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

H2: Seasonal changes affect bioturbation with a higher activity of burrowing animals during rainy seasons, when the soil is softer, and the energetic cost of digging is therefore reduced. H3: With increasing vegetation cover, the bioturbating activity of many invertebrates (includ- ing most earthworms) decreases, due to the subterranean food supply provided by fine roots in the soil independent of soil surface vegetation, while that of vertebrates increases, due to the increased availability of food and shelter. Methods Study area Our study was conducted at four sites representing a climate gradient along the coastal range of Chile (26˚S-38˚S), extending from an arid desert with a mean annual temperature of 16.8˚C and mean annual precipitation of 12 mm to a temperate humid rainforest with a mean annual temperature of 6.6˚C and mean annual precipitation of 1469 mm [37]: arid Atacama Desert, located in Pan de Azu´car National Park, semi-arid shrubland in the private reserve Santa Gra- cia, a Mediterranean forest in La Campana National Park and a humid rainforest in Nahuel- buta National Park. All approvals from the relevant authorities, i.e. the Chilean National Forest Commission (CONAF), were obtained in advance to our study and granted access to the research sites. In 2019, the year of our field campaigns, the mean temperature in the arid desert was 14.6˚C and the mean precipitation was 9.4 mm while in the humid rainforest, the mean temperature was 7.3˚C and the mean precipitation was 1885 mm [38].

To sample each research site representatively, we established 12 10 m × 10 m plots with a distance of at least 30 m between them during the first field campaign, conducted in autumn of the southern hemisphere (March to April 2019). In a second field campaign conducted in spring of the southern hemisphere (September to November 2019) we established eight addi- tional plots at each site to cover possible variation, resulting in a total of 20 plots per site. The 20 plots per research site were evenly distributed across two opposing hillsides, 10 on the north- and 10 on the south-facing hillslope. Assessment of bioturbation activity and quantity To evaluate bioturbation activity, we counted the number of all visually detectable burrow entrances on the soil surface (hole density) of each plot. We calculated the amount of soil exca- vated by burrowing animals (excavated soil volume) as an indicator of bioturbation quantity by using a caliper to measure the vertical (dv) and horizontal (dh) diameters. In addition, we defined the depth of each hole entrance (de) as the distance to the first barrier encountered by the caliper and measured this parameter. Raw data of burrow measurements can be obtained from S7 Table in S4 Appendix. Following [36, 39, 40], we calculated the (minimal) excavated soil volume assuming that the measured burrows were cone-shaped: excavated soil volume ¼ 1 3 � dv þ dh 4 � �2 � p � de: To distinguish between the burrows of invertebrates and vertebrates, burrows with a hole-

entrance diameter < 2.5 cm were assumed to be created by invertebrates and burrows with a hole-entrance diameter � 2.5 cm by vertebrates [36]. Assessment of vegetation data Vegetation cover was estimated using unmanned aerial vehicle (UAV) red green blue (RGB) images and land cover classification [41]. For each plot, we calculated the ratio of pixels 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 3 / 14