https://orcid.org/0000-0003-4599-2157
![Sample our Environment & Agriculture journals, sign in here to start your access, latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/5934/TOC-Subject-Sample-2021-Environment-Agriculture.jpg"})
ABSTRACT
Peatland is a globally important store of carbon. Peatland restoration efforts are being increasingly undertaken yet effective monitoring of landscape-scale restoration projects has been limited. A particular gap in our understanding is the length of time required before a site reaches the target state. To address this, a classification model based on remote sensing data was developed for a peatland restoration area on blanket bog in northern Scotland, UK, to evaluate whether post-restoration trajectories followed predictable trends over time. The model was trained against a chronosequence of sites within a 20 × 10 km study area that are being restored following drainage and intensive non-native afforestation. Two versions of the model were created to compare the accuracy obtainable from the suite of Sentinel-2 satellite data versus sub-metre resolution aerial imagery from GetMapping (RGB and IR). The Sentinel-2 based model greatly outperformed the aerial imagery-based model. Adding surface slope to the classification did not significantly improve the accuracy of prediction. Prediction of starting and target land covers was very robust, and both the most recent and oldest restoration sites were well predicted spatially. The main uncertainties in the model were within sites of intermediate restoration age, and sites which underwent additional treatments after the initial restoration. Using standard vegetation and wetness indices as indicators, it was possible to track the progression of areas that had been felled and rewetted towards the spectral signal of the control blanket bog locations. A further study examined the use of multiple years of satellite data (2015–2021) and including Sentinel-1 SAR imagery, and confirmed the findings obtained with only a single climatically average year, and furthermore examined the efficacy of different restoration methods. We observed consistent trends of restoration sites beginning to resemble the target hydrologically and ecologically functional blanket bog state after 10–20 years post intervention.
Acknowledgements
This work was funded through the Scottish Government Rural & Environment Science & Analytical Services (RESAS) Strategic Research Programmes 2016-2022 and 2022-2027. We are grateful to the RSPB field team at Forsinard Flows Reserve for supporting this project, e.g., with data provision and hosting of field visits. We gratefully acknowledge the contributions of two anonymous reviewers, whose comments helped greatly in improving this manuscript.
Disclosure statement
No potential conflict of interest was reported by the authors.
Data availability statement
The data that support the findings of this study are available on request from the corresponding author, [RA]. The data are not publicly available due to ongoing restoration management in the area of interest, inclusive of further follow-on treatments post restoration, and which information will be required to correctly allocate spatial classifiers for any revisit of the model training and validation.
Supplementary Material
Supplemental data for this article can be accessed online at https://doi.org/10.1080/01431161.2023.2209916
Notes
1. Initial felling treatments were generally to waste, and we adopted FTW as abbreviation for this treatment but then kept using the same abbreviation for all felling treatments so as to not confuse the reader.
2. BCFB refers to brash crushing and furrow blocking, however for the purpose of this paper, we used the abbreviation to encompass all methods of reducing surface brash and reversing the topographical legacies, as described.