Objective: The acquisition of complete and valid flux data as well as the analysis of the energy balance of an ecosystem is the basis for further understanding of the energy and material exchange processes in an ecosystem. Method: In this paper, the R language REddyProc package was selected to interpolate the flux data measured by the Eddy covariance technique. The interpolated data were evaluated using the cross-validation method and the energy balance closure. At the same time, the terms of the thermal storage in the energy balance were corrected to analyze the energy balance closure of jujube forests in the arid zone and the trend of each energy component. Result: 1) The REddyProc package could interpolate the data better with the error statistic parameters of the regression coefficient b=0.86-0.99, determination coefficient R²=0.86-0.95, index of agreement d=0.96-0.98, efficiency of simulation EF=0.84-0.92, and RMSE-observations standard deviation ratio RSR=0.28-0.40. 2) After conducting the R package interpolation, the energy balance closures of the jujube date forests were 73.45% and 73.11%, respectively. R² of the 2-year effective energy (H+LE) and available energy (R_n－G) both were 0.97. After adding the thermal storage terms, the energy balance closures increased to 77.17% (by 3.72%) and 75.86% (by 2.75%), respectively. The thermal storage terms affected some improvements in the energy balance closure of jujube forests. 3) The daily-averaged variation pattern of each energy component (the net radiation R_n, latent heat flux LE, sensible heat flux H, and soil heat flux G) was similar during the fertility and dormancy with a single-peaked daily variation according to the variation pattern of the net radiation. Conclusion: This paper analyzed the energy exchange patterns of Jujube forest ecosystems in arid areas. The year-round energy balance closure of jujube forests was in line with the range of ChinaFlux and the thermal storage term affected some improvements in the energy balance closure of jujube forests. The achievements can provide a theoretical basis for studying the energy and material cycles in jujube forests.

Objective: This study aims to construct stand canopy closure estimation models based on the extracted feature variables from the UAV-LiDAR point cloud data and the measured data of the sample plots, which would provide basic data and technical reference for rapid and accurate estimation of artificial forest canopy closure. Method: Taking the Urban Forestry Demonstration Base of Northeast Forestry University as the study area, the inversion of the forest closure for the artificial coniferous forests and broad-leaved forests was carried out based on the point cloud data obtained by multi-rotor UAV LiDAR. The height, intensity and canopy characteristic variables were calculated according to the three-dimensional coordinates and energy values of the point cloud, and the principal component analysis method was used to reduce the data dimension. Then, a stepwise regression procedure was conducted on the canopy closure obtained by the plant canopy analyzer and the processed variables, and the canopy closure estimation models for artificial coniferous forests and broad-leaved forests were established, respectively. Finally, the estimation models and inverse distance weight interpolation method were applied on the ArcGIS platform to perfom canopy closure inversion mapping. Result: Canopy characteristic variables had the most significant effect on the canopy closure estimation accuracy of coniferous forests, while intensity characteristic variables had the most significant effect on the estimation accuracy of broad-leaved forests. The canopy closure estimation accuracy of artificial broad-leaved forests (Adj R² = 0.725, RMSE = 0.005 ) was superior to that of artificial coniferous forests (Adj R² = 0.722, RMSE = 0.007 ). The canopy closure of the entire plot estimated by the estimation models and inverse distance weighted interpolation method ranged between 0.81 and 0.87, and the canopy closure of 10 testing points showed a high correlation with the measured canopy closure (r = 0.859). Conclusion: Combining multiple sets of LiDAR feature variables to estimate forest canopy closure can fully exploit the canopy structure characteristics contained in LiDAR data and improve the estimation accuracy.

