| This thesis was based on the experiments carried out on a trial area of the west ujimqin. Three kinds of landscapes with the characteristics of typical steppe were chosen. The work of the thesis was based on 3S technology and guided by the theory of landscape ecology and the principle of geological statistics. The compared experiments were performed between remote sensing images and terrestrial observations. The relationships between vegetation elements of grassland with different landforms and the thickness of chestnut soil were explored by connecting soil environment conditions and vegetation landscapes. The correlation between the evolution of vegetation coverage and that of the thickness of chestnut soil was analyzed. The correlations of the thickness of chestnut soil, soil moisture content and vegetation coverage were expressed directly using the profiles of different gradient directions. The correlation of spatial distribution pattern between vegetation coverage of grassland and the thickness of chestnut soil was analyzed by exploring quantitatively their spatiotemporal variation. Normalized Difference Vegetation Index (NDVI) from Landsat-8 TM of three samples was calculated, respectively, and then the vegetation coverage was determined. The comparison between predicted vegetation coverage and measured values was carried out to determine their similarity. The paper took Artemisia frigida as an example, which is a representative vegetation indicating the degradation of typical steppe. The correlations between the thickness of chestnut soil and the frequency and density of Artemisia frigida were analyzed. The relational models between vegetation coverage and the mean vegetation height, aboveground biomass, the thickness of chestnut soil and soil moisture content were built using multivariate regression analysis. The objectives were to interpret vegetation types using hyperspectral images and monitor the thickness of chestnut soil. The thesis provided scientific basis and essential data for constructing large scale, rapid and dynamic monitoring system of spatial pattern of typical steppe vegetation and soil. The main conclusions of the thesis were as follows.1. The correlations between the thickness of chestnut soil, soil moisture content and the vegetation elements of three kinds of landscapes in the year of 2012 and 2013 were analyzed using the method of Spearman rank correlation coefficient. The results showed that the correlations between the thickness of chestnut soil, soil moisture content and vegetation coverage in most samples were high, reaching a moderate degree, except for the corrugated high plain sample whose natural environment was affected by human in 2013. The correlations of the thickness of chestnut soil, soil moisture content and vegetation coverage were expressed directly using the profiles of two different gradient directions in three kinds of landscapes. The reasons were analyzed by combining ficldwork.2. The spatial structures of vegetation coverage and the thickness of chestnut soil in three samples were analyzed using semi-variance function. The results indicated that the vegetation coverage of three samples and the thickness of chestnut soil of corrugated high plain sample belonged to the weak rank of varaition and the thickness of chestnut’ soil of slope and hilly samples had strong variability. The degree of spatial variation of the vegetation coverage in the slope sample and that of the thickness of chestnut soil in corrugated high plain sample were strongest. The spatial structural information of the vegetation coverage in slope sample and the thickness of chestnut soil in corrugated high plain sample could be increased by decreasing the sample interval. The vegetation coverage and the thickness of chestnut soil in three kinds of landscapes had strong spatial correlation. The proportion of spatial heterogenous resulted from spatial autocorrelation accounted for more than 80% of the total spatial heterogenous. The continuities of the vegetation coverage in hilly sample and the thickness of chestnut soil in corrugated high plain sample were the best. The vegetation coverage and the thickness of chestnut soil in all the samples had anisotropy with inapparent isotropy and complex spatial distribution pattern.3. By comparing the spatial distribution map of vegetation coverage with that of the thickness of chestnut soil in the three samples, we found the spatial distribution of the two in the slope and hilly samples had large-scale spatial consistency. And the spatial distribution of the two in the corrugated high plain sample had little relevance, which was the same as the results obtained in chapter four.4. According to the quantitative investigation for the vegetation coverage and the thickness of chestnut soil, we found the percentage of the vegetation coverage was rarely more than 60%. The area of the vegetation, whose coverage was between 40% to 50%, accounted for nearly 50% of the slope sample area. The area of the vegetation, whose coverage was between 30% to 40%, accounted for more than 50% of the hilly and corrugated high plain sample area. The percentages of the sand in the three samples were all modest, but the percentages of chestnut soil, whose thickness was more than 20cm, were all small. The largest proportion occurred in the corrugated high plain sample and only accounted for 13.200%. The area of the chestnut soil, whose thickness was between 10cm to 20cm. accounted for 53.055% of the slope sample area. In the hilly sample, the differences ofarea proportions of different soil thickness were small, except for the areas of the sand and the chestnut soil whose thickness was more than 20cm. In the most areas of corrugated high plain sample, the thickness of chestnut soil was more than 10cm and the area of the chestnut soil, whose thickness was between 10cm to 20cm, accounted for 61.421% of the total sample area.5. The vegetation coverages of the three samples in 2013 were calculated using NDVl, and the calculated results with measured values were compared. The results showed that the similarities in the three samples were 93.6%、94.3% and 92.3%, respectively. It verified that the two methods for calculating vegetation coverage were highly consistent.6. When the thickness of chestnut soil was less than 10cm, both the frequency and total density were much more than that in other soil thickness conditions. The results showed that the role of soil factor for vegetation growth and species evolution in typical steppe became more and more significant as the decrease of soil thickness. When the thickness of chestnut soil was less than 10cm, vegetation growth and species evolution primarily depended on the soil; And when the thickness of chestnut soil was more than 10cm, the synergistic effects of ecological factors were dominated. Therefore, when the thickness of chestnut soil is less than 10cm, the protection of chestnut soil should be enhanced to slow the degradation of typical steppe.7. Based on the stepwise method, multiple regression models were built using the vegetation coverage of each of the three samples(Y), the vegetation coverage in the plots with Artemisia frigida (YL) and the vegetation coverage in the plots with Artemisia frigida and chestnut soil whose thickness was less than 10cm (Y1.1.) as dependent variable, respectively, and corresponding thickness of chestnut soil (x1), soil moisture constant (x2), aboveground biomass (x3) and the mean vegetation height (x4) as independent variables. The models were shown as follows.(1) The slope samlpe:Y1=9.99210.343x1+3.105x2+0.046x3:Yu-18.894+0.279x1+2.167x2+0.028x3, Y(?)=8.028+0.736x1,+3.797x2+0.036x3(2)The corrugated high plain sample:Y3=15.833+0.186x1+1.534x2+0.06x3+0.135x4;Y3L=22.9891+1.343x2+0.055x3 Y3LL-16.869+1.198.x1,+1.964x2+0.049x3(3) The hilly sample: Y2= 21.606+0.425x,+0.956x2+0.048.V,The linear relationships between the vegetation coverage of the plots with Artemisia frigida, the vegetation coverage of the plots with Artemisia frigida and chestnut soil whose thickness was less than 10cm and the thickness of chestnut soil, soil moisture constant, aboveground biomass and the mean vegetation height were inapparent. The relationships between independent variables and dependent variables were non-linear. The models needs further study.8. In the slope sample and corrugated high plain sample, the coefficient of determination (R:) of fitting models of the vegetation coverage in the plots with Artemisia frigida and chestnut soil whose thickness was less than 10cm, were all more than that of fitting models of the vegetation coverage in the plots with Artemisia frigida. It indicated that the fitting models of the vegetation coverage in the plots with Artemisia frigida and chestnut soil whose thickness was less than 10cm could explain more variation. It also proved that it was significant to divide the plots with Artemisia frigida according to the condition that the thickness of chestnut soil was more than 10cm. |
PDF Full Text Request