Search

Post-wildfire vegetation cover damage and loss can escalate the risks of secondary disasters such as flood, landslide, and water contamination, particularly in a major wildfire affected region where human settlements are situated. In assessments of the secondary disaster risks, the post-wildfire vegetation cover change plays a key role in influencing the distribution and intensity of the risks. In this work, a processing framework for mapping post-wildfire vegetation cover changes through information fusion has been generated and tested using Landsat8 and WorldView imagery data. The test site was the boreal forest region surrounding Fort McMurray, Alberta, Canada, affected by a massive wildfire in May 2016. The derived map results indicate that the fusion process in the framework is effective for generation of post-wildfire vegetation cover change information. The use of WorldView data revealed more variation details in distribution of the vegetation cover burn damages than use of Landsat data. Moreover, the uncertainty in vegetation burn severity using Landsat-based Differenced Normalized Burn Ratio (dNBR) index exists in the areas with low dNBR reading values due to the sub-pixel effect. The forest burn severity measured with dNBR index can be underestimated due to the quick herbaceous cover recovery after wildfire. These uncertainties in the post-wildfire vegetation cover mapping should be taken into consideration when the derived information is being used for risk assessments.

High-quality ground-truth data are critical for developing reliable Earth Observation (EO) based geospatial products. Conventional methods of collecting these data are either subject to an unknown amount of human error and bias or require extended time in the field to complete (i.e., point-intercept assessments). Digital photograph classification (DPC) may address these drawbacks. In this study, we first assess the performance of a DPC method developed through licensed software to estimate ground cover percentage (%) of bright lichens, a critical caribou forage in fall and winter when other food resources are scarce. We then evaluate the feasibility of replicating this workflow in an open-source environment with a modified U-net model to improve processing time and scalability. Our results indicate that DPC is appropriate for generating ground-truth data in support of large-scale EO-based lichen mapping within the boreal forests of eastern Canada. Our final open-sourced classification model, Lichen Convolutional Neural Network (LiCNN), is comparably accurate yet more efficient than the licensed workflow. Therefore, the LiCNN approach successfully addresses the mentioned shortcomings of conventional ground-truth data collection methods efficiently and without the need for specialized software.

Lichen mapping is vital for caribou management plans and sustainable land conservation. Previous studies have used random forest, dense neural network, and convolutional neural network models for mapping lichen coverage. However, to date, it is not clear how these models rank in this task. In this study, these machine learning models were evaluated on their ability to predict lichen percent coverage in Sentinel-2 imagery in Québec and Labrador, Canada. The models were trained on 10-m resolution lichen coverage (%) maps created from 20 drone surveys collected in July 2019 and 2022. The dense neural network achieved a higher accuracy than the other two, with a reported mean absolute error of 5.2% and an R2 of 0.76. By comparison, the random forest model returned a mean absolute error of 5.5% (R2: 0.74) and the convolutional neural network had a mean absolute error of 5.3% (R2: 0.74). A regional lichen map was created using the trained dense neural network and a Sentinel-2 imagery mosaic. There was greater uncertainty on land covers that the model was not exposed to in training, such as mines and deep lakes. While the dense neural network requires more computational effort to train than a random forest model, the 5.9% performance gain in the test pixel comparison renders it the most suitable for lichen mapping. This study represents progress toward determining the appropriate methodology for generating accurate lichen maps from satellite imagery for caribou conservation and sustainable land management.

Estimating the number of trees within a forest stand, i.e., the forest stand density (FSD), is challenging at large scales. Recently, researchers have turned to a combination of remote sensing and machine learning techniques to derive these estimates. However, in most cases, the developed models rely heavily upon additional data such as LiDAR-based elevations or multispectral information and are mostly applied to managed environments rather than natural/mixed forests. Furthermore, they often require the time-consuming manual digitization or masking of target features, or an annotation using a bounding box rather than a simple point annotation. Here, we introduce the Tree Convolutional Row Neural Network (Tree-CRowNN), an alternative model for tree counting inspired by Multiple-Column Neural Network architecture to estimate the FSD over 12.8 m × 12.8 m plots from high-resolution RGB aerial imagery. Our model predicts the FSD with very high accuracy (MAE: ±2.1 stems/12.8 m2, RMSE: 3.0) over a range of forest conditions and shows promise in linking to Sentinel-2 imagery for broad-scale mapping (R2: 0.43, RMSE: 3.9 stems/12.8 m2). We believe that the satellite imagery linkage will be strengthened with future efforts, and transfer learning will enable the Tree-CRowNN model to predict the FSD accurately in other ecozones.

Human disturbance of peatland environments due to geological exploration activities can significantly alter peatland greenhouse gas exchange. Here, the authors study peatland disturbance in Alberta, Canada, and estimate that induced methane emissions are approximately 4.4–5.1 kt higher per year than in undisturbed conditions.

