Modelling water table depth at rewetted peatlands with Sentinel-1 and Sentinel-2

Water table depth is the primary consideration during peatland rewetting, as a post-industrial peatland transitions from a degraded system with bare peat surfaces to a natural one. For rewetting to be successful, water table depth should be maintained in the upper 0.2 m of the soil to promote carbon...

Full description

Saved in:
Bibliographic Details
Main Authors: Eoin Reddin, Jennifer Hanafin, Mingming Tong, Laurence Gill, Mark G. Healy
Format: Article
Language:English
Published: Elsevier 2025-06-01
Series:Science of Remote Sensing
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S2666017225000446
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850116673894350848
author Eoin Reddin
Jennifer Hanafin
Mingming Tong
Laurence Gill
Mark G. Healy
author_facet Eoin Reddin
Jennifer Hanafin
Mingming Tong
Laurence Gill
Mark G. Healy
author_sort Eoin Reddin
collection DOAJ
description Water table depth is the primary consideration during peatland rewetting, as a post-industrial peatland transitions from a degraded system with bare peat surfaces to a natural one. For rewetting to be successful, water table depth should be maintained in the upper 0.2 m of the soil to promote carbon sequestration while minimising net greenhouse gas emissions. There is evidence that satellite remote sensing techniques may be effective tools at monitoring water table depth. However, these techniques have been seldom used on degraded bare peat bogs, despite their excellent potential as monitoring tools during the restoration process. The aims of this paper are to (1) systematically test the relationship between radar backscatter and water table depth (2) compare decision tree regression algorithms to evaluate the potential of multi-sensor remote sensing in peatland management, and (3) make novel estimations of site-wide water table depth using a multi-sensor approach. This paper applies multi-sensor machine learning techniques to two post-industrial harvesting degraded peatlands, which are currently undergoing rewetting. Combined, these peatlands have nearly three years (2021–2023) of water table measurements, from over 50 piezometers. These data were used to train machine learning models, resulting in R2 values ranging from 0.72 to 0.78, and RMSE values of 0.14 m and 0.12 m. Significant variation in water level throughout the year was observed, suggesting that the ability for a peatland to successfully sequester carbon may be temporally variable. With this study, we provide a timely assessment of restoration efforts at anthropologically degraded bare peat peatlands. This work proves the utility of remote sensing techniques in tracking restoration progress, and may inform future strategies in peatland restoration, rewetting, and monitoring.
format Article
id doaj-art-03823f644d2e4aaabe5c8ce9d9ad4ccd
institution OA Journals
issn 2666-0172
language English
publishDate 2025-06-01
publisher Elsevier
record_format Article
series Science of Remote Sensing
spelling doaj-art-03823f644d2e4aaabe5c8ce9d9ad4ccd2025-08-20T02:36:16ZengElsevierScience of Remote Sensing2666-01722025-06-011110023810.1016/j.srs.2025.100238Modelling water table depth at rewetted peatlands with Sentinel-1 and Sentinel-2Eoin Reddin0Jennifer Hanafin1Mingming Tong2Laurence Gill3Mark G. Healy4Civil Engineering, University of Galway, Galway, H91 TK33, Ireland; Irish Centre for High-End Computing, University of Galway, Galway, H91 TK33, Ireland; Corresponding author at: Civil Engineering, University of Galway, Galway, H91 TK33, Ireland.Irish Centre for High-End Computing, University of Galway, Galway, H91 TK33, IrelandMechanical Engineering, University of Galway, Galway, H91 TK33, IrelandDepartment of Civil, Structural, and Environmental Engineering, Trinity College Dublin, Dublin, D02 PN40, IrelandCivil Engineering, University of Galway, Galway, H91 TK33, IrelandWater table depth is the primary consideration during peatland rewetting, as a post-industrial peatland transitions from a degraded system with bare peat surfaces to a natural one. For rewetting to be successful, water table depth should be maintained in the upper 0.2 m of the soil to promote carbon sequestration while minimising net greenhouse gas emissions. There is evidence that satellite remote sensing techniques may be effective tools at monitoring water table depth. However, these techniques have been seldom used on degraded bare peat bogs, despite their excellent potential as monitoring tools during the restoration process. The aims of this paper are to (1) systematically test the relationship between radar backscatter and water table depth (2) compare decision tree regression algorithms to evaluate the potential of multi-sensor remote sensing in peatland management, and (3) make novel estimations of site-wide water table depth using a multi-sensor approach. This paper applies multi-sensor machine learning techniques to two post-industrial harvesting degraded peatlands, which are currently undergoing rewetting. Combined, these peatlands have nearly three years (2021–2023) of water table measurements, from over 50 piezometers. These data were used to train machine learning models, resulting in R2 values ranging from 0.72 to 0.78, and RMSE values of 0.14 m and 0.12 m. Significant variation in water level throughout the year was observed, suggesting that the ability for a peatland to successfully sequester carbon may be temporally variable. With this study, we provide a timely assessment of restoration efforts at anthropologically degraded bare peat peatlands. This work proves the utility of remote sensing techniques in tracking restoration progress, and may inform future strategies in peatland restoration, rewetting, and monitoring.http://www.sciencedirect.com/science/article/pii/S2666017225000446Peatland rewettingSynthetic aperture radarBackscatterMachine learning
spellingShingle Eoin Reddin
Jennifer Hanafin
Mingming Tong
Laurence Gill
Mark G. Healy
Modelling water table depth at rewetted peatlands with Sentinel-1 and Sentinel-2
Science of Remote Sensing
Peatland rewetting
Synthetic aperture radar
Backscatter
Machine learning
title Modelling water table depth at rewetted peatlands with Sentinel-1 and Sentinel-2
title_full Modelling water table depth at rewetted peatlands with Sentinel-1 and Sentinel-2
title_fullStr Modelling water table depth at rewetted peatlands with Sentinel-1 and Sentinel-2
title_full_unstemmed Modelling water table depth at rewetted peatlands with Sentinel-1 and Sentinel-2
title_short Modelling water table depth at rewetted peatlands with Sentinel-1 and Sentinel-2
title_sort modelling water table depth at rewetted peatlands with sentinel 1 and sentinel 2
topic Peatland rewetting
Synthetic aperture radar
Backscatter
Machine learning
url http://www.sciencedirect.com/science/article/pii/S2666017225000446
work_keys_str_mv AT eoinreddin modellingwatertabledepthatrewettedpeatlandswithsentinel1andsentinel2
AT jenniferhanafin modellingwatertabledepthatrewettedpeatlandswithsentinel1andsentinel2
AT mingmingtong modellingwatertabledepthatrewettedpeatlandswithsentinel1andsentinel2
AT laurencegill modellingwatertabledepthatrewettedpeatlandswithsentinel1andsentinel2
AT markghealy modellingwatertabledepthatrewettedpeatlandswithsentinel1andsentinel2