[Age estimation model for individual tree in natural Larix gmelinii forest based on random forest model]. / åŸºäºŽéšæœºæ£®æž—çš„å ´å®‰è½å¶æ¾å¤©ç„¶æž—å•æœ¨å¹´é¾„é¢„ä¼°æ¨¡åž‹.

To accurately estimate the age of individual tree and to achieve full-cycle sustainable management of natural Larix gmelinii forest in Great Xing'an Mountains of northeastern China, we constructed individual tree age prediction model using stepwise regression and random forest algorithms based on 44 fixed plots data and 280 stan-dard tree cores obtained from the Pangu Forest Farm. We analyzed the influence of stand structure, site conditions, and competition index on the accuracy of model prediction. The model was evaluated by the coefficient of determination (R2), root mean square error (RMSE), and mean absolute error (MAE). The results showed that the random forest model had the highest prediction accuracy when number of decision trees was 1500 and number of node con-tention variables was 8. The random forest model had better accuracy and prediction ability than the stepwise regression model, with R2, RMSE and MAE of 0.5882, 9.9259 a, 8.1155 a. Diameter at breast height was the most important factor affecting age prediction (83.8%), followed by tree height (34.4%), elevation (17.9%), and basal area per hectare (17.5%). The random forest algorithm exhibited better adaptability and modeling effect on constructing a predictive model for individual tree age. This research contributed to improving the accuracy of growth and harvest estimation for L. gmelinii, and could provide a reference for other scientific studies related to tree age estimation in forests.

Height-diameter models of regenerated saplings of Larix gmelinii based on dummy variable and quantile regression.

Based on data collected from 2054 saplings of Larix gmelinii forest in 55 fixed plots in 2018-2019 in Cuigang Forestry Station, Daxing'anling area, we classified the stand density index (SDI) into four classes, i.e., Class Ⅰ (SDI1<1863 plants·hm-2), Class Ⅱ (1863 plants·hm-2≤SDI2<2155 plants·hm-2), Class Ⅲ (2155 plants·hm-2≤SDI3<2459 plants·hm-2) and Class Ⅳ (SDI4≥2459 plants·hm-2) by using the quartile method. We constructed a dummy variable model and quantile regression model for the height-breast diameter of saplings of L. gmelinii with dummy variable method introduced SDI. The results showed that among the five selected representative non-linear tree height curve models, the Richards model fitted the best, with Ra2, RMSE and MAE of 0.7637, 0.8250 m and 0.5696 m. The dummy variable model including the SDI constructed based on the Richards model showed a 1.3% increase in Ra2 compared with the base model, while RMSE, MAE, and AIC decreased by 2.1%, 1.5%, and 11.2%, respectively. When the quantile τ was 0.5, Ra2 of quantile regression model was the maximum, and RMSE, MAE, AIC was the minimum, being 0.7612, 0.8294 m, 0.5657 m, and -767.19, respectively. Compared with SDI1, sapling height in SDI2-SDI4 was increased by 5.6%, 5.6%, and 11.3%, suggesting reasonable that regulation of stand density was conducive to increase the height growth of saplings in regeneration.

Compatibility predictive model for regeneration quantities of Larix gmelinii natural forest in Daxing'anling Mountains, China.

