*3.3. Dead Loads*

Weight was measured during the experimentation process for each green roof prototype (see Table 4). The results showed that both unsaturated and saturated conditions provided the following order of efficiency in terms of weight: (1) PET bottles, (2) HDPE trays, (3) rubber, and (4) gravel. Likewise, it was observed that in both conditions, the green roof with a traditional drainage layer (gravel) showed higher values than roofs made from recycled or reused materials, which provided weight reductions between 33% and 72%.


**Table 4.** Weight per area for each roofing system proposed. Value for 1 m2.

Table 5 summarizes the weight of each component per area for the four green roof systems analyzed in this study. Figure 11 illustrates the graphical distribution of layers through pie charts.


**Table 5.** Weights per area (m2) for components of the studied green roof systems.

**Figure 11.** Weight distribution for each green roof prototype. (**a**) basalt gravel, (**b**) recycled rubber, (**c**) recycled PET bottles, and (**d**) recycled HDPE trays.

#### **4. Discussion**

#### *4.1. Retention Coe*ffi*cient—C*

#### 4.1.1. Typical Rainfall

In Figure 6, it can be seen that on the days when the precipitation was less, the roof prototypes had a higher retention capacity. In roofs with drainage layers of granular materials (gravel and rubber), the performance in low-intensity rainfall provided a retention coefficient of 1.0. In the case of 1 and 4 mm rainfalls, it ranged from 0.84 to 0.94, respectively. Meanwhile, the roofs whose drainage layers consisted of the container type system i.e., bottles and trays, showed retention coefficients of 1.0 and 0.78, respectively, for 1 mm rainfall. In the case of 4 mm rainfalls, they presented retention coefficient values that oscillated between 0.53 and 0.46.

In the case of intermediate intensity rain (12 mm), it was observed that the granular-type roofs presented a decrease in the coefficient concerning their performance in low-intensity rain and an inverse effect of the container-type roofs, which improved their performance with the same type of rain. Despite this effect in intermediate intensity rain, the best performing granular types continued with an average value of 0.61. When the intensity of the rain increased to 50 and 51 mm, the behavior of all roofs showed significant reductions, oscillating their retention coefficient in average values between 0.08 and 0.29, respectively.

The described above behaviors during the typical simulated rain cycles, with significant differences between the granular roof system (gravel and rubber) and the container-type roof systems (bottles and trays), showed that the first ones started with almost a total retention that later abruptly decreased with the increase in the intensity of the rain. In contrast, the second ones on less intense days did not present good behavior, but they later stabilized and provided acceptable retention performances. This was because container types with drainage systems that used bottles and trays had holes in their design that allowed for water flow, which in low-intensity rain made drainage faster and, therefore, higher compared to granular-type roofs in which, due to the size, distribution, and specific surface of its rubber and gravel particles, water had to travel longer distances and take more time to move than due to gravity. In some cases, it was not even possible to evacuate, showing a retention coefficient of 1.0. As simulated rain intensified (and due to the same design conditions), the effects were reversed due to the storage capacity of the container-type roofs—especially tray roofs, which conserved water up to a certain level, thus allowing for the improvement of the retention coefficient. The opposite was evident in the granular-type roofs, which suffered a saturation effect in the presence of intense rain, and the incoming water came out in similar amounts as a consequence of gravity.

#### 4.1.2. Intense Rainfall

In terms of intense rainfalls, it was observed that in both cycles, the behavior of the green roof prototypes in regards to retention capacity was similar, with average values ranging between 0.30 and 0.42. The roof prototype with the highest efficiency in rain retention was the tray-type, with an average value of 0.42. In contrast, the least effective was the bottle-type, with an average value of 0.30. The lowest retention capacity of the roofs was obtained in the PET bottle arrangements due to the existing spaces inside the cells that comprised the system, allowing for the direct flow of water through the roof system. However, it was found that although the percentage of voids per square meter between the bottles was 6%, the impact on the retention coefficient was very low compared to the other systems studied.

To summarize, the results from the retention coefficient (C) were comparable to conventional intensive green roofs, showing even better performance for typical rainfall conditions. The literature has shown that C values for conventional green roofs can range, for example, from 55% to 75% [18], from 39% to 43% [19], and from 43% to 61% [20]. This study demonstrated that is possible to obtain values close to 1.0 in the case of low intensity rainfalls and close to 60% in the case of intermediate intensity rainfalls (for granular type-roofs). To provide better results using recycled materials, it is necessary to characterize all physical properties and preferable geometries to arrange materials in detail. It is important to clarify that this study was limited to use recycled materials without additional processing or modifications.