Objective: To determine the effect of initial planting density on the dominant height growth of Chinese fir plantations and its effective range for indicating site quality, as well as to provide a scientific basis for the evaluation of dominant height growth and effective density regulation of Chinese fir plantations, the impact of initial planting density on the entire growth process of dominant height of Chinese fir plantations in various production areas was explored. Method: The eastern region in the middle subtropics (Fenyi, Jiangxi) and the central region in the middle subtropics of Chinese fir (Naxi, Sichuan) were used as the research objects to analyze the response dynamics of dominant height to initial planting density of Chinese fir plantations in different production areas and different stand growth stages. A total of 30 long-term positioning observation plots of Chinese fir density test forests (ranging from 2 to 30 years old) were used. Result: Dominant height was negatively correlated with the initial planting density, the lower the initial planting density was, the higher the dominant height became. In the early stage of stand, the growth trajectory of dominant height with different initial planting densities almost overlapped, and the difference among the densities was small. With the increase of stand age, the difference of dominant height growth among initial planting densities increased. At the mature stage, the effects of initial planting densities on dominant height varied with distribution areas, which was still obvious in eastern region in the middle subtropics, but gradually weakened in central region in the middle subtropics. The mean annual increment and current annual increment of dominant height were negatively correlated with the initial planting density, and the higher the initial planting density was, the smaller the peak reached and the earlier the peak appeared. Under the same stand age, compared with the central region in the middle subtropics, the eastern region in the middle subtropics with the overall better site condition had a higher differentiation degree of dominant height by the effect of initial planting density. Initial planting density had no significant effect on the dominant height growth in the range of low and middle initial planting densities (1 667–3 333 trees·hm^?2) or high initial planting densities (6 667–10 000 trees·hm^?2) after 5 years in eastern region in the middle subtropics and during 10–19 years in central region in the middle subtropics. When the initial planting density reached or exceeded 5 000 trees·hm^?2, the initial planting density had a significant restriction effect on the dominant height growth, compared with the relatively low initial planting densities (1 667–3 333 trees·hm^?2). Conclusion: With various initial planting densities, the growth process of dominant height in stands varies dramatically. The initial planting density has no significant impact on the dominant height growth of the forest within a specific range of initial planting density variations. When the initial planting density exceeds a certain range, there is a significant difference between high density and low density. At the standard age of 20 years, the maximum difference value of dominant height is 2.46 m among the density combinations with significant differences in eastern region in the middle subtropics, which exceeds one site index class. At this time, it is not accurate to use dominant height to indicate site quality. The influence of density should be corrected for the site index prediction of the stand with initial planting density higher than 5 000 trees·hm^?2.

Objective: Urban riparian is an important ecotone between the city land and aquatic zone, serving a variety of ecological and aesthetic functions, which need to be protected, and its landscape needs to be optimized. This study analyzed the effects of color on urban riparian landscape and identified the key factors to its visual attraction, in order to provide a reference for effectively improving the visual attractiveness of urban riparian vegetation color view. Method: In this study, Chaobai River, the most important river flowing through the Beijing-Tianjin-Hebei metropolitan region, was targeted and its 1 000 m urban artificial riparian segment was selected as the research subject. A total of 46 images of 23 plots were taken by a drone in summer and autumn. The 63 color categories were defined by H (Hue) for (25, 45] (45, 80] (80, 140] (140, 220] (220, 290] (290, 345] (345, 25], normalized V (Value) and S (Saturation) for [0, 0.33], (0.33, 0.67], (0.67, 1]. According to this, 38 744 color patches of 25 color categories were extracted through a self-made program. The color patch characteristics were obtained, and their distances to the visual hotspot resulting from the eye movement experiment with 30 volunteers were used to measure the view attraction. At last, a linear mixed regression model was used to analyze the influence of color characteristics on visual attraction. Result: 1) A total of 25 color types of 4 categories, including 4 types of orange, 4 types of yellow, 8 types of yellow-green, and 9 types of green were found in urban riparian. Among them, the green and yellow-green categories accounted for more than 80% of total area proportions. 2) Visual attraction varied in different values, saturation, or hues. Color patches with high saturation belonging to (0.67, 1] or with medium value belonging to (0.33, 0.67] tend to have higher visual attraction. Meanwhile, there was a significant seasonal difference in the visual attraction of color patches, with autumn having a higher visual attraction than summer. 3) Color brightness and saturation had significant positive effects on visual attraction. Hue significantly impacted visual attraction through the interaction with color patches’ area. Moreover, this interaction varied in diverse hues. The more increase of color patch area, the higher the visual attraction only when the hue was within (80, 140]. Conversely, the visual attraction became lower when the hue range resulted in yellow-green, yellow, and orange categories. Conclusion: The gradient colorful scenery from green to yellow-green is emphasized to gain attention as a more common urban riparian vegetation landscape. Distinguishing between the base background color and foreground color is helpful to understand the visual attraction’ differentiation cross color attributes and season. Based on the influence of the color patch’s attribute and area on visual attraction, possible optimization suggestions through the color patch scale for visual attraction are proposed.

