*2.1. Green Roof Materials*

This study considered five layers to construct the roof prototypes: (i) the roof structure, (ii) the geotextile, (iii) the drainage layer (using materials aforementioned), (iv) the substratum layer, and (v) the vegetation layer. Figure 1 describes the distribution of layers in the proposed green roof prototypes. Each layer is described as follows:


**Figure 1.** Cross sections of the green roof prototypes made out of recycled or reused materials.

A detailed description of each one of the selected materials for the drain layer is summarized in Table 2, including details of appearance, shape, and typical composition.


**Table 2.** Description of materials selected for drainage layers.

#### *2.2. Green Roofing Prototypes*

Figure 2 shows the full prototype and the cross-section of the four proposed roof systems. The dimensions of the four roof structures were 1.20 m × 0.60 m (W × D). Height varied according to the type of roof, using 0.15 and 0.20 m for drainage with recycled PET bottles and HDPE trays and for basalt gravel and recycled rubber, respectively. Inside the roof structure, a drainage point was installed using a <sup>1</sup> <sup>2</sup> " galvanized pipe towards the center of the structure using a 2% slope.

**Figure 2.** Cross-section detail of roofing prototypes: (**a**) Prototype 1: basalt gravel. (**b**) Prototype 2: recycled rubber. (**c**) Prototype 3: recycled polyethylene (PET) bottles. (**d**) Prototype 4: recycled high density polyethylene (HDPE) trays.

#### *2.3. Rain Simulator Installation*

The rain simulator system was designed to provide controlled and homogeneous drip irrigation under the effective area of the roof prototypes, guaranteeing water flow and the optimization of water consumption. Figure 3 shows the rain simulation full assembly. Water retainers were installed on the roof prototypes to avoid water losses and to provide a better approximation of natural precipitation conditions. Figure 4 summarizes the instruments and materials used to calibrate the rain simulation system.

The rain simulator was calibrated using a pluviometer and by running several preliminary tests to verify the proper functionality of the system. According to the preliminary tests, it was possible to identify that the outer roof prototypes (1 and 4) received a flow rate of 0.58 L/m, while the inner roof prototypes (2 and 3) received a flow rate of 0.61 L/m. Due to restrictions of the experiment, each simulation was carried out on sets of two roofs, always in the same way, prototypes 1 and 2, and prototypes 3 and 4. Then, given the average flow rate per nozzle of 0.6 L/min, the total flow rate in two nozzles corresponded to 1.2 L/min. Based on these flows, it was found that 0.6 min was required to obtain 1 mm of precipitation on each roof. In this way, the time necessary to simulate each rain was calculated. Then, this value was multiplied by the number of millimeters.

Finally, the approximate cost for each prototype (USD per square meter) is summarized in the Appendix A at the end of this manuscript. The cost was directly estimated from the materials employed in the green roof construction (labor not included).

**Figure 3.** *Cont*.

**Figure 3.** Rain simulator system: (**a**) Overall picture of rain simulator and water supply tank. (**b**) Floating irrigation system. (**c**) Water retainers. (**d**) Spatial orientation of the rain simulation experiment with respect to cardinal points.

**Figure 4.** Calibration and rain simulation monitoring elements: (**a**) calibration test tube, (**b**) micro-spray nozzle, (**c**) timer, (**d**) submersible pump, (**e**) chronometer to measure drainage performance, and (**f**) tank for water supply for pumping.
