Browsing by Author "McDermid, Gregory J."
Now showing 1 - 10 of 10
Results Per Page
Sort Options
- ItemOpen AccessAdvancing Measurement and Modelling of Glacier Change Using Unmanned Aerial Vehicles and Structure-From-Motion(2019-09-09) Bash, Eleanor A.; Moorman, Brian J.; McDermid, Gregory J.; Marshall, Shawn; Lichti, Derek D.; Mueller, Derek R.Glaciers throughout Canada are responding to climate change with rapid changes in mass balance. There are limitations in current methods of measuring and predicting these changes in mass balance, including accessibility, spatial and temporal resolution of remotely sensed data, and cost of data acquisition. Technological developments in unmanned aerial vehicles (UAVs) and structure-from-motion (SfM) have created new opportunities to overcome these limitations. This dissertation investigated uncertainties in UAV-SfM data and used that data to understand spatial patterns and drivers of summer glacier melt. A study of glacier snow surface reconstruction in the Canadian Rockies used lidar data acquired simultaneously with UAV imagery to assess the spatial distribution of errors in the UAV-SfM data. The study revealed patterns in the errors related to snow surface illumination which must be considered when using UAVs over snow covered glaciers. Short term summer melt in the ablation zone of a glacier in the Canadian Arctic was investigated using UAV surveys. The study showed that UAV-SfM melt measurements agreed with ablation stake measurements and was a reliable method of measuring distributed melt patterns. The study found the lower limits on change detection were related to flying height and dGPS precision. A melt model was used to estimate surface melt for the three-day window where UAV-SfM measurements were collected and model results were validated against spatially distributed measurements. This study revealed patterns in model error which show that simplified melt models fail to capture important melt drivers on the glacier surface. The model errors would have cumulative effects in long term projections, which would lead to significant misrepresentation of total surface melt. UAV and SfM technologies were shown to be an effective method for gathering highly detailed information on glacier surface characteristics and change. However, the technology is not the answer to every problem and limitations still exist in its appropriate application. This work shows the utility of the data in advancing modelling efforts where site visits are not feasible. The dissertation ultimately demonstrates that studies can be strengthened using UAV-SfM data as one tool of many to address questions of glacier change.
- ItemOpen AccessConsequences of Spatial Exploitation in Complex Adaptive Social-Ecological Systems: Managing for Sustainable Freshwater Fisheries(2018-01-23) Wilson, Kyle Logan; Post, John R.; Cartar, Ralph Victor; Galpern, Paul; Pope, Kevin L.; McDermid, Gregory J.Freshwater fisheries are complex adaptive social-ecological systems structured by coupled feedbacks between fish and people (e.g., anglers). For example, fishing quality influences angler site choices, and anglers reciprocally impact fish populations at chosen sites through size-selective harvest, thus demonstrating how feedbacks between fish and anglers permeate through whole ecosystems. Overexploitation increasingly threatens these fisheries challenging management with finding robust solutions to sustain these important resources. Yet, we often lack generalization on the social and ecological processes that limit system resilience. This thesis attempts to gain some of that generalization by exploring how whole-system outcomes emerge from cross-scale interactions between fish and anglers. To do this, I used a combination of theory-driven models and empirical case studies on the rainbow trout Oncorhynchus mykiss and lake trout Salvelinus namaycush fisheries of British Columbia and Yukon to find the landscape, social, and ecological consequences of spatial exploitation on these complex systems. In general, I found that system outcomes, like sustainable or overexploited fisheries, strongly depended on both landscape contexts and the strength of social-ecological feedbacks. In doing so, I was able to generalize the kinds of freshwater landscapes at risk of overexploitation. Next, I found that spatial exploitation patterns had cascading effects across freshwater landscapes by influencing ecological processes like the demographic tradeoff between fish body size and abundance or variation in fish life histories. Additionally, I found that angler site choices were influenced by multiple characteristics, like trip contexts, travel costs and fishing quality, allowing me to better identify potential angler impacts on fish populations. I then integrated these spatial, social, and ecological processes to evaluate the kinds of policies that may improve management of the lake trout fishery and found that conservative regulations better balanced both social and ecological objectives. The results of these studies can help inform management on the feedbacks and processes that drive fishery dynamics, how a landscape of fish populations may respond to spatial exploitation, the kinds of landscapes (and populations within those landscapes) at risk of overexploitation, and the efficacy of regulations that target key spatial, social, and ecological processes to sustain freshwater fisheries.
