Canadian oil and gas companies must reclaim and monitor inactive resource extraction sites for many years after operations are completed. Land reclamation efforts are routinely evaluated through manual vegetation surveys.
To capture spatial and temporal variability across large areas, ecological surveyors set up and inspect numerous small plots, which can take a small team an entire season to survey.
The labor cost for these manual surveys is amplified by logistical difficulties, as the low-lying wetland terrain requires specialized low-ground disturbance vehicles or access via helicopter during field survey. Often, they end up sampling just a subset of all the sites due to budget and logistical constraints, making it challenging and expensive to make reliable decisions about the state of reclamation efforts.
Dr. Cassidy Rankine, from Rankine Geospatial shared:
The objective is to obtain detailed vegetation community composition inventories in these boreal peatlands — which are very difficult to access by foot — to enhance and/or replace the current ground plot survey work being done over very large areas in the boreal forest in Canada.
The Project
As a way of increasing productivity and lowering operation costs Rankine Geospatial has partnered with a local drone company to test out the MicaSense Altum on a DJI Matrice 210.
Rankine and the team at Osprey performed two separate UAV survey campaigns in July 2019 and October 2019, to capture spectral reflectance data across the different seasons that may indicate different layers of vegetation and changes in hydrology patterns.
Rankine added:
We used Altum because it offered the ability to obtain reliable spectral reflectance data with the fine spatial resolution required for standardized time series analysis — the Altum camera and the incident light sensor technology offered the best option for spatial, spectral, and temporal analysis for accurate and meaningful vegetation surveys in this kind of landscape.
Tens of thousands of images from dozens of flights were processed using Pix4D to produce sub-decimeter reflectance orthomosaics in 6 spectral bands of each site. The team also used ground control coordinates from a sub-centimeter RTK Wingtra survey orthomap that was processed in parallel to avoid the need to collect in-situ ground control points.
This high accuracy positioning of the data allowed for repeatable comparison of the orthomaps to accurately measure changes over time in the landscape vegetation at the fine spatial resolution required for peatland plant community analysis.
More work still needs to be done to prove the efficacy of this technology with further data analysis and additional data collection planned for 2020 for Dr. Rankine, Osprey Integrity, and the researchers at NAIT and C-Core. However, the preliminary results suggest this approach holds great promise for improving our capability to survey and assess the integrity of these carbon-rich landscapes for sustainable natural resource management in the great northern Boreal forests of Canada.
Source: MicaSense