| As the most important characteristic of forest ecosystem, forest biomass is the focus of many forestry and ecological problems. The accurate calculation of forest biomass data is helpful to study the productivity of ecosystem, the carbon cycle of the whole biosphere and the global climate change. Although harvest method is the most accurate method to estimate forest biomass, it costs a lot of time, manpower and material resources. At present, many studies calculate forest biomass using biomass equations, there are research using remote sensing to estimate forest biomass of large scale as well. The appearance of the traditional remote sensing method not only saves the time of study, researchers even does not need to carry out the investigation of high intensity. However, it is difficult to reach the forest community located in the high mountain area, the spatial resolution of ordinary remote sensing is too low, and its controllability is weak, it’s hard to achieve precise measurements of single wood biomass of small range.This study explore research methods of individual trees’ aboveground biomass of the two main forest types in Changbai Mountain Nature Reserve and the Wanglang Nature Reserve——broad-leaved Korean Pine mixed forest and Abies faxoniana forest through the DBH and crown area quantitative analysis of the main trees by UAV Remote Sensing Technology, thus to obtain the remote sensing estimation model of the same trees’ biomass in other regions, provide the technical support and the theory basis for biomass research of single tree level in other region of our country. It also provides a reliable source of data for forest management under the background of global climate change.Two plots were set in Changbai Mountain Natural Reserve and Wanglang Nature Reserve Separately:plot1-plot4. According to the study area location and forest types, sample plots’ size in the set were divided into:plot1 and plot2 were 100×100m, plot4 and plot3 were 32×32 m, which plot1 and plot3 were training data set, plot2 and plot4 were testing data set. The main research contents and results are as follows:1. Single tree canopy can be effectively extracted from the UAV remote sensing images through the object-oriented classification method through eCgnition software with the manual correction.2. The CA-DBH models had high fitting precision and prediction effect. Fit the CA-DBH models of the main tree species in the two study based on the sample data including 2 independent variable CA and CA2,4 model types (linear model, power function model, polynomial model and exponential model) for a total of 6 equations, respectively.The fitting accuracy of the power function equation was the highest, expressed as D=a Cb. Using the optimal equation,7 major tree species’ individual CA-DBH models were successfully fitted in the two areas. Pinus koraiensis:D=13.794C0.450 ,Tilia amurensis:D=7.527 C0.653, Fraxinus mandshurica:D=7.354C0.454, Acer mono:D=8.135C0.488, Juglans mandshurica:D=7.889C0.370,Ulmus davidianavar.japonica:D=5.317C0.641,Abies Faxoniana:D=13.541C0.422(D=DBH, C=CA). For the Changbai Mountain Nature Reserve, Broad-leaved trees’ CA-DBH model was fitted as D=7.783C0.463, All trees’ equation expressed as D=7.367C0.538Using t test to verify differences from predicted values and the observed values of the 9 CA-DBH models, and calculated the Pearson correlation coefficient (Pearson correlation coefficient). The test results show that DBH estimated and measured values of 9 models were not significant deviation (P>0.05), the Pearson correlation coefficients were all >0.800. The highest was Tilia amurensis in Changbai Mountain Natural Reserve reached 0.944. Pearson correlation coefficient is relatively high for 0.895 of Juglans mandshurica.3. Aboveground biomass of different species in the two areas was estimated using CA-AGB model: the average biomass of Pinus koraiensis was the highest 1725.41kg, followed by 994.95kg, the average biomass of Tilia amurensis, the lowest was Ulmus davidianavar.japonica 108.69kg. |
PDF Full Text Request