PA31C-2169
Detecting neighborhood vacancy level in Detroit city using remote sensing

Wednesday, 16 December 2015
Poster Hall (Moscone South)
Xiaomeng Li1, Ran Wang2 and Anni Yang1, (1)Michigan State University, Geography, East Lansing, MI, United States, (2)University of Alberta, Edmonton, AB, Canada
Abstract:
With the decline of manufacturing industries, many Rust Belt cities, which enjoyed prosperity in the past, are now suffering from financial stress, population decrease and urban poverty. As a consequence, urban neighborhoods deteriorate. Houses are abandoned and left to decay. Neighborhood vacancy brings on many problems. Governments and agencies try to survey the vacancy level by going through neighborhoods and record the condition of each structure, or by buying information of active mailing addresses to get approximate neighborhood vacancy rate. But these methods are expensive and time consuming. Remote sensing provides a quick and comparatively cost-efficient way to access spatial information on social and demographical attributes of urban area. In our study, we use remote sensing to detect a major aspect of neighborhood deterioration, the vacancy levels of neighborhoods in Detroit city.

We compared different neighborhoods using Landsat 8 images in 2013. We calculated NDVI that indicates the greenness of neighborhoods with the image in July 2013. Then we used thermal infrared information from image in February to detect human activities. In winter, abandoned houses will not consume so much energy and therefore neighborhoods with more abandoned houses will have smaller urban heat island effect. Controlling for the differences in terms of the greenness obtained from summer time image, we used thermal infrared from winter image to determine the temperatures of urban surface. We find that hotter areas are better maintained and have lower house vacancy rates. We also compared the changes over time for neighborhoods using Landsat 7 images from 2003 to 2013. The results show that deteriorated neighborhoods have increased NDVI in summer and get colder in winter due to abandonment of houses.

Our results show the potential application of remote sensing as an easily accessed and efficient way to obtain data about social conditions in cities. We used the neighborhood vacancy survey data for Detroit data (2013-2014) to validate the results of vacancy levels of local neighborhood.