Search

De belangrijkste doelstelling van dit proefschrift is het ontwikkelen van methoden voor het kwantificeren van de ruimtelijke verspreiding en omvang van Europese landschappen en habitats en hun monitoring. In een bredere context gaat het om monitoring van de biodiversiteit met behulp van remote sensing bestaande uit het analyseren van satellietbeelden en gebruikmakend van additionele Europese digitale milieubestanden, GIS technieken en veldgegevens. Deze methoden zijn verdienstelijk, maar hebben ook beperkingen vooral bij het karteren van kleine en gefragmenteerde habitats en het monitoren van geleidelijke veranderingen daarin. Daarom is het gebruik van additionele en gestandaardiseerde veldmethodieken noodzakelijk om verschillende componenten van Europese landschappen te monitoren.

Straightforward mapping of detailed heathland habitat patches and their quality using remote sensing is hampered by (1) the intrinsic property of a high heterogeneity in habitat species composition (i.e. high intra-variability), and (2) the occurrence of the same species in multiple habitat types (i.e. low inter-variability). Mapping accuracy of detailed habitat objects can however be improved by using an advanced approach that specifically takes into account and exploits these inherent patch characteristics. To demonstrate the idea, we developed and applied a multi-step mapping framework on a protected semi-natural heathland area in the north of Belgium. The method consecutively consists of (1) a 4-level hierarchical land cover classification of hyperspectral airborne AHS image data, and (2) a kernel-based structural re-classification algorithm in combination with habitat patch object composition definitions. Detailed land cover composition data were collected in 1325 field plots. Multi-variate analysis (Wards clustering; TWINSPAN) of these data led to the design of meaningful land cover classes in a dedicated classification scheme. Subsequently, the data were used as reference for the classification of hyperspectral AHS image data. Linear Discriminant Analysis in combination with Sequential-Floating-Forward-Selection (SFFS-LDA) was applied to classify the hyperspectral images. Classification accuracies of these maps are in the order of 74-93% (Kappa= 0.81-0.92) depending on the classification detail. To subsequently obtain habitat patch (object) maps, the land cover classifications were used as input for a kernel-based spatial re-classification process, in combination with a rule-set that relates specific Natura 2000 habitats with a composition range of the land cover classes. The resulting habitat patch maps illustrate the methodologys potential for detailed heathland habitat characterization using hyperspectral image data, and hence contribute to the improved mapping and understanding of heathland habitat, essential for the EU member states reporting obligations under the Habitats Directive.

This paper presents the research carried out for the spatial analysis of historic land cover changes in relation to the exact location of Natura2000 sites within the framework of the BIOPRESS project. The project has been implemented within the EC-FP5 framework to support GMES Global Moni-toring for Environment and Security'. It was the only GMES project realized under the priority theme "Land cover change in Europe. The BIOPRESS consortium consisted of eight international partners and aimed to provide the EU-user community with quantitative information on how changes in land cover and land use have affected the environment and biodiversity in Europe. Our main stakeholder was the European Environment Agency, through its Topic Centres on Biological Diversity (ETC-BD) and the Terrestrial Environment (ETC-TE). The project produced consistent and coherent sets of historical (1950 1990 2000) land cover change information in and around ca. 100 Natura-2000 sites located from the boreal to the Mediterranean, and from the Atlantic to the continental zones of Europe (http://www.creaf.uab.es/biopress/). The paper will focus on the discussion of the results obtained from analysing historic land cover changes in relation to the distance to Natura 2000 sites in order to determine the relevance of protecting precarious habitats. Results obtained already for the Netherlands point out that conservation measurements have a significant influence on the location, extent and types of land cover changes (Hazeu & Mücher, 2005).

