What are green roofs?

A green roof is a construction system that sustainably maintains a plant landscape on the roof of a building (^{Secretaría Distrital de Ambiente, 2011}). Green technologies vary in their characteristics depending on the implemented substrate and plants. Usually, it is comprised in ascending order starting with a waterproofing membrane, which aims to prevent contact between the water and roof support because pathogens could develop that affect the structural integrity of the building. This is followed by the drainage layer, which channels the surplus rainwater and prevents the passage of the material of the roof’s growth medium. Next, there is the filter layer, which prevents the roots from reaching the roof structure. The substrate is the growth medium of the vegetation and works as a nutritious material that stores water and must have a sufficient volume of air in pores to give the roots the possibility attach themselves (^{Minke, 1992}). Finally, the plant layer provides a vegetation cover (see Figure 1), and given the different weather and chemical conditions the roof’s plants are exposed to (^{Secretaría Distrital de Ambiente, 2011}), the vegetation characteristics must be analyzed in order to ensure the maintenance and sustainability of the roof.

Figure 1 Composition of a Green Roof 

Planting vegetation on the roof not only beautifies the environment (^{Sun, Lin, Sung, Ou, & Lu, 2012}), but it can also provide the benefits inherent to green roofs. Prior research has demonstrated the potential for some items to improve the conditions of these roofs to increase their effects. For example, using moss can improve the performance of rainwater management and the temperature of green roofs (^{Anderson, Lambrinos, & Schroll, 2010}).

Experiments with different green roof substrates have reached similar conclusions. The slower the rate of run-off, the higher the concentrations of all the components studied in the run-off water of the green roofs (^{Teemusk & Mander, 2011}). Additionally, the quality of the run-off from green roofs depends on the differences in the structure of the roof’s layers (^{Hashemi, Mahmud, & Ashraf, 2015}).

To date, the benefits of green roofs have been named, which are currently still under study. Implementing this technology is connected to the conditions of the buildings, where the structure plays an important role in the limitations of installing a green roof, the weight of the substrates and the planting. For example, they must be considered with respect to the load capacity of the roof (^{Wilkinson et al., 2015}). If the building does not meet the structural requirements to bear the weight, adaptations need to be made for the roof’s implementation. The inclination of the roof is an important factor, because a strong slope in areas of heavy precipitation could lead to erosion and damage to the vegetation. Additionally, wind and solar radiation can affect evaporation and, consequently, are important to consider in selecting the plants (^{Minke, 1992}).

It is worth noting that implementing these roofs is recommended in all the properties of a community that has the problems described above, given that the additional cost of a green roof is a private cost and in turn, many of its benefits are public and external to the construction itself (^{Blackhurst, Hendrickson, Asce, Matthews, & Asce, 2010}). Therefore, the cooperation of the community is required to have technologies such as green roofs adapted to the environment and fulfill their functions. In Switzerland, it is required by law for every new construction is offset with green areas (^{Kravits, 2007}). In Bogotá, every new public construction must include sustainable technology in its design (^{Consejo de Bogotá, 2009}).

Buildings that have appropriate water management, consume less energy and reduce greenhouse gases can receive the LEED certification. These sustainability strategies are changing the way in which buildings and communities are planned (^{Green Building Consul, 2017}). This helps improve green technologies that counteract the environmental problems that are currently being seen throughout the world. For example, Germany is the world leader in implementing green roofs, where 10% of buildings use this technology (^{Saadatian et al., 2013}). The increase in commercial value is a determining factor in incorporating green roofs. Tokyo has started to install green roofs to increase property value per square meter (^{Zielinski et al., 2012}). Similarly, the USA records an increase in the value of buildings with green roofs of between 6% and 15% (^{Johnston & Newton, 2004}). In turn, in the private sector, the Novartis building in Bogotá can be highlighted, which was one of the pioneers in using green roofs as an ecological emblem to encourage this model (^{Hoyos-Giraldo, 2014}).

