Robert Atchison, Kansas Forest Service, firstname.lastname@example.org (Presenter)
Kabita Ghimire, Department of Geography, Kansas State University, email@example.com
The Dust Bowl of 1930s caused the U.S. federal government to invest $13.8 million to plant more than 200,000 million trees and shrubs in the form of shelterbelts on 30,000 farms throughout the Great Plains. Many shelterbelts now exceed 70 years of age and are in need of renovation to continue conserving soil, increasing crop yields, protecting livestock and reducing energy consumption around homes. Data on condition and location of shelterbelts is lacking and is needed to target renovation initiatives. This study developed remote sensing and GIS methods to identify location and condition of shelterbelts in a seven county area in Kansas. Object-based classification used color aerial photography acquired through 2008 National Agricultural Imagery Program to classify shelterbelts. Object-based classification techniques segmented imagery into objects with similar spectral, textural and geometric properties for image classification. Condition classes were assigned using brightness value, texture analysis and normalized difference vegetative index. Finally, 12% of the shelterbelts were â€œground-truthedâ€ using the same NRCS condition criteria that determines shelterbelt eligibility for financial assistance through the Environmental Quality Incentives Program (EQIP). Locations of 1,116 shelterbelts, totaling nearly 1,051 hectares, were identified in the study area with overall accuracy of 93 percent. Twenty-seven percent of the shelterbelts were classified in good condition, 33% fair and 40% poor. Eighty-one percent of the time shelterbelts identified agreed with confirmed shelterbelt locations generated via a heads-up digitizing process. A digitized landownership layer was developed to identify landowners with shelterbelts in fair to poor condition and to invite their participation in EQIP.