Kipp&Zonen BSRN Scientific Solar Monitoring System - Manual (Page 67)

55
Table 7.1 lists the BSRN wavelengths, m axim um displacem ent from the nom inal wavelength and the
m axim um waveband (Full W idth at Half Maxim um ) in order of priority.
Nominal
W avelength
(nm)
Maximum
Displacement
(nm)
Maximum
Bandw idth
FW HM
(nm)
W avelength
Mandatory
Absorption
2
412 2 6 Yes NO
862 4 6 Yes
3 2
500 3 6 Yes O , NO
(m inor)
2
368 2 6 Yes NO
778 2 11
3
675 3 11 O
3
610 2 11 O
Table 7.1. Spectral passband inform ation for BSRN aerosol optical depth m easurem ents.
2
The absorption of NO is of m inor consequence in rem ote or rural areas, but should be accounted
for when AOD m easurem ents are m ade in urban areas.
The use of additional spectral bands, for the determ ination of atm ospheric constituents or increased
inform ation on aerosol optical properties, is encouraged. Aerosol properties, however, will only be
archived for these selected wavelengths at present.
The changing transm ission and the rapid shifting of the central wavelength in interference filters that
have been used in sunphotom eters have caused the rejection of large am ounts of data and brought
even m ore data under suspicion. The m anufacture of interference filters has im proved over the last
decade. Filters produced, using ion deposition technology, should be m ore stable than those m ade
using older technologies. The use of these filters has not yet been proven, however. Several procedures
have been found to reduce filter degradation:
(1) Sealing the radiom eter from the am bient environm ent. Filters have been found to degrade
m ore slowly when the instrum ent is tem perature controlled and the hum idity is controlled.
(2) Ensuring that the filters are exposed to solar radiation only during tim es when observations
are being m ade. This can be accom plished either by using a shutter to block radiation from
hitting the surface of the filter, or by turning the opening aperture away from the radiation
source.
(3) Filters have been found to degrade from the outer edge inward toward the centre. Therefore,
the use of oversize filters will prolong their useful life.
Even with these precautions, the transm ission and waveband properties should be checked regularly.
The use of the “ratio-Langley” m ethod (Section 7.4.2.2) provides a m eans of checking for variations
between wavelengths that m ay be caused by changes in the optical properties of the interference filters.
If changes do occur, the instrum ent should be carefully checked to determ ine the cause of the change
(m ulti-detector instrum ents m ay be showing changes in sensor responsivity). In cases, such as above,
and during tim es when the instrum ent is in an optical laboratory, the interference filters should be carefully
rem oved from the instrum ent and their optical properties m easured using a high-quality bench
spectrom eter. If changes in the bandpass are found so that it no longer m eets the requirem ents set
out in Table 7.1, the filter m ust be replaced. The filter transm ission m ay also vary with age. The principal
investigator m ust determ ine whether the change in transm ission is enough to require the filter be changed.
The tolerance provided by m ost filter m anufacturers is such that obtaining identical filters over a period
of years is unlikely. Therefore, if possible, spare filters should be purchased at the sam e tim e as the