Radar remote sensing to support tropical forest management shared by Eugène Dongmo, Geomatician

J.J. Van Der Sanden

Informations

• Publié par: dongmo9
• Publié le: 19 février 2019
• Nombre de lectures: 6
• Langue: English
• Poids de l'ouvrage: 14 Mo

Extrait

Radar remote sensing to support tropical forest management

Promotoren: Dr.ir. R.A. Feddes  Hoogleraar bodemnatuurkunde, agrohydrologie en grondwaterbeheer  Dr.ir. R.A.A. Oldeman  Hoogleraar bosteelt en bosecologie Co-promotor: Dr.ir. D.H. Hoekman  Universitair hoofddocent bij het departement omgevingswetenschappen

J.J. van der Sanden

Radar remote sensing to support tropical forest management Proefschrift ter verkrijging van de graad van doctor op gezag van de rector magnificus van de Landbouwuniversiteit Wageningen, Dr. C.M. Karssen, in het openbaar te verdedigen op dinsdag 9 december 1997, des namiddags te 16.00 uur in de Aula

The Tropenbos-Guyana Series publishes results of research projects carried out in the framework of the Tropenbos-Guyana Programme. The Tropenbos-Guyana Programme operates within the framework of the international Programme of the Tropenbos Foundation. The multi- disciplinary Tropenbos-Guyana Programme contributes to conservation and wise utilisation of forest resources in Guyana by conducting strategic and applied research and upgrading Guyanese capabilities in the field of forest-related sciences.

J.J. van der Sanden Radar remote sensing to support tropical forest management Tropenbos-Guyana Series 5 Tropenbos-Guyana Programme, Georgetown, Guyana ISBN 90 - 5485 - 778 - 1 Keywords: remote sensing, radar, tropical rain forest, forest resource assessment, forest resource monitoring, sustainable forest management ©1997 Tropenbos-Guyana Programme, J.J. van der Sanden All right reserved. No part of this publication, apart from bibliographic data and brief quotations in critical reviews, may be reproduced, re-recorded or published in any form including print, photocopy, microfilm, electronic or electromagnetic record, without written permission. Printed by Ponsen & Looijen bv, Wageningen Cover: C-, L-, P-band colour composite radar image showing the West Pibiri compartment in Mabura Hill, Guyana. Images acquired by the NASA/JPL AIRSAR system. P-band total power image shown in red, L-band total power image shown in green and C-band total power image shown in blue. Image processing by Martin Vissers.

The investigations reported in this text were carried out at the Department of Water Resources of the Wageningen Agricultural University, with funding from the Netherlands Remote Sensing Board (BCRS) and within the framework of the Tropenbos-Guyana Programme. Department of Water Resources, Wageningen Agricultural University, Nieuwe Kanaal 11, 6709 PA Wageningen, The Netherlands, http://www.slm.wau.nl/whh/rs/ . Netherlands Remote Sensing Board (BCRS), P.O. Box 5023, 2600 GA Delft, The Netherlands, http://www.minvenw.nl/rws/mdi/bcrs/en/ . Tropenbos-Guyana Programme, Lot 12 E Garnettstreet, Campbellville, Georgetown, Guyana, http://www.tropenbos.nl/ . Publication was made possible by the financial support of the Tropenbos-Guyana Programme and the Wageningen Agricultural University.

