We present an evaluation of different deep learning and machine learning approaches for tree health classification in the Black Forest, the Harz Mountains, and the Göttinger Forest on a unique, highly accurate tree-level dataset. The multispectral UAV data were collected from eight forest plots with diverse tree species, mostly conifers. As ground truth data (GTD), nearly 1500 tree polygons with related attribute information on the health status of the trees were used. This data were collected during extensive fieldwork using a mobile application and subsequent individual tree segmentation. Extensive preprocessing included normalization, NDVI calculations, data augmentation to deal with the underrepresented classes, and splitting the data into training, validation, and test sets. We conducted several experiments using a classical machine learning approach (random forests), as well as different convolutional neural networks (CNNs)—ResNet50, ResNet101, VGG16, and Inception-v3—on different datasets and classes to evaluate the potential of these algorithms for tree health classification. Our first experiment was a binary classifier of healthy and damaged trees, which did not consider the degree of damage or tree species. The best results of a 0.99 test accuracy and an F1 score of 0.99 were obtained with ResNet50 on four band composites using the red, green, blue, and infrared bands (RGBI images), while VGG16 had the worst performance, with an F1 score of only 0.78. In a second experiment, we also distinguished between coniferous and deciduous trees. The F1 scores ranged from 0.62 to 0.99, with the highest results obtained using ResNet101 on derived vegetation indices using the red edge band of the camera (NDVIre images). Finally, in a third experiment, we aimed at evaluating the degree of damage: healthy, slightly damaged, and medium or heavily damaged trees. Again, ResNet101 had the best performance, this time on RGBI images with a test accuracy of 0.98 and an average F1 score of 0.97. These results highlight the potential of CNNs to handle high-resolution multispectral UAV data for the early detection of damaged trees when good training data are available.

This study evaluated the CO₂ sequestration potential with combustion ashes in the aqueous phase. The aim was to provide a cost-effective carbon sequestration method for combustion unit operators (flue gas cleaning) or biogas producers (biogas upgrading). Therefore, two separate test series were executed to identify the carbonation efficiency (CE) of bottom wood ash (1) at different mixing ratios with water in batch experiments and (2) under dynamic flow conditions. It was furthermore evaluated whether subsequent use of the carbonated wood ash for soil amendment could be possible and whether the process water could be passed into the sewage. The batch test series showed that different mixing ratios of wood ash and water had an influence on the CE. The flow series showed that the mean CE varied between approximately 14% and 17%. Thus, the ash proved to be suitable for carbonation processes. The process water was dischargeable, and the carbonated wood ash has potential for chalking, as no legal thresholds were exceeded. Therefore, wood ash carbonation could be used as a low-tech CO₂ sequestration technology. Compared to existing energy consuming and cost intensive carbon capture and storage technologies, sequestration with ash could be beneficial, as it represents a low-tech approach.

The project PROTECT^FOREST deals with improvements in early bark beetle (e.g., Ips typographus and Pityogenes chalcographus) detection to allow for fast and effective response to initial infestation. The removal of trees in the early infestation stage can prohibit bark beetle population gradation and successive timber price decrease. A semiconductor gas sensor array was tested in the lab and attached to a drone under artificial and real-life field conditions. The sensor array was able to differentiate between α-pinene amounts and between different temperatures under lab conditions. In the field, the sensor responded to a strong artificial α-pinene source. The real-life field trial above a spruce forest showed preliminary results, as technical and environmental conditions compromised a proof of principle. Further research will evaluate the detection rate of infested trees for the new proposed sensor concept.

Water leaching and hydrothermal carbonization are tools to upgrade biomass residues for combustion. Water leaching reduces the ash content, which increases the heating value, while hydrothermal carbonization (HTC) increases heating value and ash content and decreases the organic fraction of the fuel. The objective of this study is to investigate, whether low temperature hydrothermal treatment (HTT) of biomass residues can combine the positive aspects of both treatments and avoid the negative aspects: a decrease of the ash content (leaching) instead of an increase (HTC) and a strong increase of the heating value by deoxygenation (HTC) instead of a low increase by ash reduction (leaching). Sawdust, conifer bark, black locust bark, green cuttings, and leaves were treated at 150 °C, 170 °C and 185 °C in a batch reactor. The impact of the three treatments on combustion properties, inorganic element reduction, and organic compound formation in the liquid phase were investigated. All biomass residues, except sawdust treated at 150 °C, showed increased heating values and reduced ash contents. The intensity of the heating value increases, and the reduction rates of selected ash elements were discussed, including a comparison of literature values and a short review on organic reaction and formation processes at low temperature and pressure conditions.

Forests are increasingly exposed to natural disturbances, including drought, wildfires, pest outbreaks, and windthrow events. Due to prolonged droughts in the last years in Europe, European forest stands significantly lost vitality, and their health condition deteriorated, leading to high mortality rates, especially, but not limited to, Norway spruce. This phenomenon is growing, and new regions are being affected; thus, it is necessary to identify stress in the early stages when actions can be taken to protect the forest and living trees. Current detection methods are based on field walks by forest workers or deploying remote sensing methods for coverage of the larger territory. These methods are based on changes in spectral reflectance that can detect attacks only at an advanced stage after the significant changes in the canopy. An innovative approach appears to be a method based on odor mapping, specifically detecting chemical substances which are present in the forest stands and indicate triggering of constitutive defense of stressed trees. The bark beetle attacking a tree, for example, produces a several times higher amount of defense-related volatile organic compounds. At the same time, the bark beetle has an aggregation pheromone to attract conspecifics to overcome the tree defense by mass attack. These substances can be detected using conventional chemical methods (solid-phase microextraction fibers and cartridges), and it is proven that they are detectable by dogs. The disadvantage of classic chemical analysis methods is the long sampling time in the forest, and at the same time, the results must be analyzed in the laboratory using a gas chromatograph. A potential alternative novel device appears to be an electronic nose, which is designed to detect chemical substances online (for example, dangerous gas leaks or measure concentrations above landfills, volcanic activity, etc.). We tested the possibility of early-stage stress detection in the forest stands using an electronic nose Sniffer4D and compared the individual sensors in it for detecting the presence of attacked and dead trees. Our results indicate the promising applicability of the electronic nose for stress mapping in the forest ecosystem, and more data collection could prove this approach.

European pellet production will be a future challenge due to two effects: (1) the share of hardwood species in Europe will increase and (2) the pellet market will face raw material shortages. Therefore, we investigated the blending of conifer sawdust with black locust sawdust. Twenty-one physical and chemical pellet quality parameters were recorded, including combustion emissions. Our statistical evaluation showed a strong linear correlation (p>0.8 or p<−0.8) of the share of black locust with nine quality parameters. Fifty-three percent of the overall variation in the data was explained by the major principal component, which included the share of black locust. The cause of the decreasing pellet quality with increasing share of black locust sawdust was attributed to the heat conductance in the dye, which was affected by the hydrophobicity and rigidity of the black locust saw dust. A share of 25% black locust in blends with conifer sawdust is proposed as the limit to meet the A2 standard criteria in the European DIN EN ISO 17255-2. This would allow a black locust sawdust consumption of app. 6 mio t per year in Europe, which is far above the estimated abundance and indicates a high potential for hardwood sawdust as an alternative feedstock for pellet production in general.