The reports contained in the OTTER — FED special issue can be considered building blocks for future 
applications of these advanced sensor systems. Relative to active microwave or RADAR remote sensing, Salas 
et al. (1994) explored the variations in tree dielectric properties as a function of species and time. Weishampel 
et al. (1994) considered the nature of backscatter signals as a function of spatial scale and Lang et al. (1994) 
assessed the causes of strong backscatter signals from red pine plantations. Ranson and Sun (1994) employed 
multitemporal SAR images to conduct forest land cover classification in Maine. Smith and Goltz (1994) 
presented a new modeling approach to the simulation of forest canopy thermal patterns which should be of 
considerable value in thermal infrared remote sensing. 
Studies in the solar reflective region considered both detailed spectral structure and bi 
directional reflectance patterns. Rock et al. (1994) measured detailed spectra of leaf optical properties and sample 
branch stacks for selected species and age classes from the FED MAC site. Lawrence et al. (1994) compared 
remotely sensed reflectance spectra from three airborne spectroradiometers (AVIRIS, AS AS and SE-590). 
Levine et al. (1994) considered possible relations between underlying soil properties and variations in spectral 
vegetation index measurements at the Howland, Maine site. Deering et al. (1994) presented detailed ground 
measurements of canopy spectral, bi-directional reflectance properties. Ranson et al. (1994) explored the 
information content of ASAS measurements for well-characterized subsites. 
Although this research indicates progress accomplished with a range of possible alternate 
sensor systems, it is also evident that the field is becoming more specialized and complex. For example, in 
primary reports to date from the FED MACs, there is no consideration of possible synergies in combining 
observations from the solar reflective, thermal infrared and/or microwave regions. Given the difficulties in 
exploring within any one of these EM regions, it is not surprising that such interactions are not yet considered. 
We anticipate that, with the preservation and publication of these observations on floppy disks or public on-line 
data sets, there will be many opportunities for new discoveries in remote sensing. Ongoing work at the Goddard 
Space Flight Center includes developing a data base (Geographic Information System - GIS) of georefercnced and 
registered image products and field data from the FED MAC study area. In addition to supporting our research 
and that or our collaborators with this data base, we hope to shortly make data in a subset GIS available for 
remote access over the Internet. These spatial data should complement the browse data base and smaller data sets 
being distributed on diskettes for personal computers. They will be suitable for testing ecosystem process 
models, model integration approaches, and remote sensing algorithms, and examining scaling questions and 
ideas for sensor fusion. 
ONGOING ANALYSIS AND MODELING 
Data from the FED MAC and other research at the same study site are being used as checks on model predictions 
of potentially observable attributes (e.g., above-ground biomass, thermal profiles, species composition), and as 
potential sources for extracting biophysical properties of forest canopies, soils, and hydrologic parameters used 
for model inputs. Point to point variation in natural and managed landscapes complicates direct comparisons of 
remote sensing data with results of models of radiation scattering. Effects of soil and management history on 
vegetation can account for much of the variation between large patches (e.g., Levine et al. 1994), whereas a 
dynamic model of the forest population can approximate variability seen within patches (Knox et al., in 
preparation). 
Component Models 
The FED model framework integrates existing models of forest growth and succession (e.g., ZELIG - Smith and 
Urban, 1988; HYBRID - Friend et al. 1992), soil processes (e.g., Levine and Ciolkosz, 1988; GAPS - Riha and 
Rossiter, 1993; RESIDUE - Bidlake et al., 1992), and radiation scattering (e.g., thermal - Smith and Goltz, 
1994; reflective - Verhoef, 1984; microwave - Sun et al., 1991). The forest succession submodel (ZELIG) is a 
spatially explicit individual tree simulator or "gap model." The model simulates the establishment, annual 
diameter growth, and mortality of each tree in a ca. 0.10 ha plot. Simulations can start and stop at any point 
within the life cycle of a forest and reflect changes that are caused by gradual or catastrophic events. In addition, 
the HYBRID model calculates individual short-term photosynthesis and transpiration in a population of trees. 
This makes it suitable for coupling to canopy thermal models and for predicting stand-level gas flux and energy 
exchange. HYBRID is less suitable for long-term simulations of forest structure because of its heavy 
computational demands. The soil process submodel is based on mechanisms operating during the genesis of a 
soil. Included in the submodel are short term processes such as water flux, gas flux, ion concentrations and 
decomposition. Longer term processes such as sesquioxide formation, organic matter, cation exchange capacity, 
water holding capacity, bulk density and soil structure are also considered. We are implementing it as a set of 
parallel models, each focusing on one cluster of soil processes. The radiation interaction models consider the 
energy environment within and external to the canopy and include solar radiation as a modifier to plant growth 
and energy (optical and microwave) as remote sensing signal. The forest succession and soil process models 
require initial information on available species, and characterization of soil horizons. Radiation scattering