Land take, urban sprawl and economic activities lead to habitat fragmentation, decreasing the resilience of ecosystems. Monitoring fragmentation supports policy actions that aim to ensure remaining habitats can support biodiversity. Fragmentation affects all areas of Europe, even very sparsely populated ones. Moreover, in the EU plus the United Kingdom, 27% of land is considered highly fragmented where habitats are less than 0.02km2 on average. However, policy measures to protect certain areas seem to be effective in preventing fragmentation, particularly in protected areas.
Figure 1. Landscape fragmentation by degree of urbanisation and MAES ecosystem type, 2018, EU-27 and the UK
Landscape fragmentation is the physical disintegration of continuous habitats into smaller units or patches, most often caused by urban or transport network expansion. This has a wide range of environmental, social, climate change adaptation and mitigation, and biodiversity implications.
The EU biodiversity strategy for 2030 aims to protect and restore nature, including by tackling fragmentation. Fragmentation also impacts implementation of the EU strategy on green infrastructure and achieving the long-term objectives of the EU common agriculture policy, namely the sustainable management of natural resources, climate action and balanced territorial development.
Large parts of Europe have become fragmented because of the expansion of urban and transport infrastructure. On average, every km² in the 27 EU Member States plus the UK (EU-27+UK) comprises around 1.4 habitats, indicating an average habitat size of 0.68km2. Moreover, 27% of land in the EU-27+UK is considered highly fragmented, where habitats are less than 0.02km2. As distance from city centres increases, the extent of landscape fragmentation drops rapidly. In villages, average habitat size is around 0.12km2. This increases to 0.8km2 in rural areas and 5.3km2 in mostly uninhabited regions. The proportions of strongly fragmented habitats is more persistent, with habitats smaller than 0.02km2 accounting for 79% of land in suburbs, 61% in villages and 53% in rural areas. Even in mostly uninhabited areas, more than 20% of land is covered by habitats of less than 0.02km2.
Other than urban ecosystems, croplands are the most fragmented ecosystem type, with an average habitat size of 5km2. Habitat size in grasslands is on average 8km2 but forests are more continuous, with an average habitat size of 27km2. Coastal ecosystems are under increasing pressure from urban sprawl, with transport infrastructure and other construction jeopardising wildlife movement. Indeed, the average habitat size in coastal ecosystems is around 0.4km2; in inland areas, average habitat size increases to 0.9km2. Policy measures that safeguard protected areas seem to be effective, however: while average habitat size in non-protected areas in the EU-27+UK is around 0.6km2, average habitat size in protected areas is on average 20km2.
Figure 2. Landscape fragmentation in European countries, 2018, EU-27 and the UK
The extent of landscape fragmentation varies considerably by country in the EU-27+UK region, being highest in Malta, followed by the Netherlands, Belgium, Germany and Luxembourg. Malta has the most fragmented landscape by far, with 15 landscape objects per km2 on average, which is around double the extent of landscape fragmentation in the Netherlands and Belgium, four times that of Germany and greatly above the EU-27+UK average of 1-3.5 landscape objects/km² (95% confidence interval). Moreover, the average landscape object size in Malta is around 0.06km2, considerably below the EU-27+UK average of around 0.68km2.
Although, on average, the extent of landscape fragmentation is highest in Malta, Luxembourg and Belgium have the largest area of highly fragmented habitats, that is, areas with average habitat sizes of less than 0.02km2. About 90% of the landscape is highly fragmented in Luxembourg and around 84% is highly fragmented in Belgium.
In Finland, the Baltic countries and Sweden, habitats are much more contiguous than in other parts of Europe, with habitat sizes of at least 2.8km2, which is far greater than the EU-27+UK average.
Supporting information
This indicator measures landscape fragmentation due to transport infrastructure and sealed areas. Unlike the previous indicator on fragmentation status, this updated version uses the TeleAtlas® Multinet data set to ensure the statistical comparability of the time series. While the Open Street Map data set is a valuable source of the street network available for the general public, there are still inconsistencies in this data set for some regions of Europe, which render it secondary to the TeleAtlas data set.
As in the previous version, this indicator is based on the effective mesh size method (Jaeger, 2000). For some species, the effective mesh size (meff) can be interpreted as the area that is accessible when beginning to move from a randomly chosen point inside a landscape without encountering anthropogenic barriers such as transport routes or built-up areas. The combination of all barriers in a landscape is referred to as the fragmentation geometry (FG) hereafter. However, it should be stressed that for many species that can fly, or are effective dispersers in other ways, man-made structures may not act as barriers.
