Estimation and spatial variability of organic matter in deep soil of farmland at county scale

The county area is the basic unit for implementing green development of agriculture. In the farmland soil, not only the organic matter of the plough layer will affect the soil fertility, but also the role of deep organic matter can be neglected. Therefore, it is of great significance to accurately estimate the deep organic matter content of the farmland based on the county scale. This study selected the farmland in Yongji City, Yuncheng City, Shanxi Province as the research area. According to the soil types, 8 plots were determined, 3 vertical sections for each plot, and each vertical section was sampled by layer (10 cm for 1 layer). Three samples were randomly selected from each layer, and soil samples were mixed in the same layer. A mixed sample was formed, and a total of 18 layers of 180 cm depth were obtained, and a total of 144 samples were mixed. The organic matter content of each layer of soil was determined by the potassium dichromate volumetric method. A model for estimating the content of deep organic matter in the surface layer (0-20 cm) was established. Based on variogram and spatial autocorrelation, a total of 6 soil organic matters were analyzed from 0 to 30 cm, 30 to 60 cm, 60 to 90 cm, 90 to 120 cm, 120 to 150 cm and 150 to 180 cm. The spatial variability and clustering characteristics were tested and the correlation test was carried out. The Kriging interpolation method was used to predict the organic matter content of the farmland in the study area. The results showed that: 1) The content of soil organic matter decreased with the increase of depth and decreased with negative index (R2=0.80, P<0.01), and the rate of decline of soil organic matter content in the range of 0-60 cm was greater than that of 60-180 cm. The organic matter content data of each soil layer accorded with the normal distribution (P>0.05), which was moderately mutated. The degree of variation of organic matter in each layer was different, ranging from 35.89% to 47.84 %. 2) The organic matter content at any depth could be estimated by the surface organic matter content, and the fitting accuracy R2 =0.90 (P<0.01), the error was less than 16%, accounting for 49.6%, and between 16% and 40%, accounting for 44.1%. 3) The index model was the best model to reflect the spatial structure of organic matter in this region (R2>0.80, RSS<0.001). The sill (C0/(C0+ C1)) of each soil layer in the study area was between 61.54% and 72.45%, which was moderately spatially correlated. The random factor contributed a lot to the spatial structure variation of organic matter content. 4) The global Moran index of Moran's I was 0.26, and the spatial distribution of organic matter content was positively correlated, and passed the 0.01 significance test. The organic matter content of farmland in the study area had high value clustering (High-High), low-valued aggregate (Low-Low), high value surrounded by low-valued (High-Low), and low-value surrounded by low-valued (Low-High). In space, it was characterized by low concentration of organic matter in the north and high concentration in the south. 5) Kriging interpolation could better predict the organic matter content of each layer in the study area, with high prediction accuracy and good stability. The prediction results showed that the organic matter content of the farmland layer (0-30 cm) in the study area was medium; and the organic matter contents of 30-60 and 60-90 cm were lower; the organic matters at a depth of 90-120, 120-150, 150 to 180 cm were very low. It could be seen that the organic matter content of the study area was not high and the soil fertility was moderate. It was an estimation of deep organic matter at the county scale, adjusting agronomic measures, improving soil fertility, and achieving soil weight loss and efficiency. The study provides a basis for green production and efficiency. © 2019, Editorial Department of the Transactions of the Chinese Society of Agricultural Engineering. All right reserved.