Study On Comprehensive Evaluation And Spatial Variability Of Soil Fertility On Sanjiang Plain

Soil fertility is an important content in agrology. During centuries, people never stop searching its hypostasis. Based on the current new theories and advanced technologies, the soil fertility of experimental area in Sanjiang Plain was evaluated comprehensively, and the spatial variability structure of soil fertility elements was analyzed and their spatial distributions were drawn. Then, the delineation of soil nutrition management zone was studied.Attribute recognition model based on entropy weight was used to evaluate the soil fertility of experimental area in Sanjiang Plain, which contained 100 soil samples with 7 physical and chemical index, including soil total nitrogen, soil organic matter, pH etc. Spatial distribution map of soil fertility was made by ArcGIS software. Results indicate that the overall level of soil fertility in experimental area was high, 90% of the soil was fertile and lower fertile, soil nitrogen and organic matter had degressive trend, pH was small and the soil had some acidification, and the soil phosphorus and potassium was lower. Therefore it is necessary to retain and enhance the content level of N and soil organic matter, apply P and K fertilizers or straw manuring, and farm reasonably for the sake of improving soil, preventing soil degradation and enhancing the soil fertility level of the area.Classical statistics and geostatistics were used to study on the spatial variability of seven soil fertility elements in experimental area. In nonsampled areas ordinary kriging was utilized for spatial interpolation. Using coefficients of variation, all soil fertility elements, except for pH, were moderately variable. A geostatistical analysis indicated that structural factors, such as parent material, terrain, and water table, were the main causes of the spatial correlations. Strong spatial correlations were noted with organic matter, while they were moderate for others. The effective spatial autocorrelation of all soil fertility elements ranged from 146.7 to 868.5 m. The fit of the experimental semi-variograms to the theoretical models and cross-validation indicated that Kriging could successfully interpolate all soil fertility elements. Thus, the geostatistical method used on the experimental area could accurately evaluate the spatial variability of soil fertility elements.Delineation of soil nutrient management zones provides basis for variable fertilization technique and is the important link of variable fertilization management actualized in precision agriculture. After the spatial variability characteristics and structure of seven soil nutrients were analyzed, six of them were taken as the variables to delineate soil nutrient management zones. Particle swarm optimized attribute means clustering was used to delineate management zones, and three indices were introduced to ascertain the reasonable number of management zones. According to the calculation, the reasonable number for the study area was 2. As single factor variance analysis was used to analyze soil nutrient data of the practical samples in each management zone, all soil nutrients but available phosphorus had great differences among management zones on the confidence level of 99%. Zone 2 had the higher soil fertility and zone 1 had the lower. The delineation result indicated that particle swarm optimized attribute means clustering had a good performance on delineating management zones and variable fertilization management was feasible in the study area.To sum up, the evaluation of soil fertility, the analysis on the spatial variability law of soil fertility elements, and research on soil nutrition management zone not only reflected the real situation of soil to some extent and made us know the hypostasis of soil better, but also provided new methods for the problem, and scientific basis for variable-rate fertilization, reasonable farming and sampling in agricultural science.