Geochemical Characteristics Of Soil Nutrients In Typical Black Soil Region Based On GIS

In this paper,Shertala Town of Hulun Buir City is taken as the study area.On the basis of measuring and collecting the relevant data of cultivated land soil nutrients(organic matter,total nitrogen,available phosphorus and available potassium),using spatial statistics and GIS spatial and temporal analysis technology.Combined with the method,through adaptability analysis,the soil nutrient prediction model of typical black soil region was constructed and optimized;Furthermore,the optimization method of prediction model which can quantitatively eliminate the disturbance factors of cultivated land is innovatively put forward;Finally,based on the optimization model,the geochemical characteristics of soil nutrients and the quality of cultivated land in typical black soil areas in recent ten years were comprehensively evaluated.(1)Based on the applicability analysis of the prediction method,the optimization of soil nutrient prediction model in typical black soil region was realized.Based on the reflection spectrum characteristic data of soil nutrient content in the study area,the correlation analysis is carried out,and the applicability analysis of the prediction model is carried out in combination with the data analysis results and previous research results According to the four soil nutrients of soil organic matter,total nitrogen,available phosphorus and available potassium in the study area,the prediction models of soil nutrients were constructed respectively,and finally the optimization of the prediction models of soil nutrients was realized.The analysis shows that the random forest prediction method is superior to the commonly used partial least squares and support vector machine.Among them,the R~2 of the model of soil organic matter in each year is between 0.73 and 0.76,the R~2 of the model of total nitrogen in each year is between 0.67 and 0.85,and the R~2 of the model of soil available phosphorus and soil available potassium in each year is between 0.30 and 0.41,which proves that the model has high accuracy;The RPD of each year model of soil organic matter is between 1.47 and 1.72,and the RPD of each year model of total nitrogen is between1.41 and 2.32,which proves that the prediction model of soil organic matter and total nitrogen content is stable.(2)Based on the disturbance factors of cultivated land,the prediction model of soil nutrients in typical black soil region was optimized.The influence of disturbance factors was not quantitatively eliminated in the prediction process of conventional models,and it was found that the prediction of soil nutrients in the study area was greatly disturbed by crop types in cultivated land;Therefore,in this study,Bayesian maximum entropy method was used to quantitatively analyze the disturbance information of cultivated land crop types,and the optimized soil nutrient prediction model was further optimized.By comparison,it can be seen that the RMSE of organic matter model decreased by 0.29 and the relative accuracy increased by 7.85%in the optimization study of four soil nutrient models.The RMSE of total nitrogen model decreased by 0.01,and the relative accuracy increased by 3.11%.The RMSE of available phosphorus model decreased by 0.11 and the relative accuracy increased by3.27%.The RMSE of the available potassium model is reduced by 0.96,and the relative accuracy is improved by 4.25%.The four soil nutrient prediction models in the study area have been improved in accuracy and optimized in theory,which proves that the optimization method based on Bayesian maximum entropy theory to quantify disturbance factors is scientific.(3)Taking Sheltara Town of Hulun Buir City as an example,based on the optimized high-precision soil nutrient prediction model,the practical application analysis of this typical black soil region was carried out,and the geochemical characteristics of cultivated land soil nutrients in this region were completed.Quantitative prediction analysis of temporal and spatial variation characteristics.The results showed that the overall nutrient content in the study area was high,which could support the high yield of basic crops,but the soil nutrient abundance decreased significantly in recent ten years,and 44%of the cultivated land quality grades in the study area decreased from second grade to third to fourth grade.