| Soil organic carbon of Moso bamboo forest was taken as research object at the six towns include Gongchuang,Shangping,Qingshui,Xiyang,Hongtian ang Xiaotao in the Yong'an city of Fujian province.Remote sensing estimation model which choose topographic factors as cofactor was constructed on the basis of Landsat8 remote image data using CASA ecosystem carbon cycle model to estimate met primary productivity on Moso bamboo forest.Then remote sensing inversion of soil organic carbon was achieved,spatial heterogeneity of Soil organic carbon based on results of remote sensing inversion was analyzed,and it combined geostatistical analysis to reveal spatial heterogeneity of measured sample points which provided scientific basis for bamboo forest soil quality evaluation,precision agriculture and bamboo forest sustainable management.The main conclusions were as follows:(1)Remote sensing technology was applied to extract Moso bamboo forest information by using"based on the spectra of lamellar texture-oriented class" classification method based on remote sensing image data by radiometric calibration,atmospheric correction,geometric correction and so on.The accuracy of classification was 84.8%,Kappa coefficient was 0.8235.(2)Vegetation primary productivity(NPP)distribution of bamboo forest in the Yong'an was estimated on the basis of CASA model combining basic parameters include spectral characteristics of vegetation,meteorological factors and maximum light use efficiency of vegetation and so on..Then results showed that the average primary productivity of research area was 147.06gC·M2·mon-1,that the highest NPP was 153.84 gC·M2·mon-1 at the Shangping town and the lowest was 139.40 gC·M2·mon-1 at the Hongtian in the central region.Spatial distribution of soil respiration was inverted ong the basis of NPP distribution of bamboo forest by combining the Van't Hoff soil respiration model,the average strength of soil respiration was 23.08 gM2·mon-1,Qingshui town was the strongest strength of soil respiration in the east Yong'an,that was 28.12 ghf2·mon-1,Hongtian town was the lowest in the middle of Yong'an,that was 20.73 gM2·mon-1.Result was the same as the primary productivity of Moso bamboo forest.(3)Remote sensing estimation model based on soil organic carbon content of soil respiration was constructed that was SOC=0.048*A-0.018*H-0.222*?+0.351*M+42.948,A was the soil base respiration,H was the relative elevation,B was the slope and M was the surface roughness.By 20%(24 points)mode data validation,the result showed that the average relative accuracy was 84.09%and root mean square error was 4.40 g kg-1.(4)Soil organic carbon inversion was achieved by remote sensing model of soil organic carbon constructed and spatial distribution under different terrain conditions was discussed.The result showed that soil organic carbon content increased as the elevation increasing that the lowest was 15.19 g kg-1ranging Om from 200m and the highest was 32.52 g kg-1 more than 1200m;soil organic carbon content increased as the elevation increasing and showed decreasing trends with increasing of slope gradients,the highest was 26.87 g kg-1 ranging 0 from50 and the lowest was 18.72 g kg-1 more than 46°;soil organic carbon content in the shady slope(24.34 g kg-1)was slightly higher than in the sunny slopes(23.54 g kg-1);Uneven distribution of organic carbon content in the various towns,soil organic carbon content(26.38 g kg-1)was the highest in the Qingshui town,while the smallest was 20.23 g kg-1.in the Hongtian town.(5)Spatial autocorrelation,semi-variance function,fractal dimension and anisotropy of soil organic carbon were analysed by combining geostatisticas and GIS that the relationship of the sampling points within the range of distance and orientation in space was revealed.Spatial autocorrelation Morans'I coefficient decreased with increasing of spatial distance;The optimal model were index,Gaussian,spherical model;semivariance fitting determination coefficient were 0.684,0.831,0.506,0.538 and it shows the great spatial autocorrelation,nugget effect were 10.33%,13.10%,10.16%,0.12%;fractal dimension that were A layer(1.963)>B layer(1.962)>C layer(1.944)>All profile(1.938);From the spatial distribution,soil organic carbon of bamboo forest were obvious banded and inlaid plaque distribution in Yongan that organic carbon contents at the Qingshui,Gongchuan and Shangping in the east and at the Xiaotao in the west was obvious high at the Hongtian ang Xiyang in the central... |
PDF Full Text Request