Jason Duffe

Development of boreal ecosystem anthropogenic disturbance layers for Canada based on 2008 to 2010 Landsat imagery

The generation of geospatial thematic information for managing and monitoring Canadas boreal ecosystem is essential for researchers, land managers, and policy makers. Canadas boreal region is a vast mosaic of forests, wetlands, rivers, and lakes, but anthropogenic disturbances have impacted these ecosystems resulting in habitat loss, fragmentation, and threats to biodiversity. Across Canada various geospatial datasets representing anthropogenic disturbance exist for timber harvesting, hydro-electric activity, settlement, and oil and gas activities; however, these products often vary in scale, attributes, time period, and mapping technique. Driven by the need for national data as part of the 2011 boreal caribou science assessment, a standardized methodology was developed and implemented to create a single geospatial dataset representing anthropogenic disturbances across a significant portion of Canadas boreal ecosystem. The boreal ecosystem anthropogenic disturbances data are a vector disturbance dataset of individual linear and polygonal disturbance types that were manually collected through the interpretation of 2008–2010 Landsat imagery at a 1:50000 viewing scale. Summary results identified a total polygonal anthropogenic disturbance footprint of approximately 24 million ha with forest cutblocks accounting for more than 60% of mapped polygonal disturbance. Linear disturbance features across the boreal total approximately 600000 km with roads and seismic exploration lines contributing to more than 80% of the mapped linear disturbances.

Land Cover Characterization and Classification of Arctic Tundra Environments by Means of Polarized Synthetic Aperture X- and C-Band Radar (PolSAR) and Landsat 8 Multispectral Imagery — Richards Island, Canada

In this work the potential of polarimetric Synthetic Aperture Radar (PolSAR) data of dual-polarized TerraSAR-X (HH/VV) and quad-polarized Radarsat-2 was examined in combination with multispectral Landsat 8 data for unsupervised and supervised classification of tundra land cover types of Richards Island, Canada. The classification accuracies as well as the backscatter and reflectance characteristics were analyzed using reference data collected during three field work campaigns and include in situ data and high resolution airborne photography. The optical data offered an acceptable initial accuracy for the land cover classification. The overall accuracy was increased by the combination of PolSAR and optical data and was up to 71% for unsupervised (Landsat 8 and TerraSAR-X) and up to 87% for supervised classification (Landsat 8 and Radarsat-2) for five tundra land cover types. The decomposition features of the dual and quad-polarized data showed a high sensitivity for the non-vegetated substrate (dominant surface scattering) and wetland vegetation (dominant double bounce and volume scattering). These classes had high potential to be automatically detected with unsupervised classification techniques.

Assessing the utility of lidar remote sensing technology to identify mule deer winter habitat

Winter habitat for mule deer (Odocoileus hemionus) is a critical concern throughout interior British Columbia, Canada. In winter, mule deer require a food source of twigs and woody browse and face significant winter snow cover. A range of studies have established that good winter range for mule deer reduces the impact of a negative energy balance by providing adequate food, good vegetative cover, and shallow snow. Generally, sites with old Douglas-fir and moderate to high canopy cover on warmer aspects and moderately steep slopes are preferred, resulting in a suite of structural stand conditions, which can be used to map mule deer winter range habitat within the interior Douglas-fir range. The increased availability of light detection and ranging (lidar) data to management agencies and the recent adoption of lidar technology by forestry agencies allow us to assess the capacity of this technology to map some variables important to winter mule deer habitat suitability, using criteria similar to those defined using conventional aerial photography. Results indicate that lidar-derived solar radiation regime, elevation, and overstorey cover are all useful attributes in decision-tree models relating lidar to conventionally derived descriptors of mule deer winter habitat. These lidar-derived models describe up to 75% of the variance in overall stand structure and confirm that this technology is a viable tool which can be used to assess habitat throughout this region.

A Comparative Analysis of Object-Based and Pixel-Based Classification of RADARSAT-2 C–Band and Optical Satellite Data for Mapping Shoreline Types in the Canadian Arctic

