HR: 09:45h
AN: B31A-07    [PDF]
TI: How Big is the Problem? Constraints on the Extent of Cheatgrass Invasion in the Great Basin, US
AU: * Bradley, B A
EM: Bethany_Bradley@brown.edu
AF: Brown University, Department of Geological Sciences, Providence, RI 02912 United States
AU: Mustard, J F
EM: John_Mustard@brown.edu
AF: Brown University, Department of Geological Sciences, Providence, RI 02912 United States
AU: Albert, J
EM: Jeffrey_Albert@brown.edu
AF: Brown University, Department of Geological Sciences, Providence, RI 02912 United States
AB: Land use leading to ecosystem disturbance (eg. grazing, agriculture) in the Great Basin, US has been shown to enhance the spread of invasive annuals, in particular cheatgrass (Bromus tectorum) (Young et al., 1972; Mack, 1981). Cheatgrass can dominate as a monoculture or as an understory to semiarid perennial shrubs. Cheatgrass is problematic because it outcompetes native perennial bunchgrasses and shrubs is highly flammable once senesced, and cannot be grazed after senescence in mid-June. Identification of the regional extent of this invasive is necessary for effective land management and accurate ecosystem models. We have shown previously that timeseries from the 1 km resolution Advanced Very High Resolution Radiometer (AVHRR) can be used to accurately distinguish areas of cheatgrass dominance. Here, we scale down to a series of 10 Landsat TM scenes spanning 1988-2001 and find that similar timeseries analysis can be used to identify cheatgrass at 30 m resolution. Elmore et al. (2003) demonstrated that invasive annuals can be distinguished from native perennials based on their amplified response to rainfall. During wet years, cheatgrass dominated landscapes have much higher density and productivity than native shrub dominated landscapes. Amplified response can be detected with Normalized Difference Vegetation Index (NDVI) timeseries from the AVHRR satellite. We created and verified a regional AVHRR-based map of current cheatgrass extent in the Great Basin, and are refining our methodology to map invasion over the decade of the 1990s. The regional extent map was used to identify large expanses of cheatgrass dominance for further investigation at Landsat TM resolution. We selected a high density cheatgrass corridor between Lovelock and Winnemucca in northwestern Nevada and acquired a series of 10 Landsat scenes for that area. The Landsat scenes were coregistered and radiometrically aligned to a 2001 ETM+ scene. They were then converted to reflectance based on invariant ground control points and NDVI was calculated. To detect the amplified response of cheatgrass dominated areas, we ratio NDVI of a high rainfall year (1995) to a low rainfall year (1992). In the Lovelock area, 156 mm of rain fell from Oct-Apr, 1995. This was 77 mm above the 50-year mean. From Oct-Apr 1992, a below-average 51 mm of rain fell. We took a ratio of 1995/1992 NDVI values and compared areas known to contain $>$30% live cover cheatgrass based on field observation to areas known to contain only perennial shrubs. Cheatgrass dominated areas had a value of 1.59 +/- 0.02 times higher NDVI in 1995. Native perennial dominated areas had a value of 1.22 +/- 0.01 times higher NDVI in 1995. When we subtract 1995-1992 NDVI values the results are equally significant. Cheatgrass dominated areas have a difference of 0.21 +/- 0.06 NDVI, while native perennials have a difference of 0.08 +/- 0.02 NDVI. This contrast indicates that cheatgrass' amplified response can effectively be mapped at Landsat resolution. We have shown that it is possible to map cheatgrass both at a regional scale with AVHRR as well as at a local scale with Landsat TM. By mapping at the higher resolution of Landsat, it becomes possible track the spread of cheatgrass using the high rainfall years of 1988, 1995, and 1998. Land use drivers of cheatgrass invasion can be identified based on patterns and rates of spread. Finally, landscape response to recent invasion can be modeled when the timing of the invasion is constrained. Elmore, A., J. Mustard, and S. Manning, Ecol. App., v13, 2003; Mack, R., Agro-Ecosys., v7, 1981; Young, J., R. Evans, and J. Major, J. Range Mgmt, v25, 1972
DE: 1640 Remote sensing
DE: 1699 General or miscellaneous
DE: 1851 Plant ecology
SC: Biogeosciences [B]
MN: 2003 Fall Meeting