With an increase in the frequency and severity of wildfires across the globe and resultant changes to long-established fire regimes, the mapping of fire severity is a vital part of monitoring ecosystem resilience and recovery. The emergence of unoccupied aircraft systems (UAS) and compact sensors (RGB and LiDAR) provide new opportunities to map fire severity. This paper conducts a comparison of metrics derived from UAS Light Detecting and Ranging (LiDAR) point clouds and UAS image based products to classify fire severity. A workflow which derives novel metrics describing vegetation structure and fire severity from UAS remote sensing data is developed that fully utilises the vegetation information available in both data sources. UAS imagery and LiDAR data were captured pre- and post-fire over a 300 $\mathrm{m}$ by 300 $\mathrm{m}$ study area in Tasmania, Australia. The study area featured a vegetation gradient from sedgeland vegetation (e.g., button grass $0.2$$\mathrm{m}$) to forest (e.g., Eucalyptus obliqua and Eucalyptus globulus 50$\mathrm{m}$). To classify the vegetation and fire severity, a comprehensive set of variables describing structural, textural and spectral characteristics were gathered using UAS images and UAS LiDAR datasets. A recursive feature elimination process was used to highlight the subsets of variables to be included in random forest classifiers. The classifier was then used to map vegetation and severity across the study area. The results indicate that UAS LiDAR provided similar overall accuracy to UAS image and combined (UAS LiDAR and UAS image predictor values) data streams to classify vegetation (UAS image: 80.6%; UAS LiDAR: 78.9%; and Combined: 83.1%) and severity in areas of forest (UAS image: 76.6%, UAS LiDAR: 74.5%; and Combined: 78.5%) and areas of sedgeland (UAS image: 72.4%; UAS LiDAR: 75.2%; and Combined: 76.6%). These results indicate that UAS SfM and LiDAR point clouds can be used to assess fire severity at very high spatial resolution.

Point clouds captured from Unmanned Aerial Systems are increasingly relied upon to provide information describing the structure of forests. The quality of the information derived from these point clouds is dependent on a range of variables, including the type and structure of the forest, weather conditions and flying parameters. A key requirement to achieve accurate estimates of height based metrics describing forest structure is a source of ground information. This study explores the availability and reliability of ground surface points available within point clouds captured in six forests of different structure (canopy cover and height), using three image capture and processing strategies, consisting of nadir, oblique and composite nadir/oblique image networks. The ground information was extracted through manual segmentation of the point clouds as well as through the use of two commonly used ground filters, LAStools lasground and the Cloth Simulation Filter. The outcomes of these strategies were assessed against ground control captured with a Total Station. Results indicate that a small increase in the number of ground points captured (between 0 and 5% of a 10 m radius plot) can be achieved through the use of a composite image network. In the case of manually identified ground points, this reduced the root mean square error (RMSE) error of the terrain model by between 1 and 11 cm, with greater reductions seen in plots with high canopy cover. The ground filters trialled were not able to exploit the extra information in the point clouds and inconsistent results in terrain RMSE were obtained across the various plots and imaging network configurations. The use of a composite network also provided greater penetration into the canopy, which is likely to improve the representation of mid-canopy elements.

Fuel hazard estimates are vital for the prediction of fire behaviour and planning fuel treatment activities. Previous literature has highlighted the potential of Terrestrial Laser Scanning (TLS) to be used to assess fuel properties. However, operational uptake of these systems has been limited due to a lack of a sampling approach that balances efficiency and data efficacy. This study aims to assess whether an operational approach utilising Terrestrial Laser Scanning (TLS) to capture fuel information over an area commensurate with current fuel hazard assessment protocols implemented in South-Eastern Australia is feasible. TLS data were captured over various plots in South-Eastern Australia, utilising both low- and high-cost TLS sensors. Results indicate that both scanners provided similar overall representation of the ground, vertical distribution of vegetation and fuel hazard estimates. The analysis of fuel information contained within individual scans clipped to 4 m showed similar results to that of the fully co-registered plot (cover estimates of near-surface vegetation were within 10%, elevated vegetation within 15%, and height estimates of near-surface and elevated strata within 0.05 cm). This study recommends that, to capture a plot in an operational environment (balancing efficiency and data completeness), a sufficient number of non-overlapping individual scans can provide reliable estimates of fuel information at the near-surface and elevated strata, without the need for co-registration in the case study environments. The use of TLS within the rigid structure provided by current fuel observation protocols provides incremental benefit to the measurement of fuel hazard. Future research should leverage the full capability of TLS data and combine it with moisture estimates to gain a full realisation of the fuel hazard.

Characteristics describing below canopy vegetation are important for a range of forest ecosystem applications including wildlife habitat, fuel hazard and fire behaviour modelling, understanding forest recovery after disturbance and competition dynamics. Such applications all rely on accurate measures of vegetation structure. Inherent in this is the assumption or ability to demonstrate measurement accuracy. 3D point clouds are being increasingly used to describe vegetated environments, however limited research has been conducted to validate the information content of terrestrial point clouds of understory vegetation. This paper describes the design and use of a field frame to co-register point intercept measurements with point cloud data to act as a validation source. Validation results show high correlation of point matching in forests with understory vegetation elements with large mass and/or surface area, typically consisting of broad leaves, twigs and bark 0.02 m diameter or greater in size (SfM, MCC 0.51–0.66; TLS, MCC 0.37–0.47). In contrast, complex environments with understory vegetation elements with low mass and low surface area showed lower correlations between validation measurements and point clouds (SfM, MCC 0.40 and 0.42; TLS, MCC 0.25 and 0.16). The results of this study demonstrate that the validation frame provides a suitable method for comparing the relative performance of different point cloud generation processes.