Effects of lake ice on the detection of surface water from medium-spatial-resolution satellite data

Abstract

Lake ice cover has important implications for many ecosystems and human communities around the world. Additionally, lake ice is a preeminent sentinel of climate change. The study of lake ice cover is a growing field in remote sensing; yet, be that as it may, a global, temporally-dense, cloud and gap-free dataset of lake ice cover does not exist. This thesis proposes an algorithm that uses a global, daily open surface water product and temperature data to derive a lake ice class over water bodies. The main datasets were the Global WaterPack, which is a binary open water classification layer and ERA5 Reanalysis temperature data. The algorithm’s output data showed strong agreeance with the validation data retrieved from the NOAA lake ice monitoring archives. All three lake ice phenology variables (ice onset, ice breakup, and ice season duration) retrieved from the algorithm’s output fell within an acceptable range of the validation data.