A simplified and affordable approach to forest monitoring using single terrestrial laser scans and transect sampling

Abstract

Traditional forestry, ecology, and fuels monitoring methods can be costly and error-prone, and are often used beyond their original assumptions due to difficulty or unavailability of more appropriate methods. These traditional methods tend to be rigid and may not be useful for detecting new ecological changes or required data at modern levels of precision [1] . The integration of Terrestrial Laser Scanning (TLS) methods into forest monitoring strategies can cost effectively standardize data collection, improve efficiency, and reduce error, with datasets that can easily be analyzed to better inform management decisions. Affordable (sub-$20K) off-the-shelf TLS units—such as the Leica BLK360— have been used commercially in the built environment but have untapped potential in the natural world for monitoring. Here, we provide a methodology that successfully integrates LiDAR scanning with existing monitoring methods. This new method: Allows for simplified and quick extraction of forestry, fuels and ecological vegetation variables from a single TLS point cloud and quick transect sampling. Streamlines the data collection process, removes sampling bias, and produces data that can be easily processed to provide inputs for models and decision support frameworks. Is adaptable to integrate additional or new environmental measurements.

Estimation of Plot-Level Burn Severity Using Terrestrial Laser Scanning

“…Our study provides a replicable workflow and numerous variables relating to forest structure and true color RGB values that can be used as alternative criteria for refencing burn severity (e.g., dNBR or other indices) consistently in any landscape for both prescribed fire and wildfire. The TLS approach presented is beneficial in that it is fast (~4 min to scan), consistent, repeatable, provides a far more detailed depiction of forest conditions than CBI, is relatively inexpensive with regard to the cost of current forest monitoring programs, and data can also be used to inventory fuels with greater richness and speed than traditional inventorying methods [19]. In contrast to the typical 23 indirectly observed variables that are integrated into CBI calculations, the 70 TLS-based variables and 10 RGBbased variables presented in this study demonstrate how our approach can be used to characterize forest structure and spectral conditions.…”

“…Additionally, future research can utilize the structural and reflectance variables presented in the manuscript to predict dNBR and investigate the potential for TLS and RGB-based burn severity data to replace CBI. This would be extremely beneficial in enhancing manager capacity to monitor fire treatment and utilize monitoring data to inform management and can be incorporated with TLS-based fuels monitoring strategies [19,29].…”

“…Forty-three plots were randomly selected with the criteria that plots be at least 400 m apart from each other and 100 m from roads to facilitate separation of plots and prevent potential edge effects (Figure 1). Plots were marked in the field prior to burning, at which time pre-burn scans were collected with a BLK360 TLS (Leica Geosystems, Heerbrugg, Switzerland) following [19] as single return, single scans. GPS points at plot center with >2 m accuracy were collected at each plot using a Trimble GeoExplorer 6000 paired to a Tornado receiver (Trimble Inc., Sunnyvale, CA, USA).…”

“…TLS excels in collecting structure and density data in understory and midstory strata where ALS is weakest due to canopy occlusion and can be conducted quickly with minimal training and no expertise in remote sensing or fire effects and thus offers a potential alternative to CBI as a consistent, repeatable, field-based approach to creating rich threedimensional plot-level datasets of vegetation structure and density [19,20]. Kato et al [20] used linear and non-linear regression to compare dNBR data with voxelized TLS data from burned plots and found that correlations were strongest when using all data (R 2 = 0.73-0.75) but became weaker when data were binned according to CBI correlations with dNBR (R 2 = 0.05-0.63).…”

“…More recently, TLS has been used to map fuels in three dimensions [29] and changes in forest structure following wildfire [30]. Additionally, because some TLS systems collect spectral reflectance data that are appended to each return, there is potential to integrate a spectral-reflectance index with the three-dimensional data for consistent fine-scale burn severity mapping [19,31]. With costs of TLS units becoming increasingly affordable to the typical forest manager, TLSbased inventorying and monitoring methods are well poised to revolutionize monitoring efforts as data extraction and analysis approaches continue to be refined [32].…”