VEGETATION CHANGE DETECTION USING LANDSAT DATA IN NORTHERN FINLAND : A TEST CASE USING SMOOTHING AND LAPLACIAN FILTERING (15th Symposium on Polar Biology)

• Main Authors: アワヤ ヨシオ, / タナカ ノブヒコ, Yoshio AWAYA, Tuomas HAME, Nobuhiko TANAKA
• Format: Report
• Language: English
• Published: Proceeding 1994
• Subjects: Northern Finland; Polar Biology; Proceedings of the NIPR Symposium on Polar Biology; Tundra
• Online Access: https://nipr.repo.nii.ac.jp/?action=repository_uri&item_id=5235; http://id.nii.ac.jp/1291/00005235/; https://nipr.repo.nii.ac.jp/?action=repository_action_common_download&item_id=5235&item_no=1&attribute_id=18&file_no=1

It has been predicted that some boreal tree species would migrate north several hundred meters annually in response to the predicted global warming due to doubling of CO_2 by the middle of the next century. It has also been said that large parts of the boreal forest and tundra would migrate and disappear because of global warming. The average global temperature has already shown a tendency to increase over the past two decades. In the meantime, twenty years have passed since the launching of Landsat 1 and satellite data have been taken since then. The data make it possible to monitor vegetation over a long time. This paper describes the basic concepts to monitor vegetation shifts using satellite data and a system of detection. The system, consisting of a linear-weighting running-average smoothing and a Laplacian filter, was tested using Landsat Multispectral Scanner data for northern Finland. Vegetation boundaries were defined as zero isopleths of the second derivative of a Laplacian-filtered normalized difference vegetation index image. A comparison of boundary lines derived from 1972 and 1987 Landsat data showed small changes in the vegetation boundary, but no clear evidence of vegetation shifts was found.