The natural regeneration grade is an important foundation for formulating forest management measures. Traditional studies have only considered the regeneration quantities predictive model of the total stand or dominant tree species, but the consistency among the prediction results of different tree species and the total regeneration quantities of stand is not solved. That is, the regeneration prediction results at the stand level are not equal to the sum of the predicted results of all tree species. To address this, on the basis of the traditional counting model, we attempted to construct a compatibility predictive model for regeneration quantities of different tree species within the stand, which would provide a theoretical basis for the rational management and decision-making of natural forest. Based on the survey data from 96 standard plots of Cuigang Forest Farm, Xinlin Forest Farm, and Zhuangzhi Forest Farm in Daxing'an Mountains, we selected 30 basic indices from five aspects of site factor, soil factor, stand factor, tree diversity and stand spatial structure, and used Poisson model and negative binomial model as the basic models to construct the regeneration prediction models of Larix gmelinii, Betula platyphylla and other tree species. By comparing the accuracy and fitting effect of the two traditional counting models, we selected the optimal model and used the seemingly unrelated regressions to further construct the compatibility predictive model for regeneration quantities of different tree species. Poisson model was the best one for the regeneration of L. gmelinii, B. platyphylla, and other tree species. The test index RMSE of the compatibility predictive model for regeneration quantities of L. gmelinii, B. platyphylla, other tree species and total stand regeneration quantities were 388, 413, 504, and 871 trees·hm-2, respectively. The adjusted R2 was 0.389, 0.421, 0.488, and 0.407, respectively. The most influential variables for regeneration quantities of L. gmelinii, B. platyphylla and other tree species were Pielou evenness index of DBH (25.2%), herbal coverage (34.6%) and organic matter in B layer (23.2%). In this study, the compatibility predictive model system for regeneration quantities satisfied the additive logic among L. gmelinii, B. platyphylla, other tree species, and total stands, and provided a basis for accurately estimating natural regeneration.

[Structure and dynamics of co-dominant species in different succession stages of natural forests in Daxing'an Mountains, China]. / å¤§å ´å®‰å²­å¤©ç„¶æž—ä¸åŒæ¼”æ›¿é˜¶æ®µå ±ä¼˜åŠ¿ç§ç§ç¾¤ç»“æž„ä¸ŽåŠ¨æ€.

Affected by the disturbance of forest fire and logging, the primary forest in Daxing'an Mountains gradually degenerates into secondary forest. In this study, we established 16 plots in each of three typical forests, including natural Betula platyphylla pure forest (pioneer stage), natural B. platyphylla and Larix gmelinii mixed forest (transition stage) and natural L. gmelinii pure forest (top stage). The methods of population age and tree height structure, static life table, survival analysis, dynamic index and time series prediction were used to quantitatively analyze the dynamics of dominant species (B. platyphylla and L. gmelinii) and all the arbors, aiming to provide scientific basis for the restoration and development of natural L. gmelinii forest. The results showed that the abundance of young co-dominant species and total arbors in each stage was large, and that all population had strong self-renewal potential. With the progress of succession, the abundance of B. platyphylla in each age class gradually decreased, whereas that of L. gmelinii gradually increased. The mortality and disappearance rates of total arbors and B. platyphylla in the transition stage and L. gmelinii in the pioneer stage gradually increased with the increases of age class, and the survival curve was Deevey-Ⅰ type. The survival analysis results showed that the population was stable in the early stage, increased in the middle stage, and declined in the later stage. In other stages, the mortality rates fluctuated slightly, the survival curves were Deevey-Ⅱ type, and the population increased in the early stage, declined in the middle stage, and stable in the later stage. The co-dominant species and total arbors were growing in the three succession stages, among which B. platyphylla in the pioneer stage, L. gmelinii and total arbors in the top stage showed the lowest sensitivity to the environment. The results of time series prediction showed that the co-dominant species and total arbors in each stage would increase in the future. During forest succession, it was necessary to strengthen the protection of seedlings and young trees, thin the forest with large coverage, and take appropriate measures to ensure population renewal.

[Simulation of carbon sequestration and timber production in Larix olgensis plantation based on CO2FIX model]. / 基于CO2FIX模型的长白落叶松人工林碳汇和木材生产模拟.

