Kipp&Zonen BSRN Scientific Solar Monitoring System - Manual (Page 82)

For exam ple: International Pyrheliom eter C om parisons IPC VII, 24 Septem ber to 12 O ctober 1990,
22
R esults and Sym posium. W orking R eport N o. 162, Swiss M eteorological Institute, D avos and Zurich,
March 1991, 91 pages.
70
9.0 Radiation Data Reduction and Quality Assurance Procedures
9.1 Introduction
To be certain that the quality of the data obtained is of a high standard, care m ust be taken from the
initial site set-up through the selection of the instrum entation and DAS to the daily m aintenance of
the radiom eters. Once a voltage or resistance m easurem ent is taken, nothing can be done to im prove
the quality of that m easurem ent. Nevertheless, if quality assessm ent is done in near-real-tim e, any
inaccuracies found in the process can be corrected so that future data are of a higher quality. This
section will suggest a num ber of operations that can be perform ed on the data to aid in the rapid
assessm ent of the m easurem ents.
Although the BSRN Archive has carefully laid out the form at required for the measurem ents to be included
in the archive, it is recom m ended that all m easurem ents be kept in their original form (e.g., voltage,
resistance, counts etc.), either at the network observatory or the parent institution. Maintaining these
data elim inates the need for any back processing of engineering data when new, im proved or corrected
algorithm s need to be applied. Furtherm ore, unless the instrum ent can be shown to be m alfunctioning
or disconnected, data should not be rem oved from the data stream , but only flagged because of unlikely
values. The use of the daily log report of activities associated with the station is crucial when considering
the rem oval of data.
9.2 Standard Data Reduction Procedures
Local quality assurance procedures provide a m eans of assuring that the data are internally consistent
(to som e level of uncertainty) within the site. However, to ensure consistency throughout the network,
m ore is required than sim ply providing calibration traceability; the actual m eans of reducing the transducer
signals to engineering units m ust have com m on outcom es. W ithin the BSRN instrum ents m ade by
a variety of com panies are being used in an attem pt to m easure the sam e radiative and m eteorological
variables. The results of these m easurem ents can only be com pared if the differences are known not
to com e from the algorithm s chosen by individual station m anagers in the conversion of transducer
signals to engineering units. In an attem pt to overcom e this increased uncertainty, the following section
sets out protocols to be followed in the data reduction process.
9.2.1 Cavity Radiom eter and Pyrheliom eter
The reduction of all data from cavity radiom eters should be fully com patible with the W RC procedures
used in the calculation of the W RR and the conversion of electrical signals to irradiance values used
during International Radiation Com parisons and published by the W RC .
22
For pyrheliom eter signals, the conversion should be based upon the assigned responsivity determ ined
through com parison with a cavity radiom eter following the subtraction of any zero signal. The responsivity
of the instrum ent should be norm alized to the tem perature at which the calibration was obtained if
the change in responsivity over the tem perature range of the instrum ent changes by greater than a
0.5%. The irradiance would then be calculated as:
T Z
F = R C T(K) C (V - V )
where F = the irradiance in W m -2
T
R the instrum ent’s responsivity at tem perature T in µV W m
= -1 2
T(K) = the ratio of the responsivity of the instrum ent at tem perature
K to that at the calibration tem perature T
V = the signal in m v under irradiance
z
V = the zero offset voltage.
9.2.2 Pyranom eters
Pyranom eter signals should be corrected for zero offset before the responsivity factor is applied to
the transducer signal. In the sam e m anner as the pyrheliom eter, if the responsivity of the instrum ent