Search

Burn severity is an important component of the fire regime that has not yet been fully characterized for the forests of Canada. The objectives of this study were to (i) create a Canada-wide geospatial database of burn severity for wildland fires across forested regions of Canada from 1985 to 2015, and (ii) use this database to evaluate seasonal and annual trends in burn severity across Canada and regionally using two different regional units (ecozones and Homogeneous Fire Regime zones). We developed the 30 m resolution geospatial Canadian Landsat Burn Severity (CanLaBS) product from Landsat imagery, using values of pre-fire to post-fire differences in normalized burn ratios (dNBRs) for nearly 60 Mha of area burned across Canada’s forests from 1985 to 2015, complemented with estimates of pre-fire forest composition, crown closure, and biomass. Our results suggest that burn severity is generally lower in spring fires than in summer ones nationally and in almost every regional unit. We found no consistent relationship between burn severity and annual area burned across ecozones. Finally, we observed a small but significant decrease in burn severity from 1985 to 2015 across Canada, although this is regionally variable. The CanLaBS database is publicly available at https://doi.org/10.23687/b1f61b7e-4ba6-4244-bc79-c1174f2f92cd .

Satellite forest inventories are the only feasible way to map Canada’s vast, remote forest regions, such as those in the Northwest Territories (NWT). A method used to create such inventories is the k-nearest neighbour (k-NN) algorithm, which spatially extends information from forest inventory (FI) plots to the entire forest land base using wall-to-wall features typically derived from Landsat data. However, the benefits of integrating L-band synthetic aperture radar (SAR) data, strongly correlated to forest biomass, have not been assessed for Canadian northern boreal forests. Here we describe an optimized multivariate k-NN implementation of a 151,700 km2 area in southern NWT that included ca. 2007 Landsat and dual-polarized Phased Array type L-band SAR (PALSAR) data on board the Advanced Land Observing Satellite (ALOS). Five forest attributes were mapped at 30 m cells: stand height, crown closure, stand/total volume and aboveground biomass (AGB). We assessed accuracy gains compared to Landsat-based maps. To circumvent the scarcity of FI plots, we used 3600 footprints from the Geoscience Laser Altimeter System (GLAS) as surrogate FI plots, where forest attributes were estimated using Light Detection and Ranging (LiDAR) metrics as predictors. After optimization, k-NN predicted forest attribute values for each pixel as the average of the 4 nearest (k = 4) surrogate FI plots within the Euclidian space of 9 best features (selected among 6 PALSAR, 10 Landsat, and 6 environmental features). Accuracy comparisons were based on 31 National Forest Inventory ground plots and over 1 million airborne LiDAR plots. Maps that included PALSAR HV backscatter resulted in forest attribute predictions with higher goodness of fit (adj. R2), lower percent mean error (ME%), and percent root mean square error (RMSE%), and lower underestimation for larger attribute values. Predictions were most accurate for conifer stand height (RMSE% = 32.1%, adj. R2 = 0.58) and AGB (RMSE% = 47.8%, adj. R2 = 0.74), which is much more abundant in the area than mixedwood or broadleaf. Our study demonstrates that optimizing k-NN parameters and feature space, including PALSAR, Landsat, and environmental variables, is a viable approach for inventory mapping of the northern boreal forest regions of Canada.

Large-scale mapping of soil properties is increasingly important for environmental resource management. While forested areas play critical environmental roles at local and global scales, forest soil maps are typically at low resolution. The objective of this study was to generate continuous national maps of selected soil variables (C, N and soil texture) for the Canadian managed forest landbase at 250 m resolution. We produced these maps using the kNN method with a training dataset of 538 ground-plots from the National Forest Inventory (NFI) across Canada, and 18 environmental predictor variables. The best predictor variables were selected (7 topographic and 5 climatic variables) using the Least Absolute Shrinkage and Selection Operator method. On average, for all soil variables, topographic predictors explained 37% of the total variance versus 64% for the climatic predictors. The relative root mean square error (RMSE%) calculated with the leave-one-out cross-validation method gave values ranging between 22% and 99%, depending on the soil variables tested. RMSE values