Purpose, scope and limitations

This research aims to analyze a mechanism that helps to gradually reduce cases of urban flooding and contribute to the comprehensive water management of in the northern area of the Chapinero locality. This will be done through two strategies comprised of computer modeling, which aims to analyze the changes in urban flooding by implementing green roofs, and physical modeling, which aims to analyze the behavior of the characteristics of rainwater quality and retention with different green roof alternatives.

This work is an initial study that aims to achieve results related to flooding behavior in a sector susceptible to this urban problem in Bogotá, in order to observe the impact of the sustainable technology of green roofs on the flooding problems in a community. In addition, variations in rainwater quality and the retention capacity of the substrate are expected to be found in order to propose an alternative that contributes to comprehensive water management.

Moreover, the accuracy of the model is affected by the lack of accurate information with respect to leaks, incorrect drainage connections in the storm drain system or garbage problems. During the period in which precipitation simulations were carried out, there was no rainfall in the area where the laboratory that developed the experimental process was located. Therefore, the acidity of the water used for the simulations increased.

Characterization of the study area

There are urban flooding problems in various areas of Bogotá as a result of saturation of the storm drain pipes or the overflow of surrounding channels or rivers. Green roofs have the capacity to counteract said accumulations of water. After analyzing the sectors of Bogotá that suffer from floods, using a map of flooding spots prepared by the District Institute of Risk and Climate Change Management (IDIGER, for the Spanish original), in 2015, the area between Calle 94 and Calle 106, the Nueva Granada Military School and Carrera 15 Avenue was selected.

Parks and green areas act a natural rainwater filter. This is a principle to implement with the installation of green roofs. Parks and green areas are scarce in this part of the city, which means there are many impermeable surfaces that channel the run-off to the storm drains. In Figure 2, you can see the location of the affected area chosen for analysis, from right to left with Colombia, Bogotá, the locality of Chapinero and the north of the Chapinero locality.

Figure 2 Location of the affected area. 

A survey was conducted on September 12 and 13, 2015, on 40 citizens of the sector to find out the information that the surveyed citizens have with respect to the floods that occur in the locality and the mechanisms to mitigate the problem. The survey included questions such as: “How often is there flooding in this sector due to rainfall?”, “Do you know what a green roof is?”, or “Would you be willing to install a green roof on your property to counteract the levels of flooding and increase the environmental quality of Bogotá?”.

The residents of the sector are aware of the floods generated during the rainy seasons, normally between October, November and December. They are unaware of the causes of the flooding or of any sustainable mechanism to counteract this problem. The benefits and disadvantages of implementing a green roof were explained to the surveyed citizens, who were enthusiastic about the idea of implementing this technology in the area. Figure 3 shows the results of the seventh question in the survey, where the majority of the residents said they would be willing to install a green roof on their property.

Figure 3 Question of interest in the community survey. 

The community’s interest in solving the flooding problem in the sector through sustainable technology is shown, which not only counteracts the levels of flooding, but also contributes to improving the environment and works as an example for other communities that have similar problems. The city of Bogotá recorded levels of 877 mm, 506.5 mm and 872.8 mm of precipitation for 2014, 2015 and 2016, respectively (^{Observatorio Ambiental de Bogotá, 2016}), with the last months of the year having the highest rainfall.

Conceptual model

Figure 4 explains the behavior of rainwater in traditional roofs. The image on the left presents the precipitation (a) that generates an elevated rate of run-off through the traditional roofs (b). This water flow is directed to the street drains and finally sent to the storm drain pipes. As the pipes do not have the capacity to drain said volume, the system collapses and results in a flood (d). In turn, the image on the right shows the behavior of precipitation when penetrating a green roof (c). This reduces the rate of rainwater run-off, alleviating the capacity for transporting the liquid through the pipelines, and preventing the saturation of the pipes, subsequent flooding and storage of water in internal reservoirs (e) of the building.

Figure 4 Behavior of flooding with traditional roofs (left) and green roofs (right): a) precipitation; b) traditional roof; c) green roof; d) flooding, and e) rainwater reservoir. 

