According to World Meteorological Organization (WMO) guidelines, only data provided by international networks should be used in aerosol climatology studies. These networks provide a well tracked calibration procedure, good quality standards and homogeneity on the retrievals.
Several of these international networks of ground – based sun photometers have been established over the past 20 years. These are mainly intended to monitor and characterise the atmospheric aerosols, and study their effect on the Earth climate. Nowadays, two main sun photometric networks are up: AERONET (AErosol RObotic NETwork) from the NASA (Holben et al., 1998) and SKYNET (SKYrad NETwork) in Asia (Takamura et al., 2004).
AERONET is mainly distributed in North America and Europe and employs the Cimel CE318 sun photometer as the standard instrument. More than 250 units take part in the AERONET program, which adopts an original processing package (Demonstrat) to analyze the measurements (Dubovik et al., 2000; 2002). This code is not publicly available. In AERONET, the calibration is centralised and should be performed every 12 months, so the instrument must be sent back to specific sites (in United States and France) for calibration and maintenance. The calibration is performed by comparison against a primary master (for sun irradiance measurements) and by comparison against a traceable radiance sphere (for sky radiance measurements).
SKYNET is composed by more than 20 sites in Asia. It holds the Prede POM model as the standard instrument. The Prede data are processed using the code known as Skyrad.pack (Nakajima et al., 1983; 1996). This is an open source package. In SKYNET, the calibration is descentralised. It is performed by the application of the In Situ Improved Langley Plot technique (SKYIL) (Campanelli et al., 2007) for sun irradiance measurements, and by a sun scanning method for the sky diffuse radiance. The insitu techniques avoid data gaps and instrumental damage due to transport. Further differences between both networks and instruments are discussed here.
It is worth to point out that around 250 CE318 units were working out from AERONET in 2003 (Holben et al., 2003) in small independent and local networks, or even standalone, in front of 240 federated instruments). In part, this situation is due to the stringent policies imposed from the network, that generate some reluctancy from many instrument owners to lose control on their instruments or data. Moreover, centralised networks are very demanding in terms of manpower and resources capabilities, so they are reluctant to accept instruments located in well instrumentalised areas, like Europe. This situation, implies that a huge amount of data is not made available, it is difficult to reach, or it has not the quality status and homogeneity from the international networks; on the end it cannot be properly used for regional climatological studies.