122 
Figure 3. Left: change in digital signal measured from a stable light source as temperature was varied 
from 2°C (low) to 38°C (high). Room=23°C. 
Right: Ratio of signal measured at 2°C and 38°C to that at room temperature. 
The implication of this result for applications which use the SE590 to measure spectral radiance is that a 
temperature-dependent correction should be applied to data from the SE590 at least in near infra-red 
wavelengths. Further work is underway to investigate whether this effect is similar for the other CE390 sensor 
heads in the NERC-EPFS and whether applications which use two CE390 sensor heads to measure reflectance 
in dual-beam mode (bi-conical or cos-conical) are likely to be affected. 
3.2 Example: Field calibration of a Spectralon reflectance panel 
An experimental programme is in progress concerned with evaluating the angular reflectance characteristics of 
Spectralon panels used with the EPFS spectroradiometers. The procedure uses a field goniometer with a single 
broad-band (400-760nm) detector and follows that initially described by Jackson et al. (1987). The goniometer 
enables a fixed panel to be accurately positioned initially orthogonal to the Sun, then moved through 10 degree 
increments in the principal plane of the Sun. A small silicon photodiode sensor is kept at all times normal to 
the panel, at a distance of 20 cm above it. The sensor has a spectral response approximating the photopic curve, 
and has an angular field-of-view of 16°. The goniometer enables the directional flux reflected from the panel at 
0° to be measured while the illumination geometry is quickly changed from 0° to 70°. The directionally 
reflected flux measurements are then normalised and expressed relative to one particular angle of incidence. 
Spectralon is increasing becoming the standard reflectance panel for field spectroscopy, 
however it has a non-lambertian response (Jackson ct al., 1992). As a result, it may be necessary to compensate 
for this non-lambertian effect in conditions where skylight constitutes a very small component of the global 
irradiance. 
In this study the departure of the Spectralon panel from lambertian behaviour was determined 
for global (direct + diffuse) irradiation and for diffuse irradiation (skylight) alone by shading the panel from the 
direct solar beam. From these two measurements it was then possible to determine the directional response of 
the panel, i.e. its response to the direct solar beam. This is shown in figure 4. 
Figure 4 also shows the angular response of the panel to global irradiation and to the diffuse 
flux. The presence of skylight acts to reduce the non-lambertian response of the panel. This highlights the need 
to quantify the proportion of direct-to-diffuse flux when applying a correction for the non-lambertian behaviour 
of a reflectance panel in the field. In temperate latitudes the sky is rarely very clear, and skylight is usually a 
significant proportion of the total irradiation.