Data and code from: Inverse effects of soil moisture and litter quality on litter decomposition along a gradient from hyper-arid to temperate climate

Litter decomposition, a key component of the global carbon cycle, is greatly affected by climate. Our current understanding of climate-change effects on decomposition stems from 1) spatial observational studies along climate gradients, where direct and indirect effects of climate on litter decomposition are confounded; 2) reciprocal litter translocations and common garden experiments, where biotic interactions are disrupted or 3) manipulation experiments, which can be less realistic than observational studies. Experimental studies combining all of the above-mentioned methods can separate indirect from direct climate effects on decomposition, as the confounding effects of one method can be explained with another method. Additionally, this setup can directly test if observations along a spatial gradient can predict responses to climate change, i.e. the validity of the space-for-time approach. However, studies combining the different methods are still scarce. We combined a pronounced climate gradient, large- and small-scale reciprocal litter translocations (local litter and standard litter, i.e. tea), and in situ precipitation manipulation for studying soil moisture effects on litter decomposition. All our experiments (translocation of litter and tea across the gradient, slope comparisons within sites, and rainfall exclusions) indicated positive direct effects of climate on litter decomposition. However, as local litter quality decreased with increasing precipitation, litter from species of the drier sites decomposed quicker than litter from species of the wetter sites across the gradient. In other words, while the direct effects of climate favoured litter decomposition in wetter sites, its indirect effect (i.e. litter quality) favoured the decomposition of litter from species of the drier sites within each site. Synthesis: Our results highlight the need for experimental evidence from reciprocal translocations and climate manipulations in litter decomposition studies as they indicate that space cannot always substitute for time. Such evidence would improve predictions of models of the global carbon cycle that include interactions between climate and vegetation.

Methods
At each site, we selected five abundant and representative native species, making sure the species provided a considerable fraction of the litter input. (arid: Heliotropium pycnophyllum Phil., Nolana crassulifolia Poepp., Nolana mollis I.M. Johnst., Ophryosporus triangularis Meyen, Tetragonia maritima Barnéoud; semi-arid: Cordia decandra Hook. & Arn., Flourensia thurifera (Molina) DC, Lobelia polyphylla Hook. & Arn., Maytenus boaria Molina, Senna cumingii (Hook. & Arn.) H.S. Irwin; Mediterranean: Aristeguietia salvia (Colla) R.M. King & H. Rob., Cestrum parqui (Lam.) L`Hér., Jubaea chilensis (Molina) Baill., Podanthus mitiqui Lind., Quillaja saponaria Molina; temperate: Araucaria araucana (Molina) K. Koch, Chusquea culeou É. Desv., Festuca sp., Nothofagus antarctica (G. Forst.) Oerst., Usnea sp.). In addition, Lipton® green tea (Camellia sinensis, EAN Nr.: 8 722700 055525, from here on “tea”) was used as a standard litter (Keuskamp, Dingemans, Lehtinen, Sarneel, & Hefting, 2013) to separate the influence of the decomposition environment from the climate effects, and make the study comparable to other decomposition studies.

Litterbags of all species were fully reciprocally distributed along the climate gradient and placed in the independent plots on dry and wet exposures together with two tea bags per plot. Litter and tea bags were collected at three points in time to account for the temporal dynamics of decomposition: after 3, 6, and 12 months.

Only local species (species occurring at the manipulated sites) and tea were used in the drought in-situ climate manipulation experiment.

All retrieved bags were dried at 40 °C for at least 72 hours until stable weight, after which the remaining litter was weighed. Mass loss was calculated as a proportion of the initial weight: (dry weightinitial - dry weightend) / dry weightinitial.

Carbon to nitrogen ratio (C/N) were used as a proxy for litter quality.

For more details on the methodology, please refer to the associated article and supplementary material associated to this dataset.