Objective: The aim of this paper is to explore the changes of cell structure and photosynthetic pigments during the formation of colorful leaves of Sasaella glabra ‘Albostriata’, as well as the relationship among leaf color, photosynthetic pigment and cell structure in different areas, so as to provide reference for clarifying the formation mechanism of striped leaves. Method: At different development stages of S. glabra ‘Albostriata’, the changes of the internal microstructure of the leaves were observed through scanning electron microscopy. The leaves were divided into yellow and green areas according to their colors, free hand sections were made, and the internal structures of the yellow and green areas of the leaves were compared under an optical microscope. Simultaneously, the contents of photosynthetic pigments and chlorophyll synthesis precursors of the leaves at each stage were determined. Result: There were significant differences in microstructure, ultrastructure, photosynthetic pigment content and chlorophyll synthesis precursor content in the leaves of S. glabra ‘Albostriata’ at different development stages, and there were also significant differences in the leaf structure and photosynthetic pigment in different zones of mature functional leaves. 1) During the development process of leaves, epidermal cells and mesophyll cells gradually increased, and the leaf vein area increased. 2）During the developmental process, chloroplasts in mesophyll cells in green stripe area developed normally and their internal structures were intact. The number of chloroplasts and grana lamellae in chloroplasts increased gradually, and their inner arrangement tended to be tight; While the differentiation degree of mesophyll cells in yellow district was relatively low. The layers of arm cells was fewer, and the mesophyll cells had large cavities and abnormal chloroplasts. 3) The leaf epidermis structure was similar in the yellow and green stripes, but the distribution position of chloroplasts in the mesophyll cells was obviously different between yellow and green stripes. Chloroplasts were distributed in all mesophyll cells of the green stripe, while they were only distributed in the mesophyll cells near the lower epidermis in the yellow stripe. 4) The content of photosynthetic pigment gradually accumulated during the development process of leaves, and the pigment content in green stripes was significantly higher than that in yellow stripes. 5) Urogen III and Coprogen III are key substances in the chlorophyll synthesis pathway during the development of bamboo leaves. With the appearance of stripes, the transformation from Urogen III to Coprogen III increased, while the conversion from Coprogen III to Proto IX was reduced. Conclusion: During the development process of bamboo leaves, the leaf thickness gradually increases, the number of epidermal cells, mesophyll cells and veins increase, and the photosynthetic pigment content accumulates. For the green stripe, the number of chloroplasts increases gradually and chlorophyll content accumulates normally with the leaf development. While the mesophyll cells in yellow stripe have large cavities and abnormal chloroplasts, with fewer layers of finger like arm cells. In general, the formation of striped leaf of S. glabra ‘Albostriata’ is related to low differentiation of arm cells, abnormal chloroplast development and obstruction of chlorophyll synthesis.

Objective: This study aims to investigate the preparation technique and analyze the functional components of instant tea from Forsythia suspensa leaves, in order to provide a new idea and scientific basis for high value-added utilization of forsythia resources. Method: The water extraction process of F. suspensa leaves was optimized with the yield of theinstant tea as an indicator. Three kinds of instant tea powder, FSL-D, FSL-G and FSL-B, were prepared from dried, fried, and fermented leaves of F. suspensa, respectively. The physical and chemical properties and functional components were tested to further identify the optimal raw materials for preparing the instant tea. Results: 1) The optimal extraction process for instant tea from F. suspensa leaves was solid-liquid ratio of 1∶20.14 g·mL^?1, extraction temperature of 90 ℃ and extraction time of 180 min, under which the average yield of instant tea from F. suspensa leaves was 15.30%. 2) The dissolving time of FSL-D, FSL-G and FSL-B was 50.04, 55.19 and 46.46 s, respectively, which all meet the industry standard. The hygroscopic property test and sensory evaluation indicated that FSL-B was less prone to agglomerate during storage and had a better taste. Active components analysis indicated that the protein content, the total phenolic content, the total flavonoids content and the total polysaccharides content of the three kinds of instant tea were all above 11%, 3.22%, 4.44%, and 47.76%, respectively. The content of forsythia glycoside A in FSL-G was relatively high, at 19.96%. 3) The antioxidant experiment of polysaccharides from F. suspensa leaves showed that the larger the molecular weight of the polysaccharides, the better the antioxidant effect. The infrared spectrum test showed that the antioxidant activity was related to the weakening of O-H stretching vibration peak, C=O bond stretching vibration peak, and C-H bending vibration peak activity. Conclusion: The high quality instant tea (FSL-B) can be obtained from the fermented F. suspensa leaves. This research not only provides theoretical basis for the study of forsythia leaf products, but also improves the economic value of forsythia, which is of great significance to promote the development of forsythia economic forest.

