Study On The Compatible Models Of Tree Carbon Content For Main Conifer Species In Northeast Part Of China

Forest plays main role in relieving the tendency of climate warming in the world. It has important significance to measure the carbon storage of the forest exactly. The methods of predicting carbon storage of the forest and the prediction accuracy also need to be improved at the same time. As the forest carbon storage is the key evaluation index to predict the forest ecological benefit, predicting the carbon content for the organs (stem, branch, foliage and root) exactly is the basis. Viewed the past researches in the forest we can conclude that traditional prediction methods of carbon storage of the forest, to a large extent, largely depends the biomass data from forest resource investigation combined with different types of carbon content coefficient to predict the carbon storage of the forest indirectly. Because of the fact that error exists in both of the biomass data and average carbon content ratio employed in the research, this prediction method must be produce conspicuous error. So measurement and direct attention should be raised to solve these problems.Taken Korean pine(Pinus koraiensis) from the plantation in Northeast part of China as an example, the research applied two methods (i.e. direct method and indirect method) to evaluate the prediction of the individual tree carbon content. Based on the non-linear simultaneous equation theory and measured data for biomass and carbon content of the organs of Korean pine, the particle established unitary compatible biomass and carbon content models which employed diameter at the breast height as the only explanatory variable. In the meanwhile, the prediction accuracy difference between the two methods was compared and analyzed. In the direct method, compatible carbon content model is established directly to predict the carbon content of the different organs. Whereas, the indirect method predicted carbon content of the organs by establishing compatible biomass model combined with different types of carbon content conversion coefficient. In indirect method, three types carbon ratio were employed,(1) the carbon content conversion factor0.5which is widely used in the research over the past decades;(2) the average carbon content percentage of individual trees measured from the sampled trees; and (3) the average carbon percentages of tree components measured from the sampled trees. The results showed that:the coefficients of determination (R2) for compatible biomass and carbon content model were0.76-0.99. Fitting efficiency (EF) of the two models were0.80-0.98. The carbon content evaluation prediction of direct method were91.03%,80.02%,70.24%,87.10%,93.08%for stem, branch, foliage and root, respectively. Compared with the direct method, the prediction accuracy of indirect method for the organs employing average carbon content conversion0.5decreased1.39%,1.5%,0.13%,1.09%and2.2%respectively. Whereas, the decrease range of indirect method using two other carbon content ratio were within0.3%. To be concluded, the direct method was the best method to predict carbon content for Korean pine from the plantations.Unitary and bivariate carbon content compatible models of the6conifer tree species in northeast region were established by direct method. On the whole, the coefficients of determination (R2) of the organs and the total of the6tree species except for the branch and foliage of Larix gmelini plantation reached0.70for unitary and bivariate model; the fitting efficiency (EF) for most of the organs and the total got0.8. Based on the test results, the ME%for most of the organs and the total of unitary and bivariate model were within±15%, and the MAE%were within30%. The prediction accuracy of Larix gmelini plantation、natural Larix gmelini and Pinus sylvestris plantation was good and most of organs and the total amount was above80%. Whereas, the branch of natural Korean pine、natural Abies nephrolepis and natural Picea koraiensis had poor prediction accuracy and the other organs and the total amount performed better. Compared with unitary model, the coefficients of determination and prediction of bivariate mode increased.To sum up, the compatible unitary and bivariate models hold the ability to predict the individual tree carbon content and can be used at the larger regional area. Compared with unitary model, the prediction accuracy of the stem and the total amount increases only a small range. Because of the larger measure error for tree height, the bivariate model reduces the flexibility. So the unitary compatible model should be widely used in larger region.