China has a long tradition of managing planted forests. Different species of Populus, Eucalyptus, Larix, Cunninghamia and Pinus are planted to satisfy the local demand for wood products and provide ecological services at the same time. Evidence of the greater resilience of natural forests provides the motivation to develop asymmetric planting patterns, which is the focus of this study. We present a new method for designing plantation patterns that follow those observed in natural ecosystems and to maintain some regularity for operational convenience. Based on the uniform angle index, we analyzed the spatial structure of six natural forests in different regions of China. The uniform angle index describes the degree of spatial uniformity of the n nearest neighbors of a given reference tree. Accordingly, we identified all possible patterns of a neighborhood group within a regular planting pattern and developed a method to optimize planting point arrangements that contain some randomness as well as a minimum degree of regularity. (1) There are 13 types of structural units in a regular planting, including seven random units, five even units and one cluster unit; (2) Five near-natural arrangements are presented with a minimum proportion of 50% of random units. These five arrangements represent a combination of regularity for operational convenience and asymmetry. The new planting patterns developed in this study are expected to increase the asymmetric competition and resilience of these important ecosystems. Some experimental plantings, based on our findings, have already been established, e.g., in Pinus tabulaeformis plantations in Tianshui, Gansu Province, and in a Populus deltoides plantation in Fangshan near Beijing.

Negative conspecific density dependence is one of the principal mechanisms affecting plant performance and community spatial patterns. Although many studies identified the prevalence of density dependent effects in various vegetation types by analyzing conspecific spatial dispersal patterns (spatial patterning) of forest trees, interactions between individuals and heterospecific neighboring trees caused by density-dependent effects are often neglected. The effects of negative density dependence lead us to expect that neighbourhood species segregation would increase with increasing tree size and that larger trees would be surrounded by more heterospecific neighbours than would smaller trees. We studied four mapped 1-Ha plots on Changbaishan Mountain in North-eastern China and used marked point pattern analysis to explore whether trees of different sizes exhibited differences in neighbourhood species segregation; we also determined whether larger trees were more likely to have heterospecific neighbours than smaller trees were. Our results show that bigger trees generally have higher species mingling levels. Neighborhood species segregation ranged from lower than expected levels to random or nearly random patterns at small scales as tree size classes increased under heterogeneous Poisson null model tests. This study provides some evidence in support of negative density dependent effects in temperate forests.

One of the aims of sustainable forest management is to preserve the diversity and resilience of ecosystems. Unfortunately, changes in the soil microbial communities after forest management remain unclear. We analyzed and compared the soil microbial community of a natural Quercus aliena var. acuteserrata forest after four years of four different management methods using high-throughput sequencing technology. The forest management methods were close-to-nature management (CNFM), structure-based forest management (SBFM), secondary forest comprehensive silviculture (SFCS) and unmanaged control (CK). The results showed that: (1) the soil microbial community diversity indices were not significantly different among the different management methods. (2) The relative abundance of Proteobacteria in the SBFM treatment was lower than in the CK treatment, while the relative abundance of Acidobacteria in the SBFM was significantly higher than that in the CK treatment. The relative abundance of Ascomycota was highest in the CNFM treatment, and that of Basidiomycota was lowest in the CNFM treatment. However, the relative abundance of dominant bacterial and fungal phyla was not significantly different in CK and SFCS. (3) Redundancy analysis (RDA) showed that the soil organic matter (SOM), total nitrogen (TN), and available nitrogen (AN) significantly correlated with the bacterial communities, and the available potassium (AK) was the only soil nutrient, which significantly correlated with the composition of the fungal communities. The short-term SBFM treatment altered microbial bacterial community compositions, which may be attributed to the phyla present (e.g., Proteobacteria and Acidobacteria), and the short-term CNFM treatment altered microbial fungal community compositions, which may be attributed to the phyla present (e.g., Ascomycota and Basidiomycota). Furthermore, soil nutrients could affect the dominant soil microbial communities, and its influence was greater on the bacterial community than on the fungal community.

Forest structure is largely determined by nearest neighbor interactions, and reveals detailed interpretations of the current situation and development potential of the stand. A structure-based method is needed to analyze the relationship of nearest neighbor tree groups and therefore to characterize the forest spatial heterogeneity in a comprehensively and systematically way. A natural mixed forest of Quercus aliena var. acutiserrata and Pinus tabulaeformis in the Xiaolong Mountains, China was taken as an example to demonstrate the telescope method (i.e. N-variate distributions). Each tree with DBH ≥ 5 cm was investigated and located, and the frequencies of N-variate distributions flexibly combined with four structure parameters, uniform angle index, mingling, dominance and crowding were calculated using Excel pivot tables and the Winkelmass software. The telescope method could systematically interpret the forest structural characteristics at different resolutions. Especially, the quadrivariate distribution provides the most detailed and comprehensive spatial structure information. Based on the vertical projection dimension reduction and marginal probability distribution function, the continuous recursive process well revealed the inherently quantitative relationships among different distributions. This could be conducive to flexibly optimizing and reconstructing forest structure, and effectively selecting the trees to be removed.

Patterns of tree distribution are an important attribute of forest structure and directly affect the health and stability of forest ecosystems. This paper studied the effects of forest management on the forest distribution pattern of Larix kaempferi plantations with an aim to improve the quality of the spatial structure of plantations by providing scientific grounds for near-nature distribution pattern adjustment. We set up 15 long-term positioning and monitoring plots of 20 m × 20 m of L. kaempferi plantations in Xiaolongshan, Gansu Province, China, based on the Clark and Evans index, and applied the structure analysis method of the neighboring tree relationship to determine the standard angle of the uniform angle index of regularly distributed plantation forests. The changes in forest distribution patterns were compared before and after randomized management (R1 (dumbbell-shaped random unit), R2 (torch-shaped random unit) and R1:R2 = 1:2 models) and underlayer tending and unmanaged control (CK). The results showed that (1) Under different management modes, the distribution patterns obtained by the Clark and Evans index and the mean values of the uniform angle index were entirely the same; all stand patterns were evenly distributed. (2) Regardless of randomized management or underlayer tending, the number of trees in the even distribution was the largest, accounting for more than 40%. (3) The mean values of the uniform angle index of most stands increased obviously after randomized management, among which the R1 and R2 stands showed an upward trend, while the R1:R2 = 1:2 stands showed a slight decrease. However, the mean values of the uniform angle index of the stands with underlayer management showed a decreasing trend. (4) The proportion of random trees in most stands increased significantly after randomized management, increasing by 1%~19% compared with that before adjustment, while the proportion of random trees in the underlayer tending treatment decreased by 2%~10%. Our results suggest that (1) The method based on the modified uniform angle index standard angle could effectively improve the accuracy of the distribution pattern of individual trees in plantation forests and further improve the accuracy of the overall stand pattern quantitatively. (2) Randomized management significantly increased the proportion of random trees in the stands, which accelerated the evolution of the forest distribution pattern of the L. kaempferi plantation to a random distribution.