Objective: Acer griseum is a plant native to China that is endangered and also an excellent ornamental tree species. However, the seeds of A. griseum have deep physiological dormancy, which hinders its population restoration and production application. This study aims to better understand the seed dormancy mechanism of A. griseum and to find ways to alleviate dormancy, so as to provide reference for the research of seed dormancy mechanism of A. griseum and other plants and production practice. Method: The germination process of A. griseum seeds in natural condition was dynamically observed to study the natural dormancy duration. Low temperature stratification, detachment of embryos, and different concentrations of gibberellin solution were conducted to explore methods to break the dormancy of A. griseum seeds. Wheat seeds were treated with extracts from various parts of dormant A. griseum seeds, and the germination vigor, germination rate, and germination index of wheat seeds in each treatment group were analyzed to identify the location of substances that inhibited the germination of A. griseum seeds. Finally, metabolomics techniques were used to explore the specific substances that could affect the dormancy of A. griseum seeds. Result: 1) In natural condition, the dormancy duration of A. griseum seeds was approximately 29 months. Low temperature stratification for 3–12 months, detachment of embryos, and 200, 400, and 800 mg·L^–1 concentrations of gibberellin solution did not break the dormancy of A. griseum seeds. 2) Wheat seeds treated with extracts from the wings and outer seed coat of dormant A. griseum seeds had higher germination vigor and germination rate than those in the control group. Wheat seeds treated with extracts from the embryos of dormant A. griseum seeds had significantly lower germination vigor, germination rate, and germination index than those in the control group (P < 0.01). The germination vigor, germination rate, and germination index of wheat seeds treated with extracts from the dormancy-broken A. griseum seed embryos had no differences compared to those in the control group. 3) The metabolomics analysis showed that there were 289 and 215 significantly different metabolites in the dormant and dormancy-broken A. griseum seed embryos, respectively. These significantly different metabolites were matched through secondary spectra, with 146 in the positive ion mode and 109 in the negative ion mode; Among the 20 metabolites with the highest multiple of difference (|log₂FC|), 19 showed a significant decrease in content after seed dormancy was broken, mainly distributed in fatty acyl groups and benzene and its substituted derivatives. A class of significant differential metabolites, phenolic phosphates, were identified through a first-order spectra. These metabolites included fatty acyls and phenolic compounds, and they played a key role in affecting the seed dormancy of A. griseum. The substances that inhibited the seed germination of A. griseum were fatty acyl 2-hydroxyoctanoic acid, jasmonic acid, trans-2-octenoic acid, and 1, 11-undecanedicarboxylic acid of the fatty acyls and methylgingerol, xanthurenic acid, and methyl vanillate of the phenols. Conclusion: The dormancy of A. griseum seeds belongs to deep physiological embryo dormancy, with a dormancy period of about 29 months. The seed wings, outer seed coat, and inner seed coat of A. griseum have no inhibitory effect on its germination, and the seed wings and outer seed coat contain substances that promote seed germination. The substances that inhibit the germination of A. griseum seeds exist in the embryo. The dormancy of A. griseum seeds is not caused by a single endogenous substance, but may be the result of the interaction or superposition of multiple substances. Breaking the dormancy of A. griseum seeds using gibberellin may not be effective, and more research is needed to find effective methods to break the dormancy.