Based on datasets from plot survey and bibliographic of Larix olgensis plantations in Maoer Mountain, the CO2FIX model was used to quantitatively simulate the effects of different rotations (30, 40, 50, 60 years), site indices (12, 16, 20 m), and initial densities (2500, 3333, 4444 trees·hm-2) on the stand level carbon flows among different carbon pools (i.e., biomass carbon pool, soil carbon pool, and product carbon pool). The results showed that the CO2FIX model had high reliability for simulating the processes of L. olgensis plantation, with the average relative errors of stand biomass and volume between analog and measured values being 6.4% and 3.7%. Under the baseline conditions of initial density of 3333 trees·hm-2, site index of 16 m and rotation of 40 years, the carbon stock of total and sub-pool of L. olgensis plantation changed periodically with rotation. The total stand carbon stock and volume for L. olgensis plantation increased with the extension of rotation, the improvement of site index, and the increase of initial density. The stand carbon stock and volume would be increased by 12.2% and 31.2%, 36.7% and 67.8%, respectively, when the reference rotation was correspondingly extended by 10 and 20 years. However, if the reference rotation was shortened by 10 years, stand carbon stock and volume would be correspondingly decreased by 20.9% and 40.4%, respectively. When the initial density was set as 3333 and 4444 trees·hm-2, stand carbon stock and volume were increased by 27.8% and 50.9%, 27.4% and 49.1%, respectively. When the site index was under the range of 12 to 20 m, stand carbon stock and volume could be increased by 36.0% and 40.3%, 39.3% and 44.2%, respectively, with each increase of 4 m in site index. During one rotation, 271.57 t C per hectare could be fixed into L. olgensis plantation. At the end round of the rotation, 27.47 and 56.75 t C were transferred to soil and wood product carbon pools. Therefore, when the site condition was good, the management model with a higher initial density (4444 trees·hm-2) and longer rotation (60 years) would be more beneficial to maximizing the carbon sink and timber benefits of L. olgensis plantation.

[Classification of forest types in Cuigang Forest Farm based on time series data of Landsat 8]. / 基于Landsat 8æ—¶é—´åºåˆ—æ•°æ®çš„ç¿ å²—æž—åœºæ£®æž—ç±»åž‹åˆ’åˆ†.

To explore the practical role of enhanced vegetation index (EVI) time series data in improving the accuracy of forest type recognition could promote the deep application of optical remote sensing data in forest resources investigation and monitoring. With Cuigang Forest Farm of Xinlin Forestry Bureau in Daxing'anling as the object, we constructed six classification schemes, using random forest algorithm with spectral feature, texture feature and EVI time series feature. The data sources were 20-view Landsat 8 OLI time series data from 2014 to 2018, 56 fixed plots data from 2017-2019, and the 2016 Class II survey data. Our aims were to realize the classification of forest types in Cuigang Forest Farm and to evaluate the accuracy of different classification schemes. The results showed the EVI values of Larix gmelinii forest, Betula platyphylla forest, coniferous-broadleaved mixed forest, coniferous mixed forest and broadleaved mixed forest were significantly different in non-growing seasons (36-111 days and 287-367 days), with the EVI value of mixed conifer forest being significantly higher, and that of mixed broadleaf forest being always lower than the other four forest types. In the early growing season (111-143 days), the EVI value of B. platyphylla forest were higher than L. gmelinii forest, which could effectively distinguish the two forests. Among the six classification schemes, spectral feature, texture feature, and EVI time series feature had the highest classification accuracy, with a Kappa of 0.82 and a classification accuracy of 86.1%. The comparison results showed that the overall accuracy of adding vegetation index time series feature was improved by 14.3% compared with that of spectral feature. The random forest algorithm with combined spectral, texture and EVI time series features could effectively classify forest stand types in Cuigang Forest Farm, with good recognition accuracy and confidence.

[Clarifying the factors affecting Larix gmelinii forest regeneration based on structural equation model]. / åŸºäºŽç»“æž„æ–¹ç¨‹æ¨¡åž‹çš„å ´å®‰è½å¶æ¾å¤©ç„¶æž—æ›´æ–°å½±å“å› ç´ .

Based on data from 49 plots of natural Larix gmelinii forests in Cuigang Forest Farm of Xinlin Forestry Bureau, Daxing'anling Mountains, China, we used 37 measurable variables that mainly focused on stand non-spatial structure, stand spatial structure, species diversity, soil condition, and site condition to construct the structural equation model of natural regeneration densities and size diversities (i.e., height and ground-diameter). The direct, indirect, and total influence coefficients of each path were quantified to extract the critical and controllable factors that influence regeneration density and diversity of natural L. gmelinii forests, which would help implement sustainable forest management. The results showed that the effects of various latent variables on rege-neration density were following an order as: stand non-spatial structure (-0.410) > species diversity (0.380) > soil condition (0.250) > site condition (0.249) > stand spatial structure (0.197), while the order were changed as: soil condition (0.778) > site condition (0.748) > stand spatial structure (0.684) > stand non-spatial structure (0.287) > forest diversity (0.105), when evaluated on the regeneration diversity. Generally, the critical and controllable factors affecting rege-neration quantity and diversity were soil pH, total potassium concentration, species diversity, tree height diversity, uniform angle index and stand volume per hectare. In the management, suitable thinning treatments or replanting broadleaved trees were recommended for optimizing and adjusting species composition, species diversity, soil pH and nutrition, which would promote natural regene-ration.

