Scholarlink

<p><i>Acropora cervicornis</i> and <i>Acropora palmata</i> have long provided structural complexity to Caribbean reefs, yet they have declined by over 90% in recent decades. Identifying places where wild populations of at-risk corals can survive remains essential for conservation. Coral Gardens, Belize, has historically been a refugium for endangered <i>Acropora spp</i>., but mass mortality transpired in 2023. This study aimed to determine whether coral loss at a small long-term field monitoring site was representative of regional trends. We developed a supervised Random Tree classification model to identify the spectral signature of <i>Acropora spp</i>. in satellite images and predict live coral cover across 25&#xa0;km² from 2011 to 2023 to compare with field data from 2012 to 2025. Model data predicted a drop in live <i>Acropora spp</i>. of over 90% from 2011 to 2023 coincident with total loss at the field site. Field validation of the model became a search for survivors in June 2025. Live or recently dead <i>Acropora spp</i>. were found at 28 of 42 predicted sites, and live colonies were found at 16 previously undocumented locations. This methodology provides a tool for monitoring shallow-water <i>Acropora spp</i>. and locating survivors. Our findings underscore the vulnerability of even persistent coral refugia.</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

Supervised satellite classification and field validation document endangered Acropora collapse, and identify surviving colonies in Belize

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年6月2日

查看全文
  • Lisa Greer,
  • Raja Das,
  • Harris Foad,
  • Karl Wirth

摘要

Acropora cervicornis and Acropora palmata have long provided structural complexity to Caribbean reefs, yet they have declined by over 90% in recent decades. Identifying places where wild populations of at-risk corals can survive remains essential for conservation. Coral Gardens, Belize, has historically been a refugium for endangered Acropora spp., but mass mortality transpired in 2023. This study aimed to determine whether coral loss at a small long-term field monitoring site was representative of regional trends. We developed a supervised Random Tree classification model to identify the spectral signature of Acropora spp. in satellite images and predict live coral cover across 25 km² from 2011 to 2023 to compare with field data from 2012 to 2025. Model data predicted a drop in live Acropora spp. of over 90% from 2011 to 2023 coincident with total loss at the field site. Field validation of the model became a search for survivors in June 2025. Live or recently dead Acropora spp. were found at 28 of 42 predicted sites, and live colonies were found at 16 previously undocumented locations. This methodology provides a tool for monitoring shallow-water Acropora spp. and locating survivors. Our findings underscore the vulnerability of even persistent coral refugia.