Forests suffer from heavy losses due to the occurrence of fires. A prediction model based on environmental condition, such as meteorological and vegetation indexes, is considered a promising tool to control forest fires. The construction of prediction models can be challenging due to (i) the requirement of selection of features most relevant to the prediction task, and (ii) heavily imbalanced data distribution where the number of large-scale forest fires is much less than that of small-scale ones. In this paper, we propose a forest fire prediction method that employs a sparse autoencoder-based deep neural network and a novel data balancing procedure. The method was tested on a forest fire dataset collected from the Montesinho Natural Park of Portugal. Compared to the best prediction results of other state-of-the-art methods, the proposed method could predict large-scale forest fires more accurately, and reduces the mean absolute error by 3–19.3 and root mean squared error by 0.95–19.3. The proposed method can better benefit the management of wildland fires in advance and the prevention of serious fire accidents. It is expected that the prediction performance could be further improved if additional information and more data are available.

Studies of local adaptation in populations of chinaberry (Melia azedarach L.) are important for clarifying patterns in the population differentiation of this species across its natural range. M. azedarach is an economically important timber species, and its phenotype is highly variable across its range in China. Here, we collected M. azedarach seeds from 31 populations across its range and conducted a common garden experiment. We studied patterns of genetic differentiation among populations using molecular markers (simple sequence repeats) and data on phenotypic variation in six traits collected over five years. Our sampled populations could be subdivided into two groups based on genetic analyses, as well as patterns of isolation by distance and isolation by environment. Significant differentiation in growth traits was observed among provenances and families within provenances. Geographic distance was significantly correlated with the quantitative genetic differentiation (Q_ST) in height (HEIT) and crown breadth. Climate factors were significantly correlated with the Q_ST for each trait. A total of 23 climatic factors were examined. There was a significant effect of temperature on all traits, and minimum relative humidity had a significant effect on the survival rate over four years. By comparing the neutral genetic differentiation (F_ST) with the Q_ST, the mode of selection acting on survival rate varied, whereas HEIT and the straightness of the main trunk were subject to the same mode of selection. The variation in survival rate was consistent with the variation in genetic differentiation among populations, which was indicative of local adaptation. Overall, our findings provide new insights into the responses of the phenological traits of M. azedarach to changes in the climate conditions of China.

Pruning is a regular and essential urban tree maintenance practice aimed at sustaining overall health, ecosystem services, and public safety. However, knowledge of post-pruning recovery dynamics remains limited, which in turn hinders accurate assessments of growth and ecological functions. To address this, we examined recovery dynamics of Acacia auriculiformis, a common urban species. Tree height and crown radius were recorded monthly for 12 months after pruning. Trees were classified into two size groups based on diameter at breast height (DBH, trunk diameter measured at 1.3 m above ground): medium (DBH < 45 cm) and large (DBH ≥ 45 cm). A generalized linear mixed model (GLMM), appropriate for repeated measures and non-normal data, was fitted using a Tweedie distribution and a log-link function to model the recovery pattern. Results showed continuous growth over time, with medium-sized trees presenting significantly higher crown radius growth than large trees (p = 0.006), while height growth did not differ (p = 0.788). The best model for height included time (AIC = −846.4), whereas crown recovery was best modelled by time and size class (AIC = −1586.6). These findings demonstrate that, in this study, medium-sized A. auriculiformis generally recover faster, especially in crown expansion. This exploratory study suggests that tree size may influence post-pruning recovery and can provide a reference for subsequent differentiated management studies. The morphological modeling further provides preliminary quantitative evidence for annual recovery dynamics in urban A. auriculiformis.