Broad-scale studies and regional comparisons of Indonesia’s coral reefs are critical given the relative lack of information about these large, diverse, and threatened ecosystems. Most studies on reef benthic composition and distribution have largely focused on rather short transects spanning relatively small areas. Here, we quantify the shallow large-scale spatial distribution of benthic communities in four locations in Eastern Indonesia, i.e. the Spermonde Archipelago in South Sulawesi, and Ambon, Halmahera and Lucipara in Maluku Islands. Shallow coral reef ecosystems (0.5-2 m depth) were assessed using a georeferenced photoquadrat survey method that can cover larger proportions of reefs, as well as benthic assemblages in deeper reefs ($\sim$8 m depth) recorded with photoquadrats. A total of 27.8 km of shallow and 4.8 km of deeper reef ecosystem transects were surveyed in this study. We found that reef benthic communities at the study locations were mainly characterized by low to moderate cover of live hard corals and relatively high cover of dead corals or abiotic components. Ninety percent of the surveyed reefs at both depths exhibited < 50% cover of live hard corals, whereas 88% of shallow and 78% of deeper reef sites consisted of > 25% dead corals or abiotic components. A combination of georeferenced photoquadrats and CoralNet machine learning rapidly identified benthic composition across a large area, which is useful for mapping or other large-scale applications. This information can support decision-makers for better management of resources and marine spatial planning, crucial for an archipelagic state with tens of thousands of islands such as Indonesia. Our broad-scale reef surveying, focusing on under-studied areas of Indonesia, can contribute to this effort, as well as others seeking to better map coral reef habitats, as an important foundation for their conservation amidst growing threats.

鉴于相对缺乏有关这些庞大，多样和受威胁生态系统的信息，对印度尼西亚珊瑚礁进行大规模研究和区域比较至关重要。关于珊瑚底栖生物组成和分布的大多数研究主要集中在跨越相对较小区域的较短的样带上。在这里，我们量化了印度尼西亚东部四个地区（即南苏拉威西岛的斯佩蒙德群岛，马鲁古群岛的安汶，哈马黑拉和露西帕拉）底栖生物群落的浅层大规模空间分布。浅珊瑚礁生态系统（深度为0.5-2 m）采用地理参考的光四足动物调查方法进行了评估，该方法可以覆盖更大比例的珊瑚礁以及更深的珊瑚礁中的底栖生物群（$〜$深度为8 m），并用光四极杆记录下来。在这项研究中，总共调查了27.8 km的浅水礁和4.8 km的深水礁生态系统样带。我们发现，研究地点的珊瑚底栖生物群落的主要特征是活硬珊瑚的覆盖率低至中度，死珊瑚或非生物成分的覆盖率相对较高。在两个深度的被调查珊瑚礁中，有90％的珊瑚覆盖率低于50％，而浅层珊瑚礁的88％和深层珊瑚礁的78％则由25％以上的死珊瑚或非生物成分组成。地理参考光四方体和CoralNet机器学习的结合可在大范围内快速识别底栖生物成分，这对于制图或其他大规模应用很有用。这些信息可以帮助决策者更好地管理资源和海洋空间规划，对于拥有数万个岛屿（例如印度尼西亚）的群岛国家至关重要。我们的大规模珊瑚礁调查着重于研究不足的印度尼西亚地区，可以为这一工作以及其他寻求更好地绘制珊瑚礁栖息地的图景做出贡献，这是在日益严峻的威胁下进行珊瑚礁栖息地保护的重要基础。