Straightforward mapping of detailed heathland habitat patches and their quality using remote sensing is hampered by (1) the intrinsic property of a high heterogeneity in habitat species composition (i.e. high intra-variability), and (2) the occurrence of the same species in multiple habitat types (i.e. low inter-variability). Mapping accuracy of detailed habitat objects can however be improved by using an advanced approach that specifically takes into account and exploits these inherent patch characteristics. To demonstrate the idea, we developed and applied a multi-step mapping framework on a protected semi-natural heathland area in the north of Belgium. The method consecutively consists of (1) a 4-level hierarchical land cover classification of hyperspectral airborne AHS image data, and (2) a kernel-based structural re-classification algorithm in combination with habitat patch object composition definitions. Detailed land cover composition data were collected in 1325 field plots. Multi-variate analysis (Ward's clustering; TWINSPAN) of these data led to the design of meaningful land cover classes in a dedicated classification scheme. Subsequently, the data were used as reference for the classification of hyperspectral AHS image data. Linear Discriminant Analysis in combination with Sequential-Floating-Forward-Selection (SFFS-LDA) was applied to classify the hyperspectral images. Classification accuracies of these maps are in the order of 74-93% (Kappa=0.81-0.92) depending on the classification detail. To subsequently obtain habitat patch (object) maps, the land cover classifications were used as input for a kernel-based spatial re-classification process, in combination with a rule-set that relates specific Natura 2000 habitats with a composition range of the land cover classes. The resulting habitat patch maps illustrate the methodology's potential for detailed heathland habitat characterization using hyperspectral image data, and hence contribute to the improved mapping and understanding of heathland habitat, essential for the EU member states reporting obligations under the Habitats Directive.

Habitat monitoring of designated areas under the EU Habitats Directive requires every 6 years information on area, range, structure and function for the protected (Annex I) habitat types. First results from studies on heathland areas in Belgium and the Netherlands show that hyperspectral imagery can be an important source of information to assist the evaluation of the habitat conservation status. Hyperspectral imagery can provide continuous maps of habitat quality indicators (e.g., life forms or structure types, management activities, grass, shrub and tree encroachment) at the pixel level. At the same time, terrain managers, nature conservation agencies and national authorities responsible for the reporting to the EU are not directly interested in pixels, but rather in information at the level of vegetation patches, groups of patches or the protected site as a whole. Such local level information is needed for management purposes, e.g., exact location of patches of habitat types and the sizes and quality of these patches within a protected site. Site complexity determines not only the classification success of remote sensing imagery, but influences also the results of aggregation of information from the pixel to the site level. For all these reasons, it is important to identify and characterize the vegetation patches. This paper focuses on the use of segmentation techniques to identify relevant vegetation patches in combination with spectral mixture analysis of hyperspectral imagery from the Airborne Hyperspectral Scanner (AHS). Comparison with traditional vegetation maps shows that the habitat or vegetation patches can be identified by segmentation of hyperspectral imagery. This paper shows that spectral mixture analysis in combination with segmentation techniques on hyperspectral imagery can provide useful information on processes such as grass encroachment that determine the conservation status of Natura 2000 heathland areas to a large extent. A limitation is that both advanced remote sensing approaches and traditional field based vegetation surveys seem to cause over and underestimations of grass encroachment for specific categories, but the first provides a better basis for monitoring if specific species are not directly considered.

A major challenge is to develop a biodiversity observation system that is cost effective and applicable in any geographic region. Measuring and reliable reporting of trends and changes in biodiversity requires amongst others detailed and accurate land cover and habitat maps in a standard and comparable way. The objective of this paper is to assess the EODHaM (EO Data for Habitat Mapping) classification results for a Dutch case study. The EODHaM system was developed within the BIO_SOS (The BIOdiversity multi-SOurce monitoring System: from Space TO Species) project and contains the decision rules for each land cover and habitat class based on spectral and height information. One of the main findings is that canopy height models, as derived from LiDAR, in combination with very high resolution satellite imagery provides a powerful input for the EODHaM system for the purpose of generic land cover and habitat mapping for any location across the globe. The assessment of the EODHaM classification results based on field data showed an overall accuracy of 74% for the land cover classes as described according to the Food and Agricultural Organization (FAO) Land Cover Classification System (LCCS) taxonomy at level 3, while the overall accuracy was lower (69.0%) for the habitat map based on the General Habitat Category (GHC) system for habitat surveillance and monitoring. A GHC habitat class is determined for each mapping unit on the basis of the composition of the individual life forms and height measurements. The classification showed very good results for forest phanerophytes (FPH) when individual life forms were analyzed in terms of their percentage coverage estimates per mapping unit from the LCCS classification and validated with field surveys. Analysis for shrubby chamaephytes (SCH) showed less accurate results, but might also be due to less accurate field estimates of percentage coverage. Overall, the EODHaM classification results encouraged us to derive the heights of all vegetated objects in the Netherlands from LiDAR data, in preparation for new habitat classifications.

