May the forest be with you: leveraging GEDI’s spaceborne lidar data for tropical ecosystem applications

Here, we provide an overview of the use of light detection and ranging (lidar) for tropical ecosystem applications, with a particular focus on the Global Ecosystem Dynamics Investigation (GEDI). We summarize how data from GEDI measures vegetation vertical structure and give a step-by-step description of how to obtain spatially-subset GEDI Level 2A data from the NASA EarthData Search web portal. We then provide an example of how to characterize the structure of various vegetation classes in Ucayali, Peru. These vegetation classes include: (1) old-growth lowland forest, (2) young lowland vegetation regrowth (‘Purma’)”, (3) secondary lowland forest, (4) mature oil palm plantations, and (5) cacao plantations (monocrop and agroforestry). We interpret the structural height metrics from GEDI among each of these vegetation classes, identifying edge effects as a possible influence on our results. To address this issue, we conducted a final analysis of the data with an area of 35m diameter footprint (25m of the original diameter area of the beam, and 10m as a conservative additional buffer) and excluded any observations that did not completely overlap with each land cover polygon. When we removed edge effects, no observations remained in the cacao data set and fewer observations remained in the forest stage data set. Nonetheless, the overall structural patterns shown in the relative heights of each forest stage remained very similar. We recommend that future projects utilizing spaceborne lidar for tropical ecosystems consider adopting the techniques and best practices we describe here, including refined noise filtering and explicit consideration of edge effects.

本文概述了光检测与测距（light detection and ranging，lidar）在热带生态系统研究中的应用，重点关注全球生态系统动态调查（Global Ecosystem Dynamics Investigation，GEDI）项目。本文首先梳理了GEDI数据如何反演植被垂直结构，并分步说明了如何从NASA EarthData Search门户网站获取空间裁剪后的GEDI 2A级数据。随后以秘鲁乌卡亚利地区为例，展示了如何表征不同植被类型的结构特征。本次研究涉及的植被类型包括：(1) 原始低地森林；(2) 年轻低地植被恢复地（‘Purma’）；(3) 次生低地森林；(4) 成熟油棕种植园；(5) 可可种植园（包括单作与农林复合两种模式）。本文针对上述各类植被类型，解析了GEDI获取的结构高度指标，并指出边缘效应可能是影响研究结果的潜在因素。为解决该问题，本研究采用直径35米的足迹范围（其中25米为激光束原始直径覆盖区域，额外增加10米作为保守缓冲区）对数据进行最终分析，并剔除了未与各土地覆盖多边形完全重合的观测样本。剔除边缘效应后，可可种植园的观测样本完全消失，森林演替阶段的观测样本也有所减少。尽管如此，各森林阶段相对高度所呈现的整体结构模式仍基本保持一致。最后，本文建议未来将星载激光雷达应用于热带生态系统研究的项目，可参考本文所述的技术方法与最佳实践，包括精细化噪声滤波以及明确考量边缘效应的影响。