Growth Process-based Clustering Of Forest Stand Types And Their Potential Productivity Estimation In Jilin Province

Site quality evaluation is not only an element task of forest management,but also an important research content on forestry.At present,most of the site quality evaluation focus on the application of site index to evaluate the forest site quality which fails to calculate the maximum production capacity?potential productivity?,and the majority of the research object is artificial pure forests.There is less evaluation on the site quality of natural forests,especially complex forests,such as natural mixed forests and uneven-aged pure forests.Based on stand protential productivity,site quality evaluation is built on an assumption that a same stand type has a same growth process in a same site.However,uniform approach of dividing stand types with same growth progress has been of nonexistent until now.Therefore two scientific issues are raised through this research,one is how to guarantee the stand with same forest type to have a same growth process,the other is how to estimate potential productivity.First,partition of forest stand type based on growth process and estimation of potential productivity were carried out.Then introducing the concept of stand growth types and growth process groups,a set of classification method on stand type in Jilin province with considering growth process was proposed.Finally growth models in mixed forest were established,and potential productivity of all growth process groups were evaluated.The main contents and results are as follows.1.In this study,a total of 41 067 data at forest stand level in Jilin province,which were collected from National Forest Continues Inventory which had 3 559 permanent sample plots and forest inventory investigation for management which had 14 476 permanent sample plots,was used.A classification method of stand growth types of mixed forest was developed.In this method,the Two-way Indicator Species Analysis?TWINSPAN?method was first applied to divide the mixed forest and obtain the initial classification.Mixed model with random effects and fitted equation with different group were then established to select basic model for the growth of stand height and basal area.Then,according to the initial classification,parameter clustering and curve clustering were applied;and the classification scheme was determined based on intra-and inter-class distance.Finally,stand growth types of mixed forest were formed after fine-tuning according to expert opinions and forestry knowledge on common forest types coupled with mixed forests.2.The results showed that:?1?From the perspective of stand growth process,the clustering performance of considering both stand height and basal area growth is the best,followed by clustering considering only the basal area growth.For that only considering height growth is the worst.?2?From the perspective of clustering objective,the curve clustering result is obviously better than that of parameter clustering.?3?From the perspective of clustering model,the performance of fitted model with different group is slightly better than the mixed model with random effects,but the stand characteristics of groups in mixed forest using the two models are almost the same.?4?From the perspective of clustering method,the average value method is the best,followed by the Ward method and the K mean method,while the center point method is the worst.?5?From the perspective of criteria for classification of species composition in mixed forest,the result of the tree species composition greater than 65%?named pure forest?is slightly better than that of 85%.3.Following the clustering method from part 1,the 28 975 data from the mixed forests in Jilin province?when criteria of species composition was set as 65%?was divided into 8 stand growth types according to coniferous mixed forest,coniferous broad-leaved mixed forest and broad-leaved mixed forest.Their stand characteristics are as follows:?1?Picea asperata/Abies fabri-Pinus koraiensis coniferous mixed forest,with Picea asperata,Abies fabri,and Pinus koraiensis as the dominant species,accounting for more than 60%of the stand,accompanied by broad-leaved species including Betula costata and Tilia tuan.?2?Larix gmelinii-Pinus sylvestris coniferous mixed forest,with Larix gmelinii,Pinus sylvestris and Pinus thunbergii as the dominant species,accounting for 50%-70%of the stand,accompanied by broad-leaved species including Betula platyphylla and Quercusspp..?3?Pinus koraiensis broad-leaved mixed forest,with P.koraiensis as the dominant species,accounting for 30%-50%of the stand,accompanied by broad-leaved species including T.tuan and Quercusspp..?4?Larix gmelinii broad-leaved mixed forest,with L.gmelinii as the dominant species,accounting for 40%-60%of the stand,accompanied by P.sylvestris,B.platyphylla and Quercus spp..?5?Picea asperata/Abies fabri broad-leaved mixed forest,with A.fabri and P.asperata as the dominant coniferous species?accounting for 50%-80%?and accompanied by B.costata and T.tuan.?6?Quercus spp.broad-leaved mixed forest with Quercus spp.and T.tuanas the dominant species,accounting for70%-90%.?7?Fraxinus mandshurica/Phellodendron amurense/Juglans mandshurica broad-leaved mixed forest,with Juglans mandshurica,Fraxinus mandshurica,Phellodendron amurense,Ulmus pumila and B.costata,as the dominant species,accounting for 70%-90%.?8?Poplar-birch broad-leaved mixed forest,with B.platyphylla,Populus spp.and Salix spp.as the dominant species,accounting for 60%-80%.4.The difference of stand growth process in the 8 types can be quantified.The mean value per year between 10 and 150 of the difference of stand height in 2 types of coniferous mixed forest is 1.7?0-2.1?m,and basal area is 1.23?0-1.36?m²ha^-1.The mean value per year between10 and 150 of the difference of stand height among 3 types of conifer-broadleaf mixed forest is1.1?0.1-1.6?m,and basal area is 1.17?0.17-1.61?m²ha^-1.The mean value per year between 10and 150 of the difference of stand height among 3 types of broadleaf mixed forest is 0.7?0.6-0.9?m,and basal area is 1.12?0.36-1.24?m²ha^-1.5.Based on the method of using stand potential growth for site quality evaluation,the basal area and volume growth models of 8 stand growth types were established.Then the potential productivity of each kind of stand growth types in mixed forest with 5 or 10 site levels was calculated.It is found that the potential productivity of the mixed forest with the baseline age can be effectively used to evaluate the forest site quality.The corresponding volume potential productivities of the baseline age for the site quality at the highest and lowest level are(unit:m³ha^-1year^-1)4.70 and 3.01 for type 1,4.64 and 2.93 for type 2,5.31 and 3.15 for type 3,5.90 and 2.96 for type 4,6.24 and 3.02 for type 5,7.91 and 4.92 for type 6,8.03 and 4.91 for type 7,8.25 and 4.63 for type 8.This further validates the reliability and applicability of the classification to group method.