- ItemOpen AccessHydrological Landscape Analysis of a Sinuous Depression, Yaxnohcah, Mexico(2020-01-30) Milley, David Steven; Reese-Taylor, Kathryn; Oetelaar, Gerald A.; McDermid, Gregory J.Yaxnohcah is a large site in Campeche, Mexico with evidence of continual occupation from the early Middle Preclassic into the Postclassic. In 2014, the Yaxnohcah Archaeological Project commissioned a high resolution LiDAR scan of the region, which has allowed for accurate modeling of surface hydrology and significantly contributed to our understanding of hydrological landscape modification at the site. One feature of particular interest was an irregularly shaped, deeply etched sinuous depression located in the Bajo Tomatal, just south of the narrow drainage that connects it with the Bajo Laberinto. The aim of this research was to ascertain whether this sinuous depression is a cultural or natural feature, and if cultural, what it can tell us about how the ancient lowland Maya at Yaxnohcah modified the hydrology of their natural landscape to sustain urban settlements in the dense and inhospitable rainforests of the Yucatan. In 2017 and 2018, I modeled and analysed the hydrology of the sinuous depression, and in the 2018 field season performed excavations of the feature. The excavations showed that, while the sinuous depression may originally have been a natural feature, it was extensively modified, with clear evidence for considerable refurbishment during the Postclassic and data suggesting an earlier date for initial construction. Furthermore, the modelling indicates that the sinuous depression formed part of system of hydrological features that was accretional developed throughout the Preclassic into the Early Classic as control mechanisms for redirecting, buffering, and capturing water around the Brisa complex. This system underwent considerable refurbishment during the Postclassic.
- ItemOpen AccessLiDAR Characterization of Boreal Understory(2021-05-13) Losada Rozo, Silvia Alejandra; McDermid, Gregory J.; McDermid, Gregory J.; Goldblum, David; Chamer, Laura E.The understory vegetation layer contributes considerably to the physical structure of boreal forests. This research sought to understand the relationships between field- and LiDAR- (light detection and ranging-) derived measures of boreal understory structure. As well as how environmental factors may influence discrepancies that can arise between these derived measures. Five attributes to map and characterize the boreal understory vegetation were selected: mean understory height, percent cover, density, complexity, and volume. Percent understory cover showed limited bias in LiDAR-derived estimates of compared to field measurements, in northeastern Alberta. However, LiDAR was shown to underestimate understory mean height and volume, and to overestimate understory density and complexity. Generalized linear model regression analysis were used to understand the influence of external environmental factors on these error patterns. Explanatory variables for these models included canopy openings, bole density, canopy complexity, and ecosite type. It was found that canopy openings reduced errors in understory mean height, percent cover, and volume. Higher bole density was strongly associated with increased errors in understory mean height and volume, and had weak influence on errors in understory percent cover, complexity, and density. More complex canopies were seen to slightly increase the errors in understory volume and did not influence errors in the remaining attributes. Finally, ecosite had a strong influence on errors in understory mean height, complexity, and volume. In the final phase of this research, a series of predictive maps of understory structure were developed across a 4300-hectare study area in the central mixed-wood subregion of the boreal forest, with independent-validation coefficients of determination ranging from 0.41 - 0.59.
- ItemOpen AccessMapping shrub biomass in a boreal continental fen(2019-01-24) He, Annie; McDermid, Gregory J.; Post, John R.; Strack, MariaBiomass estimation is a heavily explored topic in the literature, as biomass information can provide valuable insight into understanding an ecosystem’s health. However, few studies exist on quantifying aboveground biomass (AGB) in peatlands. This thesis summarizes how allometric equations and unmanned aerial vehicles (UAV) can be used to map AGB across a 2-hectare peatland site. Prior to using UAVs to measure AGB, accurate field measurements are required to calibrate and validate the UAV-AGB model. We developed allometric equations for three dominant shrub genera found in boreal peatlands and found that equations based on shrub genus were not significantly different from a pooled equation of all shrub genera. The UAV study revealed that UAV-derived volume was the best predictor of AGB (R2=0.885) and was subsequently used as the dependent variable for our AGB model. This thesis reports the findings revealed through the process of estimating AGB using allometric equations and UAVs.
