Scale dependency of lidar‐derived forest structural diversity

Lidar-derived forest structural diversity (FSD) metrics—including measures of canopy height, vegetation arrangement, cover (CC), complexity and leaf area density—are increasingly used to describe characteristics can be infer many ecosystem functions. Despite broad adoption, the importance spatial resolution (grain extent) over which these metrics are calculated remains largely unconsidered. Often researchers will quantify FSD at grain size process interest without considering scale dependency or statistical behaviour metric employed. We investigated appropriate inference for eight lidar-derived metrics—CC, relief ratio, foliar height diversity, index, mean median outer rugosity (RT)--representing five categories—canopy CC, density, complexity. Optimal was determined using representative elementary (REA) concept whereby REA is smallest extent. Structural were increasing (from 5 1000 m) from aerial lidar data collected nine different forested ecosystems including sub-boreal, broadleaf temperate, needleleaf dry tropical, woodland savanna systems, all sites part National Ecological Observatory Network within conterminous United States. To identify each metric, we changepoint analysis via segmented piecewise regression identifies significant changepoints both magnitude variance metric. find that a between 25 75 m sufficiently captures across multiple types 30–150 metrics. However, differences evident among with higher necessary characterize CC in evergreen forests, deciduous broadleaved forests. These findings indicate range sizes inferences drawn this set metrics, informing use research management applications.