• Home
  • Current congress
  • Public Website
  • My papers
  • root
  • browse
  • IAC-05
  • B1
  • 6
  • paper
  • Land Cover Changes over 40 Years Detected from Remote Sensing Data

    Paper number

    IAC-05-B1.6.02

    Author

    Mr. Jan Kolar, Charles’ University, Faculty of Sciences, Czech Republic

    Year

    2005

    Abstract

    Land cover changes provide quantitative information on how changing land cover/use has affected the environment and biodiversity in Europe. Such information helps fulfil the decision makers demands to quantify an assessment of the current and future pressures on biodiversity. This paper describes method used in a European Biopress project for mapping land cover changes that occurred between 1950 and 1990 as well as results obtained by interpreting historical aerial photographs and comparing to CORINE land cover based on Landsat satellite images interpretations. Sample set has been prepared consisting of about 70 square windows 30 km x 30 km distributed around Europe primarily over selected nature protected sites. Digital historical aerial photos were pre-processed in several steps leading to georeferenced mosaic of the window. DEM has been used to make orthoproduct more precise. Land cover changes were determined through comparison of overlayed Corine Land Cover database from 90’s (CLC90) polygons with aerial photos in background. Not all detected differences have been declared to be a land cover changes. It has been found that the both geometric and thematic qualities of the CLC90 data vary among countries and also among windows in one country. Also some more complex classes were understood and coded differently (example: continuous/ discontinuous urban fabric). Typically, automatic classification applied in some countries resulted in dataset with significantly smaller polygons then those interpreted manually. The Landsat scenes were used for verification of the geometric and thematic quality of CLC90. The CLC 90 was not corrected during the interpretation. However, when clear mistake in CLC90 was detected, its value was ignored, and newly determined class was taken into matrix of changes. Another feature influencing final result is quality of geometrical transformation of orthorectified aerial photos. Different type and accuracy level of input data influence mutual position of raster photo and overlayed vector polygons. All mentioned aspects are almost impossible to quantify but they significantly determine quality of final results. 73 windows from 17 countries were processed. Total area of interpreted area is 5 782 983 hectares. The changes were found on 614 902 hectares, which is 10.6

    Abstract document

    IAC-05-B1.6.02.pdf

    Manuscript document

    IAC-05-B1.6.02.pdf (🔒 authorized access only).

    To get the manuscript, please contact IAF Secretariat.