1 Weather – Month 9999, Vol. 99, No. 99 2001). Electrical devices are responsible for a large proportion of the energy consump- tion of most buildings. Illumination repre- sents 20–30% of the energy consumption in office buildings and 10% of the total energy consumption across the residential sector (Chirarattananon et al., 2002). Bartzokas et al. (2003) investigated the distributions of sky luminance in Bratislava and Athens during the winter using global and diffuse illumination and solar radiation data col- lected at 5-min intervals over 5 years. Gueymard (2008) developed the Reference Evaluation of Solar Transmittance model with two bands (REST2), to estimate solar radiation, illumination and photosyn- thetic active radiation in a cloudless atmos- phere. In this model, air mass and turbidity are used as parameters that affect the illu- mination level. Gueymard (2008) claimed that the REST2 model estimates are more accurate than those of other models when compared with measured data. Cucumo et al. (2008) proposed the use of diffuse efficacy values with local constants for the Arcavacata region and calculated the illumi- nation level of a surface under all sky condi- tion categories, obtaining valuable results. As no daylight database exists for Izmir, day- light availability is evaluated by comparing three models: the Aydinli model, the Perez model and the IESNA model (Kandilli and Ulgen, 2008). If measurements of illumina- tion are not available, these models may be helpful in calculating the theoretical day- light values for specific locations and times. Daylight is one of the most important resources in reducing the demand for electric- ity used for illumination in green and energy- efficient buildings. However, although global solar radiation measurements have been recorded for a long time, often by very large measurement networks, daylight illu- mination measurements are not made by the State Meteorological Organisation or other companies in Turkey. The absence of this data may cause problems for designers and engineers who work on green buildings. The aim of this study is to analyse and pre- sent the spatial relationship between global solar radiation and illumination levels and to evaluate maps of daylight illumination potential in and around Istanbul. Meteorology’s observation park in Ankara- Kecioren by Hasdemir (1985), and the data recorded by these instruments are avail- able for the period 1980–1982, inclusive. The relationship between daylight illumi- nation and global solar radiation values was established, and the previously measured solar radiation data were converted into illumination values. Hasdemir (1995) also investigated the ratio of hourly variation in global illumination to the hourly variation in extra-terrestrial illumination. Treado and Kusuda (1981) collected solar radiation and illumination data at the National Bureau of Standards facility near Washington, DC. They investigated the relationship between these variables under various sky conditions, and were able to produce empirical equations with which they estimated the diffuse and total illumi- nation as a function of the diffuse and total solar radiation. Chandra (1996) calculated the illumination values for a few cities in India using the luminous efficacy (the meas- urement of useful light) data for different solar altitudes and atmospheric conditions. In addition, as Muneer (1995; 2004) sug- gested, solar radiation prediction models may also be used for daylight illumination predictions, because the parameters of solar radiation and outdoor illumination exhibit similar behaviour. One of the most widely accepted models was developed by Perez et al. (1990). In this model, sky conditions are divided into eight different clusters for the calculation of illumination values under a variety of scenarios. Different coefficient values were estimated for each sky condi- tion category. The resultant model (now known as the Perez model) is considered to be one of the most successful daylight illumination models in the literature. Li and Lam (1999) compared the original and local- ized for Hong Kong versions of the Perez model, resulting in a model that performed better when the local empirical coefficients were used. Additionally, they developed daylight illumination models and validated them using observational data recorded in Hong Kong (Li and Lam, 2000). Illumination and cloudiness of the sky are the primary factors that determine the abun- dance of daylight on a given day (Mardaljevic, Bihter Durna 1,2 and Ahmet Duran Şahin 1 1 Meteorology Department, Istanbul Technical University, Istanbul, Turkey 2 Meteorology Department, Ondokuz Mayis University, Samsun, Turkey Introduction The Sun is the world’s primary source of renewable energy. Recently, passive solar energy systems, especially those related to daylight illumination, have been more regu- larly incorporated into building projects by engineers and architects. The use of day- light illumination as a new source of energy savings is becoming more commonplace in energy planning for green buildings (Wang et al., 2009; Feng and Li, 2011; Sharp et al., 2014). Indoor and outdoor illumination levels are related to electricity consumption at the level of individual houses and for cities as a whole, as well as across countries and the world in general. Unfortunately, up to now, daylight illumination in Istanbul has not been given much attention by scientists, architects or engineers. This lack of information could result in higher than expected electricity consumption. It has been documented by Vlachou et al. (1998), Santamouris et al. (2015) and Leung (2018) that the generation and consumption of large amounts of electricity increases the effects of global warming and reduces the availability of energy to subse- quent generations. The use of renewable energy sources is a core element of climate change management and mitigation, and the evaluation of daylight illumination data is therefore an important avenue of study. Daylight illumination is of interest to researchers in various fields and numerous models have been developed for theoreti- cal and practical purposes. The Standard Sky Luminance Model was developed by the International Commission on Illumination (Commission Internationale de l´Eclairage; CIE, 2003), and it uses different coefficients to represent various sky conditions. According to CIE (2003), the sky coverage is divided into 15 types. A pyranometer and a luxmeter were installed at the General Directorate of Mapping of daylight illumination levels using global solar radiation data in and around Istanbul, Turkey