A Multisource Data Approach for Change and Disturbance Mapping of Ontario's Clay Belt Toward More Accurate Carbon and Emissions Estimation

Abstract

This article is the first of a two-part study on disturbance-informed land use/land cover changes in the Clay Belt region of Northern Ontario, Canada. Despite the drive to convert forests to agricultural land, detailed information on land use changes and the resulting impacts on soil carbon and greenhouse gas (GHG) emissions in the region is lacking. Therefore, this work aims to address the information gap by estimating the amount of land cover changes. The study is driven by the urgent need to develop suitable methodologies for detecting and mapping land cover dynamics in Northern Ontario for forest and agricultural lands. Predominant land cover classes in the study area are mapped in order to quantify the changes from 2002 to 2022. Drawing on nascent technology and tools such as machine learning and Google Earth Engine's cloud computing, the satellite images from Landsat and Sentinel are used to examine the trend of land cover and land use changes. This study proposes a reliable methodology of multisensor fusion with data free of cloud contamination—a method which can be deployed anywhere for large-scale monitoring—yielding high accuracy results for regional or national accounting of ecosystem carbon stocks and GHG emissions.

Keywords

• Change detection
• classification
• Google Earth Engine (GEE)
• land cover