processes,1 rather than simply the physical storm itself, made the Katrina disaster what it was.
Second, the storm’s impacts were not evenly distributed, but instead exhibited important spatial variations and patterns. Although many of the damaged New Orleans neighborhoods near Lake Pontchartrain have lost 25-50 percent or more of their pre-Katrina populations (Figure 2), other neighborhoods in the same area have registered much lower population losses, while some neighborhoods have even experienced population growth since the storm. Explaining such spatial variations and patterns requires understanding how past decisions and practices, including possible socioeconomic or racial biases, affected different places (e.g., Did policy differences between neighborhoods produce different impacts within the city of New Orleans? Why have some neighborhoods gained population since the storm?). A better understanding of such spatial relationships could help residents and decision makers anticipate and mitigate the human toll from the next storm.
Third, the processes that led to the Katrina disaster operated at multiple and interlocking geographical scales, meaning that processes that led to the outcomes were operating at different scales, with each process possibly affecting the others. Mapping Katrina’s impacts at different geographical extents and resolutions highlights the potential for added insights when the same outcome is viewed at different scales. A Gulf Coast–scale map of county-level population change from 2005 (before the storm) to 2007 shows broad but nuanced patterns of post-Katrina population movement (Figure 3). In this map the localized pockets of modest population loss, and population gain, manifest in Figure 2 are not evident. By contrast, the regional view may provide insights that the finer-scale picture cannot provide, such as the areas where some of the displaced Gulf Coast residents may have moved—namely, areas north and west of the damaged areas. A geographical sci-