Spectral Measurements of Solar Irradiance and Radiance in Clear and Cloudy Atmospheres - Short-Wave Solar Radiation in the Earth's Atmosphere - page 83

between the aircraft transverse axis and the horizon) and the yawing angle (the
anglebetweentheaircraftlengthwiseaxisandthehorizon)-randomlyvary
under the bumps conditions. These random variations lead to corresponding
variations of the observation geometric parameters yielding the additional
random error. The bumps influence mostly the roll angle, its variations can
reach 10 ◦ . Moreover, the bumps become significant during a flight below a cer-
tain altitude. The experience of the IL-14 aircraft flights has shown that this
altitude is 300 m for a flight above a water surface and 500m for a flight above
other surfaces. The experiments described here have been accomplished under
conditions of weak bumps (according to the aircraft classification).
An analysis analogous to the above-mentioned one indicates that the in-
fluence of the roll variations on the observational accuracy is maximum just
for the flight azimuths 90 and 270 ◦ that leads to similar estimations of the
uncertainty (not systematic uncertainty but the random one).
It is evident that it is possible to neglect the influences of the pitch and
bumps during themeasurement of the upwelling irradiance because there is no
principal direction of the upward radiation. The case of observations above the
water surfaces with the mirror reflection could be a certain exclusion, however
themaximumpeak of themirror reflectionwill be shown to turn out (Sect. 3.4)
rather “smearing” due to the ripple. Hence, this peak weakly influences the
upwelling irradiance when the opal glass slightly deviates from the horizon.
The random variations of the pitch, roll and yawing angles lead to the random
variations of the viewingangleduring the radianceobservations.However, with
taking into account the weak dependence of the surface reflection properties
upontheviewingangleitispossibletoneglectthecorrespondinguncertainties.
In some cases, the maximum peak of the mirror reflection could cause the
uncertainty to increase. We should point out that all the above-considered
uncertainties could be neglected under overcast conditions because there is no
direct solar radiation in this case.
The influence of the illumination unevenness together with the ground
surface heterogeneity links with the time interval of the spectrum registration
(7 s). While the observations were made, the aircraft was flying at about 400m
and appearing above the other surface areas. It is obvious that in a clear
atmosphere, the illumination unevenness of the upper opal glass and of the
ground surface is negligible at such a distance. However, while flying below the
heterogeneous clouds, the downwelling irradiance transmitted by clouds can
vary. Moreover, if the surface type varies within the distance of 400m (forest,
tillage, marsh, etc.) the ground surface unevenness influences the accuracy
of the upwelling radiance. Thus, observations of upward radiation can be
accomplished only above areas with a large extent of homogeneous surfaces.
In view of the observational experience, there are only three kinds of such
surfaces: sanddesert, water and snow. Practically, the observationswere carried
out only in cases when the influence of the mentioned factors did not exceed
10%. It was controlled by the visual estimation of the output signal variations
at fixed wavelength within the VD spectral region.
We would like to mention that the uncertainty linking with surface hetero-
geneity decreases fast with flight altitude because the increasing of the surface