Objective: The paper aims to screen SSR loci with high polymorphism and strong universality in Camellia, evaluate the variety discrimination power (VDP) of different SSR loci combinations and determine the minimum numbers of SSR loci in combination for variety discrimination at different Camellia groups, so as to improve the efficiency of variety discrimination in practice. Method: In this study, a total of 111 Camellia varieties were used to screen and evaluate SSR loci and the variety discrimination power of combinations of SSR loci by combining the methods of gel electrophoresis, fluorescent capillary electrophoresis, loci polymorphism analysis and loci combination variety discrimination power (VDP) analysis. Result: A total of 17 SSR loci were screened and obtained, accounting for 30.9% of the total number of selected loci. In 111 Camellia varieties, 166 alleles were detected with the 17 SSR loci. The number of genotypes per locus ranged from 6 to 35, with an average of 23.353. The observed number of alleles (N_a) ranged from 3 to 15, with a mean of 9.765. The polymorphic information content (PIC) ranged from 0.447 to 0.835, with a mean of 0.696. The range of missing data rate of loci was 0–4.50%, with a mean of 0.90%. When the number of differential loci=1, the maximum R-VDP of 17 SSR loci in all varieties (clones), Sect. Camellia varieties and Sect. Chrysantha clones were all 1, and the number of SSR loci was able to be reduced to 11, 10 and 5, respectively accounting for 64.7%, 58.8% and 29.4% of the screened loci in the different groups. When the number of differential loci=2, the maximum R-VDP values of 17 SSR loci in all varieties (clones), Sect. Camellia varieties and Sect. Chrysantha clones were 0.964, 0.952 and 1.000, respectively, and the number of SSR loci was able to be reduced to 12, 12 and 7, respectively accounting for 70.6%, 70.6% and 41.2% of the screened loci in the different groups. When the number of differential loci=3, the maximum R-VDP values of 17 SSR loci in all varieties (clones), Sect. Camellia varieties and Sect. Chrysantha clones were 0.946, 0.929, and 1.000, respectively, and the number of SSR loci was able to be reduced to 8, 7 and 8, respectively, accounting for 47.1%, 41.2% and 47.1% of the screened loci in the different groups. When the number of differential loci was ≤ 6, the maximum R-VDP of 17 SSR loci combinations for the tested varieties (clones) of Camellia could maintain a relatively stable level. Conclusion: In this study, 17 SSR loci with high polymorphism and strong universality have been screened and verified. The combinations of these loci show well variety discrimination power in the tested Camellia groups. The minimum numbers of SSR loci in combinations for variety discrimination have been determined in different variety (clones) groups of Camellia, and a set of analyses and evaluation system for variety discrimination power of combinations of SSR loci in Camellia have been constructed. These results provide useful information and tools for variety identification and development of related standards.

Objective: In recent year, a leaf gall inducer has been found on Eucalyptus trees in China. It produces small herpetiform galls on leaves and causes a large number leaves fell. In this study, we classified and identified the invasive pest, and determined the type of pest in order to provide reference for the prevention and control of this invasive pest. Method: Totally 120 leaves from Eucalyptus variety Leiling No. 1 with galls were collected in Hepu County, Beihai City, Guangxi, in January and February 2023, and incubated in two transparent plastic baskets. The small wasps that emerged from galls were collected and counted daily, as well as identified. Result: Totally 198 small wasps that form galls were raised. After taxonomic study, the pest was identified as Ophelimus maskelli (Ashmead) (Hymenoptera: Chalcidoidea: Eulophidae). This was the first discovery of the pest in China. The species is native to Eucalyptus trees in New Zealand and Australia. A detailed description of the pest was given in the present paper with color photos, as well as a diagnosis method was provided for distinguishing the species from other species in this genus. Conclusion: The pest O. maskelli found in Guangxi could induce galls on leaves of many Eucalyptus species, causing a large number of leaves shedding and affecting the growth of Eucalyptus trees, and now has become a severe invasive pest in China. The more attention should be paid for controlling the new pest. Fortunately the pest parasitoid Closterocerus chamaeleon (Girault) (Hymenoptera: Eulophidae) was also found during the present investigation. The parasitoid has high parasitism and could be used for biological control of the pest.

