sgsR: a structurally guided sampling toolbox for LiDAR-based forest inventories

Establishing field inventories can be labor intensive, logistically challenging and expensive. Optimizing a sample to derive accurate forest attribute predictions is a key management-level inventory objective. Traditional sampling designs involving pre-defined, interpreted strata could result in poor selection of within-strata sampling intensities, leading to inaccurate estimates of forest structural variables. The use of airborne laser scanning (ALS) data as an applied forest inventory tool continues to improve understanding of the composition and spatial distribution of vegetation structure across forested landscapes. The increased availability of wall-to-wall ALS data is promoting the concept of structurally guided sampling (SGS), where ALS metrics are used as an auxiliary data source driving stratification and sampling within management-level forest inventories. In this manuscript, we present an open-source R package named sgsR that provides a robust toolbox for implementing various SGS approaches. The goal of this package is to provide a toolkit to facilitate better optimized allocation of sample units and sample size, as well as to assess and augment existing plot networks by accounting for current forest structural conditions. Here, we first provide justification for SGS approaches and the creation of the sgsR toolbox. We then briefly describe key functions and workflows the package offers and provide two reproducible examples. Avenues to implement SGS protocols according to auxiliary data needs are presented.

Automated summary

Establishing field inventories can be challenging and expensive. The use of airborne laser scanning (ALS) data as a forest inventory tool is improving understanding of vegetation structure across forested landscapes. The sgsR package provides a toolbox for implementing structurally guided sampling (SGS) and optimizing allocation of sample units and sample size.