[Spatial distribution pattern and scale effect of secondary forests in Daxing'anling, China]. / å¤§å ´å®‰å²­æ¬¡ç”Ÿæž—ç©ºé—´åˆ†å¸ƒæ ¼å±€åŠå ¶å°ºåº¦æ•ˆåº”.

We examined the spatial distribution patterns and their scale effects of different tree species (Larix gmelinii, Betula platyphylla and others) and different size classes of trees (1-5) of natural L. gmelinii secondary forest (LF), natural B. platyphylla secondary forest (BF) and the mixed secondary forest of both species (MF) in Daxing'anling. The results showed that among the three forest types, LF was the only one type reaching a good state of regeneration, while other two forest types were poorly regenerated. For different forest types, the abundance of seedlings and saplings in the regeneration layer were significantly different from that of the tree layer, and the diameter distribution (except for LF and BF) and height distribution of trees in each forest type were not reasonable, indicating that all the three forest types belonged to unstable communities. At species level, the spatial distributions of main species in each plot were mainly clumped. The five indicators used in this study varied significantly with the scales, which mainly focused on the linear increases (40%), the power increases (22%) and the negative quadratic polynomials (20%), respectively. For different size classes, significant clumped distributions were observed for the regeneration levels (1-3), while the spatial distribution of tree layers (4-5) usually fluctuated distinctly among various distribution patterns. The scale effects of different size classes were mainly dominated by the linear increases (44%), the power increases (15%) and the negative quadratic polynomials (12%). For each forest type and sampling scale, the cluster degrees of trees decreased significantly with increasing tree sizes. Within each forest type, the pattern size of non-dominant species was significantly larger than that of dominant species, while the pattern size of regeneration layers was significantly larger than that of tree layers.

[Seed rain dynamics of natural secondary forest in the central part of Greater Xing'an Mountains, China.] / å¤§å ´å®‰å²­ä¸­éƒ¨å¤©ç„¶æ¬¡ç”Ÿæž—ç§å­é›¨åŠ¨æ€.

We monitored seed rain in three forest types for two years (2018-2019) in the central part of Greater Xing'an Mountains, and analyzed the seasonal seed rain dynamics, deciduous dynamics, seed rain thousand-grain weight, inter-annual variation of seed rain, and spatial pattern of seed rain of the main tree species in different forest types. The results showed that seed rain of Larix gmelinii and Betula platyphylla in each forest type showed unimodal distribution. The leaf litter amount of coniferous and broad-leaved species (L. gmelinii, Pinus sylvestris var. mongolica, Picea koraiensis, B. platyphylla and Populus davidiana) also showed obvious seasonal dynamics, with a peak in the middle and early September. In coniferous and broadleaved mixed forests and coniferous forests, the 1000-grain weight of L. gmelinii at the peak period was significantly greater than that at the initial and final stages. The 1000-grain weight of B. platyphylla in the three forest types did not show obvious seasonal variation. Seed rain of L. gmelinii and B. platyphylla showed obvious inter-annual change. The year 2018 was a large seed set year, whereas 2019 was a year with small seed set. The spatial pattern of all seed rain was mainly dominated by aggregated distribution in both years, which was consistent with that of seedlings and saplings.

[Effects of selective cutting on the distribution pattern and inter-specific association of arbor seedlings in Tilia amurensis secondary forest]. / æ‹©ä¼å¯¹ç´«æ¤´æ¬¡ç”Ÿæž—ä¹”æœ¨å¹¼è‹—åˆ†å¸ƒæ ¼å±€åŠç§é—´å ³è”æ€§çš„å½±å“.