Objective: In this study, an assessment framework for forest fire spread risk (including occurrence hazard, spread hazard, and spread risk) was established, to carry out multi-scale (subcompartment, village, township) forest fire spread risk assessment in Yanqing District of Beijing based on GIS technology. The study aims to provide a basis for forest fire management in Yanqing District, which could minimize the occurrence hazard of forest fire and the loss caused by the spread of forest fire. Method: Based on the subcompartment data of the 2019 forest resources second-class survey, forest fuel data, population data, 30-meter resolution digital elevation data and meteorological data, the forest fire occurrence hazard assessment index system was constructed from two aspects: difficulty in wildfire sources management and ease degree of ignition. The forest fire spread hazard assessment index system was constructed from four aspects of forest fuel, meteorological factor, topographic factor and fire behavior. The subjective and objective weighting method of AHP and entropy weight method were used to determine the index weight, and the weighted comprehensive evaluation method and the area weighted average method were used to calculate the multi-scale occurrence hazard and spread hazard index. According to the natural break point classification method, the grade is divided. Based on the risk assessment framework of “hazard-exposure-vulnerability”, the forest fire spread risk was evaluated from two aspects of destroying forest resources and threatening life safety. Result: 1) The comprehensive weight of difficulty in wildfire sources management reached 0.771 3, which was the main factor affecting the forest fire occurrence hazard. The comprehensive weight of fire behavior was 0.403 2, which was the main factor affecting the forest fire spread hazard. 2) The forest fire occurrence hazard in Yanqing district gradually decreased from plain to mountainous area, and the high occurrence hazard areas accounted for 0.89%, 0.52% and 7.48% in subcompartment, village and township scales, respectively, mainly distributed in the plain area around the urban area. The spread hazard gradually increased from plain to mountainous area, and the high spread hazard areas accounted for 46.19%, 58.73% and 35.67% respectively in subcompartment, village and township scales, mainly distributed in two western forest farms (Badaling Forest Farm and Songshan Forest Farm) and two eastern mountainous towns (Qianjiadian Town and Zhenzhuquan Township). 3) The forest fire spread risk (forest resource) in Yanqing district was not extremely high risk area in subcompartment, village and township scales, and the medium-high risk areas accounted for 0.80%, 2.85% and 1.79% in subcompartment, village and township scales, respectively. The forest fire spread risk (population) was had no extremely high or medium-high risk areas in village and township scales. There were no extremely high risk areas in subcompartment, only 9 medium-high risk areas in subcompartment, and the medium risk areas accounted for 34.97%, 51.19% and 42.25% in subcompartment, village and township scales, respectively. Conclusion: The high occurrence hazard subcompartments were mainly in Poplar, Robinia pseudoacacia and Willow plantations. The high spread hazard subcompartments were mainly in Quercus mongolica natural forest, Pinus tabuliformis plantations and Platycladus orientalis plantations. The medium-high spread risk (forest resource) subcompartments were mainly in Pinus tabuliformis and Larch plantations. The medium spread risk (population) areas were mainly distributed in four eastern mountainous towns (Zhenzhuquan Township, Qianjiadian Town, Liubinbao Township and Sihai Town).

Objective: This study aims to document differences in movement pattern and habitat use between high water years and normal water years of Siberian Cranes wintering in Poyang Lake, to benefit crane conservation and adaptive habitat management. Method: Water level data of Xingzi Hydrological Station and tuber biomass of food plant Vallisneria spp. for Siberian Cranes were used to define the different water years. The years of 2016 and 2017, and 2020 and 2021 were defined as high water years, and the first two years were referred as group-W1 and the later as group-W3. While, the years of 2018 and 2019 were defined as normal years or referring as group-W2. Thirteen Siberian Cranes were tracked with satellite transmitters in Poyang Lake from 2016 to 2022. Differences were determined and analyzed with ArcGIS 10.7 in movement distances, home range areas and habitat use of Siberian Cranes between the high water and the normal groups. Result: 1) The Siberian Cranes in the whole winter and in mid-winter moved significantly less during normal years (W2) than those of high water years W1, or W3, or combined. 2) 90% KDE in the three winter periods during the high water years (W1 and W3) were significantly larger than those in the normal years (W2). In mid-winter, home ranges of 90% and 50%KDE in the high water years W1 was significantly larger than those in the normal years (W2) and high water years W3. 3) In normal years (W2), the overall use of wetland in daytime accounts for 89.0% ± 12.3% (n = 9). The use of farmland by the cranes in the whole winter, and in early and middle winters in high water years were significantly higher than those in normal water years (W2). The wetland use in high water years was significantly lower than in normal years, in spite of accounting for 50.4% ± 20.5% (n = 15). There was no significant difference in farmland use or wetland use by the cranes among groups in high water years. In normal years (W2), the overall use of wetland in nighttime accounts for 94.9% ± 8.7% (n = 9); In high water years, the overall use of wetland in night time about 84.5% ± 18.4%（n = 15）. The use of farmland by the cranes in middle winters in high water years (W1 and combined) were significantly higher than those in normal water years (W2). Conclusion: In high water years when natural food shortage occurs, the Siberian Cranes can cope with the impact through increasing moving distance, enlarging home range and utilizing farmland more, to meet their energy need and improve their fitness. Nonetheless, the natural wetland is still the main foraging habitat and roosting habitat for Siberian Cranes wintering at Poyang Lake either in normal water years. In high water years, even the overall use of farmland in day time accounted for 50%, the natural wetland is still the main roosting habitat. The adaptive management of natural wetlands therefore should be strengthened in order to improve the function and resilience of wetlands as whole at Poyang Lake under different hydrological conditions and ensure the health and safety of 98% of global wintering population of Siberian Cranes.