Abstract

van der Sanden, J.J., 1997.Radar remote sensing to support tropical forest management. Wageningen, The Netherlands (Wageningen Agricultural University), Tropenbos-Guyana Series 5, doctoral thesis, 330 p. This text describes an investigation into the potential of radar remote sensing for application to tropical forest management. The information content of various radar images is compared and assessed with regard to the information requirements of parties involved in tropical forest management at the global, national and local spatial levels. The study distinguishes between the use of radar remote sensing for application to forest resource assessment and forest resource monitoring. Both assessment and monitoring are essential components of procedures for sustainable forest management. The radar data studied are of tropical forest areas near the township of Mabura Hill in Guyana and the city of San José del Guaviare in Colombia. Mabura Hill is comprised of differing intact, primary forest types and forests that have been subjected to industrial selective logging. San José del Guaviare, on the other hand, is characterised by the presence of secondary forests and a variety of non-forest cover types. The available radar data set includes high resolution airborne radar images with differing wavelengths (i.e. X-, C-, L- and P-band) and polarizations, time-series images acquired by the first European remote sensing satellite ERS-1 and a collection of low altitude, nadir-looking, X-band scatterometer measurements. The study makes use of three fundamentally different information sources from the radar return signal: its strength or backscatter, polarization and phase, and spatial variability or texture. Results show that backscatter values computed from L- and P-band radar data and textural attributes computed from high resolution X- and C-band radar data make modest to good and complementary bases for region-based classification of tropical land cover at the level of primary forest types. Textural attributes and backscatter values computed per region from mono-temporal ERS-1 images make modest bases for classifying at the levels of primary forest, logged-over forest, secondary forest and non-forest and poor bases for classifying at the level of primary forest types. Roads are usually the most easily observable indicators of foregoing and/or forthcoming (selective) logging and other human activities in ERS-1 images. Detection of change in road networks by means of ERS-1 images would make a good first step in forest resource monitoring at the national spatial level, in particular. Textural attributes enable the ranking of forest types according to the degree of canopy roughness. Specific textural attributes also allow for quantification of canopy architectural properties. Despite differences in measurement scale, the canopy roughness of the land cover types studied was found to appear similarly in the texture of the available spaceborne and short wavelength airborne radar images. Keywords: remote sensing, radar, tropical rain forest, forest resource assessment, forest resource monitoring, sustainable forest management.

vii

viii

Preface

Research into the use of radar remote sensing for application to forestry first started for me in 1984 as a three months project during my M.Sc. studies at the Wageningen Agricultural University (WAU). It has, with a number of diversions, now resulted in this doctoral thesis. From the beginning, my enthusiastic and knowledgeable guide has been dr.ir. Dirk H. Hoekman. As well as guiding me through the somewhat treacherous field of radar remote sensing, he introduced me to the art of external fund acquisition. Together we succeeded in obtaining the funds on the basis of which I was employed by the WAU for a period of over eight years. To Dirk Hoekman, my co- promotor, I want to direct my first and special acknowledgements. My promotors Prof.dr.ir. R.A. Feddes and Prof.dr.ir. R.A.A. Oldeman are also much appreciated for the supervision of my work and their constructive criticism during the realisation of this text. The investigations described were carried out within the framework of the Dutch National Remote Sensing Programme (NRSP) under the responsibility of the Netherlands Remote Sensing Board (BCRS). I wish to express my gratitude to the BCRS for the financial support given and the confidence trusted in me. The Tropenbos Foundation, the Tropenbos-Colombia team and Demerara Timbers Ltd. are acknowledged for their kind cooperation and logistical support during field visits. I am especially grateful to the members of the Tropenbos-Guyana team who provided me with transportation, accommodated me in their field camps and houses, guided me through Georgetown and the Mabura Hill study area, and taught me their knowledge of the forest. Moreover they shared with me their home-cooked diners, military rations, imported potato chips and local beer. The Tropenbos-Guyana Programme is acknowledged for publishing this thesis in its series. The radar data studied were made available by the European Space Agency (ESA) and the National Aeronautics and Space Administration (NASA). I owe these organisations many thanks for initiating their respective South American radar campaigns and for including the areas of interest to this study in the flight plans. I also want to acknowledge the Colombian and Guyanese aviation authorities for providing the clearance for the various radar over flights. Without the cooperation of these space organisations and national authorities this study would not have succeeded. Part of the work presented in this text was carried out during my six months stay as a visiting scientist with the Canada Centre for Remote Sensing (CCRS). Dr. Frank J. Ahern is much appreciated for giving me this opportunity. To him and other CCRS employees and visiting scientists I want to express my gratitude for sharing with me their ideas and knowledge. I especially want to thank my helpful roommate and PCI EASI/PACE instructor Ron Pietsch. The mentioned producer of image analysis software is acknowledged for providing me with the source code of the module for textural processing by means of a moving window. I wish to acknowledge my colleagues, initially at the Department of Land Surveying and Remote Sensing and later at the Department of Water Resources of the WAU,

ix