General characteristics of solar data

The Bureau of Meteorology maintains a network of high quality ground stations which provide one minute statistics on a range of solar parameters. This network is complemented by a system which provides global solar irradiance and derived products from satellite imagery. Many of these data can be obtained online via the links on the introductory page of this solar information portal.

Ground observations include direct, diffuse and global solar irradiance and terrestrial (longwave) irradiance. All measurements provided from this network are traceable to SI units. As a result, uncertainties are also provided for each set of minute statistics and enable specific uncertainties to be estimated for other derived quantities. Data quality assurance and quality control of the station measurements are such that the initial target of 95% uncertainties for any minute exposure quantity of 3% or 900 Jm-2 (which ever is the greater) have been achieved and improved upon.

As part of the tiered approach, the Bureau derives daily solar exposure over the Australian region from geostationary satellite data with an areal resolution of 25 km2. The algorithm is satellite-specific and as satellite characteristics change, the derived data are adjusted to the ground network data. Studies indicate that, for monthly satellite-derived global exposures, data have an uncertainty of approximately 7%. Since 2010 the frequency of satellite-derived global and direct solar exposure data has expanded from daily to hourly.

As one example of testing the satellite method an intercomparison was undertaken using pyranometer data from 9 network sites from July and August 1997. On average the model agreed with the measurements to within 0.17% (around 0.04 MJ/m2 on a typical clear day) and the majority of measurements agreed within 6% (around 1.5 MJ/m2 on a typical clear day). The satellite method tends to slightly over-estimate the radiant exposure in wet, cloudy conditions, and under-estimate it in dry conditions. On the basis of these and subsequent intercomparisons it is concluded that the satellite model provides useful daily global solar exposure estimates in all conditions, with an error of 7% or better in clear sky conditions and up to 20% in cloudy conditions.

To put these numbers into perspective, one can imagine using the measured radiant exposure at a pyranometer location to estimate the radiant exposure at a point some distance away. The accuracy of the estimation will decrease as we move away from the radiation station. The further we go, the less reliable the estimate will be. In a typical agricultural area such as that around Wagga Wagga, the satellite method becomes more accurate than using the surface station values at a distance typically 40 km from the pyranometer.

Both the satellite and ground-based data are provided with relevant metadata to assist in interpretation of the data and their fitness for purpose. Ground data observed since 1993 are also provided with observation-specific uncertainties. There are periods of missing data, which may be caused by a variety of factors including equipment failure, loss of the transmitted signal from satellite, and temporary site closures.