The distribution pattern and interspecific associations of arbor seedlings were examined in two Tilia amurensis forest stands in Zhangguangcailing, with one without man-made interference (natural recovery secondary forest) and the other one being 20-year-old after selective cutting (selective cutting secondary forest). The results showed that the importance value of the seedlings of the soft broad-leaved (heliophile) and tolerant tree species was 7.9 and 64.5 in natural recovery secondary forest, and was 3.9 and 68.9 in selective cutting secondary forest, respectively. The spatial distribution pattern of arbor seedlings in natural recovery and selective cutting secondary forests showed clustered distribution at the scale of 0-18 m and 0-15 m, respectively. Most of tree species pairs were negatively correlated at large scale (14-45 m) in the natural recovery secondary forest, whereas tree species pairs at all the scales were mostly uncorrelated in the selective cutting secon-dary forest. Most of the tree pairs of T. amurensis seedlings (57%) with other tree species showed negatively correlation at large scale (31-45 m). In the natural recovery secondary forest, the proportion of negatively correlated tree pairs was higher than 60%. In selective cutting secondary forest, Acer mono seedlings were negatively correlated with other tree species at more scales (6-45 m). In the natural recovery secondary forest, the pairs of A. mono and other arbor seedlings were positively correlated at small scale (0-5 m), but not at large scale (31-45 m). Therefore, selective cutting accelerated the succession of the secondary forest of T. amurensis, which induced random distribution of the arbor seedlings at the large scale (31-45 m), promoted a more coordinated inter-specific relationship, and adjusted the spatial competition between T. amurensis seedlings. Both stands were under succession, and thus suitable artificial management should be carried out to promote tree regeneration and community restoration.

[Spatial structure optimization simulation of main forest types in Great Xing'an Mountains, Northeast China]. / å¤§å ´å®‰å²­ä¸»è¦æ£®æž—ç±»åž‹æž—åˆ†ç©ºé—´ç»“æž„ä¼˜åŒ–æ¨¡æ‹Ÿ.

Based on the data from four 1 hm2 permanent plots in main forest types [namely natural Larix gmelinii forest (LF), natural Betula platyphylla forest (BF), coniferous-broadleaved mixed forest (CBMF) and coniferous mixed forest (CMF)] in Great Xing'an Mountains, a comprehensive cutting index of individual tree (T), based on the commonly used spatial structure parameters [i.e., mingling (M), neighborhood comparison (U), uniform angle index (W), and competition index (CI)] and non-spatial structure parameters [tree vigor index (DC), tree stability index (DH)], was constructed using combined AHP and entropy evaluation method. The cutting process was simulated by Excel VBA to determine the best tending intensity on the basis of systematic comparison of comprehensive T-value under different tending intensities (10%, 20%, and 30%) of different forest types. The results showed that, in the initial state, the mean values of W were all 0.57, indicating a typical cluster distribution. The mean values of U ranged from 0.50 to 0.51 and the dominant degree of overall growth of trees was in a typical mean state. The mixed degree of four main forest types was generally low, with the mixed forest being obviously higher than the pure forest. The mean competition index within the stand was above 2.0, indicating higher competition pressure. The stability and growth vigor index of LF were significantly higher than those of other stands. Overall, the management urgency of BF was significantly higher than that of other stands. With regard to T-value growth rate between adjacent tending intensities, the optimal cutting intensity was 30% for LF forest and 10% for other types. The relative growth rates were 9.7%, 7.9%, 6.6% and 3.9% respectively. However, from the perspective of T-value and canopy density with different tending intensities, the optimal cutting intensity of BF was 20%, and the others were all 30%, in which the T-values were increased by 28.9%, 16.4%, 17.5% and 9.2% respectively. After simulated harvesting, stand structure was improved in various degrees and the mixed degree of tree species was increased. The horizontal distribution pattern of stand tended to random distribution. The dominance degree of dominant tree species was increased. The competition pressure of trees was decreased. DC of trees was slightly lower and the DH of trees was improved.

[Spatial pattern and regeneration characteristics of main woody species in natural secondary forest in Maoershan, Northeast China.] / å¸½å„¿å±±å¤©ç„¶æ¬¡ç”Ÿæž—å† ä¸»è¦æœ¨æœ¬æ¤ç‰©ç©ºé—´æ ¼å±€åŠæ›´æ–°ç‰¹ç‚¹.

