A5048445086- Flood Risk Assessment and Management
- Synthetic Aperture Radar (SAR) Applications and Techniques
- Climate change and permafrost
- Arctic and Antarctic ice dynamics
- Soil erosion and sediment transport
- Cryospheric studies and observations
- Remote Sensing and LiDAR Applications
- Methane Hydrates and Related Phenomena
- Soil Moisture and Remote Sensing
- Data Mining Algorithms and Applications
- Combustion and flame dynamics
- Fluid Dynamics and Turbulent Flows
- Machine Learning and Data Classification
- Hydrological Forecasting Using AI
- Ocean Waves and Remote Sensing
- Remote-Sensing Image Classification
- Image Enhancement Techniques
- Meteorological Phenomena and Simulations
- Coastal wetland ecosystem dynamics
- Peatlands and Wetlands Ecology
- Bayesian Methods and Mixture Models
- Soil Geostatistics and Mapping
- Remote Sensing in Agriculture
- Gene expression and cancer classification
Environment and Climate Change Canada
2017-2025
University of Ottawa
2013
Random Forests variable importance measures are often used to rank variables by their relevance a classification problem and subsequently reduce the number of model inputs in high-dimensional data sets, thus increasing computational efficiency. However, as result way that training predictor randomly selected for use constructing each tree splitting node, it is also well known if too few trees generated, rankings tend differ between runs. In this letter, we characterize effect (ntree) class...
To better understand and mitigate threats to the long-term health functioning of wetlands, there is need establish comprehensive inventorying monitoring programs. Here, remote sensing data machine learning techniques that could support or substitute traditional field-based collection are evaluated. For Bay Quinte on Lake Ontario, Canada, different combinations multi-angle/temporal quad pol RADARSAT-2, simulated compact RADARSAT Constellation Mission (RCM), high low spatial resolution Digital...
In this study, a new method is proposed for semi-automated surface water detection using synthetic aperture radar data via combination of radiometric thresholding and image segmentation based on the simple linear iterative clustering superpixel algorithm. Consistent intensity thresholds are selected by assessing statistical distribution backscatter values applied to mean each superpixel. Higher-order texture measures, such as variance, used improve accuracy removing false positives an...
There is limited research focusing on Interferometric Synthetic Aperture Radar (InSAR) applications in the Great Lakes coastal wetlands with large water level fluctuations. In this study, we investigated potential of using C-band SAR data to characterize marsh wetland and monitor changes along coast Lakes. InSAR analysis was conducted Radarsat-2 Sentinel-1 collected at Long Point, Ontario, Canada over period 2016–2018. Observations indicated that both backscattering coefficients coherence...
Wetland managers, citizens and government leaders are observing rapid changes in coastal wetlands associated habitats around the Great Lakes Basin due to human activity climate variability. SAR optical satellite sensors offer cost effective management tools that can be used monitor over time, covering large areas like providing information those making policy decisions. In this paper we describe ongoing efforts dynamic wetland vegetation, surface water extent, level change. Included...
Spaceborne Synthetic Aperture Radar (SAR) instruments are effective tools for monitoring and mapping wetlands. With the availability of SAR providing various polarization options, scope this study is to evaluate new compact wetland multitemporal change detection using simulated RADARSAT Constellation Mission (RCM) data. A series fully polarimetric (FP) images were collected over a test site located in Ontario, Canada, used simulate RCM (CP) The data evaluated results compared those from FP...
Detailed information on the land cover types present and horizontal position of land–water interface is needed for sensitive coastal ecosystems throughout Arctic, both to establish baselines against which impacts climate change can be assessed inform response operations in event environmental emergencies such as oil spills. Previous work has demonstrated potential accurate classification via fusion optical SAR data, though what contribution either makes model accuracy not well established,...
Most mapping methods for Arctic land cover are pixel-based techniques low resolution data, and have limitations in heterogeneity over complex polygonal tundra terrain. In this study, we developed a hybrid object-based approach coastal using very high optical satellite imagery by combining results from semi-automatic water/land separation, texture analysis based on local binary pattern (LBP), image classification via Random Forests (RF). The method was applied study site Tuktoyaktuk,...
Differences in topographic structure, vegetation and surface wetness exist between peatland classes, making active remote sensing techniques such as SAR LiDAR promising for mapping. As the timing of green-up, senescence, hydrologic conditions vary differently comparison with upland full growing-season time series imagery was expected to produce higher accuracy classification results than using only a few select images. Both interferometric coherence, amplitude difference datasets were...
Random Forest variable importance measures such as mean decrease in accuracy or Gini index are often used to reduce the number of predictor variables for a given classification problem. However, previous studies suggest that ranking biased, particularly presence highly correlated variables. As result, selected after might not achieve highest possible accuracy. Here, we introduce new metric based on simple statistical measure association, which can be interpreted (statistically) low order...
We investigated the potential of using Synthetic Aperture Radar (SAR) imagery from three different frequencies: X-, C-, and L-band, to characterize coastal wetlands in Great Lakes. Three sets SAR data acquired over Bay Quinte, Ontario, Canada between 2016 2018 Radarsat-2, TerraSAR-X, ALOS-2 satellites were processed small baseline subset (SBAS) Interferometric (InSAR) techniques provide maps surface changes marshes swamps. Results showed that backscatter coherence sensitive sensor...
Arctic amplification is accelerating changes in sea ice regimes the Canadian with later freeze-up and earlier melt events, adversely affecting wildlife communities that depend on stability of conditions. To monitor both rate impact such change, there a need to accurately measure deformation, an important component for understanding motion polar climate. The objective this study determine spatial-temporal pattern deformation over landfast using time series SAR imagery. We present...
There is limited research of InSAR applications in very dynamic wetlands the Great Lakes or for other Canadian wetlands. In this study, we investigated potential using C-band SAR data monitoring marsh flow dynamics Lakes. Results from observations Long Point, Ontario, Canada indicate that consistent coherence Radarsat-2 and Sentinel-1 was observed cattail phragmites dominated areas, which enabled generation reliable measurements water level changes.
This study produced a high-accuracy remotely piloted aircraft system (RPAS) imagery classification method for identifying the invasive reed Phragmites australis ( Cav.) Trin. Ex Steud subsp. using random forest (RF) machine learning. RPAS was collected in spring and fall of 2019 fixed-wing equipped with visible spectrum camera (eBee X, S.O.D.A. 3D; senseFly) Long Point, Ontario, Canada. Imagery used to produce separate early late season classifications bi-temporal which from both dates. The...
Abstract. Arctic amplification is accelerating changes in sea ice regimes the Canadian with later freeze-up and earlier melt events, adversely affecting wildlife communities that depend on stability of conditions. To monitor both rate impact such change, there a need to accurately measure deformation, an important component for understanding motion polar climate. This paper presents Interferometric Synthetic Aperture Radar (InSAR) monitoring landfast deformation as result thickness measured...
The statistical properties of the velocity and scalar fields, including all three derivatives, were measured simultaneously in nearly-homogeneous, uniformly sheared turbulence with two passively superimposed, non-homogeneous, namely a thermal mixing layer plume heated line source. probability density functions sub-Gaussian both fields expectations conditioned on values non-linear. conditional dissipation rate was strongly anisotropic could not be surrogated by any its parts along axes.