Earth Observation News

Reading time: about 3 minutes.


Land Cover Change detected by CORINE for European Protected Areas

CORINE Land Cover is the official pan-European land cover monitoring programme being operational since the late 1980s and covering four survey periods for 1990, 2000, 2006 and 2012.

Ariane Walz and Luisa Gedon, Institute of Earth and Environmental Science | University of Potsdam

We investigated land cover changes as detected by the CORINE programme for all protected areas (PA) of the European Union and showed how ECOPOTENTIAL PAs compare with the total of European Protected Areas. We used the “Nationally Designated Areas” database, which is the official source of information on PAs from European countries to the World Database of Protected Areas and includes a total of 85´319 individual protected areas in its Geodatabase for 2015 (version 11). For the land cover, we use the CORINE Land Cover Change, which is based on an improved technique to detect changes in land cover. The “mapping change first” technique has been applied by most countries since 2006, and it covers changes in land cover in a higher resolution than the initial wall-to-wall land cover surveys. Changes in land cover are mapped, if they affect an area >5 ha with a width >100 m. We convert land cover changes to the six main “land cover flows” (LCF) identified by Feranec et al. (2010) which prevents double-counting of changes from and to a specific land cover. First results show the most important land cover flows within PAs include LCF4 “Reforestation” and LCF5 “Deforestation”. LCF4 contains areas changing away from “forest and semi-natural vegetation”, and LCF5 contains areas changing towards “forest and semi-natural vegetation”. Both flows can be caused either by natural processes or by human intervention. Comparing these flows from within PAs and from a 1 km buffer around PAs shows considerably higher changes around the PAs. Assuming natural processes are similar in intensity within PAs and their direct surroundings, this indicates considerably less human intervention within PAs. However, we can still not estimate the degree of human intervention within PAs due to limitations in the categorical resolution of CORINE land cover. Both flows occur to a large extent within PAs of IUCN Category V (“Protected Landscape/Seascape”), where the human intervention is allowed (Fig 1). Changes in Category V exceed to a remarkable degree the rates of changes observed in Category VI (“Protected area with sustainable use of natural resources”), where human intervention is explicitly needed to maintain the current state of ecosystems. In absolute terms, land cover changes add up also for Category II (“National Park”) which is due to the usually large extend of these PAs.


Fig 1. Sum of LCF4 and LCF5 for different IUCN categories in (A) absolute area of change/year and (B) normalized by size of PAs.

Plotting LCF4 and LCF5 over the size of PAs shows that land cover change occurs at all sizes (Fig 2). Normalising land cover change with the size of the individual PAs indicates a high vulnerability of very small PAs. If they are affected by a land cover change, very large proportions of the PA can be transformed according to CORINE. The limited set of ECOPOTENTIAL PAs covers well the range of observed rates of changes except for very small, highly impacted PAs.


Fig 2. Distribution of LFC4 and LCF5 across the size of PAs (1) in absolute area of LCF and (B) fraction of PA, in red: ECOPOTENTIAL PAs.

In summary, CORINE confirms that changes in “forest and semi-natural vegetation” dominate land cover change in European Pas. It shows that changes in close proximity to PAs occur at considerably higher rates, which can give us some indication on both the functioning of the protection and the prevailing pressures on PAs. The data further indicate a high sensitivity of small PAs to land cover change. Major short-comings of CORINE Land Cover Change could be identified mainly in the spatial and categorical resolution, including the ability to distinguish natural processes from human intervention in the dataset which is of particular interest for PAs. ECOPOTENTIAL PAs all show land cover changes, and their change rates cover a large range of observed change rates for large European PAs. Further refinement of these findings will deepen our knowledge about the applicability of CORINE land cover for the management of PAs. Furthermore, we might be able to compare the results from CORINE land cover change with the targeted reconstruction of past changes from LANDSAT for the ECOPOTENTIAL PAs. After all, this will enable us to provide guidance to improve large-scale land cover monitoring towards the needs of PA management.

Source: ECOPOTENTIAL Newsletter