the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 18 Nov 2025
PeatDepth-ML: A Global Map of Peat Depth Predicted using Machine Learning
Abstract. Peatlands are major carbon stores that are sensitive to climate change and increasingly affected by human activity. Accurate assessment of carbon stocks and modelling of peatland responses to future climate scenarios requires robust information on peat depth. We developed PeatDepth-ML, a machine learning framework that predicts global peat depths using a comprehensive database of peat depth measurements for training and validation. Building on an existing framework for mapping peatland extent, we incorporated new environmental datasets relevant to peat formation, revised cross-validation procedures, and introduced a custom scoring metric to improve predictions of deep peat deposits. To evaluate model sensitivity to sampling bias inherent in the training data, we applied a bootstrapping approach. Model performance, assessed using a blocked leave-one-out approach, yielded a root mean square error of 70.1 ± 0.9 cm and a mean bias error of 2.1 ± 0.7 cm, performing as well as or better than previously published models. The global map produced by PeatDepth-ML predicts a median peat depth of 134 cm (IQR: 87–187) over areas with more than 30 cm of peat. Like other regression-based models, PeatDepth-ML tended to predict toward mean training depths. An area of applicability analysis suggests the model has good applicability globally with the exception of some coastal and several mountainous regions like the Andes and the highlands of Borneo and New Guinea. Predictor selection was highly sensitive to training data subsets that arose from the bootstrapping approach, occasionally resulting in regional variations in accuracy. The bootstrapping approach and our area of applicability analysis thus clearly demonstrates the prime importance of quality training data in data-driven approaches like PeatDepth-ML. Using our predicted peat depth map, together with peatland extent and literature-derived estimates of bulk density and organic carbon content, we estimate global peat carbon stocks at 327–373 Pg C, consistent with previous global estimates.
- Preprint
(32593 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2025-5363', Anonymous Referee #1, 17 Dec 2025
- AC1: 'Reply on RC1', Joe Melton, 21 Feb 2026
-
RC2: 'Comment on egusphere-2025-5363', Anonymous Referee #2, 08 Feb 2026
This paper presents another attempt to generate a global map of peat thickness and to estimate associated carbon stocks. While this topic is important for improving our general understanding of peatland conditions, several substantial aspects need to be carefully considered to strengthen the manuscript.
- This paper lacks novelty. A similar study was conducted at 1 km resolution by Widyastuti et al. (2025). The authors need to better position this work within the existing literature. Machine learning has already been widely used to map peat thickness.
- The percentage of non-zero data appears to be too small, potentially leading to underestimation of peat thickness predictions. I suggest maintaining a more balanced proportion between peat and non-peat data (e.g., 50:50), both in the training dataset and in the bootstrap samples. This balance may produce more reliable results.
- The resulting global peat thickness map (Fig. 7) appears to predict peat thickness in some areas that are not peatlands, including parts of Europe, the Amazon, South America, Indonesia, Papua New Guinea, and New Zealand. It is unclear how the global SOC stock was calculated, for example, whether non-peatland areas were included in the calculation. Since the peat fraction is considered in the analysis, it may be better to display only areas with peat fraction > 0. This could provide a clearer comparison with previous studies. In addition, including brief highlights or discussions of regional peatland characteristics could help support and validate the resulting maps.
- Overall, the paper focuses heavily on comparing methods and results with previous studies. While such comparisons are important for evaluating model performance, the authors should also discuss how their findings contribute new insights to peat carbon research.
Specific comments:
Line 17: Please specify the spatial coverage used for the carbon stock calculation.
Introduction: The study’s novelty is not clearly demonstrated. Numerous studies have applied machine learning to predict peat depth (Crezee et al., 2022; Deragon et al., 2023; Rudiyanto et al., 2018; Widyastuti et al., 2025). Since global peat thickness and carbon stock mapping have already been conducted, the authors should clearly explain what distinguishes this study.
Lines 93–97: I suggest moving Figure A1 to the main text, as it is referenced frequently. It could potentially be presented together with Figures 1 and 2.
Line 140: While climate influences peat formation, it is unclear how the two temperature datasets mentioned contribute to peat formation processes. Please clarify.
Line 473: Which results demonstrate improvement over previous studies?
