Context and objectives

The preservation of biodiversity is an increasingly important topic in European policy. Natural areas are constantly undergoing changes, mostly driven by socio-economic changes, and many of these changes lead to a reduction of the biodiversity. The following main research questions will be dealt with in this project:

• Which variables should be considered to be retrieved from hyperspectral and multispectral remote sensing, taking into account accuracy requirements of the ecological models?
• How is the uncertainty of remote sensing derived variables propagated into the ecological models?
• To what extent can uncertainty in model output be reduced by replacing present model input data by remote sensing derived variables?
• How can data derived from in situ (field) studies be extrapolated to scales that are relevant for ecological applications (regional, national and European scale)?
• How do differences in RS supported energy balance model formulations (i.e. SEBS and SEBAL) effect estimates of evapotranspiration.
• What is the dependence of the directionality of estimations of elements of the energy balance on different scales.


Methodology

• Campaign management and organisation: Airborne field campaign, field visits, definition of sampling strategy for ecological variables and field spectrometer measurements (ASD FSFR).
• Inventory of all existing field data of the Millingerwaard and Doode Bemde. Coordination between partners to identify parameters that additionally need to be acquired during the campaign. Adding of all data acquired in the Millingerwaard and Doode Bemde in a descriptive manner. Compiling a metadata in an inventory.
• Preprocessing and processing of the airborne spectrometer and field data, including quality analysis, uncertainty estimation and final product generation.
• Generation of higher order data products in the different integrated thematic working groups. Application of models and/or classification schemes.
• Intercomparison of different results. Decision on further combining derived products. Deciding on final data quality and potential improvements. Summary and reporting of findings to involved parties.


Results

Eco-hydrology: Identifying, developing and mapping of proxy for wetness conditions for valley areas.
Flooding and vegetation management: Potential assessment of limnological water parameters using quantitative models to derive Chl a and b content, DOC and suspended matter in turbid and still waters.
Spectral scaling, spectral libraries and spectral unmixing: Integration of field spectral samples into an object-relational data base for the further use in continuous classification and abundance map generation.
Modelling and radiative transfer: Using leaf optical properties models linked with canopy models (e.g. PROSPECT, LEAFMOD and SAIL, GeoSAIL) to simulate the canopy HDRF.
Habitat and vegetation mapping (spatial uncertainties): Land use and land cover classification using combined continuous and discrete classification approaches in combination with the development of spatial metrics to describe patterns of landscapes and to quantify the ecological impact of spatio-temporal dynamics of terrestrial ecosystems.
Species and diversity fragmentation: The use of hyperspectral data will enable an improved detection and identification of land use and land cover changes.