Research On The Spatial And Temporal Variation Of The Soil Nutrient And Maize Yield And Their Correlation In The Maize Main Production Area

The vastness of our national territory, large population and topography determine the nature of people living condition is more complex. Each area exists phenomena including of limited agricultural production materials, unbalanced distribution of agricultural production resources, un-abundance of per capita agricultural resources, and so on. These phenomena show that the agriculture is one of the main factors which determine the rapid development of the national economy.Under the influence of geographical conditions and meteorological factors，the agricultural production has some complex characteristics including of dynamics, uncertainty, time, space, and so on. The traditional technical means can not grasp the rural condition correctly and timely which leads to the passive situation of the agricultural production is in the behind of the informationization chronically. Computer-related technologies are effective ways to achieve the sustainable development of agricultural informationization. Along with the society’s progress and the development of information technology, traditional means of agricultural data processing have been unable to meet the development of agricultural informationization and digitization. The emergence of advanced and computer-related technology, such as3S technology, artificial intelligence and data mining broadens the vision and touch of humanity and provides new means of observing things.Corn is one of main grain crops in China and it is the main crops in the northeast. The maize production is very important in the whole grain production. Understanding of corn planting and production has great significance for yield prediction, soil nutrient analysis and precision fertilization. It has beening the hot spot that taking use of information technology to deal with agricultural knowledge decision and understand timely important rural conditions including of soil nutrient content, fertility level and expected output of grain. Taking agricultural informationization as the guidance and aiming at force situation of arable land in the corn production process, this paper takes the corn yield cared mostly as the research object, uses GPS and GIS technology, combines with temporal sequence algorithm, spatial clustering analysis and data mining technology combined with time and space, researches the correlation between the soil nutrient content and the yield, explores the method of dividing the level of soil fertility, predicts the maize yield and explores new methods for maize yield prediction.The research of this paper analyzes firstly the correlated research of the composition of grain yield, the progress of agricultural informationization, the development situation of precision agricultural, the spatial and temporal variation of the soil nutrient and the spatial and temporal correlation of crop production both at home and abroad. Then this paper indicates the main research content, purpose, significance and related work and introduces data mining theories and methods such as the time series, spatial fuzzy clustering analysis, correlation analysis and the neural network. Supported by the national863high-tech project of “Research and Application of Corn Precision Operating System” undertaken by the information technology institute of the Jilin Agricultural University, the paper takes the farmland in GongPeng town YuShu City JiLin province as the experimental area and conducts the following research work:(1) Collect and reorganize data. By collecting data of soil nutrient in corn-growing areas and maize yield during the experiment and obtaining maize yield data in YuShu City from1990to2010by consulting the yearbook database of JiLin Province, it provides the basis data for building soil nutrient evaluation model, the correlation analysis model between soil nutrient and maize yield and maize yield prediction model.(2) Analyse the spatial variability of soil nutrient and the relevance between soil nutrient and yield and discuss the effects of spatial variability of soil nutrient on the yield aiming at the spatial characteristics of soil nutrient data.(3) Research the comprehensive evaluation of cultivated land in black soil region of Jilin province according to the study on the correlation between maize yield and soil nutrient variation. Aiming at the area space attribute of the research object, this paper uses spatial fuzzy c-means clustering algorithm to evaluate soil nutrients in solving the region of the maize precise fertilization issues. At first, it uses the eight-connected method to have a spatial analysis on research area and applies the analysis result in the spatial fuzzy clustering analysis to build the soil nutrient classification model. Applying this method to soil nutrient evaluation, it considers both the attribute problem of the soil nutrient data and the regional characteristics of the soil nutrient data so it is suitable for the evaluation of soil nutrients in the precision agriculture. Through GIS analysis results we can understand visually the farmland throughout the region and provide the theoretical basis for crop fertilization and management. （4）The prediction of maize yield is an important part in the implementation process of precision agriculture. Traditional statistical methods do not take into account time correlation of maize yield. This paper predicts the corn yield from time perspective. Taking20consecutive years maize production as objects and basing on the research algorithm of the time series algorithm, it builds a corn yield time-series model and analyses trends of the corn yield from the time point and takes related forecasts. Experimental results show that the effect of applying ARIMA (Autoregressive Integrated Moving Average Model, denoted ARIMA) model to predict maize yield and the actual values is very well and state that using the time series algorithm can predict future trends of maize yield well. It provides new ideas and methods for maize yield prediction analysis.(5) Based on spatial variability of soil nutrients and the time prediction of maize yield, explore the method of fitting time and space and predict maize yield from the perspective of time and space. This paper proposes a research method on maize yield with the space-time correlation based on data mining which uses the method of time series to build time child prediction of maize yield. By means of the learn feature of BP neural network algorithm, study the planting area of adjacent plots soil nutrient effect on maize yield and then build the space child prediction of the maize yield. Combined with the linear regression fusion time child prediction and space child prediction, produce the final prediction model.This paper is based on the research of the spatio-temporal correlation on corn yield and soil nutrient of data mining and takes the data mining methods and soil nutrient evaluation, spatio-temporal variability of soil nutrients and maize yield relevance into an effective integration according to certain spatial variability and time variability both in soil nutrients and maize yield. The proposition of the research method can provide the theory basis to evaluate soil nutrient, implement precise fertilization and predict yield.