Skip to main content

Land Fragmentation

Habitat fragmentation occurs when natural habitat is broken up by non-natural land uses. For this layer, land fragmentation is expressed as the Natural Land Cover Pattern Index (NLPI), which classifies natural or semi-natural landcover into six spatial pattern classes: core, edge, perforation, islet, margin, and core-opening. Natural landcover pixels from Climate Change Initiative Land Cover (CCI-LC) raster maps are classified as ‘core’ habitat. Non-natural land cover pixels within core habitat are classified as ‘core openings’ as these represent openings from within natural habitat. External interfaces between core and non-natural landcover are classified as ‘edges’ (i.e. exposed to outside-in habitat loss pressures), while internal interfaces between core and core-openings are classified as ‘perforations’ (i.e. exposed to inside-out habitat loss pressure). Collections of isolated natural landcover pixels that are too small to contain core habitat are classified as ‘islets’, while similarly small collections of pixels are classified as ‘margins’ when they are connected to core habitats.

* Analysis technique: Soille, P., Vogt, P. (2009). Morphological segmentation of binary patterns. Pattern Recognition Letters, 30: 456–459. doi:10.1016/j.patrec.2008.10.015 * Software: Vogt, P., Ritters, K. (2017) GuidosToolbox: universal digital image object analysis. European Journal of Remote Sensing, 50: 352-361. doi:10.1080/22797254.2017.1330650 * Product sheet:… * Underlying data: Land Cover CCI. (2017). Product User Guide Version 2.0…


Although related, the index strictly presents a pattern of landcover, rather than the process of habitat fragmentation. Caution is therefore needed so as not to overinterpret process from pattern. Classification of the Natural Land Cover Pattern Index depends on the edge width used for analysis (See figure 3 in Soille & Vogt, 2009). Landscape fragmentation analysed using land cover products at a different spatial resolution would result in different classifications of the same area. The CCI-LC data used in this product was in the WGS84 projection, so area statistics should not be carried out without first reprojecting the original landcover data using an equal-area projection and repeating the analyses.

Select for mapping