Prediction Of Spatial Distribution And Management Zone Of SOM In A Mollisol Watershed Of China

The research of spatial prediction of soil organic matter(SOM)is of great significance mastering the current situation of spatial distribution of SOM,implementing precision agriculture and protecting the regional ecological environment.In this study,the Mollisol Watershed of China was selected as the study area,and the soil organic matter was taken as the research object.Discussing deeply the spatial variability characteristics of soil organic matter in the study area by constructing semi-variogram model by GS+ and calculating Moran index by Arc GIS.On this basis,with land use types as the auxiliary variable,predicting the spatial distribution of SOM in the surface soil of study area by using Bayesian Maximum Entropy(BME)method,and compared with the Co-Kriging(CK)which used soil total nitrogen and land use type as a supplementary variable,then discussed feasibility and precision of the two methods,in order to improve the spatial interpolation and mapping accuracy of soil properties in typical black soil areas.Reasonable soil nutrient management zone is one of the preconditions to realize precision agriculture.This paper researches the management zone of regional soil organic matter by using fuzzy c-means clustering method,determines the optimum number and area of management zone,and puts forward relevant suggestions or measures.the results show that:(1)Soil organic matter content varies from 6.30 g/kg to 57.30 g/kg,with an average of 24.04g/kg.There are significant differences in soil organic matter content among different land use types.Among them,paddy field soil has the highest organic matter content,forest soil has the lowest organic matter content,and the variable coefficient of all sampling sites is 44.80%,and all of them belong to the moderate degree of variation.The original data conforms to normal distribution and is suitable for interpolation modeling analysis.(2)The variation characteristics of soil organic matter in the study area conforms to the spherical function model,and has a good spatial structure.The spatial variability of soil organic matter in the study area is less affected by some random factors such as human activities,but more affected by structural factors.The global Moran's I index of soil organic matter in the study area is-0.0739,P value and |Z| value don't reach the level of spatial significance,and spatial autocorrelation isn't significant.(3)The correlation analysis of soil organic matter and soil properties shows that w(SOM)was positively correlated with w(TN)in the study area.Variance analysis of soil organic matter and land use types in the study area shows that there is a certain correlation between land use types and SOM.Therefore,soil TN and land use type are selected as auxiliary variables in CK interpolationand land use type as auxiliary variables in BME interpolation.(4)The spatial distribution of soil organic matter obtained by CK method and BME method have roughly the same trend.The distribution of soil organic matter decreases stepwise from west to east,and the content of SOM varies little in the North-South direction.But,the details of the interpolation patches using BME method is more prominent,it avoids the smoothing effect of traditional Kriging interpolation,and the interpolation results conform to the actual situation of the study area.CK method and BME method have good estimation accuracy for w(SOM).Among them,the average error,average absolute error and root mean square error of w(SOM)obtained by BME method are 3.22,3.86 and 4.01,which are lower than CK method.Therefore,BME method is superior to CK method in making full use of auxiliary information and spatial prediction.(5)By using the fuzzy c-means clustering method,the optimum number of management zone of soil organic matter in the study area is 3,and the content of organic matter in zone 1,2 and 3shows a decreasing trend.Meanwhile,variance analysis is used to verify the reasonableness and effectiveness of the zoning results,which provides a scientific reference for the next management zoning of soil organic matter.