Abstract:
Abstract: As the most basic soil data, soil map plays an important role in the development and utilization of soil resources and in the construction of agricultural production. The soil map of the second national soil survey in China has been used as the basic data of soil in resources and environment management, global change research, eco hydrological simulation and other fields. But it is found that the classification accuracy of the second national soil survey on soil types and subtypes is not high. Moreover, with the continuous progress of the society, the more requirements for the accuracy and timeliness of soil maps are put forward by various undertakings, such as precision agriculture, environmental management, land management, eco hydrological simulation and so on. With the continuous development of remote sensing technology, remote sensing survey and mapping has gradually replaced some conventional soil surveys, and it has become one of the universal soil digital mapping methods. All the methods on remote sensing survey and mapping of soil types are based on the relationship between soil and environment, but the characteristics of soil environment in different regions and its extraction methods on remote sensing are different. In recent years, domestic and foreign experts have launched a lot of researches on remote sensing survey and mapping methods, but the research on the model and method for remote sensing survey and mapping of soil types and subtypes in semi-arid sand region have rarely reported. The purpose of this paper is to solve the problem of remote sensing survey and mapping of soil types and subtypes in semi-arid sand region. Based upon soil genesis theory and soil environment condition in semi-arid sand region, we took Horqin Left Back Banner as an example, analyzed the relationship between soil type distribution and environmental factors in semi-arid sand region with the method of field survey and expert knowledges. And then put forward a method for remote sensing survey and mapping of soil types or subtypes in semi-arid sand region based upon multi-temporal Landsat 8 OLI remote sensing images. The results indicated that remote sensing survey and mapping of major soil types or subtypes in semi-arid sand region, for example swamp soil, saline-alkali soil, meadow soil, aeolian sandy soil and their subtypes, could be realized with the help of the modified normalized difference water index (MNDWI), salt index (SI), normalized difference moisture index (NDMI) and normalized difference vegetation index (NDVI) extracting from multitemporal Landsat 8 OLI remote sensing images. Using the proposed method survey and mapping soil types and subtypes in Horqin Left Back Banner, and then used statistical methods to check the accuracy of digital mapping results by soil type data observed in the field. The results showed that the overall accuracy of digital mapping results took percentages of 72.84% and the Kappa coefficient was 0.6678. This study provides approaches and references for digital soil mapping in semi-arid sand region, and provides possibilities to solve the problem of low aging and low precision on sudden basic data in semi-arid sand area by using the method of digital soil mapping.