Community Research and Development Information Service - CORDIS


The central hypothesis of this study is that remote sensing indicators of forest cover conditions can be used to measure and map the impact of long-term cover disturbances. This was tested over Papua New Guinea, with NOAA's Advanced Very-High Resolution Radiometer (AVHRR) sensor data of the year 1992 and two Landsat thematic mapper (TM) classifications.
A scale of forest-cover disturbance levels was defined through a set of variables which represent land-cover attributes, anthropogenic and natural factors. The relationship between this scale and the channels AVHRR 1, 2 and 3, the normalized difference vegetation indice (NDVI), and the surface temperature (Ts) was analysed. The combination of Ts in dry and wet seasons, the NDVI in dry season and its local spatial variance in wet season, yields the highest separability of disturbance levels. A forest-dover disturbance map was then produced using an unsupervised classification over part of Papua New Guinea, with an overall accuracy of 79%. The legend of this map was documented with the Landsat TM classifications and the forest-cover disturbance scale. This approach can be applied to monitoring and controlling land cover and aiding land management.

Additional information

Authors: ESTREGUIL C, JRC Ispra (IT);LAMBIN E F, Université Catholique de Louvain, Département de Géographie, Louvain-la-Neuve (BE)
Bibliographic Reference: Paper presented: 9th Conference l'Association Quebecoise de Teledetection, Quebec (CA), April 30 - May 3, 1996
Availability: Available from (1) as Paper EN 39789 ORA
Follow us on: RSS Facebook Twitter YouTube Managed by the EU Publications Office Top