Environmental Refuges during Summertime Heat and Elevated Ozone Levels: A Preliminary Case Study of an Urban “Cool Zone” Building

Abstract

The combination of extreme heat waves and ozone pollution is a major health hazard for urban populations in the summertime, particularly for the most sensitive groups such as children, the elderly, the unsheltered, and those with pre-existing health conditions. The “Cool Zone Program”, operated by the Salt Lake County Aging and Adult Services, identifies areas in the county and Salt Lake City facilities where members of the public can escape the summer heat, hydrate, and learn about available programs. We measured indoor and outdoor temperature and ozone for a pilot study at a designated Cool Zone location during the 22 August–6 September 2019 period and found that the building provided substantial heat relief and protection from more than 75% of the outdoor ozone. We observed a nearly 35 min delay for the outdoor ozone to be reflected on the indoor readings, providing an action window for ventilation scheduling changes to protect against the highest ozone levels during the day. Our findings show that it is critical to re-think and formulate action plans to protect vulnerable populations from excessive heat and pollution events during the summer.

1. Introduction

The combination of extreme heat waves and ozone pollution episodes is a major health hazard for urban populations in the summertime, particularly for the most sensitive groups such as children, the elderly, the unsheltered, and those with pre-existing health conditions [1,2,3,4]. In addition, recent studies found that higher vulnerability to urban heat islands is noted among minority populations [5,6]. It is critical that urban residents, particularly those who are most at risk from extreme heat waves and urban pollution, be able to mitigate these potential health risks [7].

Ozone exposure in the urban environment is complex and varies as widely as a function of time of day and location [8,9]. Exposure is often higher in suburban locations away from a city center than in highly trafficked regions where ozone destruction by nitrogen oxide titration is associated with vehicle pollutant emissions [10]. Malashock et al. [9] found that 37% of the deaths globally from ozone pollution were associated with urban exposure, while 56% were in suburban regions less than 1 h commute from more densely populated urban areas, despite fewer people living in the suburban regions. The amount of time spent outdoors (and how rigorous the physical activity that occurs when a person is outdoors is) and the levels of ozone observed during the period of outdoor exposure are key factors in determining the resultant health impacts [11,12].

Indoor locations with air conditioning will be much cooler, and indoor locations typically have much lower ozone than outdoor locations, as ozone chemically reacts with surfaces and air indoors, lowering the ambient concentrations [13]. As reviewed by Nazaroff and Weschler [14], “the ozone infiltration factor”, which is a “building’s time-averaged indoor ozone concentration in the absence of indoor sources normalized by the time-averaged outdoor ozone concentration”, can be determined for each building but will be a function of various factors related to the “physical and chemical environment of a building” [14]. This review of 2000 buildings shows a “general indoor to outdoor concentration ratio of about 25%”, meaning that indoor ozone is generally around one-fourth of the ozone concentrations outside [14]. While fewer health studies were conducted to show the impacts of indoor ozone exposure, even the generally lower concentrations observed indoors were still associated with negative health outcomes [15,16].

Extreme heat waves in urban areas are an increasing public health threat in the era of anthropogenic climate change [4], with an observed doubling of exposure of urban populations globally to heat waves between 1983 and 2016 [17]. Consequently, the societal need for “cooling centers” to prevent heat illnesses has become increasingly pronounced and an active area of research as urban residents, particularly the homeless and unemployed, become increasingly negatively impacted by seasonal heat waves [18,19,20]. A number of pilot studies were conducted in recent years investigating this emerging topic area—heat exposure and heat stress for various populations (e.g., health [21], agricultural [22], and construction [23]). In this manuscript, we provide a pilot study supporting the development of refugees for residents from the combined effects of excessive heat and pollution events during the summertime in Salt Lake City, Utah.

The “Cool Zone Program” is operated by the Salt Lake County Aging and Adult Services in partnership with The County Library, Salt Lake County Parks & Recreation, and the Salt Lake City Library in Utah, USA [24]. “Cool Zones are areas in county and Salt Lake City facilities where members of the public can escape the summer heat, hydrate, and learn about available programs” (Figure 1) [24]. To our knowledge, no study examined the combination of both temperature and ozone inside and outside “Cool Zone” buildings during summertime heat waves. In this case, the compounding impacts of both the high temperatures and ozone pollution are of particular concern. Additionally, the very high ozone concentrations outdoors [25] observed during summertime heat waves create a sort of “worst case scenario” when combined with additional pollution sources such as wildfires [26]. This combination of environmental exposures highlights the need to document the decrease in ozone exposure that can be realized by obtaining vulnerable populations, such as individuals experiencing homelessness, out of the ambient air and into cooler buildings with much lower ozone. In this study, we document the indoor and outdoor temperature and ozone concentrations during summertime heat waves in Salt Lake City, UT, USA, during the summer of 2019.

