Department of Fish and Wildlife Resources, University of Idaho, Box 441136, Moscow, Idaho 83844-1136, USA.
Ecol Appl. 2011 Mar;21(2):577-88. doi: 10.1890/09-2155.1.
LiDAR remote sensing has been used to examine relationships between vertebrate diversity and environmental characteristics, but its application to invertebrates has been limited. Our objectives were to determine whether LiDAR-derived variables could be used to accurately describe single-species distributions and community characteristics of spiders in remote forested and mountainous terrain. We collected over 5300 spiders across multiple transects in the Bavarian National Park (Germany) using pitfall traps. We examined spider community characteristics (species richness, the Shannon index, the Simpson index, community composition, mean body size, and abundance) and single-species distribution and abundance with LiDAR variables and ground-based measurements. We used the R2 and partial R2 provided by variance partitioning to evaluate the predictive power of LiDAR-derived variables compared to ground measurements for each of the community characteristics. The total adjusted R2 for species richness, the Shannon index, community species composition, and body size had a range of 25-57%. LiDAR variables and ground measurements both contributed >80% to the total predictive power. For species composition, the explained variance was approximately 32%, which was significantly greater than expected by chance. The predictive power of LiDAR-derived variables was comparable or superior to that of the ground-based variables for examinations of single-species distributions, and it explained up to 55% of the variance. The predictability of species distributions was higher for species that had strong associations with shade in open-forest habitats, and this niche position has been well documented across the European continent for spider species. The similar statistical performance between LiDAR and ground-based measures at our field sites indicated that deriving spider community and species distribution information using LiDAR data can provide not only high predictive power at relatively low cost, but may also allow unprecedented mapping of community- and species-level spider information at scales ranging from stands to landscapes. Therefore, LiDAR is a viable tool to assist species-specific conservation as well as broader biodiversity planning efforts not only for a growing list of vertebrates, but for invertebrates as well.
激光雷达遥感已被用于研究脊椎动物多样性与环境特征之间的关系，但在无脊椎动物中的应用有限。我们的目标是确定激光雷达衍生变量是否可用于准确描述偏远森林和山区蜘蛛的单一物种分布和群落特征。我们在德国巴伐利亚国家公园（Bavarian National Park）使用陷阱收集了超过 5300 只蜘蛛，这些蜘蛛分布在多个样带上。我们使用激光雷达变量和地面测量值来检查蜘蛛群落特征（物种丰富度、香农指数、辛普森指数、群落组成、平均体型和丰度）以及单一物种的分布和丰度。我们使用方差分解提供的 R2 和偏 R2 来评估与地面测量值相比，激光雷达衍生变量对每个群落特征的预测能力。物种丰富度、香农指数、群落物种组成和体型的总调整 R2 范围为 25-57%。激光雷达变量和地面测量值都为总预测能力贡献了>80%。对于物种组成，解释的方差约为 32%，明显高于随机预期。对于物种分布的检查，激光雷达衍生变量的预测能力与地面变量相当或优于，可解释高达 55%的方差。在开阔森林栖息地中与阴影有强烈关联的物种的预测能力较高，并且这种生态位在整个欧洲大陆的蜘蛛物种中都有很好的记录。我们实地调查中激光雷达和地面测量值的相似统计性能表明，使用激光雷达数据可以不仅提供高预测能力，而且成本相对较低，还可以对从林分尺度到景观尺度的群落和物种水平的蜘蛛信息进行前所未有的绘图。因此，激光雷达是一种可行的工具，可以帮助进行特定物种的保护以及更广泛的生物多样性规划工作，不仅适用于越来越多的脊椎动物，也适用于无脊椎动物。