Based on the survey data from a 1 hm2 plot of the second forest in Maoershan, we divided the stand into four size degrees, including seedlings (tree height H<30 cm), saplings (H 30-130 cm or H>130 cm meanwhile DBH< 5 cm), medium trees (DBH 5-10 cm) and big trees (DBH≥10 cm). The spatial pattern and spatial association of all woody plants and three dominant tree species in four size degrees, as well as the basic forest regeneration, were chosen as study objects, to explore the renewal trend and provide theoretical basis for the sustainable forest management. The O-ring point pattern analysis method was used to analyze the spatial distribution and spatial association. The dominant trees were planted Pinus koraiensis and three common broadleaved tree species (Fraxinus mandshurica, Juglans mandshurica, Phellodendron amurense), and Ulmus pumila. The results showed that the regeneration of the stand was quite common, with 2894 seedlings·hm-2 and 3337 saplings·hm-2 . For the whole forest, the seedlings in the stand were positively associated with large trees and medium trees and saplings respectively at the scales of 0-25 m, 0-15 m and 0-40 m, and negatively associated at the subsequent scales. The saplings were positively associated with large trees and medium trees at 0-15 m, 0-25 m respectively, and had no significant association at the subsequent scales. Medium trees showed no significant association with big trees at all the scales. For the dominant trees, the saplings and seedlings of Pinus koraiensis showed negative association at the scale of 0-40 m, while the saplings and seedlings of the three common broadleaved tree species or Ulmus pumila were mainly non-associated. There was a negative association between the saplings of Pinus koraiensis and each size degree of the three common broadleaved tree species or Ulmus pumila at most scale, whereas saplings of the three common broadleaved tree species and Ulmus pumila showed a negative association with non-conspecies only at 0-20 m scale. In the medium and large size degrees, there was no significant association between each two of the three dominant tree species at the scale above 20 m. The secondary forest in Maoershan is still under the process of succession. During the process of forest resource management, measures should be taken to promote the renewal, and the unreasonable spatial pattern should be adjusted in time.

[Spatial point patterns and their association dynamics of forest landscapes in Maoershan region, Northeast China between 1983 and 2016.] / 帽儿山地区1983­2016å¹´æ£®æž—æ™¯è§‚ç©ºé—´ç‚¹æ ¼å±€åŠå ¶å ³è”åŠ¨æ€æ€§.

Researches on the dynamics of spatial point patterns and their associations of forest landscape has important implications for maintaining forest stability and making forest management decisions. Based on the four period datasets of forest resource inventory in Maoershan region in 1983, 1993, 2004, and 2016, the O-ring statistics within Programita software (version 2010) was used to quantitatively analyze the dynamics of spatial point patterns and associations. The results showed that the cover percentage of soft-wood broadleaved mixed forest (SBM) significantly decreased, the percentage of natural Quercus mongolica (NMO) increased in the beginning, and then became lower. The percentage of hard broadleaved mixed forest (HBM), Larix gmelinii (RLG) and Pinus sylvestris var. mongolica (RPS) plantations increased dramatically from 1983 to 2016. During the study period, the scale of clumped distribution for the SBM, HBM, NMO all significantly decreased with the increasing spatial scale. The clumped distribution ranging from 0-7 km decreased to 0-3 km, and then tended to be random or uniform distribution with the further increases of spatial scale. The patterns of RLG showed aggregated distribution within the small scale and tended to be random or uniform distribution with the increasing scale. However, the range of scale of uniform distribution increased and random distribution decreased. The patterns of RPS showed clumped distribution within small scale (0-4 km) and the range of scale tended to be larger. On the whole, the spatial association between most of the landscape types within the small scale showed negative association. With the increases of scale, these landscape types gradually become non-association or positive association. However, only a few landscape types (e.g. RLG and RPS in 1983) showed positive association in a small scale. With the increases of scale, all the relationship tended to be non-association. The dynamics of the association of forest landscapes showed different regularity. The dynamics of spatial point patterns and associations of Maoershan region were the comprehensive results of the forest management during 1983 and 2016.

[Forest landscapes' spatial point patterns and associations based on Ripley L and O-ring functions].

