Environmental Loss Assessment via Functional Outlier Detection of Transformed Biodiversity Profiles

Diversity is vital across various fields like ecology, business, and medicine. From a statistical standpoint, determining diversity presents consistent methodological hurdles, regardless of the specific context. For instance, in ecology, while biodiversity is widely acknowledged as beneficial for ecosystems, there's no universally accepted measure due to diversity's multidimensional nature. Recent research has introduced functional data analysis to address diversity profiles, which are inherently complex and multidimensional. However, a notable limitation is the need for a precise strategy to identify anomalous ecological communities. This study proposes a novel approach to biodiversity assessment using a functional outlier detection system by extending the functional boxplot and outliergram to the context of suitable transformations of Hill's numbers. This research holds significance in identifying early warning signs preceding biodiversity loss and the presence of potential pollutants or invasive species in ecological communities.