Is litter decomposition enhanced in species mixtures? A meta-analysis

Abstract

Litter decomposition is a key process in the carbon balance of soils. Commonly, plant litters occur in mixtures where the species differ in quality traits such as the nutrient concentration and organic carbon quality. Many studies explored if mixing litters retards or speeds up litter decomposition compared to species decomposing alone, with varying results. To identify consistent trends with an overarching quantitative synthesis, we test in a meta-analysis whether on average across studies, the mass loss of mixed litters of two plant species is faster than the average mass loss of single litters. We hypothesise that larger trait divergence of the litter quality of the species in a mixture results in a faster mass loss of the mixture than expected based on the single species. Furthermore we hypothesise that part of the variation in litter mixture mass loss can be explained by experimental design and environmental factors. Explanatory variables used were chemical litter trait dissimilarity in the C, N, P, lignin, cellulose, phenolics concentration as well as soil properties, ecosystem, climate, the duration of litter decomposition and the experimental design. Interactions were studied if supported by mechanistic hypotheses. In the majority of studies and on average, we found that the mass loss of mixed litters is equal to the weighted average of the mass loss of the constituent single litters. None of the hypothesised explanatory variables was consistently associated with litter mixture effects on the mass loss and explained variation in mass loss of significant models was invariably only a few percent of all variation. While further data exploration might elucidate further, interactive, patterns, many of these could not be explored due to lacking data. This meta-analysis therefore refutes the notion that mixing litters in general enhances rates of decomposition. We conclude that the effects of litter mixing are in many cases predictable from the decomposition rates of the individual species. According to our results, any interactive effects (positive or negative) between litter species are contextual, and cannot be generalized and predicted beyond the context in which the results were obtained.