| Heavy metal pollution has been an important problem in the field of environmental geochemistry all along. Due to the application of modern analytical technologies and computer technology, the analytical accuracies of heavy metals become higher and higher and the analytical limits of detection become lower and lower, which contribute to the research of heavy metal pollution in the area of information dimension. However, the research of soil heavy metal pollution in the area of spatial-temporal dimension still relies on the conventional geochemical methods, which are based on the raster sampling and chemical analysis and hence are time-consuming and relatively expensive. Therefore, it is urgent to find a rapid and cost-effective way for investigating soil heavy metal pollution.The remote-sensing geochemistry approach, which can allow for synoptic and repetitive coverage of large areas, has been widely used in the area of environmental geochemistry and pedology. To our knowledge, however, until now examples of the use of remote-sensing geochemistry approach investigating heavy metals pollution for agricultural soils have not been reported. The major difficulty is that the contents of heavy metals in soils are very low, and hence it's very difficult to detect their spectral features with the existing technologies. It is one of the difficulties to explore the spectral responses of heavy metals in soils and research the relationships between the soil spectra and heavy metals.Soils in suburban environment come easily in contact with humans and hence have an important influence on public health. Serious heavy metal pollution is one of the important properties of many suburban soils. The heavy metal pollution in the suburban soils can affect the ecological functions of soils. On the other hand, the pollutants can influence the townsman's health by means of food chain or the inhalation of soil particles. Therefore, the problem of heavy metal pollution in suburban agricultural soils has come to attention by the public. Jiangning and Baguazhou areas are located in the south and north of the Nanjing city, respectively. Both areas are the important base for food and vegetable supply of Nanjing, and hence they are closely linked to the people's health.Based on the background mentioned above and supported by the project named"biogeochemical survey for the land of Jiangsu province", which was funded by Geological Survey of Jiangsu Province, the two important areas, Jiangning and Baguazhou, were chosen as the study areas to research the feasibility of the remote sensing approach for predicting heavy metal pollution in agricultural soils. One side, this study can contribute to the research of soil heavy metal pollution in the area of spatial-temporal dimension. On the other hand, this study can widen the research scope of remote-sensing geochemistry, and consequently promote the developments of the new subject, remote-sensing geochemistry. Moreover, this study has the function of enhancing the food security, protecting people's health, and advancing the sustainable development of Nanjing area.Remote-sensing geochemistry is a kind of comprehensive subject, the basal theory, technologies and applications of which are extensive. This study mostly focused on the basal research. Soil spectral features are the basis for the identification of soil constituents using the remote-sensing geochemistry approach. The primary objective of this study was to try to find the relationships between soil spectra and soil heavy metals using all kinds of technologies and methods.In this dissertation, the spectral features of soils and heavy metals were the main line, and building the relationships between soil spectra and the contents of heavy metals was the major aim. Based on these objects, the spectral responses of soil heavy metals were systemicly researched. Before the relationships between soil reflectance spectra and soil heavy metals were explored, the contents and chemical speciations of heavy metals in Jiangning and Baguazhou areas were firstly studied. In addition, the spatial variations and structures of heavy metals were also examined using the geostatistics and spatial autocorrelation method. It is the premises for exploring the feasibility of predicting heavy metals in soils using the remote-sensing geochemistry approach to understand the spectral features of soils and heavy metals. Therefore, the reflectance spectra of soils and heavy metals were measured, and then the spectral responses of soils with different pollution levels were researched by means of modeling pollution artificially in the laboratory. Based on these results, the relationships between reflectance spectra and heavy metals in soils of both Jiangning and Baguazhou areas were researched furtherly. The contents of heavy metals of the two suburban areas were successfully predicted using reflectance spectra. Moreover, the mechanism by which to predict the spectrally featureless heavy metals was discussed. Besides reflectance spectra, many other bands can be used by remote sensing technologies. However, due to the atmospheric effects, the electromagnetic waves within the shorter bands, such as X-ray orγ-ray, are not suitable for the earth surface. The feasibility of the spectra within the thermo-infrared region for mapping soil heavy metal pollution was explored. Due to the lack of suitable images of Nanjing area, the potential of remote sensing for monitoring heavy metal pollution in suburban soils of Nanjing area was investigated using the simulated sensors based on the results of the band selection. The results derived from the simulated sensors could