In recent years, forest fires have become an important issue in Central Europe. To model the probability of the occurrence of forest fires in the Lower Silesian Voivodeship of Poland, historical fire data and several types of predictors were collected or generated, including topographic, vegetation, climatic, and anthropogenic features. The main objectives of this study were to determine the importance of the predictors of forest fire occurrence and to map the probability of forest fire occurrence. The H2O driverless artificial intelligence (DAI) cloud platform was used to model forest fire probability. The gradient boosted machine (GBM) and random forest (RF) methods were applied to assess the probability of forest fire occurrence. Evaluation the importance of the variables was performed using the H2O platform permutation method. The most important variables were the presence of coniferous forest and the distance to agricultural land according to the GBM and RF methods, respectively. Model validation was conducted using receiver operating characteristic (ROC) analysis. The areas under the curve (AUCs) of the ROC plots from the GBM and RF models were 83.3% and 81.3%, respectively. Based on the results obtained, the GBM model can be recommended for the mapping of forest fire occurrence in the study area.

Forest fires are becoming a serious concern in Central European countries such as Austria (AT) and the Czech Republic (CZ). Mapping fire ignition probabilities across countries can be a useful tool for fire risk mitigation. This study was conducted to: (i) evaluate the contribution of the variables obtained from open-source datasets (i.e., MODIS, OpenStreetMap, and WorldClim) for modeling fire ignition probability at the country level; and (ii) investigate how well the Random Forest (RF) method performs from one country to another. The importance of the predictors was evaluated using the Gini impurity method, and RF was evaluated using the ROC-AUC and confusion matrix. The most important variables were the topographic wetness index in the AT model and slope in the CZ model. The AUC values in the validation sets were 0.848 (AT model) and 0.717 (CZ model). When the respective models were applied to the entire dataset, they achieved 82.5% (AT model) and 66.4% (CZ model) accuracy. Cross-comparison revealed that the CZ model may be successfully applied to the AT dataset (AUC = 0.808, Acc = 82.5%), while the AT model showed poor explanatory power when applied to the CZ dataset (AUC = 0.582, Acc = 13.6%). Our study provides insights into the effect of the accuracy and completeness of open-source data on the reliability of national-level forest fire probability assessment.

The goal of this study was to develop a GIS-based Decision Support Model for selecting the best timber harvesting systems on steep terrain. The model combines multiple layers, each representing an important factor in mechanized logging. These layers are used to create a final map that functions as a spatially explicit Decision Support Model that helps decide which machines are best suited for different forest areas. A key idea of this study is to consider not only operational criteria (slope, ruggedness, wetness, and road accessibility), but also a fundamental silvicultural aspect, i.e., the assessment of tree growth classes to enable the integration of silvicultural deliberations into timber harvest planning. The data used for this model come from orthophoto image and a Digital Terrain Model (DTM). The operational factors were analyzed using GIS tools, while the silvicultural aspects were assessed using the deep learning algorithm DeepForest and tree growth equations (allometric functions). The model was tested by comparing its results with field data taken in a Norway Spruce stand in South Tyrol/Italy. The findings show that the model reliably evaluates operational factors. For silvicultural aspects, it tends to underestimate the number of small trees, but provides a good representation of tree size classes within a forest stand. The innovation of this method is that it relies on low-cost, open-source tools instead of expensive 3D scanning devices.

Pest resistance of trees should be taken into account in future forest strategy planning and predicting risks of defoliation. The gypsy moth (GM) (Lymantria dispar L.) is a serious forest pest with outbreaking population dynamics. To estimate defoliation risk of the most widely distributed tree species in Balkan Peninsula and Europe (Turkey oak Quercus cerris L., European beech Fagus sylvatica L. and hornbeam Carpinus betulus L.), we carried out laboratory feeding trials and investigated their acceptability and suitability for GM development. We determined morphological and chemical attributes of these hosts as well as larval host preference, growth and nutritional indices. Preference, growth, and efficiency of food conversion into biomass were ranked in the order: Turkey oak ˃ European beech ˃ hornbeam. Hornbeam was the most avoided and showed the lowest conversion efficiency although, comparing to optimal oak host, its leaves were less tough, contained more water and exhibited similar values of nitrogen (index of protein content) and C/N ratio (index of investment into carbon based plant defense). We suggest that hornbeam and beech leaf chemical profiles should be further studied to reveal specific compounds that impose high metabolic cost to GM larvae. Moreover, additional research are needed to understand how intermediate hosts in natural populations affect GM outbreaks.

The objective of this study is to create a GIS-based decision support model for steep terrain timber harvesting. The tool integrates multiple raster layers, each representing a critical aspect of mechanized timber harvesting, and uses them to generate a composite map that aids in decision-making about suitable areas for specific machine configurations. The innovation lies in two pivotal aspects: first, the tool integrates silvicultural considerations alongside operational criteria; second, it employs a new methodology that factors in these silvicultural criteria without depending on resource-intensive point cloud data. Instead, the methodology integrates aerial imagery in the form of an orthofoto, and the deep-learning algorithm DeepForest for the evaluation of forest stand characteristics. The procedure was tested in South Tyrol, Italy, using remotely sensed data and validated against ground-based inventory data. While the results indicate an underestimation of stem numbers and a tendency to underrate small tree dimensions, overall performance statistics are promising. This makes the procedure a valuable tool for forest management planning, effectively integrating operational and silvicultural considerations.