In 1994, the Organisation for Economic Co-operation and Development (OECD) developed a simple territorial scheme that identifies types of regions based on population density applied at two hierarchical levels. As there is no commonly agreed definition of rural areas at European level, the OECD typology is considered as an easy and acceptable approach for identifying rural areas. However, this typology used is exclusively based on population densities and is highly sensitive to the size of the geographic units and the classification thresholds.
Over the years, attempts have been made to review and improve the OECD approach and also alternative methodologies have been proposed. However, the current methods based merely on population distributions, do not allow for detailed and quantified geographical analysis and do not reflect two main characters differentiating rural from urban areas: the “natural” (non-artificial) surface and the accessibility/remoteness.
The objective of this study was to improve the characterization of rural areas at commune level (Local Administrative Unit – LAU 2) by introducing the criteria of accessibility (peripherality) and ‘natural (nonartificial) space in the OECD methodology. The assessment was carried out at LAU2 (and NUTS3) level for 3 Member States (Belgium, France and Poland), testing different thresholds.
Firstly, as indicator of peripherality, the travel time by road network to urban centres has been selected by using the speed limit of each category of roads (based on the EuroRegionalMap dataset) and two impedance factors, a congestion index (Urban Morphological Zones)and a slope index (DEM,100m). In order to discriminate the communes on the basis of the peripherality index, two time breaks have been tested: 30 and 60 minutes. A criterion based on the total population per commune (Eurostat SIRE database, census per commune 2001) has been used to select the urban centres and the thresholds of 50,000 and 100,000 of inhabitants have been tested. The origin/destination cost matrix solver was applied, using centroids of LAU2 as destinations/facilities.
For the final selection of the optimal thresholds, it was opted that extreme situations should be excluded: the threshold of 50,000 inhabitants for an urban centre and the travel time period of 30 minutes appeared to be the most appropriate criteria to evaluate the accessibility to cities. A sensibility analysis was followed out to evaluate the impact of the integration of a 100 m DEM and a congestion effect which showed that the congestion effect impact on the classification is significantly more important than the one related to the slope effect.
The peripherality analysis was done for three countries (BE, FR and PL) considering them as being “isolated” countries. A border-effect analysis was carried out (for Belgium), to assess the impact of the urban centres of the neighbouring countries and it appeared that the accessibility of communes close to borders is indeed influenced by the neighbouring cities.
Secondly, the land cover criterion to assess the ‘natural’ (non-artificial) surface of a LAU2, was used based on the methodology of Vard et al.(2005) that states that a commune will be classified as “rural” if at least 90 % of its area is covered by forest, agricultural or natural areas (Corine Land Cover 2000). Finally, the peripherality index and the land cover indicator were integrated in the OECD methodology. The rural typology contains 4 classes as only one threshold of population density (150 inhab./km²) is used and only two characteristics are combined (population density with remoteness/accessibility or population density with land cover) because there are correlations between some categories of the 3 characteristics (population density, land use and remoteness/accessibility).