Research On The Biomass Model Of Larix Kaempferi Plantation In Northern Sub-tropical Alpine Area Based On Sample Plots And Remote Sensing

Forest biomass plays an important role in the global climate change, regional and global sustainable development, which is the entire forest ecosystem's energy and nutrient sources and the base of biological productivity, the net productivity and the carbon cycling. Estimation of the biomass of Larix kaempferi in sub-tropical alpine area based on plots and remote sensing is the main object of this research. Many surveied data from standard plot had been collected and the independent variables which have little relativity have been chosed to establish the predictable model by the stepwise regression mothed.Then based on TM data,topographic maps,administrated maps,DEM data, various vegetation index was checked to establish the predictable model for forest biomass. By correlation analysis and error inspection,the optimal model for Larix kaempferi was selected and the forestry biomass distribution pattern was simulated. Main results are as follows :(1)The total biomass of Larix kaempferi plantation increased with the age increased and the biomass densities of young, immature, near-mature and mature plantation were 18.53t/hm-2,86.13 t/hm-2,101.84 t/hm-2 and 141.76t/ t/hm-2 respectively. The net productivity during the young and immature period increased with age increased but it decreased with age after near-mature period. The net productivity of Larix kaempferi plantation during the young, immature, near-mature and mature period were 3.71t/hm-2.a, 5.74t/hm-2.a, 4.07t/hm-2.a and 4.05t/hm-2.a respectively. With the increase of crown density, the proportion of tree biomass increased but the the proportion of shrub and herbal biomass decreased.(2) Curve fitting precision of individual tree biomass model with D2H as parameters was higher than that of individual tree biomass model with D as parameters. Power function B=4.6×(D2H)0.827 was appropriate for the forecast of the Larix kaempferi plantation biomass in sub-tropical alpine area, of which theγvalue and F value were 0.979 and 926.1 respectively, reaching a significant level. The compatible biomass model was established based on algebra relationship model. Forecast accuracy of trunk (96.3%) and skin (92.1%)was higher than that of leaf (86.3%) and root(88.4%). The estimation error of shrub and herbal biomass were closely related with the measurement accuracy for samples as well as sampling error but has little relationship with sample size.(3) Stepwise method was used to select the effective independent variable from the following factors: Canopy density (C), Age (A), DBH (D), Average height (H), Standing stock (M), Stand density (N) and Basal area (B) . The biomass prediction model was established as the form B=-20.01+36.87C+0.36M-0.62D+2.51H+0.34A (precision was 94.4%). The compatibility problem between gross and the sum of components was resolved with the method of simultaneous equations. Among the systematic compatible models, the estimation accuracy of tree biomass was 95.4%, higher than that of shrub and herbal (83.2% and 80.09%). The estimation accuracy of truck biomass was the highest wih 96.89% and the second was crown with 95.47%. The estimation accuracy of root was the lowest among all the models.(4) TM images at the same time was selected to establish the remote sensing predictable model. The factors of remote sensing and measurement were set as independent variables and the practical values of biomass were set as dependent variable. 15 plots were used to check the prediction model. As there were more information in the vegetation index, the coefficient of correlation between biomass and vegetation index was higher than that between biomass and individual bands(TM3, TM4, TM5 and TM3 + 4 + 5 were positively related to the biomass and the TM4 was highest with 0.814, TM3 was lowest). Vegetation index NDVI has the highest correlation with biomass(γ=0.912). The model B=33.6+199×NDVI+17.5×NDVI2 was eatablished with NDVI for the biomass of Larix kaempferi.Under this model , the range of biomass in this areas was 25 t·hm-2～211t·hm-2. The biomass in this areas was divided into four grades and forest biomass distribution chart was established as a result. Comparative analysis was conducted with 20 independent sample plots. The result showed that the value based on TM data method was heigher than the true value while the value based on sample plots method was lower than the true value.