ALIAS: a smart classifier for monitoring Ailanthus altissima expansion in hydrologically sensitive areas

Abstract

Introduction Biodiversity loss is a growing global concern, threatening the stability of natural ecosystems. This phenomenon is driven by a combination of anthropogenic and natural factors, including urbanization, deforestation, and, most critically, climate change. Rising global temperatures and shifting precipitation patterns are altering wildfire regimes, leading to more frequent and intense fires. These disturbances create open niches and reduce competition, facilitating the establishment of invasive alien plant species. Characterized by high adaptability, rapid growth, and aggressive competition strategies, invasive plants can significantly disrupt local ecosystems. Moreover, they can profoundly impact hydrological processes by modifying evapotranspiration rates and reducing water retention capacity. Among the most concerning invasive species, Ailanthus altissima (also known as tree of heaven) has rapidly spread across continents, establishing itself in both urban and natural environments (Soler & Izquierdo, 2024). Native to China, Ailanthus has become a widespread invader due to its remarkable adaptability and aggressive colonization strategies. Indeed, Ailanthus thrives in a wide range of environmental conditions, demonstrating resilience to extreme temperatures, diverse soil types, and high levels of air pollution (Iverson et al., 2019; Tarantino et al., 2019). One of the most problematic traits of Ailanthus is its extraordinary regenerative capacity. Even when cut or burned, the species can resprout vigorously, making eradication efforts particularly challenging (Badalamenti et al. 2015). Its seeds are primarily dispersed by wind but also spread via water, animals, and human activity, further accelerating its expansion. Ailanthus exhibits a strong dependence on water availability, employing deep root systems and highly efficient water uptake strategies (Petruzzellis et al., 2019). This allows it to thrive even in water-limited environments, such as Mediterranean ecosystems, where it can outcompete native vegetation for scarce water resources. By monopolizing soil moisture, Ailanthus alters the local ecohydrological balance, reducing groundwater recharge, increasing evapotranspiration rates, and potentially exacerbating drought conditions. Given its rapid spread and ecological impact, Ailanthus altissima has been the focus of multiple management strategies, but effective control remains a challenge, requiring integrated and long-term approaches to mitigate its expansion. Materials and methods The present work is based on high-resolution PlanetScope satellite imagery to detect and monitor the spread of Ailanthus altissima. PlanetScope provides multispectral data (i.e., 8 bands) at a spatial resolution of 3 meters, enabling fine-scale vegetation analysis. To automate the detection of Ailanthus, ALIAS (Ailanthus Locator and Identification Algorithm Suite) was developed, a machine learning-based classifier based on the Support Vector Machine (SVM) model. The model was trained using a dataset of reference points, distinguishing between Ailanthus and native vegetation. ALIAS was specifically calibrated to detect Ailanthus altissima in two key environments where the species tends to establish: (i) transportation corridors, where wind generated by vehicles facilitates seed dispersal, and (ii) hydrologically connected areas, such as riparian zones, where Ailanthus exploits water availability for rapid growth. Training samples were collected through a combination of field surveys and visual interpretation of high-resolution images. The model was validated in the Vallone Piano della Corte Nature Reserve (Sicily, Italy), a protected area sensitive to the Ailanthus altissima invasion. Over the past four decades, Ailanthus altissima has progressively expanded, forming dense clusters on the south-facing slope, while the north-facing slope remains dominated by native forest stands (i.e., Quercus pubescens). To further assess