中文导读

蝙蝠夜间活动、飞得快、很难看见，科学家怎样才能知道一个国家的林地里到底有哪些蝙蝠？文章介绍英格兰刚公布结果的“全国林地蝙蝠调查”：先说明蝙蝠为什么难研究，以及它们靠回声飞行和捕食的特点；再交代调查由谁组织、在多少片林地、用什么方式录音；接着解释为什么这么多录音必须依靠人工智能处理，又为什么还要请专家抽查结果；随后列出调查中发现的稀有蝙蝠和阔叶林的重要性；最后说明这批数据填补了哪些空白，以及研究团队下一步的计划。阅读时可留意 yet、however、instead 如何引出转折或替代，以及 may、could 等词如何表明结论仍有待研究。

正文

Bats are hard to study. They come out at night, move quickly and are often difficult to see. Yet they make sounds as they fly, and they use the echoes to find their way and catch insects. A new survey in England has listened to those sounds on a huge scale, and its organisers say it would not have been possible without artificial intelligence.

本段译文

蝙蝠很难研究。它们在夜间出没，行动迅速，而且常常很难被看见。然而，它们在飞行时会发出声音，并利用回声辨别方向、捕捉昆虫。英格兰的一项新调查以巨大的规模“聆听”了这些声音，组织者表示，如果没有人工智能，这是不可能做到的。

The National Woodland Bat Survey was led by the Bat Conservation Trust and Forest Research, and it was paid for by the UK government. Between 2021 and 2024, researchers placed small sound recorders in 310 woodlands across England. At each site, three devices recorded for nine nights in a row. In total, the team collected 9,734 nights of recordings, or about 90,000 hours of sound.

本段译文

“全国林地蝙蝠调查”由蝙蝠保护信托（Bat Conservation Trust）和英国林业研究所（Forest Research）牵头，由英国政府出资。2021 年至 2024 年间，研究人员在英格兰各地的 310 片林地中放置了小型录音设备。在每个地点，三台设备连续录音九个晚上。团队总共收集了 9734 个夜晚的录音，约合 9 万小时的声音。

That is far too much sound for people to check by hand. According to the BBC, it would have taken more than 30 years to analyse without AI. Instead, the team worked with scientists at University College London and the University of Edinburgh, who helped develop open-source tools that can pick out bat calls and suggest which species made them. The system found more than 98 million calls.

本段译文

这么多声音，人工逐一检查根本不现实。据 BBC 报道，如果不用人工智能，分析这些录音需要 30 多年。于是，团队与伦敦大学学院和爱丁堡大学的科学家合作，后者帮助开发了开源工具，能够找出蝙蝠叫声并推测是哪种蝙蝠发出的。该系统共识别出 9800 多万次叫声。

The researchers did not simply trust the computer, however. A group of 19 bat call experts checked a sample of the results. Kate Jones of University College London said AI had “fundamentally changed the scale at which we can monitor wildlife”. She also said the experts‘ work meant the results could be carefully checked.

本段译文

不过，研究人员并没有简单地完全相信电脑。一个由 19 位蝙蝠叫声专家组成的小组抽查了部分结果。伦敦大学学院的凯特·琼斯表示，人工智能“从根本上改变了我们监测野生动物的规模”。她还说，专家的工作使这些结果能够得到仔细核查。

The survey found every bat species or species group in Britain except one, the serotine, which is hard to identify from sound. It also recorded some of the country’s rarest bats, such as the barbastelle and the grey long-eared bat. Barbastelles were found in parts of Cornwall, Devon and Dorset where few records of them existed before. Common pipistrelles were found in all 310 woodlands, and bat activity was highest on average in broadleaved woods.

本段译文

调查发现了英国所有的蝙蝠物种或物种类群，只有一种例外——大棕蝠（serotine），它很难仅凭声音识别。调查还记录到该国一些最稀有的蝙蝠，例如宽耳蝠（barbastelle）和灰长耳蝠。在康沃尔、德文和多塞特的部分地区，过去几乎没有宽耳蝠的记录，这次也发现了它们。普通伏翼蝠在全部 310 片林地中都有发现，而蝙蝠活动平均在阔叶林中最为频繁。

These results matter because the survey fills a gap. For seven species, it is now the only national survey in England away from places where bats are already known to rest. Lia Gilmour of the Bat Conservation Trust said bats can show the quality of the habitats they depend on. The team now plans to study how climate, woodland cover and hedgerows may shape where bats live. The answers could help people manage woodlands better as the climate changes.

本段译文

这些结果之所以重要，是因为调查填补了一项空白。对于七个物种而言，这项调查是目前英格兰唯一一项在已知栖息地以外开展的全国性调查。蝙蝠保护信托的莉娅·吉尔摩说，蝙蝠能够反映它们所依赖的栖息地的质量。团队接下来计划研究气候、林地覆盖率和树篱可能如何影响蝙蝠的分布。研究结果或许能帮助人们在气候变化的情况下更好地管理林地。

新闻来源

• BBC News · Largest woodland bat survey a ‘nice story of AI for good’ (30 Oct 2026)
• Forest Research · AI helps reveal Britain’s bats in largest-ever national woodland bat survey (30 Oct 2026)
• UCL News · AI helps reveal Britain’s bats in largest-ever national woodland bat survey (30 Oct 2026)

✳更多文章 →