【Objective】This study aims to explore the influence of unmanned aerial vehicles(UAV) flight altitude mechanism on the accuracy of monitoring larch caterpillar (Dendrolimus superans) insect pests, and provide an important reference for ground UAV remote sensing monitoring of forest pests.【Method】 The areas known for frequent occurrences of D. superans in Da Hinggan Mountains were selected and multispectral remote sensing images collected by UAV at different flight altitudes were used as the basic data. This study obtained the canopy spectral indexes and texture features of 386 healthy, mild, and severely damaged trees by D. superans. Analysis of variances and continuous projection algorithms were used to extract the spectral features sensitive to pest severity. The pest severity monitoring model was constructed using random forest and support vector machine algorithms, and expounded the influence of flight altitude on monitoring accuracy.【Result】(1) The accuracy of overall (mild + severe) monitoring of the spectral indexes and texture features decreased with an increase in flight altitudes. However, the accuracy of mild and severe monitoring of trees damaged by D. superans exhibited different trends. (2) The pest monitoring accuracy of the combination of spectral indices (MTVI 2, GNDVI 2, DVI, GMI 1 and GNDVI) + texture feature (MEA 3) was the best, and the overall accuracy and Kappa coefficient were 92.3% and 0.891, respectively. However, the overall and accuracy of mild monitoring decreased with an increase in flight altitudes, where the decline rate was 0.04%/m and 0.03%/m, respectively, and the accuracy of severe monitoring increased (the rise rate was 0.03%/m). 【Conclusion】 The flight altitudes significantly impacted the accuracy of UAV ground pest monitoring. There was a difference in the rate and trend between the accuracies of mild and severe monitoring. The rate of change in the accuracy of mild monitoring with flight altitude was faster than that of the accuracy of the severe monitoring. Thus, an early identification of pests using a high-precision UAV remote sensing, adaptable to various flight altitudes, is needed to monitor pest severity and improve the expected effects.
【Objective】This study aimed to build a birdsong classification model with strong generalization integrating multi-view features and maximizing feature information to promote profound research on bird species diversity protection and ecological environmentally-intelligent monitoring.【Method】Using 16 types of birdsong audio data as the research objects, the short-time Fourier transform (STFT), wavelet transform (WT) and Hilbert-Huang transform (HHT) feature extraction methods were used to generate three types of birdsong spectrograms to constitute multi-view feature data, and as the input of the convolutional neural network (CNN), the base classifiers STFT-CNN, WT-CNN, and HHT-CNN for different views were trained. The multi-view bagging ensemble convolutional neural network (MVB-CNN) and multi-view stacking ensemble convolutional neural network (MVS-CNN) models were constructed using bagging and stacking integration methods, respectively. With the powerful feature extraction capability of CNN, the multi-view cascaded ensemble convolutional neural network (MVC-CNN) model was proposed to cascade and fuse the deep features extracted from different views through CNN. The classification results were obtained by using a support vector machine (SVM). 【Result】The accuracy rates of the base classification models WT-CNN, STFT-CNN, and HHT-CNN constructed in this study were 89.11%, 88.36%, and 81.00%, respectively; the accuracy rates of the ensemble models MVB-CNN and MVS-CNN were 89.92% and 93.54%, respectively; and the accuracy of the multi-view cascade ensemble model MVC-CNN was 95.76%. The accuracy of the MVC-CNN model improved by 6.65%-14.76% over the single-view-based classification model and by 5.84% and 2.22% over the MVB-CNN and MVS-CNN models, respectively.【Conclusion】The MVC-CNN model proposed in this study fully combined the advantages of multi-view features of birdsong, effectively improving the birdsong classification effects with a greater stability and better generalizational ability, providing a technical solution for multi-view birdsong classification researches.
