Abstract: Reservoir impoundment has a great impact on engineering and geological environment around the reservoir, and would even give rise to landslide. By adopting global and local spatial autocorrelation coefficients and their analysis method in spatial economics, we established the spatial weight matrix in consideration of 3-D spatial coordinates of measuring points for Xiangjiaba downstream high slope. Furthermore, we calculated the global and local spatial autocorrelation coefficients of deformation data during impoundment, and compared the aggregation behavior of high-slope deformation during different impoundment periods so as to determine key measurement points. The results showed that the global and local spatial autocorrelation coefficients could effectively detect the aggregation behavior of high-slope’s deformation. Apparent aggregative deformation in downstream slope would be induced during impoundment, in particular, the initial period of impoundment, which should be monitored in emphasis.

Key words: reservoir impoundment, downstream high-slope, visual deformation, spatial weight matrix, spatial autocorrelation coefficient