Geoscience Reference
In-Depth Information
measured upward or downward light fluxes to the instrument receiver. The
ends of the light conductor jutting out from the fuselage for 30 cm were pro-
vided with opaque-skinned integrators (over all directions of the hemisphere)
made of opal glass MS-23. The edges of glasses were specially manufactured
to provide cosine dependence (Sect. 1.1). Figure 3.1 demonstrates the relative
deviation curve of the real light conductor instrumental function
f
Ω
ϑ
ϕ
)from
the desired cosine dependence obtained in the laboratory. As the direct radi-
ation provides the essential part of the total flux under clear sky conditions,
Fig. 3.1 illustrates that the systematic uncertainty of the downwelling irradi-
ance measurements does not exceed 2% for solar zenith angles less than 50
◦
,
but for higher angles the uncertainty are increasing. These uncertainties were
taken into account either during the observations or during the data processing
accomplished over two stages.
The first stage of processing of the observational results is called
an initial
processing
and obtained the radiance and irradiance spectra on the basis of
the output signal of the instrument. During the initial processing, the begin-
ning point of the spectrum
(
,
λ
1
was defined through the logical search of the
special benchmark (i. e. the square pulse formed by the mechanical system of
scanning). The first point after the benchmark was assumed as a spectrum
beginning point. Then the background (the value of the dark current) was
defined by the mean value of the signal at several points after the benchmark.
This value was subtracted from the signal magnitude at every wavelength. Note
that the constancy of the background was ascertained during the repeated lab-
oratory measurements. Further, the joining of the spectrum parts (UV, VD,
NIR) was accomplished by excluding the overlap regions while making use of
the known numbers of points at the beginning and at the end of every part.
As graduating scales
λ
i
of different samples of the instrument varied, it was
necessary to carry out the linear interpolation over the spectrum from the
initial scale
λ
i
to pool the data. In the capacity of the
united scale the following set of wavelengths was chosen: 330-410 nm with the
step 1 nm and 412-978 nm with the step 2 nm (365 points in whole), moreover
the joining regions were 410-412 nm (the end of UV and the beginning of the
VD regions) and 698-700 nm (the end of the VD and the beginning of the NIR
regions).
The transformation of
λ
i
to the united scale
the obtained spectrum to energetic units:
mWcm
−2
m
−1
sterad
−1
for the radi-
ance was the concluding step of the initial processing. The calibration was
conducted in the laboratory by measuring the signal from the standard lamp
SI-8 (Kondratyev et al. 1973a, 1974), whose spectrum was known in energetic
units. The special source of the high level stabilization of the electric current
andvoltagewasusedforthelampsupply.Duringthecalibrationofthein-
strument, the registration of the lamp SI-8 irradiance was carrying out using
thelightconductor.Thecalibrationresultwastheratioofthecalculatedlamp
energy (incoming to the instrument input slit) to the output signal of the in-
strument. This ratio was the factor, by which the output signal was multiplied
during the measurements (look at the theoretical normalizing of the instru-
mental functions in Sect. 1.1). The accuracy of the calibration was defined
µ
m
−1
for the irradiance and mWcm
−2
µ
Search WWH ::
Custom Search