Automated Mapping of Artificial Drainage in Peatlands Using Deep Learning and Very High-Resolution Aerial Imagery

Peatlands, which cover a significant proportion of the wetland ecosystems globally, play a vital role in maintaining biodiversity and regulating water and climate. However, these ecosystems are currently undergoing degradation as a result of human activities, particularly the draining of peatlands for agricultural purposes, peat extraction, and forestry. Irish raised bogs, which constitute over half of the EU's oceanic raised bogs, have been extensively drained for various land-use activities. Efforts are being made to conserve these ecosystems by implementing measures such as rewetting, restoration, and rehabilitation. However, this requires the identification and accurate mapping of artificial drainage ditches. This study uses a U-net-based convolutional neural network to develop a very high-resolution map of the artificial drainage network in Irish raised bogs, covering an area of 523,000 hectares. The map also quantifies drainage in different land-use categories, such as industrial and domestic peat extraction. The results of this study will aid in implementing conservation activities, such as drain blocking to promote rewetting and improve carbon and greenhouse gas emission accounting at the national scale.