Temperate forests are under climatic and economic pressures. Public bodies, NGOs and the wood industry are looking for accurate, current and affordable data driven solutions to intensify wood production while maintaining or improving long term sustainability of the production, biodiversity, and carbon sequestration. Free tools and open access data have already been exploited to produce accurate quantitative forest parameters maps suitable for policy and operational purposes. These efforts have relied on different data sources, tools, and methods that are tailored for specific forest types and climatic conditions. We hypothesized we could build on these efforts in order to produce a generic method suitable to perform as well or better in a larger range of settings. In this study we focus on building a generic approach to create forest parameters maps and confirm its performance on a test site: a maritime pine (Pinus pinaster) forest located in south west of France. We investigated and assessed options related with the integration of multiple data sources (SAR L- and C-band, optical indexes and spatial texture indexes from Sentinel-1, Sentinel-2 and ALOS-PALSAR-2), feature extraction, feature selection and machine learning techniques. On our test case, we found that the combination of multiple open access data sources has synergistic benefits on the forest parameters estimates. The sensibility analysis shows that all the data participate to the improvements, that reach up to 13.7% when compared to single source estimates. Accuracy of the estimates is as follows: aboveground biomass (AGB) 28% relative RMSE, basal area (BA) 27%, diameter at breast height (DBH) 20%, age 17%, tree density 24%, and height 13%. Forward feature selection and SVR provided the best estimates. Future work will focus on validating this generic approach in different settings. It may prove beneficial to package the method, the tools, and the integration of open access data in order to make spatially accurate and regularly updated forest structure parameters maps effortlessly available to national bodies and forest organizations.

The exploitation of the unprecedented capacity of Sentinel-1 (S1) and Sentinel-2 (S2) data offers new opportunities for crop mapping. In the framework of the SenSAgri project, this work studies the synergy of very high-resolution Sentinel time series to produce accurate early seasonal binary cropland mask and crop type map products. A crop classification processing chain is proposed to address the following: (1) high dimensionality challenges arising from the explosive growth in available satellite observations and (2) the scarcity of training data. The two-fold methodology is based on an S1-S2 classification system combining the so-called soft output predictions of two individually trained classifiers. The performances of the SenSAgri processing chain were assessed over three European test sites characterized by different agricultural systems. A large number of highly diverse and independent data sets were used for validation experiments. The agreement between independent classification algorithms of the Sentinel data was confirmed through different experiments. The presented results assess the interest of decision-level fusion strategies, such as the product of experts. Accurate crop map products were obtained over different countries in the early season with limited training data. The results highlight the benefit of fusion for early crop mapping and the interest of detecting cropland areas before the identification of crop types.

Forests are one of the key elements in ecological transition policies in Europe. Sustainable forest management is needed in order to optimise wood harvesting, while preserving carbon storage, biodiversity and other ecological functions. Forest managers and public bodies need improved and cost-effective forest monitoring tools. Research studies have been carried out to assess the use of optical and radar images for producing forest height or biomass maps. The main limitations are the quantity, quality and representativeness of the reference data for model training. The Global Ecosystem Dynamics Investigation (GEDI) mission (full waveform LiDAR on board the International Space Station) has provided an unprecedented number of forest canopy height samples from 2019. These samples could be used to improve reference datasets. This paper aims to present and validate a method for estimating forest dominant height from open access optical and radar satellite images (Sentinel-1, Sentinel-2 and ALOS-2 PALSAR-2), and then to assess the use of GEDI samples to replace field height measurements in model calibration. Our approach combines satellite image features and dominant height measurements, or GEDI metrics, in a Support Vector Machine regression algorithm, with a feature selection process. The method is tested on mixed uneven-aged broadleaved and coniferous forests in France. Using dominant height measurements for model training, the cross-validation shows 7.3 to 11.6% relative Root Mean Square Error (RMSE) depending on the forest class. When using GEDI height metrics instead of field measurements for model training, errors increase to 12.8–16.7% relative RMSE. This level of error remains satisfactory; the use of GEDI could allow the production of dominant height maps on large areas with better sample representativeness. Future work will focus on confirming these results on new study sites, improving the filtering and processing of GEDI data, and producing height maps at regional or national scale. The resulting maps will help forest managers and public bodies to optimise forest resource inventories, as well as allow scientists to integrate these cartographic data into climate models.

This paper investigates tree species classification using the Sentinel-2 multispectral satellite image time series (SITS). Despite its importance for many applications and users, such mapping is often unavailable or outdated. The value of using SITS to classify tree species on a large scale has been demonstrated in numerous studies. However, many methods proposed in the literature still rely on a standard machine learning algorithm, usually the random forest (RF) algorithm. Our analysis shows that the use of deep learning (DL) models can lead to a significant improvement in classification results, especially in an imbalanced context where the RF algorithm tends to predict the majority class. In our case study in central France with 10 tree species, we obtained an overall accuracy (OA) of around 95% and an F1-macro score of around 80% using three different benchmark DL architectures (fully connected, convolutional, and attention-based networks). In contrast, using the RF algorithm, the OA and F1 scores obtained were 92% and 60%, indicating that the minority classes are poorly classified. Our results also show that DL models are robust to imbalanced data, although small improvements can be obtained by specifically addressing this issue. Validation on independent in situ data shows that all models struggle to predict in areas not well covered by training data, but even in this situation, the RF algorithm is largely outperformed by deep learning models for minority classes. The proposed framework can be easily implemented as a strong baseline, even with a limited amount of reference data.

This article studies the monitoring of oak dieback in forests of the Centre-Val de Loire region (France), where drought-induced dieback has become a major concern due to climate change. The main objective of the study is to evaluate the applicability of multispectral satellite time series for operational monitoring of forest dieback. Using in situ data collected from 2017 to 2022 on approximately 2700 oak plots, a multiyear mapping of the analyzed region was performed using the random forest algorithm and Sentinel-2 images. Our results show that it is possible to detect oak dieback accurately (average overall accuracy = 80% and average balanced accuracy = 79%). A spatial cross-validation analysis also evaluates the performance of the model on regions that were never encountered during training, across all years, resulting in a slight decrease in accuracy ( $\sim$ 5%). The study also highlights the importance of measuring the stability and performance of the classification model over time, in addition to standard cross-validation metrics. A feature analysis shows that the shortwave infrared part of the spectrum is the most important for mapping forest dieback, while the red-edge portion of the spectrum can increase the stability of the model over time. Overall, both in situ data and model predictions showed evidence of forest decline in many areas of the study region. Our results suggest that large areas of forest can decline over short periods of time, highlighting the interest of satellite data to provide timely and accurate information on forest status.