3.2.2. Solar Radiation Calibration Facility
Routine Operations
The Solar Radiation Calibration Facility (SRF) is responsible for all calibration activities within the STAR group. Characterized and calibrated broadband sensors were routinely supplied to the four CMDL baseline observatories plus three additional sites at Kwajalein, Bermuda, and the BAO. Each site has a complement of broadband solar radiation sensors (pyranometers and pyrheliometers) that require regular recalibration, exchange, and upgrade support. The environmental extremes at the CMDL monitoring sites also can compound the difficulty of maintaining reliable suites of monitoring instrumentation. An important SRF goal is to maintain the continuity of field site operations and constantly seek ways to enhance the reliability of the monitoring efforts and quality of the data.
Reference Cavity Comparisons
Annual participation in comparisons of cavity radiometers were conducted
in 1996 and 1997. The comparisons were held during October at the National Renewable
Energy Laboratory (NREL) facility in Golden, Colorado. The results are summarized
and available in NREL publications and establish annual checks of the SRF cavity
ratios to peer instruments, World Radiometric Reference (WRR) reproducibility,
and operational behavior of cavity electronics and stability. Results of such
comparisons establish and help maintain confidence in the SRF reference radiometers.
Hardware and Operational Enhancements
Implementation of operational protocols consistent with BSRN guidelines was a priority during 1996 and 1997. A focus of activity within the CMDL surface radiation monitoring effort was the improvement of sun tracking capability and the implementation of broadband solar component (direct beam and diffuse sky) measurements at all sites. A measurable improvement in tracking accuracy using existing automatic solar trackers was achieved by using a combination of a more accurate algorithm for computation of solar position plus more precise leveling of the solar tracker during installation and the use of a Global Positioning Satellite (GPS) module to provide a continuous, accurate time setting of the solar tracker control computer. The implementation of these three items resulted in tracking accuracies of one tenth of a degree over the course of a day. At remote monitoring sites this tracking accuracy reduces the demand on site personnel and contributes significantly to data quality. Diffuse-sky irradiance measurements using tracking shade disk systems mounted on automated solar trackers were initiated at the CMDL BSRN sites in Kwajalein and Bermuda during 1996, and MLO in May 1997.
Solar Radiation Site Upgrades
Pyranometers mounted on trackers equipped with shade disk systems enable more accurate measurement of the total solar irradiance because of the partitioning of the total global irradiance into two parts: the direct beam and the diffuse sky components respectively. The direct beam component is typically measured with a pyrheliometer mounted on the tracker and the diffuse is measured with a pyranometer shaded by the tracking disk system. During clear sky conditions, when a pyranometer is shaded, the measurement errors due to departures from ideal cosine response are minimized. When this measurement is summed with the direct beam measurement, which can be quite accurate if a self calibrating cavity radiometer is used (±0.3% absolute), the resulting measure of solar irradiance is the best achievable with currently deployable instrumentation.
The Bermuda instrument upgrade was completed in February 1996. Kwajalein was upgraded during April 1996. Environmental and mechanical difficulties with the Kwajalein tracker unit led to a tracker exchange in May 1997. Additional mechanical problems were experienced with the Kwajalein tracker in late 1997 and another exchange of trackers is scheduled for January 1998. The MLO upgrade was completed during May 1997. However, MLO has had diffuse measurement capability for many years using older methodology. An automated solar tracker fitted with a tracking disk system was installed to modernize the MLO installation. The original diffuse system at MLO is also being maintained to provide additional redundancy for this important measurement. During November 1997 the MLO suite of solar sensors, including the tracker, was moved to the new MLO observatory building roof platform.
At the end of 1997 all CMDL radiation measurement sites were equipped for continuous measurement of diffuse-sky irradiance using tracking disk systems mounted on automated solar trackers. The next upgrade is the implementation of direct-beam measurements using self-calibrating all-weather cavity radiometers to further reduce the uncertainties. These upgrades are scheduled for 1998 and 1999. Development and testing of all-weather cavity radiometers for CMDL radiation field sites is in progress.
