关键词: 声景生态学, 声学指数, 海南长臂猿, 被动声学监测, 随机森林

Abstract: Hainan Bawangling Nature Reserve, the only national nature reserve in China that protects Nomascus hainanus and their living environment, has rich acoustic soundscape resources consisted of sounds from N. hainanus, birds and other animals. Exploring the soundscape composition and acoustic activities of N. hainanus habitat would shed light on the daily behavior of N. hainanus and provide technical support for ecological conservation. Using passive acoustic monitoring techniques, huge amount of sound data were collected in the protected area. The soundscape and its daily variations were characterized by analyzing the distribution of the soundscape power in different frequency ranges. Four widely used acoustic indices, including ACI (acoustic complex index), ADI (acoustic diversity index), BI (bioacoustic index), and NDSI (normalized difference soundscape index), were extracted to evaluate N. hainanus sounds. The Random Forest method was employed to classify sounds with the four acoustic indices as predictor variables and the presence or absence of N. hainanus sounds as response variables. The results showed that: (1) The soundscape power in different frequency ranges reflected different acoustic community information, with significant temporal variations, indicating the dawn chorus and dusk chorus activities of birds, N. hainanus and other animals. Among them, 1-2 kHz was consistent with the N. hainanus dawn chorus, and 2-5 kHz was in line with the circadian rhythm of birds and other animals. (2) The changing trends of acoustic indices with and without N. hainanus calls were significantly different, with the largest difference of NDSI and the least of ADI. (3) Four acoustic indices gave better classification results for the sound data of N. hainanus, with a crossvalidated AUC value of 0.9764. This study effectively reveals the activity characteristics of animals in the experimental area based on soundscape, indicating the feasibility of using acoustic index to evaluate the call of N. hainanus, and thus lays the foundation for acoustic monitoring of Hainan Bawangling area and N. hainanus.

Key words: soundscape ecology, acoustic index, Nomascus hainanus, passive acoustic monitoring, random forest