Long and short term population dynamics of acacia trees via remote sensing and spatial analysis: Case study in the southern Negev Desert

Abstract

Monitoring vegetation dynamics in hyper-arid zones is important because any decrease in the already sparse vegetation cover in these areas could considerably affect the entire ecosystem. The new generation of high spatial resolution satellite (HSR) sensors is suitable for monitoring trees in arid regions because of the distinct and separate objects that trees represent in these environments. The main limitation of modern HSR sensors is the lack of a historical archive that would otherwise enable meaningful landscape change detection, especially in arid regions where tree population dynamics are naturally very slow. This study uses spatial analysis to gain information regarding the long- and short-term dynamic processes affecting the acacia tree population in Wadi Ktora, in the southern Arava Valley, Israel. The data is extracted from a single HSR aerial photograph composed of three spectral bands in the visible and infrared spectrum (green, red, and near infrared).A map of individual acacia trees that is extracted from a colour infrared aerial photograph of Wadi Ktora from 2010 enables the examination of spatial distribution patterns for both tree size and foliage health. Tree size distribution is used as an indicator of long-term (decades) hydrologic spatial processes affecting the acacia population. The tree health distribution is used as an indicator for short-term (months to a few years) hydrologic spatial processes, such as the paths of recent flash floods events. Comparing the distributions of tree size and normalized difference vegetation index (NDVI) enables differentiation between the long-term and short-term processes that brought the population to its present state.Using spatial statistic grouping, the distribution of the trees in the wadi (ephemeral stream) is divided into three distinct categories: (1) large trees with high NDVI values, (2) large trees with low NDVI values, and (3) small trees with medium NDVI values. Using the resulting classification, we divided the wadi into three sections, each representing a unique combination of long- and short-term hydrologic processes affecting the acacia trees. We suggest that the lack of spatial correlation between tree size and health status is a result of spatio-temporal changes in the water supply.