies
  2. Lake water bodies
  3. Coastal water bodies
  4. Transitional water bodies
  5. Ground water bodies
  6. Artificial water bodies
  7. Monitoring Stations
  8. Discharge points

9.  Point and non-point pollution sources

  1. Protected areas
  2. …

5.  New Coding structure

The hydrological code should be composed of different segments, which together uniquely identify a hydrological feature. The code should start with the outlet of a river system, i.e. the sea region into which the river system drains. The river basins and sub basins should be identified using the principles of the Pfafstetter coding. The Pfafstetter code will be preceded by a code segment that identifies the primary catchments and inter-catchments at the coast applying the same principles as the Pfafstetter approach. This commencement code will be determined on European level on the basis of the basins and subbasin that have been reported under article 3 of the WFD. The commencement code will be followed by the actual Pfafstetter code segment. An entity type code segment would define the type of feature that is referenced by the code. These are e.g. river segments, lakes and basins.

The code would consist of the following segments:

  1. Sea Region

2.  Pfafstetter commencement code

  1. Pfafstetter code of river and associated (sub) basin
  2. Entity type

5.  Lake identifier

Code segments with text in Italics are optional, the others are mandatory.

6.  Code segments

The hydrological code is composed of a number of code segments that, together, identify a geographic feature uniquely. This chapter describes the code segments more in detail.

6.1.  Sea Region

The identification of the sea region constitutes the first segment of the entire code. The water framework directive contains two maps that delineate sea regions. The first map defines the EcoRegions for rivers and lakes and the second maps the EcoRegions for transitional and coastal waters. As the first one is mainly based on national boundaries for assigning codes it is not suitable for delineating sea regions. The second map distinguishes between 6 sea regions, excluding the Black Sea. It is therefore proposed to use the third map of sea regions which includes the Black Sea and that distinguishes some sea region further in order to exploit all figures except for the “0”. In this case, the Celtic Sea and the English Channel would define a separate region. The Mediterranean Sea would be split in a Western and Eastern part.

Figure 1: EcoRegions map

Figure 2: Sea regions commencement code

The coding of the sea region would be mandatory for all hydrologic reference features. The length of the code segment is set to 1, providing the possibility of identifying a maximum of 9 sea regions.

Table 1: Sea regions

Number / Sea Region
1 / Atlantic Ocean
2 / Norwegian Sea
3 / Barents Sea
4 / North Sea
5 / Baltic Sea
6 / Celtic Sea and English Channel
7 / West Mediterranean Sea
8 / East Mediterranean Sea
9 / Black Sea

6.2.  The primary catchment commencement code

The second code segment would identify the primary catchment or a coastal catchment within the respective sea region. The commencement code will be defined on European level based on the basins and sub-basins of the national datasets that have been reported within the article 3 of WFD. The determination of the code should follow the same logic as the Pfafstetter code. The major primary catchments within each sea region should be assigned with the figures 2, 4, 6 and 8 in clockwise rotation. The coastal region between these primary catchments should be assigned with odd numbers 1,3,5,7 and 9.

It is proposed to assign a maximum of three digits for this code segment.

Figure 3: Commencement code with major primary catchments

6.3.  River basin

The Pfafstetter approach is recommended for providing a structured hydrological code segment, identifying river basins and river sub basins. The Pfafstetter system follows a systematic approach as it is derived from topological relationships of the underlying drainage system. The numbering schema is self-replicating from the largest to the smallest drainage system. With Pfafstetter codes it is possible to identify all nested sub-basins within the larger basin and the “parent” basin from a sub-basin. All upstream sub-basins or river segments as well as all downstream segments are identifiable at each location of the river network


Having identified the major drainage system and the coastal areas the rivers and catchment areas could be assigned with the Pfafstetter code. It is suggested to start with the longest flow path within a drainage system and to subdivide basins and interbasins according to the size of their drainage area. Using the river name for defining the river stem is somehow arbitrary and makes the coding system more difficult to define and manage. The name should be assigned as an attribute to the river. The volume of the water flow is difficult to measure and subject to change.



After the recognition of the longest flow path in a given basin, the four most important tributaries (and sub-basins) are identified according to their drainage area size. They receive numbers 2, 4, 6, an 8 in upstream order. Interbasins receive odd numbers from 1 to 9, also in upstream order. Sub-basins and interbasins are then further subdivided according to the same principle (see Figures 2 and 3). For more detailed descriptions of the system see WFD GIS Guidance Document, Britton 2002, Verdin & Verdin, 1999. It is important to remind that there is a one to one relationship between a river segment and its catchment. This approach requires the existence of a consistent dataset of catchment areas at European level.

Figure 4 The Pfafstetter principle of coding sub-basins and interbasins, illustrated for a hypothetical basin with code 2. Code 2 indicates the north-eastern primary catchment system within a sea region.

Figure 5: Further breakdown of sub-basin 26 according to Pfafstetter.

The feature code should identify all features reported within the Water Framework Directive at an appropriate level of detail. Hence, it is proposed to apply a criterion of maximum size for a river basin based on the various reporting sheets for the basins and sub-basins. The maximum size for sub-basins is 1000km² within the territory of EU25 and 4000km² outside EU25. Smaller catchments should be combined and larger catchments should be split, if possible. Applying the size criterion prevents specifying a certain stream order level for the application of the Pfafstetter code, which would require the existence of a common hydrological dataset at a large scale.

Lakes should be included in the Pfafstetter coding system. If they are connected to the river network, they should receive the Pfafstetter code of the river segment to which they drain (the outlet). If they are not connected to the river network, they should receive the code of the lowest level river (sub-) basin or interbasin in which they are located. In case that more than one isolated lake is located within one sub- or interbasin, lakes should be distinguished by a specific identifier. This specific identifier should be appended after the entity code and before the country code to be able to code international lakes. A maximum of 2 characters are reserved for this segment. The code segment should only contain digits.

For the implementation of a Pfafstetter system, the river network must be fully connected also within the lakes. These hypothetical river segments are often called continua. As a consequence, water bodies within a lake can be identified according to the codes of the underlying river segments, if required.