The Crested Ibis (Nipponia nippon) is an endangered animal with an extremely high ecological, humanistic, and scientific value. However, this species still faces survival challenges, due to rapidly shrinking foraging grounds, the serious interference of human behavior, and increased habitat requirements. Geographical environment is a significant factor affecting Crested Ibis behavior-pattern analysis and habitat protection. The spatial and temporal trajectory contains habitat location and period information, a vital record of the Crested Ibis' habits, and the basis of all research. Nevertheless, there are only a handful of studies on the missing trajectory data and fusing multiple sources of environmental data-research methods. We studied the spatial and temporal habitat use of the tracked Crested lbis by fusing multiple data-sources in China. This paper adopts the LSTM (long short-term memory) model to supplement the missing trajectory data and perform cluster mining; and a random forest model is used to predict the habitat of the Crested Ibis with high fitting accuracy (R2 = 84.9%). The results show that the Crested Ibis distribution-pattern is characterized by high altitude and proximity to woodland and rivers. Additionally, the habitat dependence on the village implicates human agricultural activities in positively impacting its reproduction. This paper provides a complete method for analyzing Crested Ibis' spatial and temporal trajectory by fusing multi-source data, which is crucial for protecting the survival and reproduction of the Crested Ibis.
Keywords: anniversary activity; behavioral analysis; habitat prediction; night roosting points; trajectory data clustering; trajectory data supplementation.