Computer model

Precipitation reaches the building roofs by gravity. The number of millimeters of precipitation is estimated through the intensity-duration-frequency curve (IDF curve) of the study area. Through Equation 1, the scale of the rainfall intensity is measured with different durations (1 hour, 2 hours, 3 hours, 4 hours and 5 hours) and periods of return (3 years, 5 years, 10 years, 25 years, 50 years and 100 years):

Where i = intensity in millimeters per hour; C1 y C2 = parameters of the most suitable equation for the method of the square minimums; X0 = duration in minutes; d = duration of precipitation in minutes; C1, C2, y X0 are values extracted from the coefficients of the IDF curves calculated by the Institute of Applied Hydrology of Empresa de Acueducto y Alcantarillado de Bogotá. Subsequently, a hyetograph is created with the duration of the precipitation and its corresponding intensity.

Conventional roofs as well as green roofs are modeled as sub-basins to assign a treatment to each roof for precipitation, whether in the form of run-off (conventional roof) or infiltration (green roof). Modeling these elements depends on the area, the slope and the width of the roof, which is obtained through an analysis of maps with the Google Earth tool, understanding the width as the characteristic dimension of the water flow due to run-off. The difference that exists between the modeling of the conventional roof and the green roof lies in the impermeable area and infiltration. For the conventional roof, the impermeable area is the total area, with a Manning coefficient of 0.012 (Environmental Protection Agency, Water Supply and Water Resourses Division, ^{Efr, & M. Inc., 2004}), because of the relatively smooth materials that do not allow infiltration. In turn, for the green roof, there is no impermeable area, because the installation of the green roof is assumed in the whole area available, with a Manning’s n coefficient of 0.24 (^{Environmental Protection Agency et al., 2004}) due to the dense plant layer. Finally, characteristic values of the substrate are assigned with respect to the height of suction, conductivity and initial deficit to consider the infiltration action of precipitation in the green roof.

For modeling the rainwater drainage system, the guidelines of lengths, levels and materials of the plans granted by Empresa de Acueducto y Alcantarillado de Bogotá (EAAB) were followed. The drainage of the area is comprised of PVC pipes as well as a stretch of box culvert. Additionally, characteristics were assigned related to the shape (circular, box culvert), depth, length and roughness (PVC, concrete). There is an inspection chamber (manhole) in the joint of each duct that receives water from the roofs.

Table 1 summarizes the characteristics of the roofs, drainage and precipitation implemented for modeling in the SWMM.

The existing relationship between the roofs, drainage system and rainwater is described below. Figure 5 presents a systemization of the model carried out. In the first instance, the rain meter is the idealization of the precipitation process. Said element assigns the aforementioned hyetograph to each one of the modeled areas to simulate contact between rain and the surface of the roofs in the study area. Once the precipitation is in the areas, it runs off the roof (case: conventional roof) or is infiltrated into the roof (case: green roof). Two models: the first without considering green roofs and the second taking them into account. Each area that the roof covers is connected to an inspection chamber through a discharge line, which channels the precipitation from the roof to the chamber. Subsequently, the rainwater is transported by gravity through the pipes to the discharge points at the end of the drainage system’s lines.

Figure 5 Relationship between the elements of the computer model. 

Green roof prototype

The prototype is comprised of an unwoven polypropylene geotextile in the roof’s drainage layer. Humus, rice husk, sawdust and sand are used as an alternative substrate material. The Sedum, Asparagus plumosus and Soleirolia species are used in the plant layer. Additionally, a rain simulator is designed, which provides rainwater to the prototype, extracting the liquid from an underground tank through a connection with a centrifuge pump. Likewise, a sand filter is designed, which the rainwater passes through once it is filtered by the green roof’s substrate. Figure 6 shows an overhead view of the prototype with its provided number in order to understand the location of each combination of plants and substrates present in Table 2.

Figure 6 Numerical Assignment of the Drawers of the Green Roof Prototype. 

Finally, the rainwater absorbed by the green roof’s substrate passes through a sand filter, comprised in descending order of activated charcoal (10 cm), washed sand (10 cm), fine gravel (6 cm) and thick gravel (5 cm). The purpose of the filter is to reduce the turbidity with which rainwater is extracted from the green roof.

