Quantifying the spatial impact of an invasive Acacia on ecosystem functioning using remote sensing

Abstract

Remote sensing technology is increasingly applied to map the occurrence of invasive plant species, yet its use to map their ecological impact remains limited. Furthermore, invader-induced changes beyond the canopy, as well as the environmental context, are rarely considered. This study aimed to assess the impacts of an invasive tree on ecosystem functioning at the landscape scale using remote sensing, taking into account both spatial effects and environmental heterogeneity. Specifically, we investigated a coastal Mediterranean dune ecosystem invaded by the N-fixing tree Acacia longifolia (Andrews) Willd. (‘Acacia’). Four vegetation indices were calculated as proxies of ecosystem functions, and these indices were used to compute functional diversity in terms of spectral Rao’s Q for assessing impacts by Acacia based on airborne hyperspectral data. Vegetation cover and topographic indices derived from airborne LiDAR (Light Detection and Ranging) were used to account for spatial heterogeneity. For seven sites, we employed Generalized Linear Mixed Models to model the effects of environmental variables and Acacia-related variables on proxies of ecosystem functions. Significant impact of the invader was found beyond the invaded area augmenting to 50 % total impact on ecosystem functions. These spatial impacts are particularly prevalent at rather early stages of invasion (∼20 % invader cover at landscape level). Consequently, the impact of invaders is underestimated when spatial effects are ignored, but it is overestimated when environmental heterogeneity is neglected. Furthermore, functional diversity decreases due to invasion, though it reaches its maximum at the edges of invader stands, where Rao’s Q index captures spectral effects of both the invader and the native vegetation. Thus, we highlight that both 2D and 3D remote sensing data complement each other in remote sensing-driven impact assessments. We envision that advancements in remote sensing of ecosystem structure and functioning in terms of increasing availability of high spectral, spatial and temporal data as well as enhanced methods for data analysis will facilitate tracing the context-dependent and function-specific spatial effects of invasive species especially at early stages of invasion to enable timely management.