Spatial Patterns of Post-Wildfire Neighborhood Recovery: A Case Study from the Waldo Canyon Fire

Until recently, the recovery phase of the emergency management cycle has received relatively little attention from the natural hazards research community in comparison to the other phases of planning, mitigation, and response. However, in the prolonged aftermath of Hurricane Katrina, studies on recovery have become more common as evidence from this disaster suggests that the process is spatially uneven and temporally dynamic, and operates at finer scales than previously assumed. The heterogeneous patterns of recovery and its implications for the well-being of people and places are especially visible at the neighborhood scale. With growing empirical evidence from post-disaster environments such as New Orleans and Joplin, Missouri, studies on neighborhood recovery are becoming a useful endeavor through which to inform emergency management and city planning policies related to what happens after a disaster, why these outcomes matter, and how to systematically plan for post-disaster recovery. Despite progress made on understanding neighborhood recovery, these studies have focused primarily on posthurricane and post-tornado environments. In order to achieve a comprehensive understanding of neighborhood post-disaster recovery, other events (e.g. wildfires) must be included. However, wildfires are also notably underrepresented in natural hazards research. This Quick Response project aims to begin to address both the understudied process (neighborhood recovery) and the understudied event (wildfire). It draws attention to the need for post-wildfire neighborhood recovery studies, particularly in order to understand the implications for health outcomes of impacted residents. BACKGROUND One of the least analyzed aspects of a disaster is recovery (FEMA and APA 1998; Mileti 1999), and specifically long-term recovery and the health burdens that develop in this process (Curtis, Mills, and Leitner 2007). In particular, the spatial aspects of recovery are dynamic and poorly understood (Mills 2008). Though a number of quantitative variables have been used as proxies for recovery, such as the opening of services, (e.g. childcare centers, hospitals, grocery stores), such metrics only capture city-wide recovery of services and infrastructure. The scale is too coarse to characterize comprehensive recovery at the neighborhood level, which is believed to be a building block of city and regional recovery (Campanella 2006). For this reason, post-disaster neighborhood recovery research is needed to create an evidence base with empirical data upon which recovery planning may proceed. Collecting empirical data on damage and then the subsequent status of recovery at a fine spatial resolution is especially important in the case of wildfires. The damage pattern of these events in the wildland-urban interface (WUI) varies from contiguous areas that burn, to a seemingly random pattern where one house is destroyed amid others that appear untouched. The fine spatial scale of this damage pattern argues for an equally fine spatial scale of data collection and analysis of damage assessment and recovery. However, wildfires have received relatively little attention from natural hazards researchers. McCaffrey (2004) suggests that this oversight may be due to the timing of natural hazards research that focused on human-environment interaction occurring when wildfire suppression was the dominant management strategy; it was effective enough to render wildfires not considered as a threat to humans (509). However, as residential development has encroached on environments prone to fire (Hammer, Stewart, and Radeloff 2009), and as complete wildfire suppression is not possible in every case, all aspects of wildfires should receive greater attention. This Quick Response project proposes a new framework for studying exposure to wildfires, and in particular to the post-disaster environment which they create, as well as new methods for collecting and analyzing empirical data on damage and recovery at fine spatial and temporal scales. Theoretical and Methodological Framework. Studies on the health impacts of exposure to wildfires have primarily focused on outcomes linked to particulate matter (PM) produced by these events (Emmanuel 2000; Mott et al. 2002; Künzli et al. 2006; Vedal and Dutton 2006; Naeher et al. 2007; Delfino et al. 2009; Hänninen et al. 2009; Sastry 2009; Wegesser, Pinkerton, and Last 2009; Morgan et al. 2010; Henderson et al. 2011). To a lesser degree, the psychological effects of exposure to wildfire have also been investigated, but with a focus on exposure to the fire event, not its aftermath (McDermott et al. 2005; Jones, Ribbe, and Cunningham 2006). However, drawing on research in environmental justice and environmental health, it may be appropriate to conceptualize the term “exposure” in a broader context. For example, exposure can mean more than a physical encounter with an agent (e.g., smoke inhalation) or with an event (e.g., seeing flames). The exposure-disease framework proposes that exposure to certain toxic agents manifest physiologically in negative health outcomes. Gee and Payne-Sturges (2004) extend this framework to include issues of stress and race. They propose that the relationship between exposure and disease can also be modified by stress whereby stressors act to reduce the body’s capacity to maintain itself thereby placing it at risk for negative health outcomes related to exposure. Typically, this work is focused on health in derelict urban environments and the agents are psychosocial stressors (e.g., signs of physical and social disorder in one’s neighborhood). Observations and anecdotal information from conversations with residents undergoing recovery from natural disasters suggest that it is appropriate to extend this framework beyond urban dereliction in general to the specific case of post-disaster recovery (Curtis, Mills, and Leitner 2007). However, in order to test the