Direct 
> 
« 
"3 
OS 
2.5 
2.0 
1.5 
1.0 
0.5 
0 
0 
• ◄ Global 
Diffuse 
10 20 30 40 50 60 70 80 
Incident Angle (degrees) 
Figure 4. Relative angular response of a Spectralon reflectance panel measured under field conditions. 
3..4. Example: The accuracy of hand-held sensors versus use of a fixed support 
Field radiometers, and to a lesser extent spectroradiometers are commonly hand-held. This raises two issues 
concerning accuracy. First, there is the necessary close proximity to the operator, whose clothing can hat e an 
effect upon the signal measured. Kimes et al., 1983 have measured an error in spectral radiance of up 12% 
from this source. Similar effects have been observed in NERC-EPFS experiments: figure 6 shows hand-held 
spectra from a grass surface measured first wearing dark clothing and then a white coat. Second, the pointing 
precision of hand-held sensors is less than fixed sensors and this can introduce uncertainty into the signal being 
recorded.. The significance of this will depend upon the form of the bi-directional reflectance distribution 
function at the angle being sensed. Slight variations in the angle of a sensor viewing a surface towards the hot 
spot can be expected to introduce more error than similar variations affecting a sensor viewing a surface at 
nadir for example. 
Wavelength (nm) 
Figure 5. Effect of white clothing on hand-held measurements over a grass surface. (SE590 sensor head 
approximately 70cm from operator). 
To investigate the accuracy of data collected hand-held compared with a fixed support an experiment was 
performed using a Spectron SE590 to collect a series of spectra over two surfaces: short grass and asphalt. All 
measurements were of nadir reflectance. The mobile and fixed approaches were compared for each target 
surface by obtaining two data sets, each a series of 20 spectral triplets which comprised: (i) a scan of the 
123