NOVEL EO SYSTEMS FOR EXTENSIVE MAPPING 
OF HUMID TROPICAL FORESTS 
France Fanny Omer Gerard 
Institute of Terrestrial Ecology (ITE) - Monks Wood 
Environmental Information Centre - Abbots Ripton 
Huntingdon, PE17 2LS - United Kingdom 
ISPRS Commission VII / Working Group 5 
ABSTRACT 
Current efforts to map the distribution of humid tropical forests from remote sensing have concentrated 
on mapping the forest/non-forest boundary. Based on shifts in this boundary, the extend and rates of 
deforestation have been estimated. However, the tropical forest biome is far from uniform in 
composition, in ecological function or in its susceptibility to deforestation. Scope for extensive mapping 
of tropical forestry from remote sensing in greater detail has been limited by characteristics of the 
available data. The second Along-Track Scanning Radiometer (ATSR), to be launched by ESA in 1995, 
promises a number of advantages over current sensors, such as AVHRR. Narrower spectral channels in 
the visible and near-infrared regions will make it possible to detect subtle differences in spectral response 
from different forest canopies; near-simultaneous imaging of the same target from different view angles 
offers the prospect of reliable correction for atmospheric effects and of determining canopy structure. 
This paper describes ongoing work to develop methods to exploit these characteristics of the ATSR-2 
system to improve our capacity to map and monitor tropical forests in South America. The project is 
examining, in particular, how differences in (i) phenology, (ii) canopy roughness, and (iii) gap 
characteristics of different forest types manifest themselves through remotely sensed signatures and how 
these effects might be applied in the design of an operational tropical forest monitoring system. The 
paper will describe a programme of ground observations, designed to provide reliable areal estimates of 
forest canopy parameters of scales consistent with continental mapping systems and will give a 
preliminary report of the use of pre-launch simulations in developing models to infer these characteristics 
from remotely-sensed data. 
076