The Salt Lake County Air Pollution Bureau [27] started to examine the relationship between indoor and outdoor air pollution levels because recent studies suggest exposure to elevated pollutant levels increases the risk of obtaining or aggravating various health conditions [28,29,30]. The results from this study will help determine whether existing “Cool Zones” may also be designated as “Breathe Clean Zones” as well. The Cool Zones are distributed relatively uniformly throughout Salt Lake County, with each city having at least one Cool Zone.

As Salt Lake City and Utah suffer from an ongoing drought, air pollution is a substantial concern [31]. There were two main goals of this study:

(1)

Provide information on temperature and ozone variability within and around county senior centers and libraries.

(2)

Establish any trends (indoor vs. outdoor) from the data at each facility and provide useful information to the public about any health-related risks resulting from elevated ambient temperature and ozone concentrations.

2. Materials and Methods

2.1. Study Location and Period

The Millcreek Library center, located at 2266 E Evergreen Ave, Salt Lake City, UT 84109, USA (Figure 2), was chosen for this study because it serves a population of both young and elderly residents who typically are more vulnerable to poor air quality and high-temperature conditions. The library is ideally situated in the heart of a residential area and was designated as a “Cool Zone” where residents can escape heat waves or high temperatures. The nearest large emitters are Interstate Highways 80 and 215, which are more than 2 km and 2.5 km away, respectively. The area surrounding the Millcreek Library center has many green spaces—there are several parks in the vicinity, as well as the riparian vegetation surrounding Mill Creek. There are no point sources or other large pollution emitters near the study location.

This pilot study took place between 22 August and 6 September 2019. This time period was characterized by a weak high-pressure ridge, which led to mostly light winds and elevated temperatures, resulting in conditions conducive to increased ozone concentrations in the afternoon (Figure 3). Monday, 2 September 2019, was a USA National Holiday (Labor Day), which colloquially denotes the end of summer and marks the last period of hot weather. While the Labor Day holiday weekend is typically associated with increased travel, this would not affect the study area as it is relatively far from large highways.

2.2. Temperature and Ozone Measurements

The sensors used in this study are 2B Technologies Model 205 Ozone Monitor (Resolution 0.1 ppbv; Precision (1σ; rms noise): Greater of 1.0 ppbv or 2% of reading for 10 s average; Accuracy: Greater of 1.0 ppbv or 2% of reading) [34]. The sensors were placed inside a case with a fan to increase ventilation and reduce the temperature, as in previous studies [35]. The data were logged at 5 min intervals using a Raspberry Pi Model 3 [36]. A sensor unit was placed inside the library, and another one was located on the roof.

Ambient temperature was retrieved for the Salt Lake City airport (KSLC) using the MesoWest platform [37]. The Salt Lake City airport temperature sensors are part of the automated surface observing system (ASOS) [38]. The ASOS sites are located at primary airports and are maintained by the National Weather Service. This network is rigorously checked and quality controlled, and the sensors have radiation shields, resulting in accurate temperature readings [39].

2.3. Data Analysis

The full study temperature and ozone concentration data were aggregated diurnally to study their cycles. The outdoor unit’s temperature readings were compared with the airport readings to quantify additional warming effects of the enclosure. Outdoor ozone readings were compared with ambient temperature data to estimate the relationship between the two variables. Lastly, the indoor ozone was compared with the outdoor ozone concentrations to estimate the building infiltration.

Temperature and ozone were directly compared and were also lagged to understand delayed effects. Through the lag method, a reading (e.g., outdoor ozone) was compared against another reading (e.g., indoor ozone) delayed by a certain number of minutes. The coefficient of determination (r²) was calculated for each delay value to identify the most likely leg between the two variables. Lag analysis performed in a previous study [40] showed that fine particulate matter (PM_2.5) measured by sensors outside a school was reflected by indoor measurements following a delay of between 50 and 70 min. In this study, we used lags of 0–60 min for all comparison metrics.

3. Results

3.1. Time Series

The full study temperature and ozone concentration time series are shown in Figure 4a,b, respectively, for indoor and outdoor (“Roof”) values. Figure 4a also includes temperature readings at the Salt Lake City airport (“KSLC”), which is the nearest National Weather Service (NWS) station located approximately 15 km (9 miles) northwest of the Millcreek Library. Figure 4a shows that both the roof and KSLC temperatures follow an expected summertime pattern with low-amplitude high pressure and seasonably warm temperatures over the Western USA during this time, as was shown in Figure 3, while the indoor temperature is controlled by the heating, ventilation, and air conditioning (HVAC) system. It should be noted that the “Roof” units were inside a case for protection from the outdoor elements and to keep all the components (e.g., power supply, data loggers) in one place. Although the protective cases were ventilated using a fan, the temperature inside the case was approximately 5–10 °C warmer than the ambient air. This agrees well with the indoor unit as the library setpoint was 22.2 °C, and the average indoor temperature was 30.8 °C. The roof temperature values are, on average, 10.6 °C warmer than the KSLC readings, which is also within the expected range of temperature differential. Figure 4b shows that although the temporal structure is similar, the indoor ozone values are substantially lower than the outside values.

