(665b) Variations in Temperature in Mini-Roof Structures Employing Different Roofing Materials

The overall goal of green roof research at the University of Alabama at Birmingham (UAB) campus is to identify a suitable roofing material combination that reduces energy cooling costs. Since 2008, ambient temperature data has been recorded to better understand how the roof heats and cools on a diurnal, seasonal, and yearly basis. During this study, 15 mini-roof combinations were observed for trends in internal temperatures. Conventional roofing systems included river rock, crushed marble chips, black roofs, reflective (white) roofs, concrete pavers coated with epoxy, and an alternative roofing system involving vegetated roofs (planted with sedum plants). To investigate the thermal properties of the roofing structures, an ambient temperature probe was placed inside of each roof to record data every 10 minutes.

The University of Alabama at Birmingham (UAB) is located in the Southeastern United States. This region is characterized by its hot summers and relatively mild winters. Due to these hot summers during which the temperatures can approach 37.8^oC (100°F), along with conditions of high humidity, it is crucial to have a roof which minimizes cooling costs. These hot summers also create an urban heat island effect in the Birmingham area due to the current roofing structures and facility usage. An urban heat island is defined on the basis of a temperature difference between urban and rural area, caused by human modifications to the natural geographic landscape [Cermark et al., 1995]. The human modifications observed here are the buildings and roads added to the natural habitat. These structures retain the heat transferred from the sunlight increasing the temperature. This property results in an increased temperature in the highly modified areas (urban downtown areas) versus a less modified area (rural areas).

Historically, studies on green roofs have explored their energy performance compared with traditional roofs. Thermal performance indicated a significant reduction (~40%) of a building cooling load during the summer period [Spala et al., 2008]. Similar results were achieved for a nursery school, with reductions ranging from 6% to 49%, and reduction ranging from 12% to 87% on the last floor of the nursery school [Santamouris et al., 2007]. Wong et al. [2007] note that green roofs tend to experience lower surface temperatures than the original exposed roof, especially in areas well covered by vegetation. When green roofs are well covered by vegetation, the resulting substrate moisture will tend to keep substrate temperature lower than the original exposed bare roof. These studies determined that over 60% of the heat gain was mitigated by vegetated roof systems. Summertime data have indicated significant lower peak roof surface temperature and higher nighttime surface temperature for green roofs as compared to conventional roofs [Sonne, 2006].

The roofing materials used are all standard commercial flat roof materials. Flat roof materials were only looked at during the study, since the primary application for the roofing combinations will be on a commercial flat roof top, and not a slanted roof structure. Each mini-roof is 2.4-m (8-ft) long x 1.2-m (4-ft) wide x 1.2-m (4-ft) deep. A number of different roofing systems are being examined for their energy performance. All roofs are insulated with 5.1 cm (2.0-in) of extruded polystyrene. Then the particular roofing combination being investigated is applied over the insulation and sealed. The roofs also include a proper drainage spout, to ensure correct water evacuation, such as on a real roof.

We have also collected surface temperature measurements during the month of June on the various roofing materials used with our mini-roof systems. During this time period, we have observed roofing surface temperatures ranging from 20.6^oC to 82.2^oC. The lower surface temperature values, 20.6^oC (69.1^oF) to 48.3^oC (118.9^oF), are found in a loose rock or stone roof combination. The higher temperature values occur on the roofing combinations primarily made of a coating or membrane; these roofs exhibit temperatures roughly between 25.6^oC (78.1^oF) to 82.2^oC (180.0^oF). The vegetative mini-roofs exhibit surface temperatures ranging between 21.7^oC (71.1^oF) to 52.2^oC (130.0^oF), while their large counterpart, the pilot roof on top of Hulsey Center, exhibits higher temperatures ranging from 31.7^oC (89.1^oF) to 61.1^oC (142.0^oF). The lighter the roof, generally the cooler it is. For example, the white Firestone SBS is cooler than the black Firestone SBS on any given day. This thermal property is observed due to the reflectivity of the roof. The darker colored roofs absorb more incoming light radiation than the light colored roofs causing the dark roofs to become hotter. (Since the roofs temperatures observed are taken during the late spring and early summer, it can be inferred that overall roof temperature will increase during late summer).

When examining how to lower air-conditioning costs effectively and mitigate the urban heat island, the materials of the roofing structure come into play. Different roofing materials transfer heat into a structure in different ways, but by the same mechanisms. Heat that is being transferred into the roofing structure is primarily transferred through conduction and convection. The internal temperatures that have been investigated are affected by convection, conduction, and radiation; this is caused due to an energy transfer from the exterior surfaces of the structure to the internal surfaces [Clements and Sherif, 1998]. These are the key factors that affect the performance of the air-conditioning system. This transfer of heat is what the air-conditioning is trying to compensate for. Therefore, if a roof allows more heat to be transferred through the various mechanisms, the more work the air-conditioning unit must perform.

After acquiring the necessary raw data from the mini-roof sensors, we noticed that the temperatures cycled in a sinusoidal fashion on yearly and daily time frames, so using the general form sine function was an appropriate model to describe the data. We determined the frequency and amplitude for all fifteen mini-roof systems. The determined functions supported that most roofs are statistically different from one another from an amplitude aspect but the phase angles are statistically the same. We also have discovered that most roofs are significantly different from the standpoint of the average mean roof temperature, but the significance is most prevalent in the summer months and during the other times of the year the roofs behave in a more similar fashion to one another. The model does well in describing the average daily temperature during the year.