Objective: In the wild environment, the background of wildlife images captured by camera traps is complex, which poses a challenge for identifying wild animals in images with a large number of images and a wide variety of wildlife species. Based on convolutional neural network, this research aims to improve the existing structure and so as to implement the automatic recognition for wildlife images. Method: In this study, 2 712 wildlife images of 8 categories were taken from Wuling Mountain Beijing Nature Reserve, Miyun Districts, Beijing. The Auto Augment policy was randomly selected from 14 augmentation policies to add noise to the images. SENet and BlurPool were used to construct an improved network based on ResNeXt-50: SE-ResNeXt-50 for enhancement feature extraction, BP-ResNeXt-50 for Shift-invariance maintenance, and BS-ResNeXt-50 for both. The influences of fixed learning rate, segmented learning rate, and cosine annealing learning rate on the accuracy of the BS ResNeXt-50 network were tested on the self-built dataset. VGG16, ResNeXt-50, EfficientNet-B0, InceptionV3, DenseNet-121, and BS-ResNeXt-50 were used to train on 16 common categories of images in CCT public wildlife dataset, and the recognition accuracy of single species was compared.e influences of fixed learning rate, segmented learning rate, and cosine annealing learning rate on the accuracy of the BS ResNeXt-50 network were tested on the self-built dataset. VGG16, ResNeXt-50, EfficientNet-B0, InceptionV3, DenseNet-121, and BS-ResNeXt-50 were used to train on 16 common categories of images in CCT public wildlife dataset, and the recognition accuracy of single species was compared.eXt-50 is used to test influence of different learning rate include fixed and CosineAnnealing learning rate on collected dataset. VGG16, ResNeXt-50, EfficientNet-B0, InceptionV3, DenseNet-121, BS-ResNeXt-50 were used for training on CCT dataset, and the recognition accuracy of single species was compared. on ResNeXt-50: SE-ResNeXt-50 for enhancement feature extraction, BP-ResNeXt-50 for Shift-invariance maintenance, and BS-ResNeXt-50 for both. The influences of fixed learning rate, segmented learning rate, and cosine annealing learning rate on the accuracy of the BS ResNeXt-50 network were tested on the self-built dataset. VGG16, ResNeXt-50, EfficientNet-B0, InceptionV3, DenseNet-121, and BS-ResNeXt-50 were used to train on 16 common categories of images in CCT public wildlife dataset, and the recognition accuracy of single species was compared.eXt-50 is used to test influence of different learning rate include fixed and CosineAnnealing learning rate on collected dataset. VGG16, ResNeXt-50, EfficientNet-B0, InceptionV3, DenseNet-121, BS-ResNeXt-50 were used for training on CCT dataset, and the recognition accuracy of single species was compared. Result: The accuracy of SE-ResNeXt-50 and BP-ResNeXt-50 reached 75.16%±0.14% and 73.74%±0.13%, respectively. The enhanced scheme BS-ResNeXt-50, which integrated SENet and BlurPool, achieved an accuracy of 78.04%±0.11% when tested on a self built dataset, which was the best improved scheme. When the cosine annealing learning rate is used, the accuracy of BS-ResNeXt-50 was improved to 81.54%, which was 3.5% higher than that with the constant learning rate. The step decay learning rate achieved 79.3% accuracy, which was 2.24% less than the cosine annealing learning rate. The classification accuracy of BS-ResNeXt-50 was able to reach 95.07%, which was 1.95% higher than that of ResNeXt-50 on CCT dataset. At the same time, it was also 85.5% higher than that of VGG16, 91.38% higher than that of EfficientNet-B0, 91.38% higher than that of InceptionV3 and 93.3% higher than that of DenseNet-121. The prediction accuracy of each single category was also higher than that of the above model. In the recognition of a single category, except for the least one category, the accuracy of BS-ResNeXt-50 was 90% higher than that in other categories, and the highest category accuracy was 98.6%. Conclusion: The BS-ResNeXt-50 can more accurately complete the recognition task, and also has good generalization ability on different datasets.