There are two brief periods of data discontinuity during this study. The first is a 15 min period at approximately 15 h on 28 August at the Roof unit. This was caused by a short power outage. The second lasted approximately an hour at 10 h on 29 August, also affecting the Roof unit. This was due to a power cycle of the instrumentation and Raspberry Pi. Neither of these instances resulted in a significant data loss as, when combined, they accounted for less than 0.5% of the total study period. When the instruments were back online and recording, there was no loss in fidelity or changes to the observed patterns.

3.2. Diurnal Cycles

The diurnal cycles of temperature and ozone are shown in Figure 5. The indoor temperature does not vary throughout the day (Figure 5a), while the Roof and KSLC temperatures exhibit an expected diurnal cycle, with the hottest temperatures recorded between 3 p.m. and 5 p.m. The Roof value is generally delayed compared to the KSLC value by approximately an hour. The hottest roof temperature values are observed between 5 and 6 p.m. (also later by a few hours than the hottest air temperatures), presumably because of the solar heat being absorbed and radiated from the hot rooftop in concert with still-warm air temperatures in the late afternoon (Figure 5a).

The ozone readings (Figure 5b) show the highest concentrations in the middle and late afternoon hours, as would be expected from the diurnal cycle of urban ozone chemistry [41]. A rapid increase in rooftop ozone is observed from late morning through mid-afternoon, with a rapid decrease after 6 p.m. There is also an expected minima in roof ozone during the morning rush hour due to the destruction of ozone by nitrogen oxide titration associated with the increase in vehicle emissions during the morning commute. The amplitude of the diurnal cycles in outdoor ozone is approximately 400% of the amplitude of the diurnal cycle in indoor ozone.

3.3. Roof vs. KSLC Temperature and Lag

The Roof and KSLC temperatures were compared for the duration of the study period and are shown in Figure 6a. The 10.6 °C warmer readings of the Roof sensor are consistently demonstrated. The r² values obtained by lagging the Roof readings to the KSLC data are shown in Figure 6b. The maximum r² was found to be 0.930 at 50 min. This delay and warmer readings are likely due to the Roof case heating the temperature sensor within the 2B ozone instrument.

3.4. Outdoor Ozone vs. Temperature and Lag

The Roof ozone and KSLC temperature were compared for the duration of the study period and are shown in Figure 7a. The KSLC temperature was used instead of the Roof temperature because of the 50 min lag in instrument temperature compared to the ambient (KSLC) temperature (Section 3.3). The r² values obtained by lagging the Roof ozone readings to the KSLC temperature data are shown in Figure 7b. The maximum r² was found to be 0.372 at −5 min, meaning that the temperature readings lagged the ozone values.

3.5. Outdoor Ozone vs. Temperature and Lag

The Roof and indoor ozone levels were compared for the duration of the study period and are shown in Figure 8a. It is readily apparent that the indoor readings are only about 22% of the Roof concentrations. The r² values obtained by lagging the indoor ozone readings to the Roof data are shown in Figure 8b. The maximum r² was found to be 0.821 at 35 min. This pattern was also visible in Figure 4b and can be effectively understood to be the infiltration delay of outdoor ozone to the inside of the building.

4. Discussion

4.1. Measurement Considerations

The Roof ozone lag against the KSLC temperature demonstrates a critical component of ozone formation. Ozone is the most important photooxidant, and the concentrations of ozone pollution are well-known to be controlled by photochemical reaction rates driven by high air temperatures and high incoming solar radiation [42,43]. A modeling study conducted by Narumi, Kondo, and Shimoda [3] found “that a 1 °C increase in temperature leads to a maximal photochemical oxidant concentration of 5.3 ppb, which is an increase of 11%”. Both elevated ambient temperature and ozone concentration are the results of increased solar radiation. The fact that ozone concentrations rose approximately 5 min before the temperature readings is within the possible margin of error as this was the temporal resolution of both the temperature and ozone readings. Additionally, since the r² values did not vary substantially between lag-0 and lag-15, it reinforces the close relationship between these two variables.

The temperature lag between the Roof and KSLC sensors shows a potential concern of operating encased equipment. Although a fan was used to try to mitigate overheating concerns, as in previous studies [35], both the indoor and outdoor units recorded elevated temperatures. The 50 min lag is an approximate benchmark as the KSLC station is about 15 km away and may not reflect the specific conditions of the study site.

