A Spatiotemporal Modeling of Bird Migration Under Climate Change Using GRU and HDBSCAN

Abstract—This study investigates the influence of climatechange on bird migration patterns, utilizing machinelearning techniques to predict and analyze migratorybehaviors. Bird migration is a complex, seasonalphenomenon influenced by environmental changes thataffect survival and reproduction. Climate change,particularly rising global temperatures and alteredseasonal patterns, has increasingly impacted migratoryroutes, timings, and morphological characteristics ofbird species. This research employs a Gated RecurrentUnit (GRU) model alongside Hierarchical Density-BasedSpatial Clustering of Applications with Noise(HDBSCAN), trained on historical migration data andclimatic factors, to predict future migration patterns andassess environmental impacts. The methodology includesdata preprocessing, model training, predictionvisualization, and comparative analysis with climatedata. By providing insights into shifting migrationtrends, this approach aims to support conservationefforts, enhance understanding of climate change'secological effects, and improve strategies for preservingmigratory bird populations. The findings highlightcritical environmental variables and offerrecommendations for mitigating the adverse effects ofclimate change on avian migration.