资源环境科技发展态势分析平台

With forest fragmentation continuing in many parts of the tropics, it is likely that fragmented forests will become the rule for most remaining global tropical forests in the next decades. In this context, there is a need for practical tools to assess and monitor fragmented forests if we are to conserve tree species diversity as much as possible. Several methods, either using (1) field-based or (2) remote-sensing approaches could be used to achieve this goal. This paper aims at providing a state-of-the-art review of both approaches in order to make recommendations for rapid and cost-effective assessments of tree species diversity and forest structure, with specific applications in studies of fragmented tropical forests. Overall both methods have pros and cons, depending on the type of data needed to address the research objectives and the quantity of resources available. We suggest combining the use of both field-based and remotely-sensed methods as they can be complementary. Remote-sensing data should be used to predict and map the tree species diversity and stand structure at regional scales, while field-inventories provide accurate information at local scales and allow validation of remotely sensed data. For field-inventories, we recommend the use of small (e.g. 20 x 50 m) plots with a 10 cm DBH minimum measurement. In terms of remotely-sensed techniques, if funding is sufficient, airborne imagery seems the best regarding the quality of information (i.e. hyperspectral and hyperspatial imagery, LiDAR). If funding is limited, a cost-effective alternative providing reasonably accurate estimates would be the use of high-resolution satellite imagery such as Worldview. Ultimately, for studies where data accessible for free is the only possible option, we recommend the use of Sentinel-2, although it is relatively coarser in terms of quality. However, further research needs to be done to validate these approaches in fragmented tropical forests.