Methods

QGIS4 is a free and open-data software, which permits among others things “to create, to edit, to visualise or to publish information of different georeferenced layers”4. We used the function “Clip” to trim the thermic layer and the green roof data, in order to correspond to the city of Geneva. Indeed, we neglected the thermic data for the whole Canton (which is composed of rural areas in some parts). After reading the literature[3][5][7], we considered that the effect would be more visible in a real urban environment. Indeed, a first reason is concerning the density of green roofs, which is higher in the urban centre than in the neighbourhoods, and a second reason concerns the fact that it is better to avoid rural areas, in order to have a more homogeneous city-like domain of study. After creation of a vector grid with a resolution of 20 metres and corresponding to the size of the city, we assigned for each grid cell a thermal value and a Boolean variable for green roof presence with the attribute table of QGIS and the function “Join attributes by location”. A green roof would give to the local grid case a “1” value, while no green roof would give a “0” value. The data of the grid have then been used in the Geoda5 software, which allows to easily describe patterns with different attributes (here: thermal response and green roof presence). We used different sorts of representations for our results, which are briefly described in the next section and explained and shown in figures 1&2 at the end.