Location: Southwest Watershed ResearchTitle: Aerial photo classification for monitoring the spread of Eragrostis lehmanniana in a semiarid Arizona grassland) Author
Submitted to: Research Insights in Semiarid Ecosystems Symposium
Publication Type: Abstract Only
Publication Acceptance Date: 9/30/2010
Publication Date: 10/2/2010
Citation: Sugg, Z., Moran, M.S., Scott, C., Van Leeuwen, W., Hamerlynck, E.P. 2010. Aerial photo classification for monitoring the spread of Eragrostis lehmanniana in a semiarid Arizona grassland. Research Insights in Semiarid Ecosystems Symposium. University of Arizona, Tucson, AZ., October 2, 2010. Interpretive Summary:
Technical Abstract: Invasion by exotic grasses is a common threat to biodiversity in grassland ecosystems. Phenological variation can be a key consideration in efforts to characterize and monitor the spread of invasive grasses. This study investigated a transition from a vegetation assemblage historically dominated by diverse native grasses to a near-monoculture of non-native Eragrostis lehmanniana (Lehmann lovegrass) at the Kendall semiarid grassland site in the USDA-ARS Walnut Gulch Experimental Watershed in southeastern Arizona, USA following a multi-year drought. Supervised classification of a 1m aerial photograph of the Kendall site acquired in October 2009 was performed based on phenological differences in vegetation color, in addition to pattern and texture. Overall classification accuracy was 75%. In 2002, vegetation measurements showed no record of E. lehmanniana growing at the site, but image classification revealed that E. lehmanniana had become dominant in over 56% of the total land area of the scene in 2009. Classification revealed spatial variations not evident from transect measurements alone, including substantial areas previously dominated by native grasses that remained uncolonized by E. lehmanniana. Return visits to the site near the peak of the 2010 growing season revealed the presence of native grass species in uncolonized areas. The phenological traits of these species will have methodological implications for the appropriate timing of both transect measurements and aerial photo acquisition for achieving the different but closely related goals of characterizing the E. lehmanniana invasion and assessing the stability of native grasses.