Author(s):
Yongxin Deng* - Western Illinois University
Xianfeng Chen - Slippery Rock University
Abstract:
This paper used Pearson's r and Moran's I to detect relationships between nine local/non-local topographic attributes and two remote-sensed live fuel (vegetation) indexes NDVI, NDII, and their seasonal changes (between May, July, and September). Aggregation of data resolution from 10 m for DEM derivatives and 30 m for NDVI/NDII to 20 coarser resolutions below 1500 m supported a multi-scale analysis. Elevation, slope, and "southness" were identified to have consistent correlations (e.g. 0.2 < r < 0.6) with vegetation variables at multiple observation scales. Change of scale caused change of r and Moran's I, and scales with higher (peak) correlations exist, but at much coarser scales than the data scale. After human-controlled landscapes were masked off, Moran's I of vegetation indexes between natural mountains and the entire study area appeared to be similar in May, but increasingly different from July to September.