【Objective】The northeastern China Tiger and Leopard National Park is an important ecological security barrier in the northeast China, due to sensitivity of the vegetation to climate change and anthropogenic activities. This study aimed to explore the spatiotemporal changes in the vegetation of the region to provide a scientific basis for ecological restoration and improvement in the park management system. 【Method】Based on normalized difference vegetation index(MODIS NDVI) data of the growing season (April to October) from 2001 to 2020, Sen+Mann-Kendall trend analysis along with data supplemented by the Google earth engine(GEE) cloud platform, DEM data, meteorological data, land use data, and vegetation type data were used to reveal the spatiotemporal change characteristics of vegetation in the northeast China Tiger and Leopard National Park at different time scales. This considered the time lag effect of different vegetation types on climate change and their differences, as well as partial correlation analysis. Additionally, improved residual and relative role analyses were conducted to quantify the response mechanisms of the vegetation to climate change and anthropogenic activities, and clarify the relative role of climate change and anthropogenic activities in the evolution of vegetation under different conditions.【Result】 (1) Temporally, the growing season NDVI of the park showed a significant increasing trend at a rate of 0.003 2/a in the past 20 a (P<0.05). In different growing seasons, the order of the mean value and rate of increase of NDVI was summer >spring >autumn and spring >autumn >summer, respectively. Spatially, the NDVI trend showed clear seasonal and regional differences from 2001 to 2020; however, the improved area was larger than the overall degraded area. Since the seasonal cycle changes occurred within the growing season, the NDVI improvement area first decreased and then increased, and the main NDVI trend shifted from ‘significant improvement’ to ‘slight improvement’. (2) The spatial heterogeneity of the NDVI of the park in response to climate at different time scales was clear; however, the NDVI was positively correlated with both air temperature and precipitation, and the response of the NDVI to air temperature was stronger than precipitation. The area with a significant positive correlation between the NDVI and air temperature was larger in spring and growing seasons (P<0.05), accounting for 83.558% and 42.241% respectively, while most areas of the park showed no significant positive correlation between the NDVI and precipitation at each time scale (P≥0.05). Different vegetation types had different time lags for the maximum response of the NDVI to air temperature and precipitation, except for cultivated vegetation and meadows, wherein the response lags of other vegetation types were stronger to precipitation than to air temperature. (3) The impact of anthropogenic activity on the NDVI of the park had dual effects, in which the areas that positively promoted and negatively disturbed the NDVI trend accounted for 94.087% and 5.913%, respectively. The implementation of forestry projects was key for increasing NDVI, while the expansion of construction land was the prime factor for increasing NDVI. (4) The NDVI trend of the park was driven by both climate change and anthropogenic activities, but the mean relative role of climate change and anthropogenic activities on the NDVI trend was 32.699% and 67.301%, respectively. The mean relative role of human activities was greater than that of climate change in both the NDVI improved and degraded areas.【Conclusion】 The vegetation status of the northeast China Tiger and Leopard National Park has been generally improved from 2001 to 2020, with significant differences in spatiotemporal trends of vegetation at different time scales. Air temperature was the dominant climatic factor that promoted vegetation growth in the park at each time scale, and the effects of climate change and anthropogenic activities on vegetation changes varied significantly in terms of geography. However, both were dominated by positive effects, with the contribution of anthropogenic activities to the NDVI changes of the park being relatively higher. It is suggested that in the future, in addition to the park vegetation maintenance, efforts should be made to improve the adaptability of vegetation to climate change, along with ensuring continuous implementation of ecological restoration measures and curbing overexploitation of land resources.
【Objective】Analyzing distribution suitability of forest protection tree species in seven coastal cities of Shandong Province can comprehensively reflect the impacts of different site factors on each forest protection tree species, thereby enabling further scientific forest protection tree species planning. 【Method】The distribution suitability of the tree species in protection forests within seven coastal cities of the Shandong Province was analyzed with the LandUSEM model using survey data of ‘One map’ of forest resource management in Shandong Province, as well as the 1∶106 digital soil map and DEM data as auxiliary data. 【Result】There were 26 species which were significantly affected by geomorphological factors. Furthermore, it was found that the soil types were relatively sensitive to the influence of dominant tree species and that the suitability of dominant tree species varied greatly depending on the different forest types. The average relative weights of the seven factors on the 49 dominant tree species were as follows: landform (0.19) > soil type (0.15) = slope (0.15) = soil thickness (0.15) > forest species (0.14) > slope position (0.13) > slope aspect (0.09). In the protection forests dominated by Pinus thunbergii, P. densiflora and Robinia pseudoacacia, the relative suitabilities of the first two species were higher than that of the latter; 90.48% of R. pseudoacacia forests were not suitable or critically suitable. This suggests that Pinus thunbergii, P. tabulaeformis, Quercus acutissima, Crataegus pinnatifida, Eucommia ulmoides, Q. mongolica, Q. spp., Cerasus pseudocerasus, Cerasus pseudocerasus, Ziziphus jujuba var. spinosa and other tree species should be mixed in low mountains and hilly areas where R. pseudoacacia forests have poor distribution suitability to improve the ecological benefits of R. pseudoacacia forests. 【Conclusion】 The distribution suitabilities of tree species in the protective forests within the seven selected cities along the coast of the Shandong Province are generally good, and the LandUSEM model analysis provides reliable information on tree species distribution suitability.
【Objective】 This study provides supports for accurate site quality evaluation and reasonable management plans for Larix kaempferi by studying the effects of genetic and climate changes on plant height. 【Method】 Using height growth data from L. kaempferi trees aged from 5 to 18 years in the Changlinggang Forest Farm in Jianshi County, Hubei Province and the Logistic equation as the basic theoretical growth model, we introduced provenances and climate variables reflecting genetic effects and used repetition as random parameters to construct a tree height growth model. Using this model, we were then able to analyze the effects of genetic and climate changes on tree height growth. 【Result】 Temperature and precipitation were the main climatic factors affecting L. kaempferi tree height growth in this area. Furthermore, the fitting accuracy of the model with temperature and precipitation inputs was higher than that of the basic model. Additionally, nonlinear mixed model based on repetition as a random effect (Radj2=0.820 3) fit better than did the growth model considering genetic and climatic factors (Radj2=0.806 2) and the basic logistic model (Radj2= 0.798 9). The height growth of different provenances was in accordance with the law of ‘slow-fast-slow’ growth; however, the time to reach the fast-growing point varied, and there were significant discrepancies in height growth among different provenances at different time nodes. 【Conclusion】 Genetic and climatic factors affected the height of L. kaempferi. Furthermore, the construction of a mixed model based on genetic and climatic effects can effectively improve the fitting accuracy of a model.
