Effect of Tillage Practices on Soil Organic Carbon as Affected by Soil Texture in Southwest Michigan.
X. Hao1, Alexandra Kravchenko2, Senthil K. Subramanian2, G. P. Robertson3, and A.J.M Smucker4. (1) Dept Crop and Soil Science, Michigan State Univ, PSSB, East Lansing, MI 48824, (2) Michigan State Univ, Dept Crop and Soil Sciences, East Lansing, MI 48824-1325, (3) Crop & Soil Sciences Dept, Michigan State Univ, W.K. Kellogg Biological Station, Hickory Corners, MI 49060, (4) Dept of Crop and Soil Sciences , Michigan State Univ, PSSB, East Lansing, MI 48824
Agricultural practices are one of the major factors that affect soil organic matter dynamics by altering above and below ground organic matter inputs and decomposition rates. Soil texture plays an important role in influencing the amounts and turnover rates of Soil Organic Carbon (SOC) as well. However, few studies have investigated the interaction between these two major factors of influence on SOC. The objective of this study is to analyze the interaction between texture and agricultural practice effects on SOC in a replicated long-term experiment with a wide range of soil textures in Alfisols. Data were collected from a Long Term Ecological Research site located in southwest Michigan, USA. The experiment has been established in 1988. The three studied management practices were Conventional Tillage (CT) and No-Till (NT) with conventional chemical inputs, and conventional tillage with legume winter cover crops and zero chemical inputs (CT-cover). Four replicate plots were studied for each treatment. Approximately 100 soil samples were taken from 0 – 5 cm depth in each plot. Soil total carbon (C) and nitrogen (N) as well as sand, clay, and silt contents were measured in all samples. Data analyses consisted of linear regression analysis (PROC REG), analysis of covariance (ANCOVA), and two-segment linear regression analysis (PROC NLIN) performed using SAS software. As expected, a strong relationship between total C and texture was observed in all three studied treatments. Sand content was found to be the most useful soil textural covariate in the studied soils. Sand content explained 41, 63 and 60% variability in total C for CT, NT and CT-cover managements, whereas silt explained 41, 58 and 51%, and clay explained only 18, 31 and 21%, respectively. However, the relationship between the total C and sand or, alternatively, (clay+silt) content was better described by a two segment regression equation than by a single line regression. Compared with a single line regression, the two-segment regression between total C and (clay + silt) content improved R2 and reduced mean square error by about 15%. The results suggested that texture effects on total C differ in coarse-textured and fine-textured soils. In relatively coarse textured soils with (clay + silt) content ranging from 280 to 570 g kg-1 soil the relationship between total C and (clay + silt) was very weak in all studied treatments, and the regression slopes in different treatments were not significantly different from each other. Within this range of (clay + silt) contents, the total C contents in NT and CT-cover were not significantly different from each other, however both of them were significantly greater than the total C content of the CT treatment. In relatively fine textured soils with (clay + silt) content ranging from 570 to 930 g kg-1 soil a much stronger significant positive relationship between total C and (clay + silt) was observed in all three studied treatments, and significant differences were observed between the regression slopes (P < 0.01). The regression slope for NT was greater than the slopes for CT and CT-cover treatments, while there was no significant difference in slopes of CT and CT-cover treatments. Higher regression slope of NT indicated that the size of increases in total C that were generated after 15 years since conversion to NT was much greater in soils with higher (clay + silt) content than in coarser-textured soils. Similar regression slopes of CT and CT-cover treatments indicated that the size of increases in total C that were generated after 15 years since conversion to CT-cover was similar across the studied textural range. The ANCOVA predictions of soil total C showed that NT and CT-cover resulted in greater total C content than CT across the whole studied range of textures while the difference in total C between NT and CT-cover were more depended on soil texture. CT-cover was found to achieve similar total C levels as those of NT in soils where (clay + silt) content was less than 750 g kg-1 soil, while NT had significantly higher total C where (clay + silt) was greater than 750 g kg-1 soil.