Remote Sensing

With an area of 3,700 square kilometres - with extents of over 70km - the Iwokrama Forest presents a significant management and monitoring task. Remote sensing - satellite or airborne collection of data about various aspects of the landscape - provides perhaps the only means of doing so comprehensively and consistently. Iwokrama chose to begin its work with radar data, largely because optical imagery is faced with significant interference from conspicuous cloud cover, as well as other challenges.

The research team working in this area comprised Vijay Datadin (Iwokrama) Dr. Bruce Chapman of the National Aeronautical and Space Administration - California Institute of Technology Jet Propulsion Laboratory (JPL) and Dr. Eileen Helmer of the United States Department of Agriculture Forest Service International Institute of Topical Forestry (IITF).

The team looked at delineating understory seasonal flooding using canopy-penetrating high-resolution JERS-1 L-band imagery from the Global Rain Forest Mapping Project that is supported by the National Space Development Agency of Japan (NASDA) in collaboration with NASA JPL. Using our understanding of sensor response and some field and ancillary topographic data, we mapped flooded forest areas in a GIS using simple image classification methods. We compared the resulting map with existing digital maps of waterways and forest type based on species composition.

Some correspondence between these data showed that L-band SAR would help delineate seasonally flooded forest in a spatial decision support system for the area and support forest typing, forest ecosystem management and carbon sequestration models. However, unexpectedly high or low proportions of forest flooding mapped in various forest types highlighted the need for multi-temporal imagery that spans a greater range of flooding levels. This study is being submitted for publication and the team is proposing follow-up work.

This project was supported by ITTO PD 10/97, Department for International Development (UK) and by a partial software grant from ERDAS Inc.