==== Front PLoS One PLoS One plos PLOS ONE 1932-6203 Public Library of Science San Francisco, CA USA 10.1371/journal.pone.0287039 PONE-D-22-34453 Research Article Biology and Life Sciences Ecology Ecological Metrics Species Diversity Ecology and Environmental Sciences Ecology Ecological Metrics Species Diversity Biology and Life Sciences Agriculture Agrochemicals Fertilizers Earth Sciences Geography Human Geography Land Use Social Sciences Human Geography Land Use Biology and Life Sciences Ecology Biodiversity Ecology and Environmental Sciences Ecology Biodiversity Biology and Life Sciences Ecology Ecological Metrics Species Diversity Shannon Index Ecology and Environmental Sciences Ecology Ecological Metrics Species Diversity Shannon Index Biology and Life Sciences Ecology Plant Ecology Plant Communities Grasslands Ecology and Environmental Sciences Ecology Plant Ecology Plant Communities Grasslands Biology and Life Sciences Plant Science Plant Ecology Plant Communities Grasslands Ecology and Environmental Sciences Terrestrial Environments Grasslands Biology and Life Sciences Ecology Ecological Metrics Biomass Ecology and Environmental Sciences Ecology Ecological Metrics Biomass Biology and Life Sciences Psychology Behavior Animal Behavior Grazing Social Sciences Psychology Behavior Animal Behavior Grazing Biology and Life Sciences Zoology Animal Behavior Grazing Biomass removal promotes plant diversity after short-term de-intensification of managed grasslands Biomass removal promotes plant diversity after short-term de-intensification of managed grasslands https://orcid.org/0000-0002-2711-3326 Andraczek Karl Data curation Formal analysis Investigation Methodology Visualization Writing – original draft Writing – review & editing 1 * Weigelt Alexandra Conceptualization Methodology Project administration Resources Supervision Validation Writing – review & editing 1 2 Hinderling Judith Data curation Methodology Resources Supervision Writing – review & editing 3 ‡ Kretz Lena Investigation Methodology Writing – review & editing 1 ‡ Prati Daniel Conceptualization Investigation Methodology Supervision Writing – review & editing 3 ‡ van der Plas Fons Conceptualization Data curation Formal analysis Funding acquisition Investigation Methodology Project administration Resources Supervision Validation Writing – original draft Writing – review & editing 1 4 1 Department of Life Sciences, Systematic Botany and Functional Biodiversity, University Leipzig, Leipzig, Germany 2 German Centre for Integrative Biodiversity Research (iDiv) Halle-Jena-Leipzig, Leipzig, Germany 3 Institute of Plant Sciences, University of Bern, Bern, Switzerland 4 Plant Ecology and Nature Conservation Group, Wageningen University, Wageningen, the Netherlands Jenkins David G. Editor University of Central Florida College of Sciences, UNITED STATES Competing Interests: The authors have declared that no competing interests exist. ‡ JH, LK and DP also contributed equally to this work. * E-mail: karl.andraczek@uni-leipzig.de 29 6 2023 2023 18 6 e028703916 12 2022 26 5 2023 © 2023 Andraczek et al 2023 Andraczek et al https://creativecommons.org/licenses/by/4.0/ This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Land-use intensification is one of the main drivers threatening biodiversity in managed grasslands. Despite multiple studies investigating the effect of different land-use components in driving changes in plant biodiversity, their effects are usually studied in isolation. Here, we establish a full factorial design crossing fertilization with a combined treatment of biomass removal, on 16 managed grasslands spanning a gradient in land-use intensity, across three regions in Germany. Specifically, we investigate the interactive effects of different land-use components on plant composition and diversity using structural equation modelling. We hypothesize that fertilization and biomass removal alter plant biodiversity, directly and indirectly, mediated through changes in light availability. We found that, direct and indirect effects of biomass removal on plant biodiversity were larger than effects of fertilization, yet significantly differed between season. Furthermore, we found that indirect effects of biomass removal on plant biodiversity were mediated through changes in light availability, but also by changes in soil moisture. Our analysis thus supports previous findings, that soil moisture may operate as an alternative indirect mechanism by which biomass removal may affect plant biodiversity. Most importantly, our findings highlight that in the short-term biomass removal can partly compensate the negative effects of fertilization on plant biodiversity in managed grasslands. By studying the interactive nature of different land-use drivers we advance our understanding of the complex mechanisms controlling plant biodiversity in managed grasslands, which ultimately may help to maintain higher levels of biodiversity in grassland ecosystems. Deutsche Forschungsgemeinschaft (DFG) 433266560 - SPP 1374: Biodiversitäts-Exploratorien van der Plas Fons The work has been funded by the German Research Foundation (DFG) Priority Program 1374 ‘Biodiversity-Exploratories’ (received by: F.v.P., BEF-Loops Nr.433266560, Exploratories project phase Nr. PL 891/3-1). