The prairie region of Canada is a dynamically changing landscape in relation to past and present anthropogenic activities and recent climate change. Improving our understanding of the rate, timing, and distribution of landscape change is needed to determine the impact on wildlife populations and biodiversity, ultimately leading to better-informed management regarding requirements for habitat amount and its connectedness. In this research, we assessed the viability of an approach to detect from–to class changes designed to be scalable to the prairie region with the capacity for local refinement. It employed a deep-learning convolution neural network to model general land covers and examined class memberships to identify land-cover conversions. For this implementation, eight land-cover categories were derived from the Agriculture and Agri-Food Canada Annual Space-Based Crop Inventory. Change was assessed in three study areas that contained different mixes of grassland, pasture, and forest cover. Results showed that the deep-learning method produced the highest accuracy across all classes relative to an implementation of random forest that included some first-order texture measures. Overall accuracy was 4% greater with the deep-learning classifier and class accuracies were more balanced. Evaluation of change accuracy suggested good performance for many conversions such as grassland to crop, forest to crop, water to dryland covers, and most bare/developed-related changes. Changes involving pasture with grassland or cropland were more difficult to detect due to spectral confusion among classes. Similarly, conversion to forests in some cases was poorly detected due to gradual and subtle change characteristics combined with confusion between forest, shrub, and croplands. The proposed framework involved several processing steps that can be explored to enhance the thematic content and accuracy for large regional implementation. Evaluation for understanding connectivity in natural land covers and related declines in species at risk is planned for future research.
Differentiation of grassland/forage types and accurate estimates of their location and extent are important for understanding their ecological processes and for applying appropriate management practices. We are aiming to reveal the different spectral characteristics of six grassland/forage land covers in three ecoregions located in the Canadian Prairies, based on field data and satellite images. Three spectral indices representing productivity (Normalized Difference Vegetation Index (NDVI)), moisture content (Normalized Difference Moisture Index (NDMI)), and plant photosynthetic activity (Plant Senescence Reflectance Index (PSRI)) were used for comparison of means, comparison of coefficient of variation (CV), and analysis of variance (ANOVA). The results indicated that different grassland types show distinguishable spectral characteristics in the Moist-Mixed and Mixed Ecoregions, while it was not possible to differentiate the classes in the Fescue Ecoregion. To further investigate the within-sites and between-sites heterogeneity, we calculated the CV in a 3 × 3 window and placed them in comparative triangles to demonstrate their potential separability. Results indicated that the triangles based on the CV offered greater class separability in the Fescue Ecoregion and in the Mixed Ecoregion.
This thesis implements a pixel-based random forest classifier for differentiating rangeland, pastureland, and forage crops using multi-temporal optical and radar Earth observation (EO) data. Previous efforts to create an inventory of rangeland and forage resources across the Canadian Prairies using EO have not achieved desired accuracies due to spectral reflectance similarities amongst the classes and partly as a result of the regional variability of climate, soil type and management practices. Field data related to land cover type and dominant species composition were collected during the 2015 growing season at study areas west of Brandon, MB and near Lethbridge, AB. Landsat-8 (LS8) multispectral and derived phenological variables, as well as mid-summer RADARSAT-2 (RS2) backscatter data were integrated into a Random Forest (RF) classification. The results of this study highlight the importance of spring optical image acquisitions to differentiate between rangeland and seeded forage in multi-date optical land cover classifications, as well as low performance of mid-summer RS2 backscatter for differentiating rangeland and seeded forage alone and in combination with LS8.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.