Abstract. With increasing ship traffic and natural resource extraction in the Canadian Arctic, the chance of an oil spill occurring is steadily increasing. Oil deposited in the marine environment will impact the ocean, and as it is washed onto shore many species and their habitats. In 2009, Environment Canada initiated a project to produce baseline coastal information required for operational prioritization in the event of a marine oil spill in the Canadian Arctic. Earth observation data can potentially be used to improve on traditional mapping methods. RADARSAT-2 C-band data and geotagged videography were collected for selected sites in the Canadian Arctic. Object-based classification was compared to pixel-based classification using two of these pilot sites, Richards Island (Northwest Territories) and Ivvavik (Yukon Territory). The object-based classification achieved an overall accuracy of 74% at Richards Island and 63% at Ivvavik, whereas the pixel-based reached 73% at both sites. To test transferability, the classifiers developed at Richards Island were applied to Ivvavik. In this case, the object-based classification achieved an overall accuracy of 78% and the pixel-based reached 71%. Results showed that characterizing shorelines in the Arctic using Earth observation data for oil spill cleanup response activities is possible. Résumé. Avec l’accroissement du trafic maritime et de l’exploitation des ressources naturelles dans l’Arctique canadien les risques d’un déversement d’hydrocarbure sont en augmentation. Les dépôts d’hydrocarbures auront un impact sur l’environnement marin. En 2009, Environnement Canada a initié un projet qui visait à améliorer la connaissance des types de rivage afin de prioriser les opérations de nettoyage dans l’éventualité d’un déversement d’hydrocarbure. Des vidéos géo-référencés et des images RADARSAT-2 ont été acquis pour plusieurs sites de l’Arctique canadien. Une classification orientée-objet a été comparée à une classification pixel-à-pixel pour deux de ces sites, soit l’île Richards (Territoires du Nord-Ouest) et Ivvavik (Yukon). La classification orientée-objet a atteint une précision globale de 74% à l’île Richards et de 63% à Ivvavik. La classification pixel-à-pixel a atteint une précision globale de 73% à chacun des sites. Les méthodes de classification développées à l’île Richards ont été appliquées à Ivvavik afin d’évaluer la transférabilité. La classification orientée-objet a atteint une précision globale de 78% et la classification pixel-à-pixel de 71%. Dans l’ensemble, les résultats ont démontrés que les types de rivage de l’Arctique canadien peuvent être caractérisés à partir de données d’observation de la terre afin de supporter les opérations de nettoyage.

Assessing the Potential to Operationalize Shoreline Sensitivity Mapping: Classifying Multiple Wide Fine Quadrature Polarized RADARSAT-2 and Landsat 5 Scenes with a Single Random Forest Model

The Random Forest algorithm was used to classify 86 Wide Fine Quadrature Polarized RADARSAT-2 scenes, five Landsat 5 scenes, and a Digital Elevation Model covering an area approximately 81,000 km2 in size, and representing the entirety of Dease Strait, Coronation Gulf and Bathurst Inlet, Nunavut. The focus of this research was to assess the potential to operationalize shoreline sensitivity mapping to inform oil spill response and contingency planning. The impact of varying the training sample size and reducing model data load were evaluated. Results showed that acceptable accuracies could be achieved with relatively few training samples, but that higher accuracies and greater probabilities of correct class assignment were observed with larger sample sizes. Additionally, the number of inputs to the model could be greatly reduced without impacting overall performance. Optimized models reached independent accuracies of 91% for seven land cover types, and classification probabilities between 0.77 and 0.98 (values for latter represent per-class averages generated from independent validation sites). Mixed results were observed when assessing the potential for remote predictive mapping by simulating transferability of the model to scenes without training data.

Mapping Coastal Information across Canada's Northern Regions Based on Low-Altitude Helicopter Videography in Support of Environmental Emergency Preparedness Efforts

ABSTRACT Wynja, V.; Demers, A.-M.; Laforest, S.; Lacelle, M.; Pasher, J.; Duffe, J.; Chaudhary, B.; Wang, H., and Giles, T., 2015. Mapping coastal information across Canadas northern regions based on low-altitude helicopter videography in support of environmental emergency preparedness efforts. In the face of increasing economic opportunities in Canadas northern regions, the need to improve our state of preparedness for oil spill–related emergencies is critical. While significant efforts have been put toward documenting baseline coastal information across Canadas southern regions, there is a large information gap regarding Arctic shorelines. Baseline coastal information, such as shoreline form, substrate, and vegetation type, is required for prioritizing operations, coordinating onsite spill response activities (i.e. Shoreline Cleanup Assessment Technique [SCAT]), and providing information for wildlife and ecosystem management. Georeferenced high-definition videography was collected during the summers of 2010 to 2012 along coastlines within six study sites across the Canadian Arctic. Detailed information describing the upper intertidal, supratidal, and backshore zones was extracted from the video and entered into a geospatial database using a data collection form. This information was used to delimit and map alongshore segments in the upper intertidal zone. Almost 15,000 km of northern shorelines were mapped, including 25 shoreline types based on the upper intertidal zone. This information will feed into a larger ongoing project focused on Arctic coastal ecosystems and oil spill response planning should the need arise.