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Data AvailabilityData and Code are available in the BExIS database of the Biodiversity Exploratories program https://www.bexis.uni-jena.de/ddm/publicsearch/index (ID 31370, DOI: https://doi.org/10.25829/ BEXIS.31370-13). The raw dataset with the ID 24766 is available in the BExIS database under the URL https://www.bexis.uni-jena.de/ddm/publicsearch/index. Raw data sets with the ID 31203, 31204 will be publicly available under the URL https://www.bexis.unijena.de/PublicData/About.aspx from May 2023 on. The Data set with the ID 31180 will be publicly available under the URL https://www.bexis.unijena.de/PublicData/About.aspx from December 2022 on. Until then, data is available upon request (bexis@uni-jena.de). Data Availability Data and Code are available in the BExIS database of the Biodiversity Exploratories program https://www.bexis.uni-jena.de/ddm/publicsearch/index (ID 31370, DOI: https://doi.org/10.25829/ BEXIS.31370-13). The raw dataset with the ID 24766 is available in the BExIS database under the URL https://www.bexis.uni-jena.de/ddm/publicsearch/index. Raw data sets with the ID 31203, 31204 will be publicly available under the URL https://www.bexis.unijena.de/PublicData/About.aspx from May 2023 on. The Data set with the ID 31180 will be publicly available under the URL https://www.bexis.unijena.de/PublicData/About.aspx from December 2022 on. Until then, data is available upon request (bexis@uni-jena.de). ==== Body pmc1 Introduction Grassland ecosystems cover almost a third of the global terrestrial surface and harbour a substantial amount of biodiversity [1]. At the same time, grassland biodiversity is threatened by various anthropogenic influences, including land-use intensification [2–4], which in central European grasslands mainly encompasses the intensification of mowing, grazing and fertilization frequency [5]. However, as land-use comprises a number of different drivers, including fertilization and mowing (or grazing), we need to understand the interactive effects of these drivers to counteract their effects on plant biodiversity loss [6]. Different land-use drivers typically covary, for instance, agricultural grasslands that are frequently mown are generally also heavily fertilized [6], but we usually study their effects in isolation. So far, only a few studies investigated their interactive effects in real-world settings [6, 7], especially in a scenario of land-use de-intensification (i.e. a reduction of land-use intensity, e.g. mowing frequency, grazing or fertilization intensity). Thus, we still a have limited understanding of how different land-use drivers can be optimized in order to maintain higher levels of plant biodiversity. In grasslands, fertilization causes declines in plant biodiversity either directly or indirectly [6, 8, 9]. Following the Resource competition theory by Tilman et al. [10], the addition of a limiting resource alters interspecific competition, leading to only a few species to coexist. For example, Harpole et al. [11] found that the addition of multiple limiting resources caused plant biodiversity to decline, likely driven by reduced nutrient niche-dimensionality below ground. However, fertilization may also indirectly affect plant biodiversity via increasing biomass production of living and dead biomass, and hence also standing biomass, thereby altering light competition [8, 12, 13, but see 14], soil acidity, nutrient mineralization rates, or the activity of pathogens [15, 16]. Especially light competition is expected to be highest in summer, when standing biomass is highest, limiting reproduction and survival [17, 18]. Following that, the relative importance of indirect effects of fertilization on biodiversity mediated through changes in light availability likely differs between season. Furthermore, plant biodiversity loss driven by light competition is non-random, as slow-growing and small species are successfully outcompeted by fewer fast-growing and tall species which become dominant [12, 19, 20]. Hence, we would expect changes in both species richness but also Shannon diversity [17]. So far, there is evidence that both direct and indirect effects of fertilization are important in driving changes in biodiversity in grasslands [21]. While fertilization generally causes plant biodiversity to decline, mowing and grazing can induce more mixed responses depending on their intensity [6, 22]. Both, mowing and grazing, can promote plant biodiversity, although partly due to different mechanisms. Mowing at intermediate frequency can promote plant species richness and diversity by decreasing light competition, while concomitantly increasing competitive abilities (e.g. germination rates) of subdominant species [8, 23–26]. Furthermore, hay removal due to mowing decreases soil nitrogen pools [27, 28], thereby reducing nutrient availability in the soil and positively affecting plant biodiversity. Similar to mowing, grazing positively affects plant biodiversity under certain conditions [23, 29–31], by decreasing light competition as a consequence of biomass removal, as well as by the creation of microsites promoting seedling establishment [32, 33]. However, both mowing and grazing can also negatively affect plant biodiversity at higher intensities [6, 34, 35]. Although biomass removal may promote species richness by increasing light availability, it is possible that under dry conditions, seedling survival is reduced specifically in gaps due to higher drought stress [36, 37]. Hence, possible positive effects of biomass removal via increasing light availability may