Detecting the land-cover changes induced by large-physical disturbances using landscape metrics, spatial sampling,...

Area of Science:

• Environmental Science
• Remote Sensing
• Geospatial Analysis

Background:

• Landscape changes due to disturbances impact spatial heterogeneity and variability.
• Monitoring these changes requires robust analytical methods for remotely sensed data.
• Normalized Difference Vegetation Index (NDVI) is a key indicator for vegetation health and land cover.

Purpose of the Study:

• To integrate conditional Latin Hypercube Sampling (cLHS), sequential Gaussian simulation (SGS), and spatial analysis for monitoring landscape changes.
• To assess the effects of large chronological disturbances on landscape spatial characteristics.
• To delineate spatial patterns and variability of disturbed landscapes using remotely sensed data.

Main Methods:

• Conditional Latin Hypercube Sampling (cLHS) for sample selection from NDVI images.
• Sequential Gaussian Simulation (SGS) for generating NDVI maps.
• Spatial analysis techniques including variogram, Moran's I, and landscape metrics for pattern delineation.
• Verification of simulated NDVI maps using correlation coefficient and Mean Absolute Error (MAE).

Main Results:

• Spatial patterns of disturbed landscapes were successfully delineated using spatial analysis on NDVI images.
• The hybrid method effectively mapped the spatial patterns and variability of landscapes affected by disturbances.
• The statistics and spatial structures of multiple NDVI images showed robust behavior, validating the approach.

Conclusions:

• The integrated hybrid method provides a robust framework for quantifying landscape spatial patterns and land cover change.
• Remotely sensed NDVI data, analyzed with these methods, are valuable for understanding disturbance impacts.
• Open Geospatial techniques facilitate web-based access to results for watershed management applications.