Litter Carbon And Nutrient Chemistry Control The Magnitude Of Soil Priming Effect

Plant litter inputs can promote the decomposition of soil organic matter (OM) through the priming effect (PE). However, whereas leaf litter chemistry has long been identified as the primary driver of litter decomposition within biomes world-wide, little is known about how litter chemical traits influence the occurrence and strength of the PE. Here, we studied the effects of 15 co-occurring C-3 leaf litters of contrasting chemistry on C-4 soil respiration by analysing changes in C-13 natural abundance during early and later stages of litter decomposition (up to 125days). Besides an apparent PE of 16% in the first 3days, soil C respiration was increased by 24% on average with leaf litter addition in the initial stage of decomposition (4-26days) and by 8% at later stages (27-125days). Most interestingly, soil PE related well to initial litter chemistry and the dominant factors influencing the magnitude of the PE changed with decomposition stage. In the early stage of decomposition, litter leachate C content and litter hemicellulose concentration were positively correlated with the strength of the PE, whereas tannin concentration was negatively associated with soil PE. Together, tannin and hemicellulose explained half of the observed variation in the PE (R-2=0.58). In the later phase of decomposition, lignin and lignin:N ratios were negatively related to the PE, whereas Ca, K and Mg concentrations were positively related to the PE; lignin alone gave the best prediction of the PE (R-2=0.58) at later decomposition stages. Our findings provide evidence that the magnitude and direction of the PE is influenced by the chemistry of OM inputs and suggest that, as decomposition proceeds differently among litter of contrasting chemistry, litters can also have variable effect on soil PE through time. The predictive power of litter chemical traits on soil PE opens new perspectives for improving our mechanistic understanding of soil PE and improving our abilities to model soil C dynamics at variable scales.A plain language summary is available for this article.