The Study On Linear Prediction Model Of Soil Infiltration Model Parameters

The process of forming soil water is called soil water infiltration, which in the condition of irrigation or rainfall, water infiltrating into the soil and was stored in the soil. The infiltration process of soil moisture is essentially the process of transforming surface water into soil water to supply plant root absorb and utilize. The study of soil water infiltration parameters can provide support for scientific irrigation and also can provide guidance for rationally evaluating groundwater resources. At present, the research of soil infiltration model has been basically mature, but there are some difficulties in obtaining the model parameters of soil infiltration accurately and rapidly. In the past, the test method was mainly used to obtain the parameters of soil water infiltration. The test method is able to meet the needs of agricultural irrigation water management from the requirements of accuracy, but there are some shortcomings, for example, the difficult testing technique, the complex equipment,time-consuming and laborious etc. In order to solve the above problems, In this paper, the author try to use the method of soil transfer function to establish the linear prediction model of soil infiltration parameters, which was based on thelarge regional samples of soil infiltration parameters and soil routine physical and chemical parameters. This method can further improve the theory of soil transport and provide help for guiding the practice of irrigation.In this paper, firstly, analyzing the relationship between the coefficient K,the infiltration index a, the relatively stable infiltration rate f0 of Kostiakov three parameter infiltration model and soil physical and chemical parameters, for example, soil bulk density, soil volume containing content, soil texture, soil organic matter, soil salt content, soil temperature according to the single factor analysis method which was based on the test sample data of soil infiltration and soil physical and chemical parameters. Secondly, establishing the linear prediction model for Kostiakov three parameter infiltration model of non-frozen non-saline alkali soil, saline alkali soil and frozen soil. Finally, through the comparison the linear prediction model of the situations which including before and after the soil structure deformation, putting forward the necessity of considering the deformation of soil structure that can improve the accuracy of model prediction, then the linear prediction model of Kostiakov two parameter infiltration model and Kostiakov three parameter infiltration model was established which can be applied to the whole agricultural production cycle, and the prediction results were verified which showed that the prediction accuracy can meet the requirements of agricultural production activities. In addition,discovering that the prediction accuracy of Kostiakov three parameter infiltration model was significantly higher than that of Philip infiltration modelthrough the comparison of the prediction accuracy of these two infiltration models, so Kostiakov three parameter infiltration model was selected as research object to predict the infiltration parameters of different soil, then putting forward the linear method to predict the parameters of Kostiakov three parameter infiltration model based on the basic physical and chemical parameters of soil.It is found that the main affecting factors of non-freezing non-saline soil infiltration model parameters were soil bulk density, water content(moisture),texture and soil organic matter content. In addition to above mentioned parameters, the other main affecting factor of saline alkali soil infiltration model parameters was salt content. The main affecting factors of frozen soil infiltration model parameters were soil bulk density, water content(moisture), texture and soil organic matter content, besides that soil temperature also has a great impact on the soil infiltration parameters. Based on the conclusion of the above study,establishing the linear prediction model of non-freezing and non-saline alkali Kostiakov infiltration model parameters which considering soil structure deformation, the average relative error of each parameter can be controlled below 12%, and the prediction accuracy was higher.The average relative error of infiltration model parameters of saline alkali soil can be controlled under 15%,the prediction accuracy can meet the requirements. The average relative error of infiltration model parameters of frozen soil can be controlled fewer than 19%,the prediction accuracy was able to meet the requirements of agricultural production activities basically.In this paper, the main influencing factors of soil infiltration model parameters were considered, but some secondary factors without consideration like the farming measures, soil crust and compaction, groundwater depth. In the further study, these secondary factors can be introduced to improve the prediction accuracy. The prediction accuracy of frozen Kostiakov three parameter infiltration models was relatively low, so in order to establish a more accurate multivariate linear prediction model, a deeper study on the influencing factors of frozen soil is necessary from the aspect of quantitative indicators.