Abstract

Potential of IKONOS Satellite Data for Tropical Forest Inventory: A Case Study in Northern ailand.


S. Kaojarern, Ho Dinh Duan
Asian Institute of Technology,
P.O. Box 4,
Khlong Luang,
Pathumthani 12120,
Thailand.
Phone : 662 5245587, Fax : 662 5245597,
Email : skaoja@ait.ac.th, duanhd@ait.ac.th


Hans-Jürgen Stibig
Joint Research Centre of the European Commission (JRC),
Institute for Environment and Sustainability (IES),
Global Vegetation Monitoring Unit (GVM),
TP 263 Via E. Fermi 1,
Ispra 21020, Italy.
Phone +39 0332 789834.


Abstract
IKONOS satellite data is assessed for its potential in identification and mapping of forest cover categories over the test site in Northern Thailand. The primary objective of the study is to investigate this very high-resolution satellite image to what extent the detailed information can be mapped. The detailed information in terms of the forest types, density and structure are evaluated.

It is found that visual analysis and interpretation of IKONOS data was excellent compared to other satellite images. The field measurement of forest parameters including species, position, height, diameter, crown width and others were conducted in the three sampling plots of one hectare each. This information was invented into a spreadsheet for statistical data analysis and geographical information system (GIS) database for spatial data analysis. The information on number of trees, height, basal area, stand density, and forest components including analysis of plant community, i.e.; dominance (Do), frequency (F), and density (D) and also biomass are presented. Data processing of IKONOS images was carried out to spatially adjust and matching the spectral enhanced histogram of the three IKONOS images (long strip of about 5 km by 15 km each) into a mosaic ked image of 15 km by 15 km. The forest classification scheme follows the FAO classification with minor modification to suit the local environment. Three main forests of hill evergreen, mixed deciduous, and dry dipterocarp were found in the study area. The classified image of 16 mapping units with 8 units presenting forest cover types is demonstrated. The area for each unit is calculated. Unfortunately the classified map is not validated due to road accessibility not allow to visit every parts of the study area and a detailed map of forest cover types is not existed.