| Soil organic carbon (SOC) as the biggest "carbon sink" and "carbon source" in the terrestrial ecosystems is an important factor to influence global climate change and agriculture productivity. In addition, having a good acknowledge of the spatial variation of the soil organic carbon and choosing the best method of prediction is the prerequisite for agriculture production and environment imitation. Both of natural and human factors can bring soil organic carbon great influence, which varies in different scales and regional.In this dissertation, the cultivated soil of mountain and hilly area in three gorges reservoir area was selected as the object of study, the study area is located in Yunyang county, to be more precisely. Elevation, slope, parent materials, slope position, topography wetness (TWI), cropping systems are considered to be essential factors for soil organic carbon. Under the supports of a serious soft tools, such as, ArcGIS9.3、 SPSS18.0、Matlab7.0、GS+9.0, the effect laws of each factors and the spatial variation characteristics of soil organic carbon were gained. Category variables were introduced into the regression model, through the path analysis, the mechanism of factors on cultivated soil organic carbon (SOCD) density was discussed. We also gained the variability of SOCD by border analysis and Anisotropic analysis.At last, three different prediction methods, Ordinary kriging(OK), Regression kriging(RK) and Generalized regression neural network(GRNN), were used to generate soil organic carbon density (SOCD) map of the study area, prediction precisions were measured. The results were showed as follows:(1) The average,0-20cm, cultivated SOCD is2.91kg/m2, is just a little below the whole country’s corresponding figure (3.0kg/m2). Limestone soil (3.27kg/m2) and yellow soil (3.24kg/m2) has higher SOCD. The cultivated soil carbon storage is1838.75×106kg in the study area.(2) The correlation between elevation and SOCD was significant (0.329**). The essential impacts of elevation on SOCD is great and can not be influenced by other factors easily. Topography wetness index (TWI) also has great correlation with SOCD (0.256**), but the direct impacts is smaller than indirect impacts.(3) Areas covered by Gray-brown purple mud (shale) efflorescence and Purple sand and mud (shale) efflorescence have lower SOCD. Dolomite weathered material contain higher SOCD because of the high density vegetation during its establishment. Reddish brown thick mudstone is mainly located in sloppy area, soil erosion may be one reason for lower SOCD.(4) Different slope position has different SOCD: Valley> slope foot>ridge> slope shoulder> slope back. Correlation analysis illuminates that valley and slope foot has the positive correlation with SOCD, ridge, slope back and slope shoulder has the negative impacts on SOCD. The redistribution of soil component caused by water and soil loss is one reason for valley’s higher SOCD.(5) The cropping systems impact SOCD in this order: three harvests a year>two harvests a year>one harvests a year. Maize and rice has a similar level of impact on SOCD. Multiple rotation result in the increase in the amount of organic fertilizer and crop straw is an important cause of soil organic carbon accumulation.(6) From the global trend, in the south-north orientation, SOCD is higher at both ends and is lower in the middle area, presenting like a U-shaped. In contrast, there is no significant trend in east-west orientation. From spatial variation aspect, Anisotropic analysis illuminates that, SOCD’s spatial variability is more drastic in south-north orientation than in east-west orientation.(7) For purposes of this study, the sequence of precision of spatial prediction methods is: RK>GRNN> OK, RK presents higher precise and its suitability in the SOCD prediction in low-relief terrain at regional scale. |
PDF Full Text Request