Project Funding: 2008 - 2012

NRA: 2007 NASA: New Investigator Program   

Funded by NASA

Abstract:
Human-constructed impervious surface area (ISA) is an important indicator on human alterations to natural environments and ecosystems. ISA reduces or eliminates the capacity of the underlying soil to absorb water, and has a significant impact on the environment and human health, therefore playing an important role in land use decisions. With this project we approach ISA spatiotemporal monitoring through two interrelated tasks, namely ISA detection and spatiotemporal modeling. Our ISA detection method will be comprised of multiple processes, each designed to perform a targeted classification (e.g. distinguish dark soil from dark ISA). All processes will be tied together through a hierarchical tree structure. This approach facilitates high accuracy with a model that is easy to implement, interpret and evaluate. Our spatially explicit uncertainty metrics will allow a comprehensive yet intuitive accuracy assessment by the non-expert, therefore enhancing earth system science beyond the boundaries of remote sensing. The second part of our analysis will improve the understanding of the spatiotemporal relationships underlying ISA change. Our spatiotemporal modeling will support knowledge discovery of bidirectional ISA dependencies to transportation networks, land use change trends, census and economic data. By better understanding these interdependent relationships we will: i) improve our ability to forecast ISA impact on ecosystem functions, and ii) quantify the influence of ISA change characteristics to land use and socioeconomic factors. Through our education and outreach plan we will foster a new generation of scientists in related disciplines. The public will become more aware of the value of remote sensing and its key application to important societal problems for the twenty first century.