2.4.3. Monitoring of Performance Parameters

This subsection describes, in detail, the three main variables analyzed in this study: hydraulic performance, thermal performance, and dead load (Figure 5). Each one of these aspects is described in detail as follows:

**Figure 5.** Main variables monitored on the green roof prototypes.

Hydraulic Performance: Retention Coefficient and Drainage Capacity

The benefits of green roofs related to hydraulic performance were evaluated by analyzing the retention coefficient and drainage efficiency. Retention is defined as the volume of water that a green roof can store during and after a rain event. Such a volume of water can be calculated through the equation V = P × A × C, where P is the precipitation (mm), A is the area of the green roof, and C is a measure of the water storage of the roof system which varies between 0 and 1.0, with being 1.0 the maximum retention value [16]. Since the objective was to compare the retention capacity (storage), variable C (retention coefficient) of such equation was solved to find C = V/(P × A). Considering that the post-sowing cycle is a cycle that only occurs once after the installation of the green roof, the post-drought and stable cycles were those taken for the analysis, they should occur permanently during the lifespan of green roofs.

In addition to the water retention capacity, there is a variable that is related to the delivery of water flows, ultimately used for the design of urban rainwater drainage systems, which on cities serve on a large scale to regulate the evacuation times of rainwater to public sewerage systems. Such capacity mitigates the risk of possible flooding or stream formation in extreme events, and reduces the demand for large conduction drainage systems.

The performance of the drainage from the green roofs made with recycled and reused materials is determined by the concept of efficiency, and it is interpreted in an inverse way when the time variable is associated. Usually, a green roof is considered more efficient than those which demand less resources. In the drainage study, on the contrary, the efficiency in the performance of a green roof was recognized as the one that delivers the water flow using the highest amount of time, demonstrating that no associated side effects, such as an intolerable increase in weight or excessive saturation of the systems due to the presence of water, were generated. This condition of efficient performance was identified by having higher flow delivery times, avoiding the collapse of city collection systems for the evacuation of rainwater. For this analysis, the accumulated in milliliters of the drained water was taken for each time interval, which was directly measured on the drainage of green roof prototypes.

The volume of drained water collected in plastic containers was measured and subsequently measured in a 1000 mL cylinder for each rainfall simulations. The collection of retention and drainage information was manually carried out at the same time when precipitation was simulated. These conditions varied for each type of rainfall. In the case of intense rainfall, data were collected every 10 min for one hour, making a consistent relationship with the precipitation intervals of the IFD curve. For typical rain, the frequency of data collection was every 30 min. A pluviometer, chronometer, plastic containers, and a test tube were employed to measure water retention and drainage capacity.

#### Temperature

Temperature data were taken from the environment or project site that corresponded to the space under the tent and at the bottom of the support that contained each green roof prototype, with constant measurement of 24 h every thirty minutes. It is important to clarify that the temperature under the tent was higher than the environment temperature due to the local greenhouse effect caused by the tent. However, all the green roof prototypes were exposed to the same temperature conditions. The temperature was monitored and recorded through an data logger (HOBO ONSET—H08-004-02). One day of temperature monitoring was divided in four periods (morning, noon, sunset, and early morning), and data were collected in the following dates (mm/dd/yy): 07/14/15 for morning and sunset periods and 07/24/15 for early morning period. Therefore, the obtained results were comparatively satisfactory and consistent.

#### Dead Load

The variable of dead load was determined by the weight per square meter of each roof system, and its interpretation was that the less weight it has for each area unit, the more efficient the system was because of the lower structural requirements. To consider the weight of the substratum in a water-saturated state, as suggested by County Flat Roofing (2005) [17] and according to what has been established by the NSR-10 (Known as "Norma Sismo Resistente de Colombia from 2010"or Colombian building code 2010), wet soil weighs approximately 1750 kgf/m3. Both saturated and unsaturated loads were determined in this study.
