Asymbiotic biological nitrogen fixation in a temperate grassland as affected by management practices

Estimates of asymbiotic biological N fixation (BNF) in temperate grasslands are few with large variations. In the past six decades, European grasslands have been subjected to intensive management practices and presently it is not known how asymbiotic BNF is influenced by these practices. Our objective was to assess the impact of fertilizer application and mowing frequency on asymbiotic BNF in a Central European grassland. In 2008, we established a three-factorial experiment with two fertilizer treatments (no fertilizer application and combined nitrogen (N), phosphorus (P) and potassium (K) fertilization at 180–30–100 kg ha−1 yr−1), two mowing frequencies (cut once and thrice per year) and three sward compositions through the application of herbicides (control, monocot- and dicot-enhanced swards). Three years after the initial sward manipulation, there was no more difference in functional group composition. Between June 2011 and May 2012, we measured in-situ asymbiotic BNF using the acetylene reduction assay, calibrated with 15N2-fixation method. Across treatments, asymbiotic BNF rates in the 0–5-cm soil depth ranged from 1.7 (±0.2 SE) kg ha−1 yr−1 for fertilized plots cut once a year to 5.7 (±2.3 SE) kg ha−1 yr−1 for unfertilized plots cut thrice a year. Fertilization decreased asymbiotic BNF, suggesting that the potential positive effect of increased soil P levels might have been overruled by the negative effect of increased soil mineral N levels. Intensive mowing stimulated asymbiotic BNF, which was probably due to an increase in rhizodeposition. Our calibration of the acetylene reduction assay with the 15N2-fixation method resulted in a conversion factor of 0.61, which largely deviates from the theoretical conversion factor of 3. Furthermore, laboratory incubations under increased soil moisture and temperature conditions overestimated BNF rates compared to in-situ measurements. Thus, laboratory measurements with altered soil moisture, temperature or disturbed soil may lead to strong biases in estimates of asymbiotic BNF. Our results suggest that input of N through BNF may be considerable in temperate grasslands. We conclude that BNF studies should be conducted in-situ and that the acetylene reduction assay should be calibrated against 15N2-fixation calibration for reliable estimates.