ASSESSING EQUIPMENT AND TECHNOLOGIES FOR USE IN THE DEVELOPMENT OF DYKELAND IN ATLANTIC CANADA FOR SUSTAINABLE AGRICULTURAL PRODUCTION

Abstract

Dykelands in Atlantic Canada represent highly productive yet environmentally vulnerable agricultural landscapes, facing challenges from rising sea levels and climate change. This thesis investigates advanced technologies to enhance dykeland management and sustainability. The first study quantifies land use in Nova Scotia’s dykelands using crop inventories and property boundaries, revealing forage production as the dominant agricultural use. The second study applies a Mask R-CNN deep learning model on LiDAR-derived elevation data to map land-formed fields, achieving a mean Average Precision of 0.89. The third study integrates AI-generated field boundaries, crop data, and economic tools within GIS to evaluate profitability, highlighting significant variability across dyke systems. Finally, drone-based remote sensing over three years demonstrates the critical role of surface drainage maintenance in mitigating flood risks and optimizing crop productivity. The findings provide valuable insights for stakeholders aiming to optimize agricultural productivity while making informed decisions in the current context of rising sea levels.