Global mangrove soil organic carbon stocks dataset at 30 m resolution for the year 2020 based on spatiotemporal predictive machine learning

Aug 6, 2026

Upload by Himbo

This dataset provides global mangrove forest soil organic carbon stocks predicted for 2020 at 30 m resolution. Spatiotemporal ensemble machine learning is used to estimate soil organic carbon content and bulk density to a 1 m depth, which are aggregated to produce soil organic carbon stocks. Training relies on validated mangrove measurements of SOC (%) and bulk density from global and published sources, together with globally consistent covariate layers. Model evaluation includes spatial cross-validation, and 2020 maps report mean predictions with prediction intervals to quantify uncertainty. Downloadable outputs include raster products and cloud-optimized GeoTIFFs under a CC-BY license.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages10

Reading Time25 min

AccessFree to download

Dataset overview

Spatial resolution and target year

Modeling approach and predictors

Training data and validation

Data processing and outputs

Depth estimation and aggregation

Spatial cross-validation and uncertainty

Download options and formats

Data in Brief 50 (2023) 109621 Contents lists available at ScienceDirect Data in Brief journal [homepage:](homepage: www.elsevier.com/locate/dib)[ www.elsevier.com/locate/dib](homepage: www.elsevier.com/locate/dib) Data Article Global mangrove soil organic carbon stocks dataset at 30 m resolution for the year 2020 based on spatiotemporal predictive machine learning Tania L. Maxwell a, 1,∗, Tomislav Henglb, 1,∗, Leandro L. Parenteb, Robert Minarikc, Thomas A. Worthington a, Pete Bunting d, Lindsey S. Smart e, Mark D. Spalding a,f, Emily Landis e a Conservation Science Group, Department of Zoology, University of Cambridge, Cambridge, UK bEnvirometriX Ltd, Wageningen 6708 PW, the Netherlands c OpenGeoHub foundation, Wageningen 6708 PW, the Netherlands d Department of Geography and Earth Sciences, Aberystwyth University, Aberystwyth, SY23 3DB, UK e The Nature Conservancy, Arlington, VA, USA f The Nature Conservancy, Strada delle Tolfe, 14, Siena, 53100, Italy a r t i c l e i n f o Article history: Received 24 May 2023 Revised 26 July 2023 Accepted 19 September 2023 Available online 26 September 2023 Dataset link: Global mangrove soil carbon data set at 30 m resolution for year 2020 (0-100 cm) (Original data) Keywords: Blue carbon Carbon sequestration Coastal ecosystem Spatial modelling Mangroves ∗ Corresponding authors. a b s t r a c t This dataset presents global soil organic carbon stocks in mangrove forests at 30 m resolution, predicted for 2020. We used spatiotemporal ensemble machine learning to produce predictions of soil organic carbon content and bulk density (BD) to 1 m soil depth, which were then aggregated to calculate soil organic carbon stocks. This was done by using training data points of both SOC (%) and BD in mangroves from a global dataset and from recently published studies, and globally consistent predictive covariate layers. A total of 10,331 soil samples were validated to have SOC (%) measurements and were used for predictive soil mapping. We used time-series remote sensing data speciﬁc to time periods when the training data were sampled, as well as long-term (static) layers to train an ensemble of machine learning model. Ensemble models were used to improve E-mail addresses: [taniamaxwell7@gmail.com](taniamaxwell7@gmail.com) (T.L. Maxwell), [tom.hengl@envirometrix.net](tom.hengl@envirometrix.net) (T. Hengl) . Social media: @tania_maxwell7 (T.L. Maxwell), @opengeohub (T. Hengl) 1 Shared ﬁrst-authorship. [https://doi.org/10.1016/j.dib.2023.109621](https://doi.org/10.1016/j.dib.2023.109621) 2352-3409/© 2023 Published by Elsevier Inc. This is an open access article under the CC BY license ([http://creativecommons.org/licenses/by/4.0/](http://creativecommons.org/licenses/by/4.0/)) 2 T.L. Maxwell, T. Hengl and L.L. Parente et al. /Data in Brief 50 (2023) 109621 performance, robustness and unbiasedness as opposed to just using one learner. In addition, we performed spatial cross-validation by using spatial blocking of training data points to assess model performance. We predicted SOC stocks for the 2020 time period and applied them to a 2020 mangrove extent map, presenting both mean predictions and prediction intervals to represent the uncertainty around our predictions. Predictions are available for download under CC-BY license from 10.5281/zenodo.7729491 and also as Cloud-Optimized GeoTIFFs (global mosaics). © 2023 Published by Elsevier Inc. This is an open access article under the CC BY license ([http://creativecommons.org/licenses/by/4.0/](http://creativecommons.org/licenses/by/4.0/)) Speciﬁcations Table Subject Speciﬁc subject area Type of data How the data were acquired Data format Description of data collection Data source location Data accessibility Agricultural Sciences (Soil Science), Environmental Science, Computer Science (Computer Science Applications) Soil carbon in mangroves, remote sensing signal processing, spatio

FAQs

What does the dataset provide for mangrove forests?

It provides global soil organic carbon stocks predicted for 2020 in mangrove forests, mapped at 30 m resolution.

How are soil organic carbon stocks predicted?

Spatiotemporal ensemble machine learning predicts SOC content and bulk density to 1 m depth, then aggregates these estimates to calculate SOC stocks.

How is uncertainty handled in the 2020 predictions?

Predicted soil carbon stocks are mapped for 2020 with mean predictions and prediction intervals, reflecting uncertainty around model outputs.

Aug 6, 2026

Upload by Himbo