Ensemble modelling of Pirarucu (Arapaima gigas) distribution in biodiversity hotspot to understand its invasion risk

Abstract

Invasive species pose a severe threat to biodiversity around the world. Managing the consequences of invasion is difficult in aquatic settings, as the rate at which invaders establish typically outpaces the resources available to eradicate them. For proactive management measures to be implemented, prior knowledge of the probability of invasion is required. In this study, we created a spatial model of the probability of the Pirarucu (Arapaima gigas) invasion in the Western Ghats. The Western Ghats, one of the world's top biodiversity hotspots, is home to numerous endemic species, many of which are now threatened. An ensemble modelling approach using 10 models, including machine learning techniques such as Artificial Neural Network (ANN), Maximum Entropy (MaxEnt), Random Forest (RF), Generalised Boosted Regression Model (GBM) and Classification Tree Analysis (CTA), was adopted. The model was built using the species' occurrence data and nine climate variables. The findings revealed that southern regions of the Western Ghats have a high risk of Pirarucu invasion. Sri Lanka also has a much greater geographical area with a higher percentage of appropriate habitats for the species. The study becomes vital as this exotic species was repeatedly reported from the rivers since the extensive floods in the region in 2018. The developed model will assist managers in prioritising locations and initiating monitoring and management steps to prevent the spread before they establish in the wild. With earlier Pirarucu invasions in Bolivia, Peru and East Asia and recent climatic vagaries in the Western Ghats, the native biodiversity of the region is in grave danger of being displaced.