Based-on the data of forest resource inventory in Pangu Forest Farm of Great Xing' an Mountains in northeastern China, the spatial distribution pattern and associations of the main forest landscape types [natural Larix gmelini forest (NLG), natural Betula platyphylla forest (NBP), natural coniferous mixed forest (NCM) and natural mixed broadleaf-conifer forest (NCB)] were studied by the two main spatial point pattern analysis methods (Ripley L and O-ring functions). The results showed that the spatial distribution pattern of the four forest landscape types were all consistent with each other the whole, which were all significantly clumped at small scale, and then mainly the obvious characteristics of random distribution with the increase of scale. Spatial associations of the four forest landscape types differed significantly with the Ripley L and O-ring functions. The results of Ripley L function showed that NLG and NCB, NBP and NCB had the obvious negative correlations at small and medium scales, and then mainly showed the trend of non-correlations or even positive correlations at medium and large scales, however, there were significantly negative correlations for.the other forest landscape types at all the research scales. Unlike the results of Ripley L function, the results of O-ring function showed that the main forest landscape types were all significantly negative at small scale, no at medium scale, and positive at large scale with each other. Meanwhile, there were also significant differences for the spatial distribution patterns and associations for the same forest landscape type (or group) at the same level of scale with two different methods, and the rate of consistency of the two methods at all levels of scale mainly exist three forms, i.e., basically remain unchanged, reduced firstly and then increased, and almost always reduced, respectively.

[Primary branch size of Pinus koraiensis plantation: a prediction based on linear mixed effect model].

By using the branch analysis data of 955 standard branches from 60 sampled trees in 12 sampling plots of Pinus koraiensis plantation in Mengjiagang Forest Farm in Heilongjiang Province of Northeast China, and based on the linear mixed-effect model theory and methods, the models for predicting branch variables, including primary branch diameter, length, and angle, were developed. Considering tree effect, the MIXED module of SAS software was used to fit the prediction models. The results indicated that the fitting precision of the models could be improved by choosing appropriate random-effect parameters and variance-covariance structure. Then, the correlation structures including complex symmetry structure (CS), first-order autoregressive structure [AR(1)], and first-order autoregressive and moving average structure [ARMA(1,1)] were added to the optimal branch size mixed-effect model. The AR(1) improved the fitting precision of branch diameter and length mixed-effect model significantly, but all the three structures didn't improve the precision of branch angle mixed-effect model. In order to describe the heteroscedasticity during building mixed-effect model, the CF1 and CF2 functions were added to the branch mixed-effect model. CF1 function improved the fitting effect of branch angle mixed model significantly, whereas CF2 function improved the fitting effect of branch diameter and length mixed model significantly. Model validation confirmed that the mixed-effect model could improve the precision of prediction, as compare to the traditional regression model for the branch size prediction of Pinus koraiensis plantation.

[Retrieval of forest topsoil organic matter's spatial pattern based on LiDAR data].

Forest soil is one of the main carbon pools in terrestrial ecosystem. Its organic matter content can provide basic information for estimating soil carbon storage, and also, is an important index for evaluating the function of soil carbon sink. Based on the LiDAR data and the topsoil organic matter contents in 55 permanent plots at Liangshui National Nature Reserve, Heilongjiang Province of Northeast China in August 2009, and by using partial least squares (PLS) method, this paper retrieved the forest topsoil organic matter's spatial pattern in the Reserve, extracted and screened the variables related to the distribution of the topsoil organic matter (e. g. , intensity, counts, elevation, slope, and aspect), and analyzed and defined the correlations between the screened variables and topsoil organic matter content, with the prediction model of forest soil organic matter content established and validated. In the Reserve, the forest topsoil organic matter content was significantly and positively correlated with three variables (intensity, r = 0.765; counts, r = 0.423; and elevation r = 0.475; all P<0.001). The model prediction on the topsoil organic matter content was reliable (precision = 83.3%, R2 = 0.725, RMSE = 1.955 ). In the areas of forest edge and of low canopy stands, the topsoil organic matter content was less than 100 g x kg(-1). The majority of the study area had a topsoil organic matter content of 100-150 g x kg(-1), while a few areas had the topsoil organic matter content as high as 150-318.4 g x kg(-1).