Grazing and light modify Silene latifolia responses to nutrients and future climate

grazing is a common practice to maintain open grasslands in Germany [71]. To prevent other larger, naturally occurring grazing animals such as deer from entering the experimental site, the entire area was fenced. Small mammalian herbivores such as the European hare (Lepus europaeus) can enter the experimental area. Five of the pastures were randomly assigned to the ambient climate, the other five to the future climate scenario. For the climate manipulation, pastures under future climatic condi- tions are equipped with metal roof structures with foldable roof and side covers, as well as with a controllable irrigation system [63]. To control for any effect of the roof structure, pastures in ambient conditions have similar structures but with non-foldable roof and side covers. The intent of the climate treatment is to increase temperatures and to simulate decreasing precipi- tation in the summer and increasing precipitation in the spring and fall, consistent with pre- dictions for this region [63]. The future climate treatment increases air and soil temperature using passive night warming for which the roof and side covers are automatically closed during the night. The amount of temperature increase varies over season and years; on average the daily minimum temperature increased by 1.14˚C [63]. To simulate increasing spring and autumn precipitation the weekly precipitation at ambient blocks is used as a reference to which ten percent is added to simulated the increase in precipitation in future blocks [63]. To

simulate decreasing summer precipitation, future blocks receive 80 percent of the weekly pre- cipitation that falls on ambient blocks [63]. Experimental design We established our experiment in 2017 using a full factorial design of herbivore (sheep) exclu- sion and nutrient enrichment in each of the ten pastures (ambient and future climate). In each of the ten pastures, we established four 1.4 × 1.4 m plots that were randomly assigned to the following treatment combinations: 1) unfertilized and grazed (e.g. control), 2) fertilized and grazed, 3) unfertilized and fenced, 4) fertilized and fenced. This resulted in 40 experimental plots, arranged in four-plot groups (called “blocks” in the statistical analyses), half of which received ambient climate and the other half future climate treatment. In later March-May and June we added a slow-release NPK fertilizer mixture (Haifa Multicote 2 M 40-0-0 (40% N), Super Triple Phosphate TPS (45% P205), potassium sulfate fertilizer (50% K2O, 45% SO3)) to each plot assigned to fertilization treatment. This corresponds to an addition of 10 g N, 10 g P and 10 g K m-2 [see the protocol of 72] per growing season to experimentally relax nutrient limitation. The fertilization treatment started in May 2017, while the herbivore exclusion treat- ment started three months later, at the end of August. We used rectangular, mobile metal fences of 1.8 × 1.8 m size, 82 cm height and 10 cm mesh size to exclude the sheep from plots assigned to the herbivore exclusion treatments. The fences allowed small rodents like mice and

voles (mainly Apodemus sylvaticus and Microtus arvalis), abundant in the experimental area, to pass [Jessen & Eskelinen, personal observation,73]. In 2019, one plot originally intended for herbivore exclusion only was accidentally fertilized and one plot originally intended for herbi- vore exclusion and fertilization treatment was not fertilized. Both plots were treated as fertil- ized in the data analysis. In addition to the fertilization and herbivore exclusion treatment we implemented a light addition treatment using modern LED lamps (C65, Valoya, Finland). For this purpose, the 1.4 × 1.4 m plots were divided into two 0.7 × 1.4 m subplots, one of which was randomly assigned to light addition treatment resulting in total 80 experimental subplots. Each lighted subplot was equipped with two 120 cm long and 3.5 cm wide high intensity LED lamps that were installed horizontally to the ground and parallel to each other. The light addition treat- ment started at the same time than fertilization treatment, i.e., at the end of May 2017. In the PLOS ONE Transplant performance and trait responses to abiotic and biotic changes PLOS ONE | https://doi.org/10.1371/journal.pone.0276789 November 8, 2022 4 / 16