DIRECTIONAL RADIANCE MEASUREMENTS: 
CHALLENGES IN THE SAMPLING OF LANDSCAPES 
D.W. DEERING 
Biospheric Sciences Branch, Code 923 
NASA/Goddard Space Flight Center 
Greenbelt, MD 20771 (USA) 
ABSTRACT: 
Most Earth surfaces, particularly those supporting natural vegetation ecosystems, constitute structurally and 
spectrally complex surfaces that are distinctly non-Lambertian reflectors. Obtaining meaningful measurements 
of the directional radiances of landscapes and obtaining estimates of the complete bidirectional reflectance 
distribution functions of ground targets with complex and variable landscape and radiometric features are 
challenging tasks. Reasons for the recently increased interest in directional radiance measurements are 
presented, and the issues that must be addressed when trying to acquire directional radiances for vegetated land 
surfaces from different types of remote sensing platforms are discussed. Priority research emphases are 
suggested concerning field measurements of directional surface radiances and reflectances for future research. 
Primarily, emphasis must be given to the acquisition of more complete and directly associated radiometric and 
biometric parameter data sets that will empower the exploitation of the "angular dimension" in remote sensing 
of vegetation through enabling the further development and rigorous validation of state-of-the-art plant canopy 
models. 
KEY WORDS: Bidirectional reflectance, BRDF, anisotropy, field radiometry, canopy modeling 
1 - INTRODUCTION 
Measuring land surface parameters, such as the albedo and fPAR, and their temporal changes from satellites 
would be tractable tasks if the surfaces behaved as Lambert planes; but most Earth surfaces and particularly 
vegetative ecosystems constitute complex surfaces that are distinctly non-Lambertian reflectors (Salomonson and 
Marlatt, 1971; Suits, 1972; Kriebel, 1978; Kimes, 1983; Barnsley, 1984; Holben and Fraser, 1984; Deering, 
1989, Deering et al., 1990). The growth and senescence dynamics of the great variety of ecosystems occupying 
the several continents, coupled with the ever-changing illumination conditions (sun position and atmospheric 
conditions) can be expected to result in a wide variety of bidirectional reflectances that may be difficult to 
characterize a priori or from simple satellite measurements. More accurate and reliable estimates of the Earth 
surface albedo, which is an key parameter in global climate modeling, can only be obtained through knowledge 
of the complete pattern of multidirectional reflectances from the Earth surfaces. This is one of the reasons that 
such orbiting sensors as the pointable High Resolution Multispectral Stereo Imager (HRMSI), which is 
scheduled to be included on Landsat-7, and the Multi-angle Imaging SpectroRadiometer (MISR) instrument, 
which is slated as one of the instruments in the suite of future EOS satellite systems, have been proposed. 
The more indirectly-determined land surface parameters, such as vegetative biomass or leaf area index 
(LAI), are even more difficult to estimate from satellite (or aircraft) sensor data since some type of relationships 
between the electromagnetic energy measured by the remote sensing instrument and the biophysical parameter 
must be established. In other words, remote sensing surrogates must be developed to estimate the desired 
biophysical factors. Numerous plant canopy models have been developed to describe the bidirectional 
reflectance characteristics of Earth surface types and to attempt to provide a means to estimate important 
biophysical parameters from spectral radiometric measurements (Goel, 1988). However, they have been 
generally very poorly validated. In order to appropriately interpret data from existing satellite instruments and 
to maximally exploit future space sensors, such as will become available in the near future, a more complete 
understanding of the bidirectional reflectance properties of Earth surfaces is needed, and the need for off-nadir 
viewing ground truth data sets will become more apparent. 
The "promise" of remote sensing still remains as both an opportunity and a challenge today. Earth 
observing satellites still offer the only practical means for 1) measuring important Earth surface variables on 
regional to global scales, and 2) monitoring dynamic Earth system processes or their products (e.g., 
photosynthesis, fPAR, CO,, biomass) at these scales. Unfortunately, many of the bold "promises" of remote 
sensing that came with the first land-focused satellites in the early ’70s are still unfulfilled. Most remote sensing 
scientists no longer believe that the wavelength dynamics of surface spectra hold all of the important keys to