| Soil salinization has become more and more serious in the world with the development of 3s technology in large scale. The remote sensing has also been used for classification, analysis, simulation and prediction of soil salinization. In order to study the development of salinization in the Manas river basin under different irrigation methods, the temporal and spatial distribution of soil salinization has been analyzed with the sampling survey in field. This study would be helpful for the development of the region’s agriculture.By coupled macro remote sensing monitoring and micro experiment measure, supervised classification and decision tree model was established for extracting the salinization. ENVI, Arcmap, SPSS and Excel were used to analyze the salinization distribution under different irrigation methods. The results were listed as follows:1. Based on the experimental observation, DEM, NDVI and NDWI indexes, salinization extraction model of remote sensing was established using the supervised classification and decision tree classification methods. Four kinds of salinization, that is, severe, moderate, mild and salted salinization were obtained. Kappa coefficient is 0.86. This model develops the salinizetion extraction method with remote sensing information in large scale. It resolves some problems such as slow investigation, difficult cartography and low precision.2. The area of salinization and distribution in 1990, 2000 and 2015 were gotten based on the classification results. Further the spatial difference of salinization degree in the past 25 years was analyzed using transfer matrix. The results show that each kinds of salinization change greatly. The severe and moderate type has decreased from 1990 to 2015 surrounding basin. There was a lot of saline and mild saline land inside the cultivated land in 2015. The saline land proportion accounted for 74.69% of the arable land area, and the heavy saline has fallen from 7.2% to 0.34%, moderate salt dropped down from 21.33% to 3.87% during the past 25 years.3. The changes of different irrigation proportion indicate the salinization has decreased under long-term drip irrigation. In the past 25 years, non-saline land area increased by 73.54 percent, slightly salinization area decreased by 23.11%, moderate salinization area decreased by 82.05%, and severe salinization area decreased by 95.51 % when flooding to drip. Drip irrigation area accounts for more than 90% of the arable land in 2015. Drip irrigation method mainly distributes in the plains of the middle and lower reach. Flood irrigation mainly distributes in the piedmont alluvial fan with high altitude and cold climate.4. Through the analysis of saline soil samples under different irrigation methods, it shows that the base ions exists spatial difference with depth and irrigation methods. The recerch of spatial variability which under different depths and different irrigation shows that drip irrigation is more effective than flood and rained region on pH, conductivity and Mg2+. The intensity of CV in different irrigation and depths is not obvious. As a whole, CV in rained region is least, canal water drip irrigation is the most in 0-40 and 60-100 cm, Total salt content’s CV is the highest in 40-60 cm under canal flood.The soil salinization in the different depths is more serious under canal water irrigation than under well water irrigation.5. Through the seasonal soil samples under different irrigation methods, the change of soil salt ion in the microscopic scale in the summer and autumn was analyzed. The value of CV is between 0.04 and 0.65. The variation of the salt ions is very small under drip irrigation, decreases under flood irrigation, and increases under rained region. |
PDF Full Text Request