The use of native species in forest restoration has been increasingly recognized as an effective means of restoring ecosystem functions and biodiversity to degraded areas across the world. However, successful selection of species adapted to local conditions requires specific knowledge which is often lacking, especially in developing countries. In order to scale up forest restoration, experimental data on the responses of native species to propagation and restoration treatments across a range of local conditions are required. In this study, the restoration potential of 34 native tree species was evaluated based on nursery research and field planting experiments at a highly degraded site in a subtropical area of southwest China. We examined species performance in terms of germination rates as well as survival rates and growth over 2 years after planting. Of the 34 species examined, 25 had a germination percentage greater than 50%. Survivorship ranged from 0 to 97% across species and was greater than 50% for 20 species. Mean monthly growth increments varied between species. Pioneer species performed well, and 14 mid- and late-successional species performed reasonably well to very well in this study. However, the remaining 16 mid- and late-successional species performed poorly. These results indicate that carefully selected mid- and late-successional species can be effectively incorporated into mixed species plantings. This data can be used to inform restoration planning, helping to identify suitable species and so enhance the biodiversity and resilience of restored forests.

Livestock rearing is one of the oldest and most important types of smallholder farming worldwide. The sustainability of livestock production depends on the efficient utilization of locally available resources. Some traditional methods of raising livestock may offer valuable lessons in this regard. This study documented and evaluated local knowledge of wild forage plants in the Dulongjiang area in Southwest China in the context of rearing mithun (Bos frontalis) in order to provide a sound evidence base for tree fodder selection and the establishment of integrated tree-crop-livestock systems. The snowball technique was used to identify key informants with specific knowledge about the topic. Free listing and semi-structured interviews were conducted with 58 households. Participatory investigation and transit walks were used to investigate potential fodder species. Ethnobotanical information was collected, documented and organized. Overall, 142 wild forage plants from 58 families and 117 genera were identified. Species of the Poaceae, Rosaceae and Urticaceae families were most abundant, with 16, 14 and 11 species respectively identified as fodder plants. Our results indicated that tree/shrub forage plays a major role in the diet of mithun, unlike that of other ruminants. Mithun prefers to browse and move around the forest in search of food, particularly rough and even barbed leaves. Tree species like Debregeasia orientalis, Saurauia polyneura and Rubus species were identified as being important fodder sources. Farmers in this area have traditionally relied on common property resources such as community-managed forests and grasslands to feed their livestock. Farmers have strong incentive to raise mithuns rather than other livestock species due to Dulong people’s cultural preferences. The wide variety of plants cited by the informants demonstrate the importance of traditional knowledge in gathering information about forage resources. This diversity also offers the prospect of identifying promising species which could be used as fodder plants. Identifying such species and tree fodder species in particular could help smallholder farmers to integrate trees, livestock and crops as part of a sustainable farming system.

Surface defect inspection plays an important role in the process of industrial manufacture and production. Though Convolutional Neural Network (CNN) based defect inspection methods have made huge leaps, they still confront a lot of challenges such as defect scale variation, complex background, low contrast, and so on. To address these issues, we propose a joint attention-guided feature fusion network (JAFFNet) for saliency detection of surface defects based on the encoder-decoder network. JAFFNet mainly incorporates a joint attention-guided feature fusion (JAFF) module into decoding stages to adaptively fuse low-level and high-level features. The JAFF module learns to emphasize defect features and suppress background noise during feature fusion, which is beneficial for detecting low-contrast defects. In addition, JAFFNet introduces a dense receptive field (DRF) module following the encoder to capture features with rich context information, which helps detect defects of different scales. The JAFF module mainly utilizes a learned joint channel-spatial attention map provided by high-level semantic features to guide feature fusion. The attention map makes the model pay more attention to defect features. The DRF module utilizes a sequence of multi-receptive-field (MRF) units with each taking as inputs all the preceding MRF feature maps and the original input. The obtained DRF features capture rich context information with a large range of receptive fields. Extensive experiments conducted on SD-saliency-900, Magnetic tile, and DAGM 2007 indicate that our method achieves promising performance in comparison with other state-of-the-art methods. Meanwhile, our method reaches a real-time defect detection speed of 66 FPS.

The plant species in the mountainous regions might be relatively more vulnerable to climate change. Understanding the potential effects of climate change on keystone species, such as Rhododendron species in the subalpine and alpine ecosystems, is critically important for montane ecosystems management and conservation. In this study, we used the maximum entropy (MaxEnt) model, 53 distribution records, and 22 environmental variables to predict the potential impacts of climate change on the distribution of the endemic and vulnerable species Rhododendron purdomii in China. The main environmental variables affecting the habitat suitability of R. purdomii were altitude, temperature seasonality, annual precipitation, slope, and isothermality. Our results found suitable distribution areas of R. purdomii concentrated continuously in the Qinling-Daba Mountains of Central China under different climate scenarios, indicating that these areas could potentially be long-term climate refugia for this species. The suitable distribution areas of R. purdomii will expand under the SSP126 (2070s), SSP585 (2050s), and SSP585 (2070s) scenarios, but may be negatively influenced under the SSP126 (2050s) scenario. Moreover, the potential distribution changes of R. purdomii showed the pattern of northward shift and west–east migration in response to climate change, and were mainly limited to the marginal areas of species distribution. Finally, conservation strategies, such as habitat protection and assisted migration, are recommended. Our findings will shed light on biotic responses to climate change in the Qinling-Daba Mountains region and provide guidance for the effective conservation of other endangered tree species.

The infrageneric taxonomy system, species delimitation, and interspecies systematic relationships of Leontopodium remain controversial and complex. However, only a few studies have focused on the molecular phylogeny of this genus. In this study, the characteristics of 43 chloroplast genomes of Leontopodium and its closely related genera were analyzed. Phylogenetic relationships were inferred based on chloroplast genomes and nuclear ribosomal DNA (nrDNA). Finally, together with the morphological characteristics, the relationships within Leontopodium were identified and discussed. The results showed that the chloroplast genomes of Filago, Gamochaeta, and Leontopodium were well-conserved in terms of gene number, gene order, and GC content. The most remarkable differences among the three genera were the length of the complete chloroplast genome, large single-copy region, small single-copy region, and inverted repeat region. In addition, the chloroplast genome structure of Leontopodium exhibited high consistency and was obviously different from that of Filago and Gamochaeta in some regions, such as matk, trnK (UUU)-rps16, petN-psbM, and trnE (UUC)-rpoB. All the phylogenetic trees indicated that Leontopodium was monophyletic. Except for the subgeneric level, our molecular phylogenetic results were inconsistent with the previous taxonomic system, which was based on morphological characteristics. Nevertheless, we found that the characteristics of the leaf base, stem types, and carpopodium base were phylogenetically correlated and may have potential value in the taxonomic study of Leontopodium. In the phylogenetic trees inferred using complete chloroplast genomes, the subgen. Leontopodium was divided into two clades (Clades 1 and 2), with most species in Clade 1 having herbaceous stems, amplexicaul, or sheathed leaves, and constricted carpopodium; most species in Clade 2 had woody stems, not amplexicaul and sheathed leaves, and not constricted carpopodium.