Carbon and nitrogen dynamics in agricultural soils

• Main Author: Karlsson, Thord
• Format: Doctoral or Postdoctoral Thesis
• Language: Swedish; English
• Published: 2012
• Subjects: Soil biology; Soil Science; Northern Sweden
• Online Access: https://pub.epsilon.slu.se/8902/; https://pub.epsilon.slu.se/8902/1/karlsson_t_120523.pdf

An understanding of soil organic carbon (C) and nitrogen (N) dynamics is essential for efficient and environmentally sustainable agricultural production. This thesis includes model studies of C mineralization at regional/national level with annual time steps, N balances at field level with annual time steps, and N dynamics at 34 locations within a single field with daily time steps. The same model family (ICBM) was used in all studies. Generally, carbon stocks in mineral soils increased from southern to northern Sweden and were near steady state. Temperature, crop type, and animal density were the main drivers for C fluxes at the regional scale. The inclusion of the changes in soil N stocks in field N balance sheets were necessary for obtaining unbiased estimates of N use efficiency or N surplus. Without these changes in soil N stocks, N surplus in low input systems was underestimated and was overestimated in systems with increasing soil organic matter. The model explained 56% of the total variation in apparent net N mineralization estimated for 34 locations within one field during two growing seasons; however, for a single year, only a small proportion of the variation could be explained. The sources of uncertainty in both measurements and the model at different scales in time and space were addressed, as the precision of available measurements limited model testing and the estimation of critical model parameters. From a decision support perspective, commonly available field data are usually sufficient for models such as ICBM/N in the long-term predicting of the effects of agricultural management on C and N stocks at the field or the regional scale. However, more detailed information and models that are more complex are required for short-term applications at small spatial scales, for example for decision support in a precision farming context.