Menu
NWSAR Committee was established to ensure local interests are reflected as part of species at risk, and more specifically, Woodland Caribou management. Among the recommendations generated by NWSAR Committee’s 2017 Recommendations Report, new data collection to support Woodland Caribou management was identified as a high priority by local, regional and national stakeholders.
To support the acquisition of new and more accurate data related to Woodland Caribou, NWSAR worked with the Alberta Biodiversity Monitoring Institute (ABMI), Manning Forest Products, Boucher Bros Lumber Ltd. and Zavisha Sawmills to develop the best available information on the status of caribou habitat within two caribou ranges.
This project primarily focused on the Chinchaga and Caribou Mountains caribou ranges within Northwest Alberta, and sought to address the following questions:
1. What is the accuracy of existing human footprint data?
2. What is the state of vegetation growth on vegetated footprint types?
3. Where should habitat restoration efforts on seismic lines be targeted?
Aerial Imagery was acquired at approximately 20 townships distributed across five areas using a high quality, large-frame aerial digital camera sensor, DMC III. Aerial photos were captured using five cameras, each for a separate spectral band (Panchromatic, Red, Green, Blue and Near-Infrared). The spatial resolution of the image acquisition was under 15cm, ensuring that the resulting ground sampling distance of the orthophoto mosaic will be created by resampling to 15cm.
The average flight altitude was 3,550 meters, and the images were captured with 80% of forward overlap to increase the number of instances per each subject point during the point cloud generation process.
The five areas were selected using the following criteria:
Ground-truthing protocols were developed to ensure comparability between the photogrammetry and ground-truthing locations.
Ground-truthing was conducted at 120 locations encompassing a range of disturbance and vegetation classes distributed across the five blocks.
At each ground-truthing location, the following data was collected:
The ground-truthing sites were visited between September and December 2020, and basic field metadata was also recorded at each site (i.e. weather, date, and presence of natural disturbance).
Ground-truthing data was used to support vegetation inventory on seismic lines, to confirm human footprint interpretation during accuracy analyses, and to provide a coefficient of correction for tree height generated from point clouds.
The aerial imagery was processed to generate high-resolution multiband orthophoto mosaics and 3D point clouds for data extraction and analysis. The ultimate purpose of this processing was to generate high-resolution, three-dimensional data on ground cover, including information on vegetation height.
This involved a number of steps, including:
Orthophotos were used to verify the thematic accuracy of the most current publicly available ABMI human footprint dataset, the ABMI Wall-to-Wall Human Footprint Inventory, known as the HFI (2018).
The HFI (2018) contains 20 human footprint sublayers, based on 117 feature types. For example, a feature type “low impact seismic lines” is part of the seismic lines sublayer. ABMI compared the existing dataset to the orthophoto mosaics generated by this project and assigned each feature type to one of the following accuracy categories:
The state of vegetation on successional human footprint was assessed using two different approaches:
Areas for habitat restoration were prioritized within the Bistcho, Caribou Mountains, Chinchaga and Yates caribou ranges. ABMI mapped the current habitat disturbance and simulated the reduction in disturbance following habitat restoration. Townships were used to represent a scale whereby efficient groupings and economy of scale for restoration can be achieved.
ABMI used two scenarios:
The gain in undisturbed habitat is calculated by assuming that all conventional seismic lines and trails are restored. ABMI then subtracts the percent disturbance after all treatable features are restored from the current percent of altered habitat. This step identifies pixels that offer the highest potential to gain undisturbed habitat.
ABMI assessed the benefit to cost ratio of habitat restoration in each township (i.e. “Bang-for-Buck”), by calculating the reduction in percent disturbance and the effort, or cost, required to achieve this result.
Townships are ranked from highest to lowest “Bang-for-Buck” and grouped similarly ranked townships into five hierarchical zones of ordered priority for habitat restoration, so that there is an equal number of townships in each priority zone.
Lowest priority zones included townships with no potential benefits from habitat restoration, either because there were no treatable features within a pixel, or because all treatable areas fell within other disturbances.
The overall thematic accuracy of the Human Footprint Inventory (HFI) (2018) was 93.44%; however, accuracy varied across footprint types. Wells and harvest areas had the highest accuracy within each block, with overall accuracies of 96.09% for wells and 94.51% for harvest areas.
The overall accuracy of seismic lines was 95.22%, and “other human footprint” had the lowest accuracy. The most common source of error was “FEATURE TYPE INCORRECT, SUBLAYER CORRECT” where the detailed feature type identification was incorrect, particularly for seismic lines.
Overall. the percentage of human footprint that was either missed or not in fact human footprint was low, indicating a high level of accuracy in the ability of the HFI to capture human footprint.
This project was funded by NWSAR Committee, Manning Forest Products, Boucher Bros Lumber Ltd. and Zavisha Sawmills through the use of Forest Resource Improvement Association of Alberta (FRIAA) 2021 Forest Resource Improvement Program (FRIP) funding.