5.1.7 Hotspot Analysis

The angular distribution of radiation reflected by a three-dimensional surface that is illuminated by a directional source exhibits a sharp maximum in the retro-reflection direction. Indeed, when observed along the same direction as the incident radiation, only the directly illuminated structures are seen; no shadows are visible, thus there is a peak in the retro-reflected light. This effect is known as the opposition effect in astronomy, the Heiligenschein in meteorology, and the hotspot effect in remote sensing.

Figure 23a Schematic representation of the radiance enhancement between ± 10° from the Earth-Sun line at L-1 (Gerstl, 1999). W stands for full width at half maximum.

Located close to L-1 (4° to 15° from the Sun-Earth line), Triana will acquire images of the Earth near the solar retro-reflection direction. Such images will exhibit an angular signature, as schematically illustrated in Figure 23a (Gerstl and Simmer, 1986; Gerstl, 1988). This viewing direction is useful for the remote observation and monitoring of vegetated land surfaces because of the retro-reflection sensitivity to vegetation characteristics, in particular canopy structure, vegetation leaf structure, vegetation health and stress situations, vegetation amount, and fractional land cover.

The enhanced radiances fall within an observation cone of about 10o around the Earth-Sun line. Since the Earth occupies only 0.5° in EPIC‚s field of view, the entire Earth is within the hotspot region. Under ideal clear sky conditions, the hotspot can cause a doubling of the radiance reflected exactly in the L-1 direction (Gerstl, 1988). The characteristics of the actual orbit around L-1 will allow observations away from the retro-reflection peak, as shown in Figure 23a. Actually, the full angular region between 4 and 15 degrees will be covered as the orbit evolves, thus providing observations of the "wings" of the angular signature.

Figures 23b and 23c depict examples of the anisotropic reflection properties of vegetated land surfaces. Such anisotropic effects are correlated with scattering and absorption events and enable the retrieval of several surface parameters (described below) from the remotely sensed angular distribution of the reflected radiation.

Figure 23b Satellite-measured BRDF distribution for 865, 670, and 443 nm, from a cut through the principal plane in POLDER observations. The reflectances are composed of data from different orbits and spatial resolutions between 15 and 30 km. Although the POLDER angle scale is different, the radiance peak in the retro-reflection direction can be clearly seen.

Hotspot analysis will yield forest-canopy structure data such as canopy height and leaf-phytoelement size and shape by using pre-established correlations between canopy structural parameters and the hotspot parameters {W, C}, where W is the hotspot angular width and C the hotspot strength or magnitude (Gerstl, 1988, 1999). These are results not obtainable by classical remote sensing measurements that primarily rely on spectral signatures (e.g., the vegetation index planned for MODIS). Therefore, the angular signatures from Triana canopy hotspot measurements promise to be an ideal complement to the existing spectral index characterizations of vegetation cover. Continuous observations with Triana will allow us to establish time-series of ecological parameters for all biomes by longitude, latitude, wavelength, and season, which will form the basis data set for a new global hotspot land vegetation ecology (Gerstl, 1999).

Figure 23c Measured canopy hotspot angular distribution from 20 km above a deciduous forest in visible and near infrared wavelengths from the MODIS Airborne Simulator at solar zenith angle of -29 degrees and 50 meter spatial resolution.

Triana data coupled with Terra data will allow an estimate of the hotspot contribution to Earth radiation budget. While this is expected to be small, it may be important as we place tighter and tighter requirements on our estimates of global change.