19. Decomposition of mixtures

Organizer Bart Hoorens. First meeting 19-21 November 2006

PARTICIPANTS Helen Quested, Stockholm U, Sweden ( VISTA, traits, decomp.) Bart Hoorens, U Illinois, now Vrije U, Amersterdam (global change and decomp.) David Wardle, Swedish U, Sweden (linking aboveground and belowground) Valerie Eviner, University of California, Davis Natalia Pérez-Harguindeguy, U Córdoba, Argentina (pfts, traits and decomposition) contributor not present at first meeting: Stephan Hättenschwiler (CEFE-CNRS Montpellier, France)

21 Nov 06>

Background
In most ecosystems plant litter does not decompose as monospecific litter, but in mixtures of different litter types. Numerous studies have shown that in many litter mixtures the component species interact and cause the decomposition rate of the litter mixture to be different from what would be expected based upon the decomposition of the component species decomposed singly. Such interactions can be either positive or negative, respectively accelerating or slowing down the decomposition of the mixture. The common occurrence of synergistic and antagonistic effects during the decomposition of mixed litter implies that the overall, ecosystem-level decomposition cannot be easily extrapolated from the known decomposition of the component species.

Many papers have reported on the decomposition of litter mixtures and, relatively recently, the topic has been reviewed by Gartner & Cardon (2004, Oikos 104: 230-246) and Hättenschwiler et al. (2005, Annu. Rev. Ecol. Evol. Syst. 36: 191-218). Despite these reviews there still remain many questions regarding the decomposition of litter mixtures. For example, it still remains unclear what are the underlying mechanisms that cause litter mixture effects. This working group brings together scientists involved in litter mixture experiments in different ecosystems around the world and their data-sets.

Summary of dataset
The dataset assembled at the first meeting consists of information on single species and mixed litter decomposition from 1,113 two-species litter mixtures, along with litter quality and study site data, from 33 studies. A wide range of biomes, habitats and plant growth forms are included. We focus on patterns in, and drivers of, the direction and magnitude of non-additive interactions between decomposing litters of different species. Contrary to an earlier suggestion in the literature, positive interactions were no more frequent than negative interactions in the dataset as a whole. Overall, 40.2 % of mixtures showed negative non-additive effects, 20.5 % were additive (within 2.5 % of 0) and 39.3 % were positive. Few studies showed an overall tendency towards positive or negative non-additive interactions.

Plant functional types : Based on preliminary analyses, mixtures of litters from contrasting functional groups did not generally result in stronger or more frequent non-additive interactions. There was no evidence that mixtures of broadleaves and needle leaves decomposed faster than expected, which contradicts a long standing hypothesis. In contrast, mixtures of Sphagnum spp. and vascular plant litter did decompose faster than expected, which provides support for the hypothesis that bryophytes can act as ‘sponges’, mitigating moisture limitation on decomposition. There was also evidence that mixtures of herb and shrub litters showed greater negative non-additive effects than expected. The only other group of mixtures that showed significant non-additive interactions was grass litter mixed with other grass litter, which further supports the idea that, in general, mixtures of dissimilar growth forms or functional groups do not show greater non-additive interactions.

Litter quality as a predictor of non-additive effects : Based on preliminary analyses, single litter quality variables (mean mixture C, N, P, lignin or total phenolic concentration, and differences between the species in mixtures) did not generally explain the non-additive effects in litter mixtures, either across the dataset as a whole, or within the selected studies consisting of more than 8 litter mixtures. Since a number of litter quality measures are likely to be relevant, we explored the potential for combining litter quality variables, either via multivariate statistics or by an index of ‘Functional dissimilarity’ (Walker et al. 1999).

Future planning:- It is anticipated that the data set will be finalized by February 2007, and analysis and some writing will proceed March-September 2007. Further meetings will be considered after that.

Last updated December 2006