Figure 7 presents the green roof prototype (a). All of the substrates are comprised of a combination of garden soil and alternative materials of 10% and 5% of the weight of the soil in order to obtain a 10 cm thick substrate. Additionally, it shows the sand filter (b) implemented in the process of improving the rainwater quality.

The trials were carried out between January 30 and February 24, 2017.

Figure 7 Green roof prototype (left) and sand filter (right).  

The prototype trials include the simulation of five characteristic precipitation levels of Bogotá (30.6 , 48 , 64.56 , 86.7 and 128.4 mm, respectively), which are within the range of precipitation levels recorded in the Colombian capital, over each one of the eight green roof sections that the prototype contains. The supply of rainwater is stopped half an hour after the start of the simulation and the exit flow is measured every five minutes for another hour, recording the amount of water collected in said time through a rain meter. Additionally, the characteristics of water quality (conductivity, resistivity, suspended solids, salinity, temperature, pH, oxygen and turbidity) are measured before and after being filtered by the green roof substrate.

SWMM Modeling of the Affected Area afectación

When comparing the flood areas notified by the residents with the flooded places resulting from the SWMM, the lack of some variables that were not taken into account is acknowledged, such as garbage, incorrect drainage connections or leaks in the system. The behavior of each flooded chamber in the model is analyzed through the comparison of the level of flooding of said chambers with the results of the levels of the same chambers with different percentages of implemented green roofs. Likewise, the saturation of the current pipes is analyzed before and after the replacement of the conventional roofs with green roofs.

The results produced by the SWMM program firstly indicate the flood points for a period of return of three years. Figure 8 compares the flood points reported by the community (small circles) with the flooded areas as a result of the modeling (large circles). Eight areas of flooding are presented as a result of the model. According to the map of Figure 8, a greater amount of floods actually occurs. In the modeling carried out, only the parameters described in Table 1 were taken into account, but the existence of variables such as leaks, incorrect drainage connections or the presence of garbage, which increase urban flooding, have repercussions on the accuracy of the model and reality.

Figure 8 Comparison between the Actual Flooded Areas (points) and the Areas of Flooding according to the Model (circles). 

Additionally, the following analysis is conducted. Two critical chambers are taken into account according to the results of the previous simulation (chamber 90 and chamber 146). An inspection chamber has contributions from different areas of roofs. The analysis consists of replacing a certain percentage of conventional roofs with green roofs and observing the level of flooding. Subsequently, the percentage of green roofs increases until it mitigates the flooding in the chambers.

Figure 9 shows the ratio between the amount of green roofs implemented and the level of flooding in the chamber. The roofs that directly contribute rainwater to the chambers are replaced. Chamber 90 had an 88.9% reduction in flooding when replacing all of the roofs that had a direct impact on the chamber, for a total replacement of 4732.21 m² of roof area. The decrease in the flooding of this chamber when replacing the roofs that drain rainwater directly onto it is considerable. However, chamber 146 had a 9.5% reduction, maintaining the problem of urban flooding.

Figure 9 Behavior of flooding phenomena depending on the amount of green roofs implemented. 

In a system of rainwater drainage that works by gravity, the accumulation of rainwater tends to increase at lower levels, because the pipes must have the capacity to maintain the volumes of water that the impermeable areas in the surrounding area contribute as well as the water volumes of the impermeable areas at higher levels. The probability of flooding tends to decrease with the implementation of green roofs not only in the surrounding areas of the area of flooding, but also in areas with a higher level than the one of the areas with the problem.

With the aim to counteract urban flooding, green roofs are deemed an alternative that must involve the whole community. Collaboration and cooperation are the starting point from which work needs to be implemented.

To illustrate the discharge performance of rainwater in the pipes with conventional roofs and green roofs in the model, the rainwater drainage system of the affected area is compared at the time of maximum flooding. In Figure 10, the dark-colored ducts indicate a normal drainage capacity, while the light-colored ducts indicate the saturation of the stretch.

Figure 10 Comparison of Pipes in Saturated Status (Left) with Traditional Roofs and Pipes in Normal Status (Right) with Green Roofs.  

