Assessing Potential Urban Tree Planting Sites in the Piedmont of the United States: A Comparison of Methods

There is a growing interest of late to enhance our ability to reduce carbon dioxide levels in the atmosphere through tree growth. Although the spread of development in urban areas may, in some cases, cause a reduction in forest cover, there may be opportunities to establish additional trees in open areas within cities and metropolitan areas. We assess three freely available aerial or satellite imagery products for their ability to correctly identify open areas in six southern United States cities. A standard supervised classification process and a statistical assessment of plantable open areas are employed. Results suggest that while the United States National Agriculture Imagery Program (NAIP) aerial imagery may require more computer processing time and more computer memory, classification accuracy is acceptable for the purpose of identifying open areas where trees might be planted. Therefore, the use of NAIP for this purpose is as sufficient as Landsat 5 and 7. Given recent uncertainties in the availability of Landsat 5 and 7 imagery, and given analytical needs such as the one proposed here, improvements in estimates of urban carbon potential can be made over less intensive methods.