Measuring and reporting on forest landscape pattern , fragmentation and connectivity in Europe: methods and indicators .

Abstract

This report presents and demonstrates possible solutions to implement two headline policy indicators listed under the biodiversity criteria: the EEA/SEBI2010 Indicator 13 fragmentation and connectivity of ecosystem and the MCPFE 4.7 Indicator Landscape level forest spatial pattern. Focus is clearly on large regions assessment and on the change in the forest landscape structure (spatial pattern), not its function or quality. A brief review of knowledge enabled to select important concepts and principles to address spatial pattern processes likely to have ecological effects. It is proposed to make the assessment at local level with relatively fine-grained data and the reporting per spatial units which best capture local processes without loosing too much information. In some cases, forest losses must be disaggregated from forest gains and treated separately. Measures for MCPFE 4.7 are based on (1) the morphology of the forest cover in terms of core forest (interior forest with a 100m edge width) and forest edge, also providing an insight on connectors, and on (2) the landscape context of forest in its close (50 ha) surroundings (natural context or mixed forest-non forest interface zones with agriculture and/or infrastructure). The temporal stability of core forest (i.e. forest potentially staying in the same conditions), the increase of edges and the loss of forest in a natural context are measured. For the SEBI2010 indicator 13, fragmentation is looked upon when associated to core forest loss and each of the four spatial pattern processes (attrition, perforation, shrinkage, fragmentation/breaking-apart) that potentially contribute to four effects (sample, area, edge, isolation) on forest habitat and species is quantified. Measures on forest connectivity combine the landscape and organism dimensions; they account for the habitat availability and inter-patch functional distances. The measures were based on the application of three methods and GIS techniques. Data inputs were forest-non forest masks, the forest spatial pattern maps obtained by applying the mathematical morphology based software GUIDOS, the landscape patterns maps obtained by applying the landscape mosaic index and the equivalent connectivity area index derived from the Conefor Sensinode software. The analysis was conducted to demonstrate the methods with the only readily available, harmonized, relatively fine-grained and bi-temporal European-wide land cover data from CORINE Land Cover (100 m spatial resolution, 25 ha minimum mapping unit) of years 1990 and 2000. Forest habitat maps do not exist over large regions. For each measure, local spatial information was aggregated per province (NUTS level 2 or 3, 564 provinces in total) and results were presented on the basis of European-wide maps and tabular data. Indicator layers can be queried on line at the map viewer of the European Forest Data Centre (EFDAC): http://efdac.jrc.ec.europa.eu/. The additional delivery of a European-wide snap-shot of hot-spot provinces was proposed to identify provinces where changes in spatial pattern (particularly forest loss, loss of forest in natural context, core forest fragmentation, forest connectivity loss) were significant (both in area and proportionally to the forest). Ecological impacts of spatial pattern processes would be more likely in those provinces. With the data at hand used for demonstrating the methods, 106 hot-spot provinces were flagged. It will be now essential to further compare local change in forest spatial pattern with net forest area change, and add complementary field-based data on forest quality.