Special Projects and Measurement Intensives
A prototype version of an open window all-weather cavity radiometer system was tested during the summer of 1996. The unit was based on the Eppley model AHF system and incorporated a ventilated housing unit to protect the open cavity from debris that could enter the opening. Because of size limitations and other considerations, and after additional testing, the concept of an open, all-weather cavity system was set aside in favor of a cavity fitted with a window. The presence of a window adds complexity because of spectral cutoff considerations, particularly at the long wavelength limits. At the end of 1997 the possibility of an alternative window material was being explored.
For several weeks during the winter of 1996 and 1997 and again for 2 weeks in August of 1997, an evaluation of the solar components method of determining total solar irradiance was conducted. This method simply uses the sum of the vertical component of the direct beam and the diffuse irradiance determined from a tracking shade disk pyranometer. This method is recommended by BSRN but has not been extensively tested. During these two periods, known as Wintertime Solar Components Evaluation Experiment (WinSCEE) and Summertime Solar Components Evaluation Experiment (SumSCEE) an array of various pyrheliometers, cavity radiometers, pyranometers, and shade disk pyranometers was assembled at the SRF.
The experiment effectively created nine duplicates of a typical installation at a CMDL radiation monitoring site. A reference irradiance was created by operating absolute cavity radiometers and then constructing a component sum using the average of the cavity radiometers and then summing it with the average from the diffuse group. A set of global irradiances operationally equivalent to typical field measurements was created by using individual pairs of pyrheliometers and tracking disk shaded pyranometers. Departures from the reference irradiance were computed by differencing the reference global irradiance and the global irradiance obtained using individual pairs of pyrheliometers and tracking disk shaded pyranometers. To demonstrate the advantages of the component summation technique for determination of global irradiances, the difference between the reference and a global irradiance obtained using a single unshaded pyranometer was also computed. The measurement intensives were conducted during winter and summer periods to examine the performance of the sensors during different seasons and solar positions. The results of this study were summarized by Michalsky et al. [1998], and the advantage of the component summation technique using cavity radiometers for direct-beam measurement is evident.
In summary, the results indicated that at the 95% confidence level the uncertainty in unshaded pyranometers was about 20 W m-2, but for the sum of an arbitrary pyrheliometer and shade disk pyranometer the uncertainty dropped to 9 W m-2, and for the combination of an arbitrary cavity and a selected shaded diffuse the uncertainty drops to 5 W m-2. This demonstrated that the cavity component sum does tend to meet the BSRN specifications of 5 W m-2. These results are summarized in Figure 3.16. The process of carrying out our first WinSCEE resulted in considerable attention being paid to CMDLs Boulder measurement capabilities and resulted in extended research on the functionality of other research-grade irradiance measuring instrumentation as compared with SIO. The SCEE events also led to the identification and quantification of a phenomenon heretofore unnamed but now widely referred to as a radiation hole and to additional local collaboration on various aspects of cloud radiation properties.
Fig. 3.16. Accumulative percentage of occurrences versus irradiance errors where the irradiance error is defined as the difference between a single instantaneous test observation and the value of a reference observation. This figure is based on over 250,000 individual observations under all sky conditions during WINSCEE. Three difference measurement techniques are compared. The merit of each measurement technique, relative to the reference, is indicated by a lower error for a given percentage of occurrence such as indicated by the 95% and 99% lines.
A second experiment and measurement intensive was conducted during August 1997. The resources of the SRF were utilized to conduct a side-by-side comparison of techniques for determination of global solar irradiance. The SRF utilized absolute cavity radiometers plus tracking disk shaded pyranometers to determine the global irradiance via the component summation method. This system was operated in parallel with a pyroelectric detector radiometer system from SIO . The data from the side-by-side blind comparison are being analyzed at NASA Goddard and results are scheduled to be made available in 1998.