iroaches. For example, 
much of the knowledge 
e remote sensing data, 
used in remote sensing 
riority should be given 
data are collected from 
REMOTE SENSING OF FOREST ECOSYSTEM DYNAMICS: 
MEASUREMENTS AND MODELING 
Darrel L. Williams, K. Jon Ranson, Robert G. Knox and Elissa R. Levine 
iriance in the measured 
he priorities for further 
on high spectral resolu- 
ER on Japan’s ADEOS 
Biospheric Sciences Branch, Code 923, NASA Goddard Space Flight Center 
Greenbelt, MD 20771 USA 
comments on a draft of 
Wiley & Sons, 734 pp. 
of the landscape, Pro- 
ry 1994. 
biophysical parameters 
University Press, Carn 
al AVHRR data, Inter 
nment, 25, 295-309. 
srties spectra’, Remote 
R.VI) for EOS-MODIS, 
ues to infer vegetation 
nal ISPRS Symposium, 
oumal, 93, 403—410. 
r York, 565 pp. 
getation from satellites, 
reflectance and albedo 
¡ctance (CSAR) model, 
ution Radiometer data, 
Earth’s surface for the 
5-20,468. 
anopies, Proceedings of 
nents, in Imaging Spec- 
ABSTRACT: 
The Forest Ecosystem Dynamics Project at NASA Goddard Space Flight Center is developing an integrated 
approach to modeling of forest dynamics encompassing submodels of forest growth and succession, soil 
processes and radiation interactions. Remote sensing technology is a key element of this study in that it 
provides data for developing, initializing, updating, and validating the models. In this paper we review project 
objectives, discuss data collected and models in use, and describe a framework for studying interactions between 
the forest growth, soil process and energy interaction components. Remote sensing technology used in the 
study includes optical and microwave field, aircraft and satellite-borne instruments. The types of data collected 
during intensive field and aircraft campaigns included bidirectional reflectance, thermal emittance and 
multifrequency, multipolarization synthetic aperture radar backscatter. Synthetic imagery of derived products 
such as forest biomass and NDVI, and collections of ground data are being assembled in a georeferenced data 
base. We then use these data to drive or test multidiscipline simulations of forested ecosystems. Enhancements 
to our modeling environment permit considerable flexibility in configuring simulations and selecting results for 
reporting and graphical display. 
INTRODUCTION 
The Forest Ecosystem Dynamics (FED) project is being conducted by the Biospheric Sciences Branch within the 
Laboratory for Terrestrial Physics at Goddard Space Flight Center, the University of Virginia, and associated 
university investigators. The goal of this research is to use forest succession models, soil process models, and 
radiation scattering models, combined with ground-based and remotely sensed observations, to improve 
understanding of the dynamics of the northern forest ecosystem (Figure 1). 
This research program concerns changes within forest ecosystems at local to regional spatial 
scales (10 2 to 10 5 meters) and temporal scales ranging from daily to decadal periods (10“ 2 to 10 2 years). 
Explanations for spatial patterns and dynamics are sought among mechanisms operating at scales ranging from 
those of physiological processes to long term ecological processes (10“ 4 to 10 3 years). The nature and impacts 
of these changes, as well as the feedbacks to global climate, are being addressed through the integration of 
mathematical models using an object-oriented simulation workbench (see Levine et al., 1993). 
The initial focus of the FED project is the North American transition zone between northern 
hardwood forests and the boreal forest biome. The boreal forest is one of the earth's major vegetated ecosystems, 
accounting for nearly 20% of the terrestrial plant carbon and covering one-sixth of the Earth's land surface 
(Bolin, 1986). The northern and southern margins are especially sensitive to climate change as evidenced by the 
northward migration of boreal species since the end of the last Ice Age. 
FED MULTISENSOR AIRCRAFT CAMPAIGN 
Data to develop and verify models for the FED project come from several sources. The most important sources 
have been intensive field campaigns conducted in cooperation with the University of Maine at International 
Paper's Northern Experimental Forest at Howland, Maine, USA. Numerous investigators coordinated their 
research objectives and activities, and supporting aircraft flights in concentrated Multisensor Aircraft Campaigns 
(MACs). This approach not only made for more efficient use of aircraft hours, but it also fostered cross- 
collaboration of research activities between scientists of diverse interests and expertise. The research carried out 
under the First International Satellite Land Surface Climatology Project (ISLSCP) Field Experiment (FIFE) and 
the Geological Remote Sensing Field Experiment (GRSFE) are two of the better known examples of NASA 
MAC activities. A 1994 special issue of Remote Sensing of Environment highlights results from research 
carried out under two NASA sponsored Multisensor Aircraft Campaigns focused on forest ecosystems: the FED 
MAC study, which focused on the research site near Howland, Maine and the Oregon Transect