Mapping Post-Wildfire Neighborhood Recovery: Integrating Spatial Video with GIS for Data Collection and Analysis

Jacqueline W. Curtis, Andrew Curtis, Andrea B. Szell, Adam Cinderich, and Rachel Will
Department of Geography
GIS Health & Hazards Lab
Kent State University


Background


The recent increase in the number, size, and severity of wildfires has rightfully drawn attention to their causes in nature and society, with the aim of reducing negative impacts to life, property, and environment. Despite this attention, one question that has received too little consideration is, “What happens afterward?” Specifically, knowing that wildfires impact residential areas, how should we plan for post-wildfire recovery?


In the context of disasters, “recovery” can be a difficult word to meaningfully define. In emergency management, recovery is often described as the phase of returning to normalcy. However, this definition raises questions about the nature of  “normalcy.” If we are unclear on what normalcy is and how to measure it, how can we know when recovery is achieved?  In addition, these questions challenge us to consider whether it is a good idea to return to pre-disaster conditions, as they likely had a role in the disaster occurring in the first place (e.g., building homes in areas known to have a history of wildfire). Overall, recovery can be defined in many different ways, and these definitions are in part dependent on who is studying the process (e.g., sociologists, planners and geographers).


a

Figure 1. Images of a) cleared lots and b) rebuilding six months after the Waldo Canyon fire that occurred June 2012 in Colorado Springs, Colorado. Image Credit: Andrea Szell.



“Recovery” must be defined before  it can be systematically measured. The GIS Health & Hazards Lab in the Department of Geography at Kent State University focuses on understanding recovery using geospatial techniques. For our studies of post-wildfire neighborhood recovery in the United States, we use a definition that considers the act of returning home as an indicator of full recovery [1].


In comparison to other aspects of disasters (i.e., preparing for, mitigating, and responding to them), scant attention has been paid to recovery. This is particularly true in “developed” countries. There has been an underlying assumption that these places have the financial and political capacity to facilitate recovery and that, by extension, recovery is a linear and even process. However, in practice the process of recovery is complex and is the result of the interaction among many components. Whether or not people return (and therefore places recover) is dependent on a variety of factors that include federal policies, local issues such as the condition of physical infrastructure (e.g., roads, utilities and schools) and social infrastructure (e.g., faith-based communities, neighborhood organizations), as well as resident perceptions that the place to which they are returning is safe and healthy. The first two authors of this article have studied post-disaster recovery in a variety of settings (e.g., hurricanes, earthquakes, tornadoes and wildfires). From walking through the impacted areas and talking to residents, it is clear that neighborhood conditions are important to the decision to return and stay. For this reason, we focus on what we call “neighborhood recovery” (Figure 1).


A New Geospatial Approach For Mapping Recovery


c

Figure 2. The spatial video system. a) Preparing for data collection. b) Coding the spatial video (viewed in Contour Storyteller on the right screen) in ArcGIS 10.1 (on the left screen). These data were collected as part of a neighborhood recovery study after the Waldo Canyon Fire in Colorado Springs, Colorado. This work was supported by the University of Colorado Natural Hazards Center Quick Response Program. Image Credit: Adam Cinderich.



Despite its importance, until the mid-2000s, collecting data on recovery at such a fine spatial resolution as the neighborhood was challenging. The damage pattern of these events varies from contiguous areas that burn to seemingly random patterns in which 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. Until recently, no data collection strategy that is systematic, dynamic, and cost effective has been available. In order to address this concern, we employed a geospatial approach, “spatial video”, which linked video with coordinates acquired by a Global Positioning System (GPS) receiver. We developed and perfected this approach in post-hurricane, -tornado, and -wildfire neighborhoods across the United States [2-4]. For instance, we used spatial video to trace neighborhood recovery in post-Katrina New Orleans.


Specifically, this approach enables increased efficiency in field data collection as well as the ability to survey locations over multiple time periods. In addition, unlike existing survey methods, this approach generates archival data so that places can be revisited through video.


Though the authors have tested several spatial video systems, the current one uses an extreme sports camera, the Contour Plus, which has an internal GPS tagging the video stream. One camera is mounted on each back window of a car. Then, the impacted neighborhood is driven systematically one street at a time until video has been collected for all streets (Figure 2).