Illumination variations in non-atmospherically corrected high-resolution satellite (HRS) images acquired at different dates/times/locations pose a major challenge for large-area environmental mapping and monitoring. This problem is exacerbated in cases where a classification model is trained only on one image (and often limited training data) but applied to other scenes without collecting additional samples from these new images. In this research, by focusing on caribou lichen mapping, we evaluated the potential of using conditional Generative Adversarial Networks (cGANs) for the normalization of WorldView-2 (WV2) images of one area to a source WV2 image of another area on which a lichen detector model was trained. In this regard, we considered an extreme case where the classifier was not fine-tuned on the normalized images. We tested two main scenarios to normalize four target WV2 images to a source 50 cm pansharpened WV2 image: (1) normalizing based only on the WV2 panchromatic band, and (2) normalizing based on the WV2 panchromatic band and Sentinel-2 surface reflectance (SR) imagery. Our experiments showed that normalizing even based only on the WV2 panchromatic band led to a significant lichen-detection accuracy improvement compared to the use of original pansharpened target images. However, we found that conditioning the cGAN on both the WV2 panchromatic band and auxiliary information (in this case, Sentinel-2 SR imagery) further improved normalization and the subsequent classification results due to adding a more invariant source of information. Our experiments showed that, using only the panchromatic band, F1-score values ranged from 54% to 88%, while using the fused panchromatic and SR, F1-score values ranged from 75% to 91%.

Relating ground photographs to UAV orthomosaics is a key linkage required for accurate multi-scaled lichen mapping. Conventional methods of multi-scaled lichen mapping, such as random forest models and convolutional neural networks, heavily rely on pixel DN values for classification. However, the limited spectral range of ground photos requires additional characteristics to differentiate lichen from spectrally similar objects, such as bright logs. By applying a neural network to tiles of a UAV orthomosaics, additional characteristics, such as surface texture and spatial patterns, can be used for inferences. Our methodology used a neural network (UAV LiCNN) trained on ground photo mosaics to predict lichen in UAV orthomosaic tiles. The UAV LiCNN achieved mean user and producer accuracies of 85.84% and 92.93%, respectively, in the high lichen class across eight different orthomosaics. We compared the known lichen percentages found in 77 vegetation microplots with the predicted lichen percentage calculated from the UAV LiCNN, resulting in a R2 relationship of 0.6910. This research shows that AI models trained on ground photographs effectively classify lichen in UAV orthomosaics. Limiting factors include the misclassification of spectrally similar objects to lichen in the RGB bands and dark shadows cast by vegetation.

Lichen is an important food source for caribou in Canada. Lichen mapping using remote sensing (RS) images could be a challenging task, however, as lichens generally appear in unevenly distributed, small patches, and could resemble surficial features. Moreover, collecting lichen labeled data (reference data) is expensive, which restricts the application of many robust supervised classification models that generally demand a large quantity of labeled data. The goal of this study was to investigate the potential of using a very-high-spatial resolution (1-cm) lichen map of a small sample site (e.g., generated based on a single UAV scene and using field data) to train a subsequent classifier to map caribou lichen over a much larger area (~0.04 km2 vs. ~195 km2) and a lower spatial resolution image (in this case, a 50-cm WorldView-2 image). The limited labeled data from the sample site were also partially noisy due to spatial and temporal mismatching issues. For this, we deployed a recently proposed Teacher-Student semi-supervised learning (SSL) approach (based on U-Net and U-Net++ networks) involving unlabeled data to assist with improving the model performance. Our experiments showed that it was possible to scale-up the UAV-derived lichen map to the WorldView-2 scale with reasonable accuracy (overall accuracy of 85.28% and F1-socre of 84.38%) without collecting any samples directly in the WorldView-2 scene. We also found that our noisy labels were partially beneficial to the SSL robustness because they improved the false positive rate compared to the use of a cleaner training set directly collected within the same area in the WorldView-2 image. As a result, this research opens new insights into how current very high-resolution, small-scale caribou lichen maps can be used for generating more accurate large-scale caribou lichen maps from high-resolution satellite imagery.

Groundwater level (GWL) and depth to water (DTW) are related metrics aimed at characterizing groundwater-table positions in peatlands, and two of the most common variables collected by researchers working in these ecosystems. While well-established field techniques exist for measuring GWL and DTW, they are generally difficult to scale. In this study, we present a novel workflow for mapping groundwater using orthophotography and photogrammetric point clouds acquired from unmanned aerial vehicles. Our approach takes advantage of the fact that pockets of surface water are normally abundant in peatlands, which we assume to be reflective of GWL in these porous, gently sloping environments. By first classifying surface water and then extracting a sample of water elevations, we can generate continuous models of GWL through interpolation. Estimates of DTW can then be obtained through additional efforts to characterize terrain. We demonstrate our methodology across a complex, 61-ha treed bog in northern Alberta, Canada. An independent accuracy assessment using 31 temporally coincident water-well measurements revealed accuracies (root mean square error) in the 20-cm range, though errors were concentrated in small upland pockets in the study area, and areas of dense tree covers. Model estimates in the open peatland areas were considerably better.

Microtopographic variability in peatlands has a strong influence on greenhouse gas fluxes, but we lack the ability to characterize terrain in these environments efficiently over large areas. To address this, we assessed the capacity of photogrammetric data acquired from an unmanned aerial vehicle (UAV or drone) to reproduce ground elevations measured in the field. In particular, we set out to evaluate the role of (i) vegetation/surface complexity and (ii) supplementary LiDAR data on results. We compared remote-sensing observations to reference measurements acquired with survey grade GPS equipment at 678 sample points, distributed across a 61-hectare treed bog in northwestern Alberta, Canada. UAV photogrammetric data were found to capture elevation with accuracies, by root mean squares error, ranging from 14–42 cm, depending on the state of vegetation/surface complexity. We judge the technology to perform well under all but the most-complex conditions, where ground visibility is hindered by thick vegetation. Supplementary LiDAR data did not improve results significantly, nor did it perform well as a stand-alone technology at the low densities typically available to researchers.