Green roof prototype

In terms of the green roof prototype, the variation in water quality is analyzed depending on the type of plant used and the alternative materials mixed with garden soil, which comprises the substrate of each section. The rainwater retention capacity of the prototype’s substrates is also analyzed.

The function of the green roof lies in its capacity to retain rainwater as an alternative to reduce the probability of urban flooding from the accumulation of rainwater. This research proposes reusing rainwater in households as a measure for managing water.

When filtering through the substrate of the green roof, the rainwater’s properties are altered, as well as the conditions of the resource’s quality. The indicators that show a clear alteration are conductivity, resistivity, suspended solids and pH.

Figure 11 and Figure 12 present the variation in the conductivity and resistivity of the filtered water, respectively. Five rainwater simulations are carried out in which the properties of the rainwater before entering the green roof and once filtered by the green roof are analyzed.

Figure 11 Drawer with satisfactory performance in terms of conductivity. 

Figure 12 Drawer with satisfactory performance in terms of resistivity. 

In drawer 1, comprised of garden soil and 10% humus, it is observed that once filtered through the green roof, the conductivity of the rainwater increases and its resistivity decreases. After carrying our five simulations, the values of both conductivity and resistivity stabilized, achieving a similar value to that of rainwater before being filtered by the green roof.

When rainwater passes through the green roof, it takes in particles of garden soil and humus, which increases its capacity to conduct electricity. Therefore, its conductivity increases and its resistivity decreases. Figure 13 shows the behavior of the suspended solids in both the rainwater and water filtered by the green roof. It shows that in the first simulations, the suspended solids increase in the rainwater, although, as in the first two parameters, after five simulations of precipitation, the values tend to stabilize.

Figure 13 Drawer with satisfactory performance in terms of suspended solids. 

pH is a measure that indicates the level of alkalinity or acidity of water; it is an important characteristic to take into account in consuming drinking water, which must have a pH of “between 6 and 8.5” (^{Ministerio de Ambiente Vivienda y Desarrollo Territorial, 2000}). The results produced indicate that sawdust with the Asparagus plumosus plant and sand with the Soleirolia plant are efficient in controlling the pH of rainwater. However, these plants do not have the characteristics to resist extreme environments.

To supply rainwater to the prototype, a tank is used that stores the precipitation through a pump that pushes the liquid into the compartments. As there was no precipitation during the trial period, the acidity of the stored water increased. Therefore, pH values gradually decreased.

A similarity is shown between the pH of filtered water and the pH of rainwater. This indicates that sawdust (see Figure 14) and sand (see Figure 15) in a greater proportion could efficiently control the pH and could elevate this parameter when filtered by the roof. However, it is possible that when increasing the percentage of sand in the substrate of the roof, its drainage capacity would be affected.

Figure 14 Drawer with satisfactory performance in terms of ph (sawdust). 

Figure 15 Drawer with satisfactory performance in terms of pH (Sand). 

The behavior of water turbidity is due to the sand filter, which was backwashed for the second simulation of precipitation. From this point, the turbidity decreased, because a greater percentage of activated charcoal was included and there was no more backwashing in the rest of the simulations.

Apart from controlling the water conditions, as analyzed, green roofs must also enable the passage of rainwater at a rate that guarantees the lowest amount of run-off generated, because the main reason for adopting green roofs is to counteract the problems of urban flooding.

The alternative materials with the greatest rate of run-off are sand and humus. In turn, rice husk and sawdust obtained high levels of infiltration with respect to the other two studied materials. Figure 16 compares the characteristic performance of rice husk as an infiltration material exposed to a precipitation of 128.4 mm/min.

Figure 16 Satisfactory Hydraulic Performance of the Substrate. 

Drawer 2 (10% rice husk) presents a greater intake of rainwater. Although it could be concluded that a greater capacity of water retention is efficient, the surplus in the intake of the resource could generate internal floods on the roof, which would generate elevated run-off. Therefore, it is considered that Drawer 6 (5% rice husk) presents satisfactory infiltration conditions in the green roof prototype.