Scholarlink

<p>Reeves’ Muntjac (<i>Muntiacus reevesi</i>) are an invasive alien species in a number of European and other countries. They listed as a “Species of European Union Concern” requiring member states to survey their populations and undertake management. Muntjac are secretive and existing survey methods are inefficient at low density and cannot reliably provide confidence in species absence. We assess acoustic monitoring as a surveillance method for this species. Muntjac produce readily identifable barking calls in sequences that can last many minutes. We collected 133,227&#xa0;min of nocturnal recordings using automated sensors from areas known to contain muntjac populations in the UK. These contained 2,453&#xa0;min containing at least one muntjac call. We developed a machine learning model to automate detection with high precision (F1 = 0.98). Calls were recorded in all months with no clear seasonal pattern. Calls were produced in sequence in sessions with an average duration of over 10&#xa0;min and could be reliably detected by the sensors over a radius of around 200&#xa0;m. We recommend 107&#xa0;h of surveillance per sensor are required to provide 95% confidence that the species is absent from that site. Calling rate can be increased three-fold by the use of playback. We describe how acoustic surveillance can provide complete coverage of an area, a feature that cannot be provided by other methods such as cameras, can be efficient when the species is only present at low density, can provide confidence in absence, and how automation could further improve the cost-effectiveness of this process.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Evaluating automated acoustic monitoring for the invasive reeve’s muntjac deer (Muntiacus reevesi): detecting presence, absence and survey design

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年7月10日

查看全文
  • Peter Robertson,
  • Madeleine Bishara-Robertson,
  • Maureen Woodburn,
  • Roy Sanderson

摘要

Reeves’ Muntjac (Muntiacus reevesi) are an invasive alien species in a number of European and other countries. They listed as a “Species of European Union Concern” requiring member states to survey their populations and undertake management. Muntjac are secretive and existing survey methods are inefficient at low density and cannot reliably provide confidence in species absence. We assess acoustic monitoring as a surveillance method for this species. Muntjac produce readily identifable barking calls in sequences that can last many minutes. We collected 133,227 min of nocturnal recordings using automated sensors from areas known to contain muntjac populations in the UK. These contained 2,453 min containing at least one muntjac call. We developed a machine learning model to automate detection with high precision (F1 = 0.98). Calls were recorded in all months with no clear seasonal pattern. Calls were produced in sequence in sessions with an average duration of over 10 min and could be reliably detected by the sensors over a radius of around 200 m. We recommend 107 h of surveillance per sensor are required to provide 95% confidence that the species is absent from that site. Calling rate can be increased three-fold by the use of playback. We describe how acoustic surveillance can provide complete coverage of an area, a feature that cannot be provided by other methods such as cameras, can be efficient when the species is only present at low density, can provide confidence in absence, and how automation could further improve the cost-effectiveness of this process.