Crezee, B., Dargie, G. C., Ewango, C. E. N., Mitchard, E. T. A., Emba B, O., Kanyama T, J., Bola, P., Ndjango, J.-B. N., Girkin, N. T., Bocko, Y. E., Ifo, S. A., Hubau, W., Seidensticker, D., Batumike, R., Imani, G., Cuní-Sanchez, A., Kiahtipes, C. A., Lebamba, J., Wotzka, H.-P.,…Lewis, S. L. (2022). Mapping peat thickness and carbon stocks of the central Congo Basin using field data. Nature Geoscience, 15(8), 639-644. https://doi.org/10.1038/s41561-022-00966-7
Deragon, R., Saurette, D. D., Heung, B., & Caron, J. (2023). Mapping the maximum peat thickness of cultivated organic soils in the southwest plain of Montreal. Canadian Journal of Soil Science, 103(1), 103-120. https://doi.org/10.1139/cjss-2022-0031
Rudiyanto, Minasny, B., Setiawan, B. I., Saptomo, S. K., & McBratney, A. B. (2018). Open digital mapping as a cost-effective method for mapping peat thickness and assessing the carbon stock of tropical peatlands. Geoderma, 313, 25-40. https://doi.org/https://doi.org/10.1016/j.geoderma.2017.10.018
Widyastuti, M. T., Minasny, B., Padarian, J., Maggi, F., Aitkenhead, M., Beucher, A., Connolly, J., Fiantis, D., Kidd, D., Ma, Y., Macfarlane, F., Robb, C., Rudiyanto, Setiawan, B. I., & Taufik, M. (2025). Digital mapping of peat thickness and carbon stock of global peatlands. CATENA, 258, 109243. https://doi.org/https://doi.org/10.1016/j.catena.2025.109243
Citation: https://doi.org/10.5194/egusphere-2025-5363-RC2 - AC2: 'Reply on RC2', Joe Melton, 21 Feb 2026
Data sets
Peat-DBase version 0.9 Jade Skye https://doi.org/10.5281/zenodo.15530645
Model code and software
PeatDepth-ML: Using Machine Learning to Predict a Global Map of Peat Depth Jade Skye et al. https://doi.org/10.5281/zenodo.15530817
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 415 | 270 | 37 | 722 | 30 | 28 |
- HTML: 415
- PDF: 270
- XML: 37
- Total: 722
- BibTeX: 30
- EndNote: 28
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
The authors present PeatDepth-ML, a machine-learning framework for predicting global peat depth using a large compilation of peat depth measurements and environmental covariates. They extend existing peatland mapping approaches by incorporating additional predictors, revised spatial cross-validation, a custom metric targeting deep peat, and a bootstrapping strategy to assess sensitivity to sampling bias. Model performance is evaluated with blocked leave-one-out validation, and the resulting global peat depth map is used to estimate global peat carbon stocks, which are found to be consistent with previous studies.
I think the work is relevant for the journal and generally well-executed, though I think some revisions are in order prior to publication. I will give detailed list of comments in the following. Thank you for your work.
Detailed comments:
Lines 49 and 66: "machine learning" --> use abbreviation "ML".
Line 92, Figure A1: I think Figure A1 is quite important, presenting the peat data distributions. Why not include it in main text instead of in appendix?
Line 97: "However, grid cells with zero peat depth consistently dominate..." --> explicitly state the percentage of zero peat depth as it is the substantial majority of the data. I think it is good to state as the data is quite, though naturally, imbalanced.
Line 185: "machine learning" --> "ML"
Line 189: What were the hyperparameters which were optimized? I did not see them listed.
Line 192: "cross validation" --> "cross-validation"
Line 205: "don't" --> "do not"
Line 209: Add reference for LightGBM, maybe also fully open up the term. Lets not assume reader knows all the abbreviations by default.
Line 247: Did you mention somewhere how many predictors you had in total available for the ML runs? I would be curious to know this.
Figure 8 and A1: I am not used to horizontal histograms or distributions being presented. Was there a particular reason for this? If not, why not use standard orientation in visualization (vertical bars), which, to my experience, is more common.
Figure A1 caption: extra whitespace before ".", "...desert data ."
Line 357: Open up the abbreviations, although well-known, the RMSE, MBE, NME. They are mentioned also in appendix more specifically, but good the clarify the abbreviations, once introduced.
Line 362: Could you please elaborate on the null models a bit. Do you mean baseline models? Also on same line, notice extra period ". ."
Line 370: "BLOOCV" Did you define this abbreviation, even though clear to myself. But still, define it earlier in the text when you mention cross-validation.
Figure 9: The legend is little bit unclear for me. What is "bootstrap results", what results? Maybe rephrase more clearly, if possible.