Kipp&Zonen BSRN Scientific Solar Monitoring System - Manual (Page 79)

67
Secondly, it alleviates the potential of therm al shock to the instrum ent which occurs first when the
instrum ent is exposed to direct beam radiation and then again when the instrum ent is shaded. The
actual extent of such shock has not been m easured for all instrum ents, but m ay be significant. Thirdly,
the pair of instrum ents being used to m easure diffuse and global (the redundant m easurem ent) solar
radiation are calibrated sim ultaneously.
A sim ilar transfer m ethod of calibration can also be undertaken during days where there are periods
where the solar line of sight is clear and periods where the sun is covered by cloud. By assum ing that
the disk subtending the angular extent of the sun rem oves an insignificant am ount of diffuse radiation
during overcast conditions, the responsivity of the direct beam radiom eter can be transferred using
a sim ilar set of sim ultaneous equations of two variables and two unknowns.
By grouping all the data obtained from either of these procedures, the uncertainty due to the instrum ent
directional responsivity (cosine and azim uth error) becom es inherent in the coefficients over the multitudes
of sam ples that m ake up the calibration procedure. Conversely, by grouping sam ples with respect
to zenith angle and intensity, the cosine response and the linearity of the instrum ent can also be
determ ined.
In the calibration procedure, care m ust be taken to elim inate the zero offset com ponents associated
with the net therm al radiation of the sensor and its surroundings. ISO 9060 considers a ventilated first
class instrum ent to be one for which this negative flux is less than ± 15 W m for a net therm al flux
-2
of 200 W m . This offset becom es im portant when two instrum ents have significantly different offsets -2
and when the responsivity is transferred from the pyrheliom eter to the shaded radiom eter. If care is
not taken to elim inate the offset, it will be incorporated into the responsivity as an uncertainty in the
calibration slope. At large radiation levels the error is m inor, however, in the diffuse flux, it can be lead
to as m uch as a 20% underestim ation.
Although the zero offset is observed at night it rem ains part of the pyranom eter signal throughout the
daylight hours, especially during clear days. Several m ethods have been used to estim ate the m agnitude
of the offset, with the general consensus being that the m ost accurate m easurem ents of solar irradiance
are those that correct the the zero offset of individual pyranom eters. A m ethodology, however, has
yet to be agreed upon within the BSRN com m unity. Section 9.2.2 outlines two experim ental techniques.
Individual therm al offset calibration tests should be perform ed on site and with the instrum entation
used for the global and diffuse m easurem ents.
Not all locations, nor all instrum ents experience nighttim e therm al offsets. Black and white therm opile
instrum ents are self-com pensating with respect to infrared em issions. Pyranom eters that use different
transducers for the m easurem ent of solar radiation are also not affected in the sam e m anner.
8.4 Pyrgeometer Calibration
Absolute calibration of pyrgeometers is difficult because of the com plex interaction between the instrum ent
and the incom ing signal. This is prim arily due to the difficulty in producing an hem ispheric interference
filter to transm it the broadband infrared signal (approxim ately 4 - 50 :m ) em itted by the atm osphere
and/or earth’s surface to the therm opile detector. Two com plications to be surm ounted through
characterization and calibration are: (1) The absorptance of solar radiation by the dom e causing heating
and thus therm al em issions from the dom e to the sensor surface. (2) The variation of transm issivity
of the dom e over the wavelength range. The first is overcom e by m onitoring the dom e tem perature
and correcting for the increase in signal reaching the therm opile, while the second requires calibrating
the instrum ent in a therm al radiation regim e sim ilar to that in which the instrum ent is to be deployed.
At present no standard m ethod exists for the calibration of pyrgeom eters, but m ost characterizations
are accom plished by applying the Stefan-Boltzm ann Law to a blackbody calibration source. Therefore,
to reduce the overall uncertainty between m easurem ents m ade in various countries using different
calibration techniques, the BSRN Scientific Panel recom m ends that the prim ary calibration of pyrgeom eters
be perform ed at the WRC, or other authorized centres, following the procedures developed by Philipona
et al. (1995). W hile not yet recognized as an absolute calibration, this procedure reduces m easurem ent
uncertainty through the inclusion of varying both the cavity and dom e tem peratures of the pyrgeom eter,
as well as varying the radiative tem perature of the blackbody. All three tem peratures are varied respecting
the m ean annual tem perature of the location of the final deploym ent of the pyrgeom eter. In this m anner,
each instrum ent is characterized for a specific radiation regim e.