Following the field data collection process, data preparation and visualization were accomplished by employing a mixed approach of playing the videos through spatial video data visualization software (Contour Storyteller) and coding recovery scores in ESRI’s ArcGIS mapping software. Parcel level data were used to assess neighborhood recovery by assigning a recovery score (RS) between one and four to each property located in the impacted neighborhood. Recovery scores reflected the characteristics of a damaged structure (RS=1), a cleared lot (RS=2), an emerging structure (RS=3), or a completed structure (RS=4) (Figure 3).


d

Figure 3. Example of using the RS to map neighborhood recovery after the Waldo Canyon Fire (2012) at a) six months post-event and b) one year post-event. Note the transition of many properties from a cleared lot (yellow) to an emerging structure (light green). These maps demonstrate the dynamic spatial and temporal scales of post-wildfire recovery. Image Credit: Authors.



There are several benefits to using this system, including high image quality, small camera size, and relatively low cost. In addition, the software used to display both video and GPS path, Contour Storyteller, is free. The relative ease of download and display allows the image and GPS signal to be checked at the end of each collection day. The work done in the GIS Health & Hazards Lab focuses on maximizing the transferability of geospatial tools and approaches to a wide array of users. The potential for this type of spatial video to spread the burden of field work and disseminate video is appealing.


The Future of Using Spatial Video To Map Neighborhood Recovery


Collecting and mapping patterns of post-wildfire neighborhood recovery is just a first step. With advances in geospatial technology, this spatial video approach has only become widely available in the last five years. With this advance, the dynamic spatial and temporal scales of post-wildfire recovery are beginning to be identified. The frequency of wildfires events means that more impacted places should be studied to build a broader knowledge base to understand neighborhood recovery and why it matters. We are now working to answer this second question by integrating spatial video with residents’ narratives of their neighborhood. This approach is called a “geonarrative.” and it provides an opportunity to move from the limited perspective of researchers to a resident-informed perspective. Residents are able to ride along with us or view the spatial video data on a laptop and comment on what they see as barriers and catalysts to neighborhood recovery and how this process impacts their health and well-being. Local information is essential to creating a complete understanding of how effective post-wildfire neighborhood recovery should be designed and implemented


In the midst of catastrophic events, there is an opportunity to learn. We do not minimize the terrible toll wildfires can take on human life and property. Indeed, through using geospatial techniques, we hope to gain new understanding of this toll so as to improve the well-being of residents and the viability of places impacted by future wildfires.


References


1. Bolin, R. 1976. Family recovery from natural disaster: A preliminary model. Mass Emergencies 1: 267-277.
2. Burkett, B.C. and A. Curtis. 2011. Classifying wildfire risk at the building scale in the Wildland-Urban Interface: Applying spatial video approaches to Los Angeles County Risk, Hazards & Crisis in Public Policy 2:4 Article 6.
3. Curtis, A., and J.W. Mills. 2011. Spatial video data collection in a post-disaster landscape: The Tuscaloosa Tornado of April 27th 2011 Applied Geography 32 393-400.
4. Curtis, A., J.W. Mills, B. Kennedy, S. Fotheringham, and T. McCarthy. 2007. Understanding the Geography of Post-Traumatic Stress: An Academic Justification for Using a Spatial Video Acquisition System in the Response to Hurricane Katrina.  Journal of Contingencies and Crisis Management 15(4): 208-219.


Acknowledgements


Southern California Wildfires: This research was supported by the National Science Foundation: Small Grants for Exploratory Research under Grant 0807462, Andrew Curtis, Principal Investigator. Any opinions, findings and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation.


Waldo Canyon Wildfire: This research was supported by the University of Colorado Natural Hazards Center Quick Response Program, Jacqueline W. Curtis, Principal Investigation.


The authors would also like to acknowledge the support of the Kent State University College of Arts & Sciences Research Resources Program for providing funds for field equipment.


We also acknowledge the assistance of undergraduate students from the University of Southern California: Juliet Kahne, Gregory Elwood, Mia Costa, C.J. Windisch, Timothea Tway and Cameron Burkett.