applicability of this theoretical framework, new datasets are needed to capture the a) psychosocial stressors and b) exposure to the psychosocial stressors in a postdisaster environment. This Quick Response project employs an emerging geospatial technology, spatial video, to capture neighborhood recovery from the Waldo Canyon Fire in Colorado Springs, Colorado. With this technology, neighborhood characteristics can be captured and digitized in their real-world location in a map. Then, at different time intervals, the neighborhood survey can be repeated and the neighborhood characteristics mapped again. The result is the ability to analyze the spatio-temporal patterns of variables hypothesized to be psychosocial stressors. Specifically, in the case of this post-wildfire environment, three variables are believed to be potential psychosocial stressors to residents who return to the damaged neighborhood: 1) severely damaged parcels, 2) parcels that demonstrate stagnation in recovery (e.g., persistent vacancy, persistent blighted (damaged) lots), and 3) burned land (not residential). All of the variables are visible to returning residents and all are reminders of the wildfire. In order to calculate residents’ exposure to these characteristics, a novel application of viewshed analysis is applied. A viewshed is a geographic scale that captures the entire visible area from an observation point at a specific location; this is a common approach used in studies of visual impact. Despite their potential for meaningfully capturing visual exposure to environmental characteristics, they have not been employed in studies examining neighborhood recovery and its health implications. Therefore, this project is a first step to employing a new theoretical frame, as well as novel data collection and analysis approaches to understanding how post-disaster environments impact health outcomes. Research Questions. The specific objectives of this Quick Response project are twofold. First, the spatial video approach will be used to collect neighborhood recovery data in the study area at two intervals (6 months and 1 year post-event). Second, these data will be used to map the spatial pattern of recovery using the Recovery Score (RS) method developed and tested in other postdisaster environments including New Orleans, San Diego, Joplin, and Tuscaloosa (Curtis et al. 2010). Addressing these objectives is central to establishing the neighborhood environmental conditions, and the characteristics which can be identified as psychosocial stressors. With these data, the RS and other visible environmental variables can eventually be integrated with health self-reports to measure health outcomes. As recovery is a spatially and temporally dynamic process, fieldwork on this project is occurring at two time intervals: six months post-event and one year post-event. Consequently, two Quick response Reports will result from this work. This report, Report #1, will detail the methods used for data collection using spatial video and the Recovery Score (RS) in a Geographic Information System (GIS). Substantively, it will focus on mapping the spatial patterns of neighborhood recovery from 6 months post-wildfire by using the RS for all damaged parcels in the study and then creating viewsheds of all neighborhood parcels for calculating exposure of each parcel to severely damaged parcels and burned land. These results will be used to identify areas where residents are theoretically at risk for negative health outcomes due to exposure to the aftermath of the wildfire. Furthermore, the investigators will compare RS patterns at the 6 month postevent mark between the area impacted by the Waldo Canyon Fire (Colorado Spring, CO 2012) with patterns at the same time interval in the study areas impacted by the Witch Fire (San Diego County, CA 2007). Report #2 will serve as the comprehensive documentation of both data collection trips, the comparison of neighborhood recovery patterns at both time intervals between the San Diego study area and the Colorado Springs study area, and will build on the substantive and methodological issues presented in Report #1. Specifically, it will focus on a) comparing the spatial patterns of neighborhood recovery from 6 months post-wildfire and then from 1 year post-wildfire using the RS approach to spatial video coding in GIS, b) identifying parcels that demonstrate persistent signs of severe damage, or of stagnation, c) creating viewsheds of all neighborhood parcels for calculating exposure of each parcel to damage, stagnation ,and burned land, and d) using the results from viewshed analysis within the exposure-disease framework to hypothesize about health outcomes. A follow-up study will then collect self-reported health data from participants in order to look understand how the post-disaster neighborhood environment impacts health. By using Report #1 to establish a baseline, Report #2 will show the dynamic spatiotemporal characteristics of post-wildfire recovery, even in micro-environments (e.g., the changes that occur within a neighborhood). Based on the similar levels of damage and sociodemographic variables, the investigators expect that patterns of recovery from this study will be similar to patterns of recovery from the Witch Fire at both time intervals. However, due to the compounding factor of initial stages of the national economic crisis occurring during recovery from the Witch Fire, less stagnation and decline may be evident after the Waldo Canyon Fire. Overall, the empirical documentation provided in Reports #1 and #2 will argue for further study of post-wildfire neighborhood recovery at final spatial scales, extended temporal scales, and with a justification of why health implications of wildfires cannot be theorized as being linked only to exposure to the immediate event. In addition to the intellectual contribution of this project, Report #2 will serve as an applied outline for policy-makers, planners, and emergency management professionals of how such an approach can be implemented.