The meff value expresses the probability that any two points chosen randomly in an area are connected. Hence, meff is a measure of landscape connectivity, i.e. the degree to which movements between different parts of the landscape are possible. The larger the meff, the more connected the landscape. The indicator addresses the structural connectivity of the landscape and does not tackle functional, species-specific connectivity.
The effective mesh density (seff) is a measure of landscape fragmentation, i.e. the degree to which movement between different parts of the landscape is interrupted by fragmentation geometry. It gives the effective number of meshes (or landscape patches) per 1,000 km2, in other words the density of the meshes. The seff value is 1,000 km2/meff, hence the number of meshes per 1,000 km2. The more barriers fragmenting the landscape, the higher the effective mesh density.
The values of meff and seff are reported within the cells of a 1 km2 regular grid.
The value of meff is area-proportionally additive, hence it characterises the fragmentation of any region considered, independently of its size, and thus can be calculated for a combination of two or more regions. It has several advantages over other metrics:
· It addresses the entire landscape matrix instead of addressing individual patches.
· It is independent of the size of the reporting unit and its values can be compared among reporting units of differing sizes.
· It is suitable for comparing the fragmentation of regions with differing total areas and with differing proportions occupied by housing, industry and transportation structures.
· Its reliability has been confirmed on the basis of suitability criteria through a systematic comparison with other quantitative measures. The suitability of other metrics is limited, as they only partially meet the following criteria:
o intuitive interpretation;
o mathematical simplicity;
o modest data requirements;
o low sensitivity to small patches;
o detection of structural differences;
o mathematical homogeneity (i.e. intensive or extensive).
The calculation of the effective mesh size (meff) is based on three spatial data sets: (1) the 'landscape' extent, (2) the fragmentation geometry (FG) (landscape elements representing man-made barriers) and (3) reporting units (spatial units for which meff is calculated). The following steps are followed in computing the indicator.
Step 1: landscape extent
The 'landscape' for the calculation of meff is the seamless area of Europe. The input for this step is the Copernicus high resolution layer (HRL) on imperviousness density (IMD) from 2012.
Step 2: fragmentation geometry
Fragmentation geometries are man-made landscape elements, which divide the landscape into unconnected patches. Only anthropogenic elements are considered because the indicator addresses fragmentation of the landscape in urban areas and from transport infrastructure (road and rail).
Step 2.1: fragmentation geometry — built-up areas
Built-up areas are excluded during the 'landscape extent' preparation step. From this layer, a binary mask is created and pixels with IMD values of > 30% are deleted from the data set.
Step 2.2: fragmentation geometry — road network
The data set representing the transportation network must meet the following technical requirements:
1. It has to be methodologically stable, so that changes in time represent real changes and not the level of data set completion.
2. The nomenclature/classes of roads must be clearly defined, and consistent over time, to allow different levels of fragmentation detail.
3. It must be topologically correct, i.e. it must not contain discontinuities.
4. It must enable the differentiation of landscape elements that have a major impact on the resulting connectivity or isolation of patches, such as tunnels, overpasses, etc. (where such elements occur, landscape patches may in fact be interconnected and thus the value of fragmentation can be considerably different).
5. It must be based on regularly updated and if possible open source data streams to ensure the sustainability of the indicator.
The TeleAtlas® Multinet data set]was used to process the road network fragmenting geometry. The following road/rail classes were included (road class numbering is based on FRC attribute values):
0 — motorway, freeway
1 — major road less important than a motorway
2 — other major road
3 — secondary road
4 — local connecting road
Railroads
The line vectors were buffered according to the road classes to create polygon objects. Buffer sizes were selected according to the road class they represent. Buffering was also applied to prevent small topological inconsistencies in the TeleAtlas data set.
Table 1: Buffers applied to the various TeleAtlas road and rail classes
TeleAtlas road class
Buffer width (in m) on either side of the roads
Motorway, freeway
15
Major road less important than a motorway
10
Other major road
7.5
Secondary road
5
Local connecting road
2.5
Railroad
2
The result of step 2 is a fragmentation geometry layer that contains landscape patches (i.e. polygons representing the remaining non-fragmented areas) and gaps (no value), in locations where fragmentation geometries were deleted from the landscape.
