*4.2. Drainage Capacity*

### 4.2.1. Typical Rainfall

The behavior of the roofs in typical rain, related to the flow of evacuated water, was similar to that presented in the retention coefficient. It was observed that in the first two days of the cycle with rain of 4 and 12 mm (Figure 8), the roof prototype of recycled trays presented higher levels in milliliters per minute than the others. In the same way, it was observed that roofs of gravel and rubber presented deficient levels, showing better performance during the first days. This allowed us to conclude that for rain events equal to or less than 12 mm, the roof system that delivers lower flow rates was the gravel roof, followed by that made out of recycled rubber.

Figure 8 shows that on the third day, when the simulated rainfall was 50 mm, the condition of the roofs was the opposite, and the tray roof was the one with the highest efficiency because it showed the lowest levels of water evacuation. It was observed that the difference between the tray roof and the others was enlarged over time. This was a consequence of the maturity of the vegetal layer, which improved self-regulation and reduced flow delivery times.

When the simulated precipitation was 51 mm, the trend of less flow evacuated by the tray roof continued, and at the same time, it was observed that the gravel roof ended with the highest values. This was because in the first two intervals (30 and 60 min), its evacuation was lower because in the first moments of rain, it retained better; at the end, however, it increased the evacuation quantities.

Finally, it was observed that for 1 mm precipitation, the only roof system that showed water delivery was the tray roof, which demonstrated a high sensitivity to the initial moments of rain, in which it did not demonstrate acceptable performance. It should be noted that due to the design of the drainage layer, it was possible to store water in such a way that it was able to self-regulate flow evacuation and improve retention capacity in more extended and demanding periods of rain.

#### 4.2.2. Intense Rainfall

It was observed that the flow behavior was similar for the four green roofs, with the best performance being the tray roof and the lowest performance being the one made up of bottles. The key factor of the trays was their operation based on their modular design, while in systems with granular-type drainage, in addition to the size and distribution of particles, specific surface area and absorption of the material also had influence. The low performance of the bottle system roof was because it was made up of isolated units that formed cells and not a consolidated and consistent layer structure that facilitated the retention, self-regulation, and homogeneous life of the vegetation.

It was observed that tray roof delivered one of the highest flows at the beginning due to its design. However, this situation was later found to be the opposite, and it was the roof that, over time and in the presence of higher precipitation, behaved more efficiently in self-regulation because it evacuated the drained water more slowly. The above was reflected in lower volumes for each calculated time.

Between minute ten and minute twenty, the drainage curves showed a noticeable increase in the amount of water drained concerning the following intervals; this was because the IFD curve (see Appendix A—Figure A1b) of the selected intense rain (49.7 mm/h), as it was intended to perform in the most precise way. Therefore, and according to the conditions of such curve, in the interval of the first twenty minutes, 27.9 mm were simulated continuously over 16 min and 44 s. From minute twenty, the flows were more regulated according to the conditions of the selected IFD, which generated the stability of the drainage curves. The gravel and recycled rubber roofs behaved similarly, delivering little water at the beginning compared to the others, but they stabilized over time, and despite showing good performance, they were surpassed by the tray roof system.

The results showed that the proposed green roof systems provided good performance in terms of water drainage. In the case of bottle system and tray roofs, an intrinsic ability to self-regulate was evident, which is highly desirable. Responses in respect to the reference green roof made of gravel demonstrated that all considered recycled materials provided similar responses under the same rainfall conditions. Further modifications regarding the arrangement of layers and systems (e.g., bottles and trays) have to be studied to analyze the drainage response and optimize area and material distribution.

#### *4.3. Temperature*

In all the analyzed intervals, the behavior of the roofs tended to preserve the ideal conditions of thermal comfort, which ranged between 21 and 25 ◦C according to the Colombian Sustainable Building Code. However, it was not possible to reach this value in some analyzed segments. However, it was possible to mitigate the effect of external temperature, reducing the intense use of active cooling systems. The following observations were obtained from the temperature behavior obtained from the four different green roof systems:

(a) In the first half-hour of the morning for all the analyzed days, it was generally observed, as shown in Figure 10, that environmental temperature was lower than the temperature of the four roof systems, presenting a change due to the sunrise, where the condition was contrary and the roofs had the lowest temperatures. Before reversing this condition, the roof systems with container type drains (trays and bottles) presented lower temperatures compared to those with a granular type drainage layer (gravel and rubber). The difference between the highest and lowest temperature roof did not exceed, in any case,1◦C.


Regarding temperature, the results showed the same behavior with minimum variations with respect to the reference green roof system (basalt gravel). Therefore, the use of recycled materials, as proposed in this study, is suitable from the perspective of environmental temperature reduction. The next step in this research direction is the analysis of different vegetation layers (growing and development) using recycled materials that can provide better responses with respect to temperature and CO2 reduction.

#### *4.4. Dead Load*

According to Figure 11, the drainage layer was the most representative component when analyzing the variable weight. In green roofs with granular materials, values of 52% for rubber and 68% for gravel were found (these were related to the sum of all green roof components), while, on the contrary, the percentage of this layer only represented 6% for both cases of roof systems with a container-type drainage layer. In these, the most representative component was the substratum, with percentages of 78% for bottles and 70% for trays, in terms of the sum of all their components.

When analyzing the behavior of the drainage layers made with recycled and reused materials compared to the traditional gravel, the second one had a higher weight in all cases. For example, the recycled rubber drainage layer weighed 48% less than the gravel layer. In the case of the layers of bottles and trays, these weighed 98% and 97% less than that of gravel, respectively. Likewise, the low weight of the drainage layers of the PET bottle and HDPE tray systems must be highlighted, since, in addition to reducing the weight more than significantly, they facilitated handling at the time of construction, reducing mechanical risk factors associated with the installation work.

When analyzing the total weight of green roof systems, it was found that the system whose drainage layer was composed of gravel presented a higher percentage of weight than the systems consisting of rubber, bottles, and trays (33%, 72%, and 60% respectively), which made it a more demanding system when designing the structures of buildings due to the required increase of loads.

The composition of the weight inside the roofs showed that in the case of systems with a gravel and rubber drainage layer, the highest proportion was determined by the drainage layer, while the bottle and tray roofs showed the highest representative weight or load.

With respect to the overall sustainability of the proposed green roof prototypes from recycled material, it is important to clarify that this study did not cover the lifecycle assessment of materials involved in the construction of the systems. Therefore, the environmental impacts of materials were out of the scope of this article. Hydraulic and thermal performances were included and widely analyzed. Dead loads and initial cost were also calculated as additional parameters for the proposed green roof prototypes.

In terms of initial cost, it is important to mention that recycled materials provided a cost reduction per square meter that was greater than 10% (in the case of HDPE trays) compared to basalt gravel (Table A1). Maintenance and another lifecycle cost were not included in this article.
