| In this paper, combined by geographic information systems and neural network evaluation method, studied the spatial variability of soil nutrients under farm-scale red soil hilly region of Guangxi Lu Tang sugarcane field in1980,1999and2009, and also studied spatial-temporal variability of soil nutrients in continuous10yearsã€19years and29years of sugarcane fields. Using of neural network toolbox in MATLAB established a sugarcane Soil Fertility comprehensive evaluation model. The results were as follows:1ã€Soil properties under three years sugarcane farm-scale traditional statistical analysis showed that each of the three years of soil properties are characterized by a certain degree of variability. Which in1980except for pH, organic matter, avail phosphorus and potassium coefficients of variation are more than22%; In1999except for pH, organic matter and nitrogen coefficient of variation are less than22%, the rest are more than22%; In2009in addition to phosphorus the coefficient of variation of more than22%, the rest of the pH, organic matter, total nitrogen and potassium coefficient of variation are less than22%.2ã€The three years spatial interpolation distribution properties of soil nutrients shows a significant distribution.1980graph shows the distribution of spatial interpolation neutral soil, organic matter and other nutrient elements have distinct spatial characteristics; space for each character in1999the trend shows the distribution of soil organic matter and macronutrients have different degrees of increase, but the micronutrient levels were low;2009phosphorus and potassium content have improved to varying degrees, there is a certain degree of soil acidification.3ã€After a long sugarcane cropping, the nature of the various nutrients are present temporal and spatial variability of soil characteristics, soil acidification of sugarcane has increased by some trend. From1980to1999, after19years of continuous cropping, soil acidification aggravation. From1980to2009the29-year cropping, soil organic matter and other nutrients have a great amount of change in the degree of reduction in area. From1999to2009the10-year cropping, there is a certain degree of soil acidification, changes in the amount of soil organic matter and Avail.K fields has a certain degree of reduction.4ã€Soil Fertility BP comprehensive evaluation model are built based on neural network toolbox, It has got the Soil Fertility comprehensive evaluation of the results of1980ã€1999and2009. The results showed Soil Fertility results are belonging to one or two, soil nutrient content are high level or very high level. |