Aboveground Biomass And Uncertainty Estimation Of Main Tree Species For Different Site Classes In Jiangxi Province

Forest site productivity is the potential productivity of forest vegetation.It is to guide the forest management,development of business decision-making program and is an important indicator of sustainable forest management.Forest biomass is the amount of dry matter produced by the forest community in its life process and can be used as an indicator of the productivity of forest sites and is closely related to site type and quality.Due to the diversity or difference of site conditions,the biomass estimation results and the biomass estimation errors of the same tree species in different areas will vary with the quality of the site.Ignoring the difference of biomass estimation results and the biomass estimation error caused by the difference of site quality,The result must be rough and not accurate.In order to obtain the estimation of forest biomass and site productivity under different site conditions,the main tree species in Jiangxi Province were divided into four categories: Cunninghamia lanceolata,Pinus massoniana,Hardwood Broad and Softwood Broad.Based on the sixth and seventh national forest resources inventory,divided the sample plots into several site classes.Allomatric growth model and Biomass expansion and conversion factor(BCEF)model was used to estimate the forest biomass and the uncertainty in difference site classes.The effects of factors such as origin,age group and stand density on the estimation of forest biomass and its components were explored.The biomass increment model of main tree species in Jiangxi Province was established based on climate and site factors.The uncertainty of biomass estimation provide theoretical basis on improvement of biomass estimation model.The main conclusions are as follows:(1)The dominant tree height-DBH model of Cunninghamia lanceolata,Pinus massoniana forest,hardwood broad forest and softwood broad forest in Jiangxi Province was established by using the tree height classifying method.The site quality of Cunninghamia lanceolata forest is divided into seven classes.The site quality of Pinus massoniana forest is divided into five classes,and the site quality of hardwood broad is divided into five classes.The quality of softwood broad is divided into three classes.By calculating the mean biomass density and error estimates of Cunninghamia lanceolata and Pinus massoniana forest at different site classes,it was found that the average biomass of Cunninghamia lanceolata and Pinus massoniana dominant forest increased with the increase of site level,that is,the higher the quality of the forest site,the greater the biomass density.It is proved that the height classes of dominant tree is used as the basis for dividing the site classes.(2)The three kinds of biomass allomatric growth model and two kinds of biomass expansion and conversion factor(BCEF)model were compared by the relative root mean square error of aboveground biomass.And the Monte Carlo simulation method was used to estimate the mean aboveground biomass density and error of Cunninghamia lanceolata and Pinus massoniana in different sites to find the optimal model of aboveground biomass on the different site conditions.A.Compared with the above results,adding the tree height variable in biomass model can improve the estimation accuracy and have lower uncertainty than the model with only variable diameter at breast height.At the same time with both the DBH and the tree height exist,the Pinus massoniana has the equation with two parameters better than the equationwith three parameters in obtainning a better estimate.The Cunninghamia lanceolata equation with three parameters is better than the equation with two parameters.The variance of the biomass estimation error of the allometric model is the lowest at the site average-that is,themore intermediate site class the closer the site quality on average,the smaller the relative error of the biomass estimation will be.B.Compared with the two BCEF models,the mean biomass value and error of each site class estimated by the empirical(regression)model method is not very different.and the biomass mean values error of each site class estimated by the continuous function method have big difference.Compared with the allometric model,the relative error of the BCEF model is smaller than that of the allometric model when estimating the mean biomass dengsity of the regional scale.The regression model method of BCEF is superior to the allometric biomass model method when estimating the relative error of biomass at different site classes.The allometric biomass model method is superior to BCEF continuous function method.For large scale area with many different site conditions,The BCEF regression model is more reliable.For the Cunninghamia lanceolata and Pinus massoniana,the mean BCEF is not very different at different site classes.(3)Forest competition is the main influencing factor of tree biomass growth in Jiangxi Province.For coniferous forests,the secondary influencing factors are climate factors(frost-free days,annual mean temperature),age groups,and finally terrain factors(elevation).For broadleaf forests,site grade is a secondary factor,and finally climatic factor.The effect of site classes on the increment of biomass of coniferous forest was not significant,and the climatic factors influencing the increment of biomass of main tree species in Jiangxi Province were temperature.The terrain factor had no significant effect on broad-leaved forest,and had little effect on coniferous forest.Therefore,when estimating the productivity of coniferous forest in Jiangxi Province with the increase of biomass,the climatic factor and age group can be used as a substitute for the site index,and the site class seems to be a better choice when estimating the productivity of broadleaf forest.(4)The different origin classification had little effect on the aboveground biomass estimation and various components for the biomass allomatric growth model of Cunninghamia lanceolata.The error of the mean biomass density value of Pinus massoniana model planted forest was higher than that of natural forest,and has a greater impact in the branches and leaves.The effect of origin on the estimation error of allometric biomass model varies with tree species.With the increase of the age group,the aboveground biomass of the Cunninghamia lanceolata and the biomass estimation of the biomass of each component decreased with the increase of the age group.The influence of the age group on the estimation error of the allometric model of Pinus massoniana was the largest in the young forest model,followed by the middle age forest,mature forest,near mature forest and over mature forest.There was no significant effect on the aboveground biomass estimation error of the allometric biomass model of Cunninghamia lanceolata.For the branches and leaves,the estimation of the biomass mean value in the low density stands was higher than that of the high density and the regional level.There was no significant effect on the aboveground biomass estimation error of Pinus massoniana allomatric biomass model.For the different components,the model error of the mean biomass density was higher than that of the leaves after dividing the density,And the biomass mean value of low density and high density stands were lower than those of non-classification.The use of stand density classification is a good way to reduce the model error when calculating the mean biomass density of Pinus massoniana.(5)The mean value of mean biomass density and the error value under the three kinds of the original sampling design,three sampling intervals and multi-starting systematic sampling were not very different,and each sampling method was very good Reflecting the average aboveground biomass of Cunninghamia lanceolata and Pinus massoniana in Jiangxi Province.With the increase of the sampling interval,the sampling unit is increased and the number of samples is reduced.The absolute value and the relative value of the total root mean square error are increasing.This is due to the absolute value of the sampling error and the relative value are rising,and the model error is not significant difference.Correspondingly,the sampling error is also increasing in the total error.However,considering the difficulty and cost of sampling,the design of three different sampling spaces and the sampling design of three kinds of multi-starting points can reflect the average level of aboveground biomass,which can be used for other large-scale regional investigation.