Moving to the RADARSAT Constellation Mission: Comparing Synthesized Compact Polarimetry and Dual Polarimetry Data with Fully Polarimetric RADARSAT-2 Data for Image Classification of Peatlands

For this research, the Random Forest (RF) classifier was used to evaluate the potential of simulated RADARSAT Constellation Mission (RCM) data for mapping landcover within peatlands. Alfred Bog, a large peatland complex in Southern Ontario, was used as a test case. The goal of this research was to prepare for the launch of the upcoming RCM by evaluating three simulated RCM polarizations for mapping landcover within peatlands. We examined (1) if a lower RCM noise equivalent sigma zero (NESZ) affects classification accuracy, (2) which variables are most important for classification, and (3) whether classification accuracy is affected by the use of simulated RCM data in place of the fully polarimetric RADARSAT-2. Results showed that the two RCM NESZs (−25 dB and −19 dB) and three polarizations (compact polarimetry, HH+HV, and VV+VH) that were evaluated were all able to achieve acceptable classification accuracies when combined with optical data and a digital elevation model (DEM). Optical variables were consistently ranked to be the most important for mapping landcover within peatlands, but the inclusion of SAR variables did increase overall accuracy, indicating that a multi-sensor approach is preferred. There was no significant difference between the RF classifications which included RADARSAT-2 and simulated RCM data. Both medium- and high-resolution compact polarimetry and dual polarimetric RCM data appear to be suitable for mapping landcover within peatlands when combined with optical data and a DEM.

Landsat Super-Resolution Enhancement Using Convolution Neural Networks and Sentinel-2 for Training

Landsat is a fundamental data source for understanding historical change and its effect on environmental processes. In this research we test shallow and deep convolution neural networks (CNNs) for Landsat image super-resolution enhancement, trained using Sentinel-2, in three study sites representing boreal forest, tundra, and cropland/woodland environments. The analysis sought to assess baseline performance and determine the capacity for spatial and temporal extension of the trained CNNs. This is not a data fusion approach and a high-resolution image is only needed to train the CNN. Results show improvement with the deeper network generally achieving better results. For spatial and temporal extension, the deep CNN performed the same or better than the shallow CNN, but at greater computational cost. Results for temporal extension were influenced by change potentiality reducing the performance difference between the shallow and deep CNN. Visual examination revealed sharper images regarding land cover boundaries, linear features, and within-cover textures. The results suggest that spatial enhancement of the Landsat archive is feasible, with optimal performance where CNNs can be trained and applied within the same spatial domain. Future research will assess the enhancement on time series and associated land cover applications.

Mapping Arctic Coastal Ecosystems with High Resolution Optical Satellite Imagery Using a Hybrid Classification Approach

ABSTRACT Most mapping methods for Arctic land cover are pixel-based techniques for low resolution data, and have limitations in mapping land cover heterogeneity over complex Arctic polygonal tundra terrain. In this study, we developed a hybrid object-based approach for Arctic coastal tundra mapping using very high resolution optical satellite imagery by combining results from semi-automatic water/land separation, texture analysis based on local binary pattern (LBP), and image classification via Random Forests (RF). The method was applied for coastal land cover mapping in a study site in Tuktoyaktuk, Northwest Territories, Canada using Pleiades satellite data. Results from pixel-based Maximum Likelihood Classifier (MLC), segment-based MLC, pixel-based RF, and segment-based RF were compared with the proposed method. The hybrid method outperformed other approaches and achieved an overall accuracy of 88% for 9 classes. In particular, it has successfully identified unique land cover types of Ice-Wedge Polygons, Wetland (inundated low-lying tundra and marsh with water ponds), with both producers and users accuracy over 91%. Results from this study indicate that the developed hybrid method is suitable to be applied for mapping Arctic coastal ecosystems, and confirms the feasibility of proper use of LBP at segment level for mapping complex environment.

Semi-Automated Surface Water Detection with Synthetic Aperture Radar Data: A Wetland Case Study

In this study, a new method is proposed for semi-automated surface water detection using synthetic aperture radar data via a combination of radiometric thresholding and image segmentation based on the simple linear iterative clustering superpixel algorithm. Consistent intensity thresholds are selected by assessing the statistical distribution of backscatter values applied to the mean of each superpixel. Higher-order texture measures, such as variance, are used to improve accuracy by removing false positives via an additional thresholding process used to identify the boundaries of water bodies. Results applied to quad-polarized RADARSAT-2 data show that the threshold value for the variance texture measure can be approximated using a constant value for different scenes, and thus it can be used in a fully automated cleanup procedure. Compared to similar approaches, errors of omission and commission are improved with the proposed method. For example, we observed that a threshold-only approach consistently tends to underestimate the extent of water bodies compared to combined thresholding and segmentation, mainly due to the poor performance of the former at the edges of water bodies. The proposed method can be used for monitoring changes in surface water extent within wetlands or other areas, and while presented for use with radar data, it can also be used to detect surface water in optical images.