Science Service System
You are here : Home : Proposals_Summary

Summary of Proposal RES1201

TitleForest Degradation Mapping with SAR Data in the Republic of the Congo
Investigator Rauste, Yrjö - VTT Technical Research Centre of Finland, VTT Information Technology
Team Member
Dr. Ramminger, Gernot - GAF AG, Forestry Unit
Summary

The scientific activities planned in this AO proposal contribute to the development of operative methods for detection of forest degradation in the tropical region. The forest degradation in this project is defined as a removal of significant amounts of forest biomass in such a way that the area is still considered as forest land. The principal reasons of forest degradation may be related to selective logging and forest road construction, intensive fuel wood collection around rural populated areas, fire, etc. Forest degradation is mapped over an area in the Republic of the Congo in this study.

The produced forest degradation maps can be used to locate and quantify the forest areas that are considered as degraded. The maps can be used to stratify the service area into degradation strata where more detailed ground surveys can be performed using stratified sampling approaches. The aim of the degradation mapping activity is to test candidate methods and data sets for space-borne mapping of degradation using space borne SAR sensors. The technique that demonstrated highest performance at a pre-operative spatial extent will be further used for wide-area mapping purposes in the tropical region. The previous experience of VTT with radar data in tropical forest areas (ERS-1 and ERS-2 in Cameroon and ALOS/PALSAR in the LaoSilva project in Laos) shows that the contrast between degraded and non-degraded forest is relatively small. Therefore the techniques under development in this study are based on a change detection methodology.

Due to a wide spatial extent of the mapped area and higher radiowave penetration at L-band, SAR data from ALOS PALSAR and Envisat ASAR are used for mapping purposes. In this respect, a question of an adequate choice of training and reference data becomes of crucial importance. Conventional approaches utilize optical very high resolution (VHR) data both as training and reference data under an adopted sampling strategy. However, high resolution SAR data is expected to effectively augment or even substitute optical VHR data under some conditions. The objective of this AO proposal is the development of effective sampling techniques combining VHR and TerraSAR-X data, or using exclusively TerraSAR-X data for training and evaluation of the produced maps.

Data requirements: 9 single-pol and dual-pol very high resolution TerraSAR-X scenes and 1 dual-pol stripmap scene at specified locations.

The study described in this AO proposal will be performed as a part of GSE Forest Monitoring/REDD Extension funded by ESA, with GAF AG, Germany as the coordinator of the overall project.

Back to list of proposals

© DLR 2004-2016