provide theoretical base of sensor selection for mapping soil heavy metal pollution rapidly in the future.The main results of the thesis are listed below:(1) The new-found problem of Cd anomalies spreading along the whole Changjiang River has been laid store by all the provinces along the River. In this dissertation, two ways, geo-statistics and the sequential extraction were used to research the sources and risk assessment of Cd. Some novel results were achieved from the two ways. Firstly, the results of spatial auto-correlation showed that Cd had low value of spatial autocorrelation, which suggest that the spatial distribution of Cd was scattered and hence implied that the source of Cd may be from the anthropic input. Secondly, the results derived from the sequential extraction showed that Cd had the highest extractability, 82.3%. Considering that the total contents of Cd were far higher than the background value, it could be concluded that in Baguazhou there was high risk for Cd. Moreover, the correlation analysis showed that the residual fractions of Cd were negatively correlated with the total Cd, which further validated the conclusion that besides from the natural materials, Cd in Baguazhou area was also from anthropic input.(2) In this dissertation, the spectral responses of soil heavy metals were researched detailedly for the first time in the world. For the heavy metals discussed in this dissertation, only Ni, Cr, and Cu were spectral active, while other elements were spectral featureless. Only when the contents of Cr and Cu were as high as 4‰, the heavy metals could be explored from the soil spectra. For the agricultural soils, however, the contents of heavy metals were far lower than this value. Therefore, it was pointed that it could not predict heavy metals in agricultural soils with reflectance spectra only by means of the direct way, and the prediction for heavy metals should be changed a new method.(3) For the first time in the world, the contents of heavy metals in agricultural soils were predicted with reflectance spectra. The results showed that heavy metals were all negatively correlated with reflectance for both Jiangning and Baguazhou areas. It was very interesting that the order of the prediction accuracy was the same as the order of their correlation coefficients with Fe. It was concluded that the inter-correlation between heavy metals and total Fe in soils was the major mechanism by which to predict spectrally featureless heavy metals.(4) For the soil samples located in the northwest of Baguazhou and polluted seriously by industries, their spectral curves were significantly different from other spectra of agricultural soils. These high polluted soil samples were identified directly according to their reflectance spectra. The criterion to judge whether heavy metals were from the industrial area or not was: 1) the average reflectance was lower than 30%; 2) the absorption depth around 880nm was more than 2%; and 3) the absorption depth around 2210nm caused by clay minerals was lower than 4%. The method of reflectance spectra was a kind of complementarity for the identification of the source of heavy metals to the traditional methods.(5) Goethite, which can affect the distribution, transfer and translation of heavy metals in soils, is the major iron oxides in Nanjing area. It's very difficult to identify goethite using the traditional methods, such as XRD. In this dissertation, goethite in soils was successfully identified using reflectance spectra, and the criterion for the identification was brought forward. The strong d-d absorption band around 490 nm, which corresponds to the electron pair transition (EPT), 2(4T1g), was the diagnostic spectral feature of goethite. Moreover, the absorption depth around 490nm caused by free iron oxides in soils becomes stronger with the increase of free iron oxides. Based on the relationship between the absorption depth and the contents of free iron oxides, free iron oxides in soils can be predicted quickly.(6) The heavy metal adsorption caused by goethite is the important factor affecting the contents of heavy metals in soils. The results of Ni2+ adsorption isotherm were Langmuir function. Ni2+ adsorption was very low at the low Ni2+ contents. Adsorption became significant when the contents of Ni2+ were higher. The results derived from reflectance spectra, XPS and FT-IR showed that, H+ bonded to Fe-OH was displaced by Ni2+ during the adsorption. During the adsorption, Ni2+ was the base while goethite was the acid. Electrons were transferred from Ni2+ to goethite.(7) The results of FT-IR showed that, the FT-IR spectra of agricultural soils in Nanjing suburban area only reflected the vibrational bands due to OH and Si-O. The FT-IR spectra were very similar for the soils polluted by different heavy metal levels. Moreover, iron oxides and organic matter, which are related with heavy metals in soils, could not be identified directly from the FT-IR spectra of soils. Therefore, it was concluded that the optimal bands for investigating heavy metal pollution were VNIR band.(8) Due to the lack of suitable images of the research areas and based on the bands selection derived above, in this dissertation the relationships between the simulated HyMap, TM and Quickbird and heavy metal contents were explored. The prediction accuracy for each sensor was satisfactory and similar. It suggested that low spectral resolution was not a limitation for predicting soil heavy metal contents. Considering the costs and the problem of mixing pixels, it was thought that Quickbird was the first choice for mapping soil heavy metal pollution rapidly.
|
PDF Full Text Request