Forecasting the moisture content of live forest fuels with a transformer machine learning model and Sentinel-2 satellite data
Abstract
Live fuel moisture content (LFMC) provides an indicator of vegetation flammability, the potential for canopy dieback from drought, and habitat quality for arboreal fauna. Recent advances in artificial intelligence have increased the potential in meteorological forecasts for Earth surface forecasting. Combining weather forecasts with remotely sensed LFMC offers a promising avenue for short and medium-term LFMC forecasting (i.e. weeks to months). We adapted a transformer-based surface forecasting model to predict LFMC across the forest regions of New South Wales, Australia, up to 50 days in advance. Using a multi-modal spatiotemporal transformer (Contextformer) as a hind-cast model of LFMC with ‘minicubes’ of weather, soil, vegetation structure, elevation and Sentinel-2 reflectance data, we tested four variants to evaluate if theory-guided biophysical predictors improve performance compared to the base model. Explainable AI methods were applied to demystify model reliance, using Integrated Gradients attribution. Incorporating biophysical theory in predictor selection produced the lowest error (Global RMSE=15.68% dry matter), whereas adding only spectral bands to the base model was less successful (partly due to the structure of the forecast architecture). Within the LFMC forecast, explainable AI revealed mixed patterns in importance amongst temperature and rainfall variables due to the transformer’s ability to model multiple eco-physiological pathways to vegetation hydration. Dynamic weather variations dominate forecasted LFMC temporal trends, over spectral bands. Crucially, regression-to-the-mean effects in the LFMC forecasts caused poorer local performance (median R2≈0.21) compared to global evaluation (R2≈0.69, i.e. poorer within minicubes than amongst minicubes). When converting this hind-cast model to an operational model (i.e. true forecast), best performance could be utilised by running rolling short-term forecasts (5 to 15 days) to capture localised fuel hazard peaks and canopy dieback in the future.
Related
- Fine-grained hierarchical crop type classification from integrated hyperspectral EnMAP data and multispectral sentinel-2 time series: A large-scale dataset and dual-stream transformer method
- Evaluating in-context learning with prompting regimes for transformer-based synthetic health tabular data generation
- Enhancing Strawberry Yield Forecasting With Backcasted IoT Sensor Data and Machine Learning
- Sentiment analysis in Bangla: A three-class dataset and comparative evaluation of machine learning, deep learning, and transformer approaches
- Improved vision transformer for sports image classification via transfer learning and knowledge distillation
BibTeX
@misc{kotzur2026forecasting,
title = {{Forecasting the moisture content of live forest fuels with a transformer machine learning model and Sentinel-2 satellite data}},
author = {Ivan Kotzur and Matthias Boer and Vitus Benson and Rachael Nolan},
howpublished = {EarthArXiv},
year = {2026},
month = sep,
doi = {10.31223/x5dn6z},
url = {https://eartharxiv.org/repository/view/15265/},
note = {CC BY-ND 4.0},
}
RIS
TY - UNPB
TI - Forecasting the moisture content of live forest fuels with a transformer machine learning model and Sentinel-2 satellite data
AU - Ivan Kotzur
AU - Matthias Boer
AU - Vitus Benson
AU - Rachael Nolan
PY - 2026
DA - 2026/09/29
PB - EarthArXiv
DO - 10.31223/x5dn6z
UR - https://eartharxiv.org/repository/view/15265/
AB - Live fuel moisture content (LFMC) provides an indicator of vegetation flammability, the potential for canopy dieback from drought, and habitat quality for arboreal fauna. Recent advances in artificial intelligence have increased the potential in meteorological forecasts for Earth surface forecasting. Combining weather forecasts with remotely sensed LFMC offers a promising avenue for short and medium-term LFMC forecasting (i.e. weeks to months). We adapted a transformer-based surface forecasting model to predict LFMC across the forest regions of New South Wales, Australia, up to 50 days in advance. Using a multi-modal spatiotemporal transformer (Contextformer) as a hind-cast model of LFMC with ‘minicubes’ of weather, soil, vegetation structure, elevation and Sentinel-2 reflectance data, we tested four variants to evaluate if theory-guided biophysical predictors improve performance compared to the base model. Explainable AI methods were applied to demystify model reliance, using Integrated Gradients attribution. Incorporating biophysical theory in predictor selection produced the lowest error (Global RMSE=15.68% dry matter), whereas adding only spectral bands to the base model was less successful (partly due to the structure of the forecast architecture). Within the LFMC forecast, explainable AI revealed mixed patterns in importance amongst temperature and rainfall variables due to the transformer’s ability to model multiple eco-physiological pathways to vegetation hydration. Dynamic weather variations dominate forecasted LFMC temporal trends, over spectral bands. Crucially, regression-to-the-mean effects in the LFMC forecasts caused poorer local performance (median R2≈0.21) compared to global evaluation (R2≈0.69, i.e. poorer within minicubes than amongst minicubes). When converting this hind-cast model to an operational model (i.e. true forecast), best performance could be utilised by running rolling short-term forecasts (5 to 15 days) to capture localised fuel hazard peaks and canopy dieback in the future.
ER -