- ItemOpen AccessModelling and Mapping Fine-Scale Vegetation Biomass in Banff National Park Using Remotely Sensed Data Collected by Unmanned Aerial Systems(2020-05-12) Poley, Lucy Gem; McDermid, Gregory J.; Bender, Darren J.; Hugenholtz, Chris H.Ecological investigations and long-term monitoring programs in grassland ecosystems often require detailed information on vegetation parameters across an area of interest. However, characterization of grassland vegetation is challenging using ground-based measurements or satellite imagery due to the fine-scale heterogeneity present in grasslands. Unmanned Aerial Systems (UASs) provide a way to characterize vegetation at a high spatial resolution, bridging the gap between ground-based measurements and satellite imagery. Motivated by the need for long-term monitoring of fine-scale vegetation parameters following the reintroduction of plains bison (Bison bison bison) to Banff National Park, Canada, this research explored how UAS-derived data could be used to estimate the aboveground biomass of vegetation at remote grassland sites in Banff’s bison reintroduction zone. I assessed the factors affecting quality of UAS-based vegetation estimation in previous research to design a series of analytical experiments. I conducted UAS surveys at study sites in July 2018 using visible-light and multispectral sensors. Concurrent to the aerial surveys, I collected ground-based data on shrub and herbaceous vegetation biomass. I derived spectral and textural variables using one or more wavelengths of light from the UAS imagery and related to ground-measured biomass using linear regression models. I accurately estimated shrub biomass using an area-weighted vegetation index derived from visible-light imagery that fused spectral and structural information into one parsimonious model. For herbaceous vegetation, combining visible-light and multispectral texture information derived from vegetation indices was the best approach to biomass estimation, and I was able to quantify the relative contributions of photosynthetic and non-photosynthetic vegetation within total biomass. I then modelled the distribution of shrub and herbaceous vegetation biomass across the study site and a workflow for collecting and analyzing UAS imagery for vegetation biomass monitoring was developed. The methods and maps of grassland vegetation produced in this study will provide the data and workflow necessary to monitor fine-scale impacts of bison on vegetation in Banff National Park, supporting conservation and management of this ecologically important species. This research also increases general knowledge of remote sensing of vegetation and provides an approach to fine-scale vegetation characterization suitable for ecological investigations in grasslands ecosystems
- ItemOpen AccessModelling Whitebark Pine Distribution in the Crown of the Continent Ecosystem(2019-07-09) Blackadder, Shannon; McDermid, Gregory J.; Jacobson, Dan; Quinn, Michael S.Whitebark Pine (Pinus albicaulis; WBP) is a species facing serious threats throughout its range. To support ongoing WBP conservation efforts in the Crown of the Continent Ecosystem (CCE), this thesis builds a multi-jurisdictional species distribution model (SDM). There are many critical decisions in building SDMs and little consensus on what works best. In this thesis, I consider how parameters of model type, pseudo-absence selection methods, and the number of pseudo-absences affect model performance. To deal with imperfect input data, I first build and test these model parameters on five WBP-like virtual species (VS). The results show that there are many model parameter combinations that perform equally well, but they predicted different spatial patterns of occurrence. Ensemble models (EM) were used to combine the information in these models into a final EM that outperforms any single model. Building a number of simple and ecologically sound models to use in an EM may be more useful than searching for the best single model. Whether VS are suitable proxies for real species is still unknown, but, used with caution, they have the potential to inform modelling parameter choices in circumstances where input data are imperfect. The final model built in this study is suitable for understanding the broad-scale relative probability of WBP across the CCE, and should complement finer scale work.