Step 3: calculation of meff
The meff values are calculated for all reporting units. The reporting units are 1km2 grid cells corresponding to the EEA’s accounting grid. It should be noted that any regular (i.e. larger or smaller grids) or irregular (e.g. NUTS (Nomenclature of Territorial Units for Statistics) regions) reporting units can be chosen for the calculation as long as the spatial detail is satisfactory for the topic that the indicator is designed to support. To calculate meff, the cross-boundary calculation (CBC) procedure is used (Moser et al., 2007). In the CBC process, not only the area of the landscape patch that falls inside the reporting unit is an input to the computation, but the whole area of that given landscape patch is accounted for (see image below). Hence, the borders of analytical units themselves do not influence meff values (see detailed explanation in Moser et al., 2007).
Values of meff are positive real numbers, where 0 stands for grid cells completely covered by urban areas and infrastructure (i.e. the landscape is covered by impermeable surfaces). The lower threshold for meff is 0.000001km2 (= 1m2), and smaller values are rounded to this value. The highest possible value of meff is limited by the area of the landscape patches as well as by the area of the fragmentation geometry affecting the landscape patches. A landscape patch is defined as a continuous area with the barriers of the fragmentation geometry as boundaries. Hence, the largest meff value will be assigned to the largest continuous landscape patch with the smallest area taken up by the fragmentation geometry (see illustration in "Methodology for indicator calculation" section).
The seff values are positive real numbers. If meff = 0.000001km2, then seff = 1,000,000,000 meshes per 1 000km2. For grid cells completely covered by built-up areas and infrastructure (i.e. where meff = 0km2), the seff value is set to -2, i.e. -2 represents positive infinity.
For convenience and practical considerations, meff values of < 0.01km2 (= 10 000m2) are rounded to 0, as these values are too small to be measurable without noise on a European scale. As a consequence, the largest reported seff value is 100,000 (= 1,000km2/0.01km2) meshes per 1,000km2.
This indicator presents seff, rather than meff, values because these are more intuitive to understand as indications of fragmentation. For the assessment, seff values were grouped into five fragmentation classes (very low, low, medium, high and very high) by performing the following steps:
(1) selecting 95% of the seff value range (ignoring the upper and lower 5th percentiles);
(2) running geometric interval classification;
(3) rounding threshold values for straightforward comparisons and change detection.
The thresholds for the fragmentation classes are outlined below.
seff values
(number of meshes per 1,000km2)
Fragmentation class
0-1.5
Very low
1.5-10
Low
10-50
Medium
50-250
High
Very high
> 250 seffs
In May 2020, the EU adopted a Biodiversity Strategy to 2030, related to protecting and restoring nature. The strategy states that ‘The biodiversity crisis and the climate crisis are intrinsically linked. Climate change accelerates the destruction of the natural world through droughts, flooding and wildfires, while the loss and unsustainable use of nature are in turn key drivers of climate change’. Droughts are negatively affecting agricultural ecosystems and food security, resilience of forest ecosystems and ability of green urban spaces to protect people against heat waves. In particular, the impacts of extended droughts on ecosystems need to be assessed because they can lead to significant loss of vegetation productivity and irreversible damage to the condition of ecosystems and can lead to land degradation.
In February 2021, the European Commission presented the EU adaptation strategy package. The new strategy sets out how the European Union can adapt to the unavoidable impacts of climate change and become climate resilient by 2050. One of the objectives of the strategy is to ensure better-informed decision-making, which will be achieved by bridging knowledge gaps and further developing the European climate adaptation platform (Climate-ADAPT) as the ‘one-stop shop’ for adaptation information in Europe. Climate-ADAPT has been developed jointly by the European Commission and the EEA to share knowledge on (1) observed and projected climate change and its impacts on environmental and social systems and on human health, (2) relevant research, (3) EU, transnational, national and sub-national adaptation strategies and plans, and (4) adaptation case studies.
Methodology uncertainty
The methodology is without any major uncertainty. Some critique might arise regarding the fragmentation geometries, which were included (or not included) as barriers. This is however not a methodological uncertainty of meff and seff, but is rather a matter of consciously addressing the spatial detail of the indicator.
Data set uncertainty
Clouds are contained in the Copernicus Land Monitoring Service data layer. Corresponding Copernicus CLC data are used for the map filling (see 'Methodology for gap filling' section). Because the spatial resolutions of the HRL IMD and CLC data are different, the spatial detail of the indicator may be influenced for the cloudy area. The metadata layer is part of the indicator data set indicating HRL IMD cloud areas.
EC, 2013, Communication from the Commission to the European Parliament, the Council, the European Economic and Social Committee and the Committee of the Regions — Green infrastructure (GI): enhancing Europe’s natural capital, COM(2013) 249 final.
