Using Ndvi Image Texture Analysis for Bushfire-prone Landscape Assessment

ABSTRACT: Bushfire, as a major external ecological factor, diversifies bushland environments. Monitoring bushfire-prone landscape patterns and vegetation recovery after fires is critical for the long-term bushland management. Landscape ecology studies using remotely sensed imagery have been effective to identify the relationship between landscape patterns and ecological processes. This paper uses the Normalised Difference Vegetation Index (NDVI) as a surrogate for vegetation biomass and applies two texture statistics to quantitatively examine biomass variability in bushfire-prone landscapes. Specifically, the texture number is used to describe landscape heterogeneity and post-fire vegetation recovery, and the fractal dimension is applied to evaluate landscape pattern changes with post-fire vegetation successional stages. Results show that bushland spatial patterns are heavily affected by bushfires and human disturbances, and the fractal dimension of post-fire vegetation decreases with successional stages. It is suggested that the use of NDVI image texture analysis is a potential technique to examine preand post-fire landscape changes, and these methods can be applied using remotely sensed imagery in a cost-effective way.