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

(invertebrates: 0.00015 ± 0.00022 m3 ha-1, vertebrates: 0.012 ± 0.02 m3 ha-1, S1 Fig in S1 Appendix). Correcting the amount of excavated soil volume for the number of holes, the geo- graphic pattern revealed by the residuals was similar to that obtained based on the analysis of the raw data (Fig 2C); thus, the excavated soil volume was larger for vertebrates than for inver- tebrates, especially large at the two southern sites. All predictors for the response variable hole density were significant in the GLMM, with the fixed predictors explaining 48% and the random predictor plot number explaining 39% of the variation (AIC = 2030.7, p < 0.001, S3 and S5 Tables in S3 Appendix). The overall hole density was higher in Santa Gracia and Nahuelbuta during the field campaign from March to April than during the field campaign from September to November while in Pan de Azu´car there was no difference between the two seasons (Fig 3A). For invertebrates, hole density decreased at all sites with increasing vegetation cover. The hole density of vertebrates was Fig 2. Bioturbation patterns of invertebrates (yellow) and vertebrates (blue) in each site (Pan de Azu´car, Santa Gracia, La Campana, Nahuelbuta). (A) Median hole density based on the raw data, (B) median excavated soil volume of holes, (C) the residuals of the excavated soil volume (log10-transformed) after correcting for hole density (log10- transformed) using separate regressions for the two animal groups. Note that the x-axis in (B) and (C) was log10-scaled. Data from the field campaign from September to November were used.

https://doi.org/10.1371/journal.pone.0264408.g002 Table 2. Ordinary least squares (OLS) and reduced major axis (RMA) regression analyses of the relationships between the depth and mean diameter of the holes and between the excavated soil volume and hole density for invertebrates and vertebrates (all variables log10-transformed). invertebrate vertebrate Relation method r slope SE p r slope SE p Depth and diameter OLS 0.31 0.729 0.053 <0.001��� 0.44 0.996 0.045 0.93 (mixed model) OLS 0.32 0.628 0.057 <0.001��� 0.66 1.04 0.043 0.3 RMA 0.61 1.97 0.057 <0.001��� 0.53 1.57 0.43 <0.001��� Excavated soil volume and hole density OLS 0.55 0.933 0.10 0.5 0.66 1.76 0.13 <0.001��� (mixed model) OLS 0.69 0.955 0.10 0.64 0.81 1.77 0.13 <0.001��� RMA 0.48 1.38 0.10 <0.001��� 0.66 2.17 0.13 <0.001��� A slope of one represents an isometric relationship. Depicted are statistical method, correlation coefficient, slope, standard error (SE) and p-value (p) of the offset. Significant effects are labelled with asterisks: �:<0.1, ��:<0.01 ���:<0.001. Data from both field campaigns and all sites were used. Further information on the statistical analysis is provided in the Methods section. https://doi.org/10.1371/journal.pone.0264408.t002 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 6 / 14