Free-ranging livestock changes the acoustic properties of summer soundscapes in a Northeast Asian temperate forest

If not properly managed, free-range livestock grazing could be a great threat to forest ecosystems by causing heterogeneous and unpredictable changes in forest structure, functions, and biodiversity. Understanding these changes is important for informing land management decisions compatible with long-term biodiversity conservation. Passive acoustic monitoring (PAM) of biodiversity may provide integrative indices for assessing grazing effects. However, little is known about the responses of soundscape characteristics to forest grazing, especially in temperate forests. Here, we use bioacoustics to assess the impact of continuous livestock grazing on acoustic indices at the landscape scale in Northeast China. We compared the acoustic complex index (ACI), acoustic diversity index (ADI), acoustic evenness index (AEI), bioacoustic index (BIO), total acoustic entropy (H), and normalized difference soundscape index (NDSI) between 16 cattle-grazed forest plots and 36 ungrazed forest plots for different time scales (24 h diel cycles, dawn, day, dusk and night). We collected and analysed over 7000 h of recordings from 52 sites. Our results showed that the six acoustic indices exhibited distinct diel patterns in both forest habitats. Moreover, several indices, except for the ADI and AEI, significantly responded to grazing over 24 h diel cycles, but with contrasting patterns; compared to grazed forests, ungrazed forests had much more bioacoustics activity measured by significantly higher ACI and NDSI values and significantly lower BIO and H values. Livestock grazing was a key driver of diel acoustic patterns, resulting in the loss of nocturnal and dawn soundscapes. Our results indicate that intensive/excessive livestock grazing has a significant negative impact on soundscape quality in human-dominated temperate forest areas, which is likely triggered by shifts in faunal assemblages. Livestock, for example, may reduce the occupancy of sika deer (Cervus nippon), resulting in fewer sika deer vocalizations in grazed forests. Cattle lowing and the intensive sounds of cattle bells may also mask biological vocalizations. We argue that multiple acoustic indices are required to adequately account for the complex responses of soundscapes to large-scale livestock grazing. For forest management, we propose employing PAM to quantify the acoustic communities of the landscape from a new perspective, thereby fostering evidence-based forest management at the local scale to reconcile the livestock sector with ecosystem conservation.