In the construction and validation of all types of models and/or correlations accurate long-term and spatially wide range surface data is needed. In addition to this, and may be more important is that the accurate surface data is used to understand our natural environment, especially the atmosphere which we live in. The derived information can be utilized to develop new revenues in order to reduce the human-made hazards given to our Globe (Page 2005).
Early instruments to measure the solar irradiation were quite sophisticated but not accurate enough to derive reliable and universal models/correlations to use in the estimations. However, they were rather accurate enough to understand some very basic physically effective relations between irradiation and other measured meteorological parameters such as the temperature, relative humidity and bright sunshine hours. A detailed historical development in the measuring instruments is given in Duffie and Beckmann (1991) and somewhat extensive discussion can be found in Coulson (1975).
Angstrom (1924) used a device constructed by him to record the solar irradiation on a photographic film by means of a deflection of a mirror galvanometer. Following scientific works on the relation between the global solar irradiation and bright sunshine hours, especially for the last fifty years used mostly the measurements of two types of instruments for the irradiation: Robitzsch type actinographs and Eppley type black-and white pyranometers, and for the bright sunshine hours: Campbell – Stokes type recorders. Note also that in most of the stations Robitzsch data exist for longer than fifty years while Campbell-Stokes recorders are taking data for more than 100 years. Measurements of the beam and/or diffuse components of the solar irradiation are also carried out regularly in a small number of stations with normal incidence pyrheliometers that follows the sun and/or with a shading ring on the black and white pyranometers (Duffie and Beckman 1991; Coulson 1975).
Robitzsch type pyranometer uses bimetallic strips, one is a white reflector and the other is a black painted absorber exposed to sun, to convert the thermal expansion into the deflection of the pen of the instrument to record irradiation. This instrument needs frequent calibration which is lacking in most stations but may be the more important is the temporal variations in its calibration constants within a year which seemed also instrument dependent (Akinoglu 1992a). Errors in measurements of this instrument may be as high as 30% even for the averages of the readings and it might be far better to use the sunshine based estimation models for global irradiation instead of using the records of this instrument (Akinoglu 1992a). However, the physical relationship between the bright sunshine hours and solar irradiation measured with these instruments might be used in some manner especially in understanding the physical basis of the empirical relations.
Robitzsch type instruments (and also Eppley 180° pyranometers) were replaced with far more accurate Eppley black and white type instruments (or by Kipp and Zonen instruments which use similar principles) for around 40 years in some stations all over the world. But it must be noted that these instruments also need regular calibrations and/or maintenance to produce data less than ±1.5% error (Duffie and Beckman 1991). These instruments basically use a black (hot) and white (cold) regions exposed to sun connected by thermopile junctions producing mV output signal varying with the solar intensity impinging on it. Number of stations with such new radiation measuring instruments is still not enough in most of the countries and also the amount of accumulated data is insufficient for long term investigations of solar radiation modeling. Fully automatic weather stations monitoring all components of the irradiation together with meteorological parameters are also installed at some of the meteorological stations in the last ten years. Robitzsch type pyranographs yet are still recording the solar radiation on the Earth surface in some stations although its data is not accurate enough to drive reliable estimation models.
Controversially, long term records of bright sunshine hours are relatively accurate as the most common instrument used is very simple and quite free of human interfere to run. It is Campbell-Stokes type sunshine recorders. It basically needs the replacement of its recording paper once a day. It contains a spherical glass lens focusing the sun rays on a sensitive record strip placed behind, burning the paper whenever the sun is shining. The trace of the sun collected in the burnt portion of the paper is then read and recorded usually in hourly intervals in hours. The sum for all the hours within a day gives the value of the daily bright sunshine hours. Solar irradiation must exceed a threshold value for the burning of the strip used. As it is clear from the working principle of these recorders, the recorded hours of sunshine carries information about the solar irradiation. However, burning of the strip is accomplished whenever the sun is shinning, provided that irradiation is above the threshold value, regardless of the time within a day. Therefore, it must be noted that the data collected during sunrise and sunset has less effective contribution to the solar radiation amount within a day than the data collected at noon hours.
There are some other problems with the records of Campbell-Stokes instruments which may affect the relation between solar irradiation and the sunshine duration. The irradiance threshold value that can produce a burn can vary from 100 to 300 W/m2 (Painter 1981) or even a wider range of 16-400 W/m2 as determined by Gueymard (1993a). Humidity, frost etc. are some factors that results the loss of records (Aksoy 1999) yet the more important is the extension of the burn during intermittent strong sunshine. This can produce higher outcomes of sunshine durations than the real values (Painter 1981; Gueymard 1993a). Nevertheless, long-term reliable bright sunshine data must have been accumulated as long as 100 years or more in stations from all over the world which is relatively reliable and available to explore, not only for the estimation models but also to use in possible analysis to understand the long-term temporal variations of our atmosphere we live in. Some recent attempts on the subject can be found in references Aksoy (1997) and Chen et al. (2006). Another fact is that the network of sunshine recording sites is denser compared to the sites that records irradiation.
Normal incidence pyrheliometers following the sun are also used to record the bright sunshine hours to be utilized in the models and in the comparisons with the records of Campbell-Stokes recorder. This may aid to determine the errors that would be introduced due to variations of its threshold value with respect to the standard value of 120Wm~2 (Gueymard 1993a), set by the World Meteorological Organization (WMO).
Photovoltaic response of p-n junction is another means of detecting the solar irradiation and various types appeared in rather recent years. The sensitivity however is less than thermopile type pyranometers mainly because of their unsteady spectral response curves which have their maximums usually above the wavelength of 0.6p, m (Duffie and Beckman 1991).
Number of surface stations recording all meteorological parameters should certainly increase to understand our environment and the atmosphere as this would help the future of our Globe. These data should be used in the development and validation of the atmospheric models dynamically and to aid the satellite models of radiation estimation which has great potential of producing spatially continuous solar irradiation maps both for the average and instantaneous energy incomes. To take and reduce the data has become very easy and rather cheap with the new computer technologies and developed user-friendly software which are usually supplied free by the companies selling the instruments.
In literature, a number of data, mostly for the monthly average daily values of both the global solar irradiation and bright sunshine hours, appeared, and those published in between 1978-1989 for 100 locations were tabulated by Akinoglu (1991). These data of course averages of some number of years varying for different
locations and also the sensitivity of the instruments have certainly different values. Averaging however, suppress the random errors that are introduced due to sensitivities of the measuring instruments and hence the set given by Akinoglu (1991) can be used for model development, validation and comparison.
An important source of data is the internet site of WMO, namely World Radiation Data Center in St. Petersburg, Russia (http://wrdc-mgo. nrel. gov/). It is possible to acquire this data by getting contact with the site and any reliable data can be sent to the internet site to be included in their database.