Integrating fuzzy logic with piecewise linear regression for detecting vegetation greenness change in the Yukon River Basin, Alaska
2013
Xue, Yueju | Liu, Shuguang | Zhang, Li | Hu, Yueming
Effective methods are required to detect vegetation changes and attribute these changes to different driving forces. This article presents a new hybrid change detection method by integrating fuzzy logic with piecewise linear regression (PLR) as applied to the Yukon River Basin (YRB) in Alaska, USA, to detect the changes in spatial and temporal patterns of vegetation greenness based on Advanced Very High Resolution Radiometer (AVHRR) normalized difference vegetation index (NDVI) data from 1995 to 2004 (2000 data missing). Abrupt changes (or breakpoints), and speed and tendency of gradual changes in short-, mid-, or long-term segments, were simultaneously detected during the study period. Fire scars, precipitation, and temperature data were applied to explain the spatial and temporal patterns of breakpoints and temporal trends (i.e. slopes or change rates). Results showed no significant systematic temporal trend (p = 0.1) in more than 90% of the study area. Abrupt changes in NDVI were partly attributed to wildfire disturbances. The NDVI decline partly overlaid the fire scars provided by the Alaskan Fire Science Center in time and space. The temporal change rates of NDVI were closely related to the temporal trends of temperature/precipitation. The approach presented in this article can be effectively adapted to other study areas.
显示更多 [+] 显示较少 [-]