基于生态调查与多光谱影像的珊瑚礁遥感监测方法研究

Coral reefs are considered as a marine ecosystem that contains the highest level of biodiversity and primary productivity comparing with tropical rainforests on land. Remote Sensing is an extremely effective technique, possessing potentials for continuously monitoring ecological status of coral reefs in large scale. However, as the remote sensing imagery is influenced by the water column and its classification scheme of the benthic habitats is far from imperfect, there are still many problems in the remote sensing-based monitoring of the coral reefs, e.g. the classification results may be not ideal, and the results cannot satisfy the requirements, etc. Therefore, direct comparative study of the ecologically in-field investigation and the remote sensing observation will be conducted in this research to explore a feasible remote sensing-based method for monitoring the macro-ecological status of the coral reefs, in which Luhuitou Island and Weizhou Island are selected as the study sites. Specifically, the research firstly proposed an image pre-processing method to eliminate the influence of the water column for getting the bottom albedos on the basis of the continuity of the spatial variations. Thereafter, a classification scheme of the benthic habitats is developed according to the requirement of the Coral Reef Ecology. An image classification method corresponding to the scheme is then provided based on the fact that any individual image pixel can be composed of two different types of major benthic habitats. Subsequently, a quantitative mathematics model is discussed to convert the image classification results to the ecological parameters of the benthic habitats. This research is an important attempt to explore the integration of Remote Sensing and Coral Reef Ecology. The proposed method will contribute to the remote sensing-based monitoring of the benthic habitats in the South China Sea (SCS) and even the whole globe.

珊瑚礁是海洋中生物多样性和初级生产力极高的生态系统，遥感在其生态状况的大范围连续监测方面具有很大的潜力。然而，珊瑚礁遥感影像因其分类体系不完善和受到水体作用等因素的影响，导致其在珊瑚礁生态的遥感监测中还存在分类效果不佳、结果与需求不符等问题。因此，本项目拟选择鹿回头与涠洲岛为对象，通过影像资料与现场调查资料的定量对比，探索珊瑚礁宏观生态的遥感监测方法。具体地，本项目在以空间变化的连续性等条件为基础，提出能够消除水体的影响而得到水底反射率的影像预处理方法后，按照珊瑚礁生态的需要先进行遥感影像分类体系的完善，再根据“任意像元均可由两种主导的底质类型混合构成”的观点提出与分类体系相称的影像分类方法，并构建将分类结果转换为各种底质类型覆盖率等参数的数学模型。该方法的建立有望使得遥感监测与珊瑚礁生态学得到良好的交叉与融合，切实实现对南海乃至全球珊瑚礁生态状况的有效监测。

珊瑚礁是海洋中生物多样性和初级生产力极高的生态系统，遥感在其生态状况的大范围连续监测方面具有很大的潜力。然而，珊瑚礁遥感影像因其分类体系不完善和受到水体作用等因素的影响，导致其在珊瑚礁生态的遥感监测中还存在分类效果不佳、结果与需求不符等问题。因此，本项目拟选择鹿回头与涠洲岛为对象，按照珊瑚礁生态学的需要，通过影像资料与现场调查资料的定量对比，探索珊瑚礁宏观生态的遥感监测方法。项目组在论文《珊瑚礁遥感研究进展》中阐述了国内外珊瑚礁遥感研究与应用的进展，分析了珊瑚礁遥感研究与应用的发展趋势，并概括了我国珊瑚礁遥感的机遇和挑战。同时，项目组取得了如下的研究成果：1）提出涠洲岛珊瑚礁水深反演的方法和活珊瑚覆盖度（LCC）遥感反演的模型，为珊瑚礁生态遥感提供全新的思路与视角；2）实现了珊瑚礁地貌和底质类型的遥感分类，并评估了国产卫星遥感影像在珊瑚礁生态与地貌遥感监测中的应用潜力；3）提出遥感分析珊瑚与环境关系的新方法，发现涠洲岛LCC的下降很可能与人类活动引起的水体富营养化密切相关；4）在鹿回头开展长时间系列LCC遥感反演的评估，提供解决遥感与生态学者缺乏沟通与融合问题的基础；5）利用遥感探测了南海岛、礁地貌近30年的变化并分析了其对气候变暖的响应，解决南海珊瑚礁缺乏可靠长时间系列地貌变化分析数据的问题，分析了南海环礁和灰沙岛在气候变暖背景下的变化趋势并推测其在气候变暖条件下未来可能的发展。这些研究成果促进了遥感监测与珊瑚礁生态学得的交叉与融合，可以用于对南海乃至全球珊瑚礁生态状况的有效监测，为未来更广阔的南海乃至全球珊瑚礁的生态遥感监测提供技术积累和经验基础。