- ItemOpen AccessRemote Sensing Boreal Coarse Woody Debris(2019-09-19) Lopes Queiroz, Gustavo; McDermid, Gregory J.; McDermid, Gregory J.; Else, Brent G. T.; Reid, Mary L.Coarse woody debris (CWD) are vital components of forested environments, affecting the physical structure and biochemistry of forests, supplying habitats, nutrients and food for many organisms. Additionally, CWD is an especially important element in boreal forest management in Alberta, Canada. Large quantities of CWD can configure a fire hazard, whereas moderate quantities of CWD in linear disturbances can aid newly planted seedlings to survive and hinder the hunting effectiveness of predators of endangered caribou herds. Forest managers and ecologists can benefit from large-scale high-accuracy quantitative mapping of CWD in complex boreal environments. This thesis represents the first high-resolution remote sensing study of CWD within the context of Alberta’s boreal forest. The research conducted here tested the effectiveness of a geographical object-based image analysis (GEOBIA) workflow with random forest classification for mapping CWD logs and snags in a 4300-hectare study area in northeastern Alberta, Canada. Additionally, zero-adjusted models were selected for accurate estimation of CWD volume in the study area using Akaike’s information criterion. The developed models successfully mapped (up to 93.4% completeness and 94.5% correctness) and estimated volume of CWD (0.623 R2, 0.224 RMSE) with good accuracies. Light detection and ranging (LiDAR) data improved the distinction between logs and snags in CWD maps (~6% better distinction; significant at α 0.05), and multispectral LiDAR data improved the estimation of CWD volume occluded by superimposed vegetation (~ 0.1 higher R2 and ~0.018 lower RMSE). Models developed in a calibration area could be applied to a verification area 4 km distant from all training data without substantial differences in accuracy (average 9% drop in mapping accuracy, no decrease in R2 or increase in RMSE when estimating volume). Given the potential of emerging multispectral LiDAR technologies, it is likely that future improvements to sensors will make ever more accurate CWD predictions possible. Site managers, as well as ecologists and foresters interested in studying the spatiality of CWD can make use of the developed workflows to obtain accurate and extensive map products in forested areas.
- ItemOpen AccessUAV Remote Sensing Can Reveal the Effects of Low-Impact Seismic Lines on Surface Morphology, Hydrology, and Methane (CH4) Release in a Boreal Treed Bog(Journal of Geophysical Research: Biogeosciences, 2018-02-23) Lovitt, Julie; Rahman, Mustafizur M.; Saraswati, Saraswati; McDermid, Gregory J.; Strack, Maria; Xu, BinPeatlands are globally significant stores of soil carbon, where local methane (CH4) emissions are strongly linked to water table position and microtopography. Historically, these factors have been difficult to measure in the field, constraining our capacity to observe local patterns of variability. In this paper, we show how remote sensing surveys conducted from unmanned aerial vehicle (UAV) platforms can be used to map microtopography and depth to water over large areas with good accuracy, paving the way for spatially explicit estimates of CH4 emissions. This approach enabled us to observe—for the first time—the effects of low-impact seismic lines (LIS; petroleum exploration corridors) on surface morphology and CH4 emissions in a treed-bog ecosystem in northern Alberta, Canada. Through compaction, LIS lines were found to flatten the observed range in microtopographic elevation by 46 cm and decrease mean depth to water by 15.4 cm, compared to surrounding undisturbed conditions. These alterations are projected to increase CH4 emissions by 20–120% relative to undisturbed areas in our study area, which translates to a total rise of 0.011–0.027 kg CH4 day 1 per linear kilometer of LIS (~2 m wide). The ~16 km of LIS present at our 61 ha study site were predicted to boost CH4 emissions by 20–70 kg between May and September 2016.
- ItemOpen AccessA Workflow to Minimize Shadows in UAV-based Orthomosaics(Journal of Unmanned Vehicle Systems, NRC Research Press, 2019-01-08) Rahman, Mir Mustafizur; McDermid, Gregory J.; Mckeeman, Taylor; Lovitt, JulieShadows from buildings, terrain, and other elevated features represent lost and/or impaired data values that hinder the quality of optical images acquired under all but the most diffuse illumination conditions. This is particularly problematic in high-spatial-resolution imagery acquired from unmanned aerial vehicles (UAVs), which generally operate very close to the ground. However, the flexibility and low cost of re-deployment of the platform also presents opportunities, which we capitalize on in a new workflow designed to eliminate shadows from UAV-based orthomosaics. Our straightforward, three-step procedure relies on images acquired from two different UAV flights, where illumination conditions produce diverging shadow orientations: one before solar noon and another after. From this multi-temporal image stack, we first identify and then eliminate shadows from individual orthophoto components, then construct the final orthomosaic using a feature-matching strategy with the commercial software package Photoscan. The utility of our strategy is demonstrated over a treed-wetland study site in northwestern Alberta, Canada: a complex scene containing wide variety of shadows, which our workflow effectively eliminated. While shadow-reduced orthomosaics are generally less useful for feature-identification tasks that rely on the shadow element of image interpretation, they create a superior foundation for most other image-processing routines, including classification and change-detection.