人类的活动正以显著的速度改变着全球生态系统，导致了栖息地退化和生物多样性丧失。了解生物多样性如何受到人类影响已经显得尤为重要，也很有必要开发实用且高效的方法来进行大时空尺度的多样性监测。近期研究表明鸣叫的动物，如鸟类，可以通过声学手段来监测。根据我们前期在云南研究结果，结合广西大学在7个自然保护区建立的32个永久固定样地，本项目将重点回答以下问题： (a)不同研究地有效声学指标的一致性及其重要程度的关联性如何? (b)整合当前一些常用的指标能否提高其测量鸟类多样性的能力? (c)声学数据的预处理，特别是过滤昆虫的叫声能否增强指标的准确性? (d)研究指标是否受到森林的组成和空间结构等因素的影响? (e)引入景观尺度的栖息地质量能否提高指标的准确性? 通过回答这些问题将对改进声学指标，并为最终用以自动监测森林鸟类群落多样性提供科学依据。

Anthropogenic activities are altering the world’s ecosystems at a substantial rate often resulting in habitat degradation and biodiversity loss. To be able to protect biodiversity it is important to understand how it is being affected by anthropogenic activities. We therefore need effective and efficient tools that will allow us to monitor biodiversity over large spatial and temporal scales. Consequently, scientists are increasingly looking into alternative technologies. Recent research has suggested that vocalizing animals, such as birds, could be monitored using acoustic methods. By measuring acoustic complexity, one can make inferences about the bird community. Assuming acoustic indices can measure bird diversity effectively then a large network of acoustic sensors – which are becoming increasingly cheaper and efficient – could be deployed to monitor autonomously birds over large scales. Although the indices do have the potential to measure bird diversity there still remain several key questions regarding their performance and how they can be improved further. Here, we propose a project through which will explore some of those remaining questions regarding the indices. Building on our results from a previous study on acoustic indices in Yunnan Province, combined with the 32 permanent forest plots in seven nature reserves established by Guangxi University, we plan to answer specifically the following questions: (a) How consistent are acoustic indices across different areas and how do they rank relatively to each other? (b) Does combining some of the currently available indices help improve the measurement of bird diversity? (c) Does pre-processing the acoustic data, and specifically filtering out the insect calls, improve the accuracy of the indices? (d) Is the accuracy of the indices affected by parameters related to forest structure and composition? (e) Does including information about habitat quality at the landscape-level improve the accuracy of the indices? Answering these questions is an essential next step for improving the acoustic indices and for eventually using them to monitor autonomously the bird communities in biodiverse forests.