Community Resilience to Climate Change: Theory, Research and Practice

156 health conditions, and bedridden patients, though such groups have not been included in this study. The covariance analysis found racial and age-related disparities in distribution of UHI, CAC, and walkable access to heat refuges. Risk factors concentrate on some socio-demographic groups, especially young children. They are more likely to live in census blocks which are hotter during urban heat events, and with a smaller number of CAC units. In contrast, white populations tend to live in census blocks with less UHI effect, and more CAC units. Black/African American populations tend to have better accessibility to public heat refuges, which may prove helpful if they are concentrated in high-heat census block groups. While analyses focusing on environmental justice have found that non-white communities are disproportionately living near point sources, urban heat and the access to refuge arguably represent novel concern that may further deepen the inequities in society. This study does not offer a complete exploration of the factors which determine why certain socio-demographic groups cluster in areas experiencing higher temperatures. This is a complex question that would require a complete study of its own, though some of the likely contributing factors are known to researchers. Urban development patterns often feature lower rents in areas near large roads and buildings [57], both of which can amplify urban heat effects. Assuming individuals with limited financial means seek out lower rent, this increases their likelihood of locating in areas with higher heat stress. A second possible factor relates to socialization and, in some cases, spatial isolation of minority communities. Such groups have a history of building social capital by co-locating in neighborhoods, as well as being coercively isolated in specific locations [58,59,60]; this could result in apparently heightened heat exposure for such groups, simply due to their proximity. In the case of Portland, local development practices have typically placed large trees and other heat-ameliorating features in higher-income neighborhoods [61], exacerbating heat exposure of low-income andminority communities who have historically been excluded from these areas. These are multifaceted relationships that differ across cities, and are outside the purview of this study. However, it is useful to consider the underlying causes of physical clustering and resulting exposure. One major drawback to the analyses in this study is the geographic format of the data. The irregular polygon geometry of the census block group data relies on areal aggregation to protect the anonymity of individuals; this aggregation of population and heat data into enumeration units can ‘smooth’ the dataset, eliminating extreme highs and lows in the process of representing the data with a single mean value. This complication is difficult to avoid, as the block group geometries employed in this study are the highest resolution datasets available with the required socio-demographic information. Additionally, the Modifiable Areal Unit Problem (MAUP) may introduce error when using enumeration units such as census block groups [62]. A potential alternative to this census-based study would be to create an entirely new survey of randomly sampled households in the region. This potential new study could allow for a building-level analysis similar the one performed here for affordable housing, but for all socio-demographic variables. A survey with a sample size high enough for statistically sound inference and analysis would be time consuming and costly, however it could potentially reveal more accurate or meaningful results. This study is also lacking in a key piece of information which would provide a more complete understanding of vulnerability; though exposure and adaptive capacity have been well explored, sensitivity has not, mainly because reliable data on health, genetics, and lifestyle choices are difficult to obtain. For this reason, the definition of vulnerable populations may not be fully accurate because we do not accurately know whether individuals do not, in fact, have access to other forms of refuge (e.g., ductless heat pump, swimming pool, alternative residences, etc.). At the same time, at the population level, the present study finds significant associations between high exposure and low adaptive capacity, which provide meaningful direction for decision makers to prioritize those areas and groups that are likely to be at high risk. Next Steps for Practitioners The results of this study may serve as a guide for practitioners in Portland, Oregon, directing attention to those areas of the city most at risk of extreme heat exposure. However, socio-demographic indicators can only reveal general characteristics of a population; as such, community engagement in these priority areas will be a key strategy moving forward. These results suggest that practitioners will need to meet with community members directly to better understand what they experience during a heat wave, how they adapt, and what they perceive their needs and strengths to be. Rather than offering strictly external monetary or technological support, sustainable solutions may be reached by working with local organizations and individuals to build internal capacity. Given the diverse nature of marginalized groups exposed to extreme heat, it will be helpful for the City of Portland and Multnomah County to release heat-related materials for such an audience. Information regarding public refuges and heat safety, as well as heat wave warnings should be issued in multiple languages and formats (print, online). Messaging tailored to specific groups may also be helpful. It is further recommended that government agencies work with community organizations to disseminate information and provide refuge, as marginalized populations may be wary of government programs. Although this particular study pertains to Portland, the development of inclusive materials and interventions is a best practice for all cities. 5. CONCLUSIONS This study provides compelling evidence that extreme heat exposure is an environmental justice issue. Exposure and adaptive