Spatial variability of soil carbon storage and root production in conventional and perennial agroecosystems

Background/Question/Methods

Agricultural management may substantially influence regional carbon (C) cycling, particularly with anticipated shifts to cellulosic bioenergy production. Data quantifying ecosystem processes driving soil C storage such as root productivity of bioenergy crops and their impacts on soil organic matter (SOM) pools are critical for development of a sustainable bioeconomy. Characterizing the complex interactions between edaphic factors and the root‐SOM continuum at landscape scales is a critical gap that must be addressed to meet these data needs. In this study we use a topographic gradient to investigate how variation in soil properties influences belowground processes to improve knowledge of ecosystem C flow at landscape scales. Our objective was to determine the variation in C inputs from root production and the subsequent changes in soil aggregation and POM pools over three years in response to heterogeneity in edaphic conditions. Using root ingrowth cores, we measured annual root production to 20 cm soil depth within continuous corn, a triticale/sorghum double crop, and switchgrass. We quantified soil aggregation and particulate organic matter (POM) pools in 2009 during the first growing season and again in 2012 to determine changes in response to C inputs from root production over three years.

Results/Conclusions

Our results show that in 2009, following decades of corn-soybean production and prior to the establishment of bioenergy crops, soil properties varied across the topographic gradient. Soil aggregation was higher (P<0.05) on the toeslope and floodplain, and decreased in upslope positions. Levels of POM varied with soil aggregation across landscape positions. Root production differed among the three cropping systems, with the highest production under switchgrass and the lowest with continuous corn. Root productivity in the annual crops was not influenced by soil properties; however switchgrass root production was associated with greater soil sand content, which explained 44% of the variation. Percent sand was correlated (R² > 0.7) with low soil C and nitrogen and high bulk density that were also significant predictors of switchgrass root production. Over three years, changes in soil aggregation and POM pools were significantly correlated with greater root productivity, however these changes were not consistent across landscape positions. These results indicate the importance of both cropping system and soil properties on changes in SOM with conversion to bioenergy. Quantifying these interactions can improve C cycling modeling efforts, allowing for more accurate scaling of measurements to predictions over broad spatial scales.