LAUSR

Abstract. Sulfur (S) availability plays a vital role in driving functions of terrestrial ecosystems, which can be largely affected by soil inorganic S fractions and pool size. Enhanced nitrogen (N) input can significantly affect soil S availability, but it still remains largely unknown if the N effect varies with frequency of N addition and mowing management in grasslands. To investigate changes in the soil S pool and inorganic S fractions (soluble S, adsorbed S, available S, and insoluble S), we conducted a field experiment with different frequencies (two times per year vs. monthly additions per year) and intensities (i.e., 0, 1, 2, 3, 5, 10, 15, 20, and 50 g N m−2 yr−1) of NH4NO3 addition and mowing (unmown vs. mown) over 6 years in a temperate grassland of northern China. Generally, N addition frequency, N intensity, and mowing significantly interacted with each other to affect most of the inorganic S fractions. Specifically, a significant increase in soluble S was only found at high N frequency with the increasing intensity of N addition. Increasing N addition intensity enhanced adsorbed S and available S concentrations at low N frequency in unmown plots; however, both fractions were significantly increased with N intensity at both N frequencies in mown plots. The high frequency of N addition increased the concentrations of adsorbed S and available S in comparison to the low frequency of N addition only in mown plots. Changes in soil S fractions were mainly related to soil pH, N availability, soil organic carbon (SOC), and plant S uptake. Our results suggested that N input could temporarily replenish soil-available S by promoting dissolution of soil-insoluble S with decreasing soil pH and mineralization of organic S due to increasing plant S uptake. However, the significant decrease in organic S and total S concentrations with N addition intensity in mown plots indicated that N addition together with biomass removal would eventually cause soil S depletion in this temperate grassland in the long term. Our results further indicated that using large and infrequent N additions to simulate N deposition can overestimate the main effects of N deposition and mowing management on soil S availability in semiarid grasslands.