Highlights

• •

  DAP and lidar were evaluated for modeling fire effects.

• •

  DAP augmented with Sentinel-2 approached the performance of lidar.

• •

  Lidar and DAP heights exhibit weaker correlations as lidar cover decreases.

• •

  DAP struggled in high severity burn areas missing crown foliage.

Abstract

Wildfires have the potential to profoundly alter forest structure, which can leave land managers with severe information gaps when making post-fire forest management decisions. We assess point clouds derived from aerial lidar (16.8 points/m²) and digital aerial photogrammetry (DAP, 13.9 points/m²) for the purposes of generating post-fire forest attribute maps using 82 permanent field plots. Low-cost aerial lidar and DAP data can enable predictive mapping after fires occur, and an assessment of the utility of DAP in this context is warranted given its lower cost. We evaluated the performance of models constructed from post-fire lidar and DAP data collected over a 158,000 ha area formerly dominated by Douglas-fir in Oregon, United States for four forest attributes: live aboveground biomass, standing dead aboveground biomass, live stem density, and relative basal area change. Both data sources were augmented with Sentinel-2 pre- and post-fire images. DAP models had relative RMSEs that were 10–47 % larger than lidar models. Augmenting the models with Sentinel-2 narrowed this error differential to 3–16 %. We found that the difference between lidar and DAP heights is explained by lidar cover and canopy intensity, suggesting that consumption of canopy fuels can adversely affect the quality of DAP heights. Our results indicate that DAP models augmented with Sentinel-2 present an attractive prospect for fire footprints lacking lidar data owing to DAP’s broader spatial and temporal availability; however, uncertainties may exist for some response variables in areas with high levels of canopy fuel consumption.