Skip to main content

Lund University Publications

LUND UNIVERSITY LIBRARIES

Forest tree species classification and entropy-derived uncertainty mapping using extreme gradient boosting and Sentinel-1/2 data

Abdi, Abdulhakim M. LU orcid and Wang, Fan LU (2025)
Abstract
We present a wall-to-wall map of dominant tree species in Swedish forests accompanied by pixel-level uncertainty estimates. The tree species classification is based on spatiotemporal metrics derived from Sentinel-1 and Sentinel-2 satellite data, combined with field observations from the Swedish National Forest Inventory and auxiliary data on geomorphometry and canopy height. We apply an extreme gradient boosting model with Bayesian optimization to relate field observations to satellite-derived features and generate the final species map. Classification uncertainty is quantified using Shannon’s entropy of the predicted class probabilities, which provide a spatially explicit measure of model confidence. The final model achieved an overall... (More)
We present a wall-to-wall map of dominant tree species in Swedish forests accompanied by pixel-level uncertainty estimates. The tree species classification is based on spatiotemporal metrics derived from Sentinel-1 and Sentinel-2 satellite data, combined with field observations from the Swedish National Forest Inventory and auxiliary data on geomorphometry and canopy height. We apply an extreme gradient boosting model with Bayesian optimization to relate field observations to satellite-derived features and generate the final species map. Classification uncertainty is quantified using Shannon’s entropy of the predicted class probabilities, which provide a spatially explicit measure of model confidence. The final model achieved an overall accuracy of 85% (F1 score = 0.82, Matthews correlation coefficient = 0.81), and mapped species distributions showed strong agreement with official forest statistics (r = 0.96). Variable importance analysis revealed that the most influential predictors were optical bands from Sentinel-2, particularly those acquired in spring and summer. This study provides scalable, interpretable, and policy-relevant method for tree species mapping with integrated uncertainty that are well-suited to meet emerging legislative and environmental goals. (Less)
Please use this url to cite or link to this publication:
author
and
organization
publishing date
type
Working paper/Preprint
publication status
published
subject
pages
24 pages
publisher
arXiv.org
DOI
10.48550/arXiv.2509.18228
language
English
LU publication?
yes
id
aef547a5-a465-46d3-abd8-67c5cb7071cf
date added to LUP
2025-10-07 10:38:24
date last changed
2025-10-07 12:08:40
@misc{aef547a5-a465-46d3-abd8-67c5cb7071cf,
  abstract     = {{We present a wall-to-wall map of dominant tree species in Swedish forests accompanied by pixel-level uncertainty estimates. The tree species classification is based on spatiotemporal metrics derived from Sentinel-1 and Sentinel-2 satellite data, combined with field observations from the Swedish National Forest Inventory and auxiliary data on geomorphometry and canopy height. We apply an extreme gradient boosting model with Bayesian optimization to relate field observations to satellite-derived features and generate the final species map. Classification uncertainty is quantified using Shannon’s entropy of the predicted class probabilities, which provide a spatially explicit measure of model confidence. The final model achieved an overall accuracy of 85% (F1 score = 0.82, Matthews correlation coefficient = 0.81), and mapped species distributions showed strong agreement with official forest statistics (r = 0.96). Variable importance analysis revealed that the most influential predictors were optical bands from Sentinel-2, particularly those acquired in spring and summer. This study provides scalable, interpretable, and policy-relevant method for tree species mapping with integrated uncertainty that are well-suited to meet emerging legislative and environmental goals.}},
  author       = {{Abdi, Abdulhakim M. and Wang, Fan}},
  language     = {{eng}},
  month        = {{09}},
  note         = {{Preprint}},
  publisher    = {{arXiv.org}},
  title        = {{Forest tree species classification and entropy-derived uncertainty mapping using extreme gradient boosting and Sentinel-1/2 data}},
  url          = {{http://dx.doi.org/10.48550/arXiv.2509.18228}},
  doi          = {{10.48550/arXiv.2509.18228}},
  year         = {{2025}},
}