The indoor and Roof ozone comparison highlighted two important aspects of the protectiveness of indoor environments against ozone. Firstly, the indoor ozone concentration was approximately 22% of the Roof readings. This showed that being inside the Millcreek Library protected inhabitants from nearly 80% of the outdoor ozone. Secondly, the 35 min lag provides an infiltration metric. Since ozone concentrations follow a regular diurnal pattern, this delay can be used to reduce outdoor air intake during the highest ozone periods to mitigate some of the indoor ozone. As the indoor ozone time series reflects the outdoor ozone, albeit with the 35 min delay, it is clear that there are negligible (if any) indoor ozone sources in this building.

4.2. Limitations and Future Directions

The main limitation of this study was that despite the fan cooling the case containing the ozone instruments, the enclosure was substantially warmer than the ambient setting. The Roof unit was affected more than the indoor unit and provided a false ambient temperature reading as it was not only warmer but also delayed in comparison to a stationary weather station. Therefore, care must be taken to measure the true ambient temperature by either relying on an external temperature gauge or installing a sensor outside the case. However, the temperatures did not increase to a point where they overheated the 2B ozone sensor and impacted its performance as they were well within operating range [34].

This pilot study took place during two weeks in late August and early September 2019. Although this is still a warm period, the warmest parts of the summer (July to early August) were not measured. Despite this, we were able to find substantial differences between the indoor and outdoor temperatures and ozone concentrations. Additionally, the instruments were set to read every five minutes, which could lead to less well-resolved lag estimates. Follow-up work could focus on a longer study period, ideally a full summer, to capture a larger data set. Additionally, the ozone instrument could be set to a higher resolution (e.g., 1 min) so that even finer scale patterns may be identified, such as those associated with cloud coverage.

Future research could study additional buildings that are part of the Cool Zone program (e.g., senior centers and community recreation centers) to understand whether the results found in this study are replicable in other types of buildings. It is recommended that the study be carried out over a longer time period to capture the impact of the worst heat waves in Salt Lake County and also during the highest summertime ozone episodes. Furthermore, a longer study that includes PM_2.5 [40] and black carbon sensors could measure wildfire smoke and dust events, which are often associated with both high ozone and temperature events. This study highlights the need for combined “clean air” and “cool air” centers in tandem in cities globally where the high pollution episodes overlap with high summertime temperatures. The concept of “Wildfire Smoke Clean Air Centers” development in the state of California was recently discussed by Treves et al. [44], and the United States Environmental Protection Agency (USEPA) suggested using schools and both cooling and clean air centers [45].

As discussed by Kim et al. [46], additional research is needed to quantify the disparities noted in “cooling center preparedness” across the USA. Unfortunately, those cities with the greatest heat indices in the USA were currently found to be some of the most unprepared in terms of cooling centers. Additionally, in recent years, ozone pollution has been worsening in many parts of the world [47,48,49,50], while even in countries such as the USA, where overall ozone trends have been decreasing, the co-occurrence of high ozone along with high particulate pollution associated with wildfires is a major concern [51]. Developing targeted approaches to protect the population from the combination of extreme heat and air pollution concerns will only become more apparent in a warming climate subject to greater weather extremes and increased frequency of wildfires and associated air pollution impacts.

4.3. Policy Implications

The substantial temperature differences between the KSLC and indoor units show that the Cool Zone is protecting individuals, particularly during the early to mid-afternoon hours when the temperature difference is over 10 °C (Figure 5a). Furthermore, the indoor ozone was less than one-quarter of the Roof values, making the Millcreek Library an effective air pollution refuge as well (Figure 5b). While this study focused on only one site, it is reasonable to assume that similarly designed and operated Cool Zone buildings would offer a comparable level of protection. Although there are currently 59 operating Cool Zones in Salt Lake County (Figure 1), many have limited weekend hours, closing around 2–3 p.m., and may not be open on Sundays. If the centers close at the hottest time of the day, they are not able to provide heat relief as intended. While the Utah Department of Health and Human Services has enacted “Code Blue” alerts during cold winter days [52], there is currently no counterpart for excess heat days. This is a particularly grave concern for unsheltered individuals and community members who may not have access to central cooling systems. Therefore, it is critical to re-think and formulate action plans to protect vulnerable populations from excessive heat and pollution events during the summer.

5. Conclusions

We studied the potential environmental exposure benefits of a Cool Zone building during an elevated heat event in Salt Lake City, UT, USA. The nearly 75% reduction in ozone exposure and over 10 °C difference between indoor and outdoor temperatures showed that the study site was effective at protecting patrons. The 35 min ozone infiltration timeframe could provide an action window to reduce ventilation and protect the indoor population against the peak ozone concentrations. Although this is a pilot project, the preliminary findings are supportive of the interest in Cool Zones as a protective environment against elevated heat and air pollution and may play a substantial role in environmental justice.