Submitted to: Environmental Monitoring and Assessment
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 7/16/2010
Publication Date: 7/15/2011
Citation: Ge, S., Carruthers, R.I., Everitt, J.H., Kramer, M.G., Anderson, G.L. 2011. Multiple-level defoliation assessment with hyperspectral data: integration of continuum-removed absorptions and red edges. Environmental Monitoring and Assessment. 32(21):6407-6422. DOI:10.1080/01431161.2010.510492. Interpretive Summary: Saltcedar is a highly invasive exotic shrub that has infested valuable riparian lands throughout the western United States. Saltcedar caused damage to both agricultural systems and the natural environment through extensive water use, increased fire danger, and saltation of soils that limits natural vegetation from populating infested areas. These factors have allow extensive monocultures of saltcedar to develop along many western rivers and streams, where the associated cost of damage and control now exceed several $ million each year. In response to this problem, the US Department of Agriculture has developed and implemented an extensive biological control program using highly specific natural enemies that feed exclusively on the destructive saltcedar. As part of that control program, remote sensing methodologies were developed to help assess and provide summary data for test and implementation release areas. This paper describes an assessment of using hyperspectral data to identify four different categories of saltcedar plants, those that are 1) unaffected, 2) newly/ slightly damaged (partially defoliated), 3) highly damaged (completely defoliated) and 4) those showing signs of regrowth. Being able to use remote sensing information to properly categorize these damage classes of saltcedar foliage allows land managers to better evaluate the success of this program and to improve implementation efforts across wide areas. Follow ip remote sensing has shown that this biological control project was highly successful, helping to control saltcedar across hundreds of thousands of acreas in multiple state from California to Colorado and Texas to Wyoming.
Technical Abstract: Hyperspectral data were collected from 40 canopies of saltcedar (Tamarix ramosissima), 10 healthy canopies and 30 canopies defoliated by an introduced biological control agent, the saltcedar leaf beetle (Diorhabda elongata). These data were assessed to detect categories of defoliation in response to the process of biological control. Two important characteristics of the hyperspectral data, red edges and continuum-removed absorptions were used to discriminate four defoliation categories of saltcerdar (completely-defoliated plants, healthy plants, newly-defoliated plants, and refoliating plants) at a canopy level. The red edge positions were located in ranges of 711 ~716 nm, 706 ~712 nm, 694 ~698 nm, and 715~719 nm (ARE THESE IS THE CORRECT ORDER give the categories listed above? I put them in the same order that you used them below, however, it might make more sense to out them in the following order 1) healthy, 2) newly-defoliated, 3) totally defoliated, and 4) refoliating) for the four defoliation stages described above, respectively. These red edger positions alone could not clearly judge the four defoliation categories associated with feeding by the beetles. Only the completely-defoliated canopies had distinct red edge positions that could be differentiated from the other three types of canopy conditions. While using a classification tree to integrate the red edge positions and their derivatives with the central band depths of these five continuum-removed absorptions, it was found that only two band depths of the continuum-removed absorptions were selected, they were the red absorption between 470 nm and 524 nm and the first water absorption between 936 nm and 990 nm in the near infrared region (NIR). This implied that the continuum-removed absorptions outperformed the red edges to identify the defoliation categories. The resulting overall accuracy was 87.5%. The producer’ accuracy was 100%, 100%, 70% and 80% for the completely-defoliated plants, healthy plants, newly-defoliated, and refoliating canopies, respectively. The corresponding user’ accuracy was 100%, 90.91%, 77.78%, and 80% (again, put these all in the same order as above). Therefore, we concluded that any single spectral data based variable failed to separate the four stages but combination of the two continuum-removed absorptions locating in the red absorption and the first water absorption in NIR improved the identification of defoliated canopies associated with the dynamic defoliation process of the biological control agent.