Objective: This paper aims to investigate the relationship of the oleoresin yield and monoterpenoid composition of Pinus elliottii × P. caribaea between different locations and different stimulating pastes, and explores a mild, effective and cost-efficient stimulating pastes suitable for the hybrid pine in Guangdong. Method: Experiments with different types of stimulating pastes, such as potassium sulfate(K₂SO₄), benzoic acid mixture(BA) and ethylene glycol(CEPA), were carried out in hybrid pine plantations suitable for oleoresin tapping at three sites in Guangdong, Haifeng, Shixing and Yangjiang, respectively. The tree diameter at 1.7 m height and oleoresin yield by hybrid pine was measured per individual at three sites and four treatments, and the rate of increase of oleoresin was estimated by each stimulating paste. The samples of oleoresin were collected for monoterpene composition identification and their relative contents were identified by the GC-MS. Result: 1) Compared with the control group, K₂SO₄, benzoic acid mixture and CEPA stimulating pastes had significantly increased oleoresin yield, with CEPA having the most stable yield increase, K₂SO₄ the second and benzoic acid mixture the worst. In inter-site comparisons, Haifeng had the best yield increase with stimulate paste, Yangjiang the second and Shixing the worst. 2) The oleoresin yield of hybrid pine was positively correlated with the tree diameter at 1.7 m. When compared to the control groups, the correlation coefficient between the oleoresin yield and the tree diameter at 1.7 m was increased in the treated groups regardless of the type of stimulating pastes used at Haifeng and Yangjiang sites, while decreased at Shixing. 3) Composed with control group, the use of stimulate paste did not change the main pine oleoresin monoterpene composition and their relative content of hybrid pine, though all of the three treatment groups could generate α-Terpineol. Likewise, there were significant differences in the relative content of individual pine oleoresin monoterpene components, namely α-pinene, β-pinene, α-phellandrene, β-phellandrene and β-myrcene, in hybrid pine from different sites, and there is no detection of 3-carene. Conclusion: In Guangdong, CEPA, K₂SO₄ or benzoic acid mixture can be used to increase yield with the premise of quality unaffected. Although CEPA has the best effect on yield, the cost of raw materials is higher than K₂SO₄ and the effect is only slightly higher than K₂SO₄ at the plantations in Shixing and Yangjiang, while K₂SO₄ is the best in Haifeng. Therefore, K₂SO₄ is the most suitable stimulating paste for the hybrid pine in Guangdong Province. The proper use of stimulate paste can stimulate the oleoresin production potential of hybrid pine without affecting the monoterpene composition and content ratio of hybrid pine oleoresin.

Objective: In order to solve the problems of strong subjectivity and low efficiency that the identification of furniture wood species mainly depends on manual identification in daily life, a common furniture wood species identification model based on Mobilenetv3 convolutional neural network (CNN) was designed to effectively improve the identification speed and accuracy of wood species. It provided a scientific and effective method for the rational utilization of wood resources, the management of wood import and export trade and the determination of furniture wood types by consumers. Method: Firstly, 3 880 images of four kinds of wood images coated without wood wax oil and two kinds of wood images coated with wood wax oil were collected. The data set was divided into training set, verification set and test set according to 6∶2∶2 and 2∶6∶2, and the data of the training set was expanded to four times of the original by using operations such as rotation and flip. Then four convolution neural networks and two traditional machine learning methods were used to establish the recognition model for the wood image without coated wood wax oil. Through analysis and comparison, the optimal recognition network model Mobilenetv3 was obtained, and the parameters of the model were optimized based on transfer learning. The wood species identification model based on MobileNetv3 network was constructed by putting the wood images coated without wood wax oil together with the wood images coated with wood wax oil and forming a new data set together with the remaining two kinds of wood images coated without wood wax oil. Finally, in order to simplify the operation of the classification personnel and reduce the operation difficulty of the staff in the actual detection, we selected the above wood species recognition model and built a wood species recognition system based on PyQt5. Result: Compared with traditional machine learning methods, convolutional neural network had better recognition results, and transfer learning could significantly improve the convergence speed and classification performance of the network. In the validation set, MobileNetV3 with the best recognition performance had an average recognition accuracy of 98.13% for four wood images without coated wood wax oil and 97.25% for wood images mixed with coated wood wax oil. Conclusion: In this paper, the wood species recognition model not only recognized the recognition of wood images coated without wood wax oil, but also realized the fast and accurate recognition of the painted wood wax wood image. Apart from bring convenience to wood classifiers, the wood species recognition model also could provide a reliable identification method for consumers to select solid wood furniture and determine wood species.