Monitoring land cover and habitat change is a key issue for conservation managers because of its potential negative impact on biodiversity. The Land Cover Classification System (LCCS) and the General Habitat Categories (GHC) System have been proposed by the remote sensing and ecological research community, respectively, for the classification of land covers and habitats across various scales. Linking the two systems can be a major step forward towards biodiversity monitoring using remote sensing. The translation between the two systems has proved to be challenging, largely because of differences in definitions and related difficulties in creating one-to-one relationships between the two systems. This paper proposes a system of rules for linking the two systems and additionally identifies requirements for site-specific contextual and environmental information to enable the translation. As an illustration, the LCCS classification of the Le Cesine protected area in Italy is used to show rules for translating the LCCS classes to GHCs. This study demonstrates the benefits of a translation system for biodiversity monitoring using remote sensing data but also shows that a successful translation is often depending on the degree of ecological knowledge of the habitats and its relationship with land cover and contextual information.

At the local spatial scale, land-use variables are often employed as predictors for ecological niche models (ENMs). Remote sensing can provide additional synoptic information describing vegetation structure in detail. However, there is limited knowledge on which environmental variables and how many of them should be used to calibrate ENMs. We used an information-theoretic approach to compare the performance of ENMs using different sets of predictors: (1) a full set of land-cover variables (seven, obtained from the LGN6 Dutch National Land Use Database); (2) a reduced set of land-cover variables (three); (3) remotely sensed laser data optimized to measure vegetation structure and canopy height (LiDAR, light detection and ranging); and (4) combinations of land cover and LiDAR. ENMs were built for a set of bird species in the Veluwe Natura 2000 site (the Netherlands); for each species, 26–214 records were available from standardized monitoring. Models were built using MaxEnt, and the best performing models were identified using the Akaike’s information criterion corrected for small sample size (AICc). For 78% of the bird species analysed, LiDAR data were included in the best AICc model. The model including LiDAR only was the best performing one in most cases, followed by the model including a reduced set of land-use variables. Models including many land-use variables tended to have limited support. The number of variables included in the best model increased for species with more presence records. For all species with 33 records or less, the best model included LiDAR only. Models with many land-use variables were only selected for species with >150 records. Test area under the curve (AUC) scores ranged between 0.72 and 0.92. Remote sensing data can thus provide regional information useful for modelling at the local and landscape scale, particularly when presence records are limited. ENMs can be optimized through the selection of the number and identity of environmental predictors. Few variables can be sufficient if presence records are limited in number. Synoptic remote sensing data provide a good measure of vegetation structure and may allow a better representation of the available habitat, being extremely useful in this case. Conversely, a larger number of predictors, including land-use variables, can be useful if a large number of presence records are available.

Both science and policy require a practical, transmissible, and reproducible procedure for surveillance and monitoring of European habitats, which can produce statistics integrated at the landscape level. Over the last 30 years, landscape ecology has developed rapidly, and many studies now require spatial data on habitats. Without rigorous rules, changes from baseline records cannot be separated reliably from background noise. A procedure is described that satisfies these requirements and can provide consistent data for Europe, to support a range of policy initiatives and scientific projects. The methodology is based on classical plant life forms, used in biogeography since the nineteenth century, and on their statistical correlation with the primary environmental gradient. Further categories can therefore be identified for other continents to assist large scale comparisons and modelling. The model has been validated statistically and the recording procedure tested in the field throughout Europe. A total of 130 General Habitat Categories (GHCs) is defined. These are enhanced by recording environmental, site and management qualifiers to enable flexible database interrogation. The same categories are applied to areal, linear and point features to assist recording and subsequent interpretation at the landscape level. The distribution and change of landscape ecological parameters, such as connectivity and fragmentation, can then be derived and their significance interpreted.