Vegetation and vertebrate abundance as drivers of bioturbation patterns along a climate gradient

total including fixed and random effects, we calculated R-squared for the fitted models using the rsq command from the rsq package [49]. We additionally performed an ANOVA between all possible combinations of fixed predictors retained within the fitted model to evaluate the significance of certain combinations between predictors using the anova command and per- forming a Chi-square test [50]. Results First, we investigated the allometric relationships by examining the relationship between the mean diameter and mean depth of the burrows and between the hole density and excavated soil volume. While the respective estimates of the slope are presented herein, our focus is on the RMA slopes. For both, invertebrates and vertebrates, the slopes showed a positive allome- tric relationship (Fig 1A, Table 2) that was maintained also in the single-season analysis (S1 Fig, S2 Table in S2 Appendix). However, note that the statistical tests evaluating burrow char- acteristics and the excavated soil volume were not strictly independent, as the former parame- ter was used to calculate the latter. Hole density was always greater for invertebrates than for vertebrates (Fig 2A). For inverte- brates, hole density decreased continuously from the arid site Pan de Azu´car (raw mean and standard deviation: 14 ± 7.8 no/ 100 m-2) to the humid site Nahuelbuta (2.8 ± 3.1 no/ 100 m-2) while hole density for vertebrates was highest in the semi-arid site Santa Gracia (9.1 ± 9.7 no/ 100 m-2) and remained similar in the other three sites (Pan de Azu´car: 2.8 ± 2.9 no/ 100 m-2,

La Campana: 5.6 ± 8.7 no/ 100 m-2, Nahuelbuta: 2.2 ±2.1 no/ 100 m-2, S1 Fig in S1 Appendix). Overall, the pattern of excavated soil volume from arid to humid was hump-shaped for ver- tebrates (largest in La Campana), whereas for invertebrates we could not determine a clear geographic pattern along the gradient (Fig 2B). In each site, the soil volume excavated by verte- brates was larger. This difference between the two groups of bioturbators was especially clear in the Mediterranean site La Campana (raw mean and standard deviation for invertebrates: 0.00019 ± 0.00016 m3 ha-1, for vertebrates: 0.06 ± 0.18 m3 ha-1) and the humid site Nahuelbuta Fig 1. Relationships between burrow parameters and bioturbation parameters created by burrowing invertebrates (yellow) and vertebrates (blue). (A) Relationship between the depth and mean diameter of the holes, (B) relationship between the excavated soil volume and hole density. The regression lines are derived from the reduced major axis analysis. Note that both axes in (A) and (B) were log10-scaled. Data from both field campaigns and all sites were used. https://doi.org/10.1371/journal.pone.0264408.g001 PLOS ONE Vegetation and vertebrate abundance as drivers of bioturbation patterns along a climate gradient PLOS ONE | https://doi.org/10.1371/journal.pone.0264408 March 4, 2022 5 / 14