The above mentioned collected climatic data were used to calculate the cooling load of a representative, (reference), building for five experimental stations selected from the 23 stations described in chapter 2, (numbered 2, 7, 9, 13, 14). Station numbered 2, as already mentioned, is the reference station while the remaining four experimental stations were placed in the central area of Athens and can be regarded as urban stations. The average values of any parameter characterising these four stations are attributed to a station mentioned as "urban”. Urban stations have been selected after a detailed statistical analysis of all the collected data from the 23 stations. Analysis has shown that it is possible to group them according to different climatological criteria, (Livada et al, 2002). Based on this analysis five representative stations have been selected to be considered in the present study.

In order to satisfy all possible scenarios regarding the buildings’ height in the centre of Athens, the thermal behaviour of a reference building with a. one single floor, b. two floors,

c. three floors, d. four floors, and e. five floors, has been calculated. This building is considered to be representative of the Athens’ buildings. Definition of the representative residential building in Greece, has been performed through a detailed collection of data on residential buildings. The survey included data for more than 1000 residences located in the Western Athens Area and was performed in the frame of a major research program aiming to improve the energy and environmental quality of the region, (Klitsikas and Santamouris, 1997). In parallel, the design of the reference building is according to the statistical data collected for all residences in the country, (Hellenic Statistical Service, 1991). Thus, the average cooling load, (in kWh/m2), of the reference building was calculated for the whole summer period, (1st of May to 30th of September), of the years 1997 and 1998 using the SUMMER BUILDING energy simulation program (1995). Each
floor has an area of 200m2 in square shape. The vertical envelope of each floor is made of double brick with an air gap between the two brick layers and external insulation, in a percentage of 90% and of a concrete layer with external insulation in the remaining 10%. A description of the reference building is presented in Table3.

Simulations of the thermal behaviour of the reference building and calculations of its cooling load for the whole summer period for both years, 1997 and 1998, have been performed using hourly data of the ambient temperature collected at the selected five experimental stations presented in chapter 3. As mentioned,

The same solar radiation data have been used for all stations as a non-significant spatial variation regarding solar radiation values has been observed in Athens (Psiloglou, 1997).

Table 4 shows the mean electrical cooling load in kWh/m2 for the selected five stations as calculated using the measured climatological data for the years 1997 and 1998 The air- conditioner is assumed to be a PTAC air-conditioner with no fresh air input and an electric input ratio EIR of. 0.438 at ARI-rated conditions COP 2.25 . The calculated values are very close to the cooling loads reported by Hassid et al, (2000), for the same period as well as with data collected through different monitoring campaigns. The impact of the different parameters affecting the cooling load of the buildings like the set pint temperature, the cooling schedule, the air flow rate, and the orientation of the building has been studied and reported in details in a previous paper, (Hassid et al, (2000), and will not be repeated here.

We define as heat island energy cost, the difference between the urban station’s cooling load, which is the average value of the central stations’ cooling loads, and those of the reference one. The average heat island energy cost was calculated close to 33.2 kWh/m2 for the year 1997 and 29 kWh/m2 for the year 1998