depend on soil moisture [14, 38]. As soil moisture is most limited in summer, the relative importance of indirect effects of biomass removal on biodiversity could also be expected to differ between growing seasons. Although there is a large body of research investigating the effects of land-use on plant biodiversity in managed grasslands, there is still a limited understanding about the relative and potentially opposing influences of the different land-use drivers, i.e. fertilization and mowing/grazing. Previous studies investigating effects of land-use on plant biodiversity [6, 13, 39] found confounding effects of different land-use drivers, which could not be disentangled, due to the lack of an appropriate study design. For instance, in managed grasslands mowing is often found to be correlated with fertilization intensity [40]. Furthermore, mowing and grazing may (partly) compensate for the generally negative effects of fertilization on plant biodiversity [12, 32, but see 41], by reducing litter accumulation and thereby reducing light competition [26], but also, and in the case of mowing, by removing substantial amounts of nutrients from of the system [42, 43]. While crossed fertilization and mowing/grazing studies exist, there are typically carried out within grasslands with initially low land-use intensity, but not in systems with a history of intensive grassland management. Furthermore, few studies manipulated the intensity of different land-use drivers within the same grassland, which provides crucial insights on how biodiversity responds to reductions of different land-use drivers in a realistic setting. Thus, to truly unravel the drivers of biodiversity loss in managed grasslands and to gain insights in potential biodiversity recovery, it is necessary to study the interaction of land-use drivers by using a suitable experimental design. Ultimately, comparing the strength of direct and indirect effects of different land-use drivers across different seasons may advance our understanding on the complex mechanisms which alter plant biodiversity in managed grasslands. In doing so, we may also improve our current knowledge on which land-use drivers should, or should not, be de-intensified if we want to promote plant biodiversity. In this study, we test an hypothesis involving proposed effects of fertilization and biomass removal (by mowing/grazing) on plant biodiversity (here species richness and Shannon diversity). Specifically, we aim to distinguish between direct and indirect effects of fertilization on plant biodiversity, and how biomass removal due to mowing/grazing mediates these pathways (Fig 1). We thus aim to understand how different drivers of land-use de-intensification may, or may not, promote plant biodiversity in the short-term, while also addressing seasonal variability of these relationships. To address these questions, we designed a full factorial experiment, replicated within managed grasslands in three regions in Germany. We compare the effects of land-use drivers crossing fertilization with a combined treatment of biomass removal (mowing/grazing treatment): (a) fertilized & biomass removal, (b) unfertilized & biomass removal, (c) unfertilized & reduced biomass removal and (d) fertilized & reduced biomass removal. In particular, we tested the following hypotheses: Fertilization will cause plant species richness and Shannon diversity to decline, both directly and indirectly, by altering light competition due to increased biomass production and thereby increased standing biomass. Biomass removal partly compensates for the negative effects of fertilization on species richness and Shannon diversity, by decreasing levels of standing biomass and thereby reducing light competition or changing the competitive dominance of plant species. 10.1371/journal.pone.0287039.g001 Fig 1 Conceptual framework. Path diagram showing how the different land-use drivers fertilization and biomass removal (mowing/grazing), may directly or indirectly (via increasing or decreasing light competition due to changes in biomass production/standing biomass) affect plant biodiversity (represented here by species richness and Shannon diversity) in managed grasslands. Pathways are sequentially numbered referring to supporting evidence, 1: [46, 47], 2: [10, 44], 3 and 4: [16], 5: [11, 15], 6: [47], 7: [13, 14, 48], 8: [28, 33]. Alternatively changes in soil moisture can mediate pathways between biomass removal or fertilization and plant biodiversity [17] (not shown here). For further information also on seasonal variability of relationships, see the introduction. Dashed arrows indicate negative and straight arrows positive relationships between two variables. 2 Material and methods 2.1 Study area We studied the interactive effect of fertilization and mowing/grazing on plant biodiversity in 16 commercially managed (unsown) grasslands located in three regions in Germany as part of the Biodiversity Exploratories project (www.biodiversity-exploratories.de; [44]). The regions are distributed across the south-east (UNESCO Biosphere Reserve Schwäbische Alb—Swabian Jura), the centre (National Park Hainich-Dün) and the north-east of Germany (UNESCO Biosphere Reserve Schorfheide-Chorin). Due to various climatic and edaphic differences, these regions vary in several environmental conditions, such as temperature and soil fertility (deep layers of organic soils) being highest in north-west, or annual mean precipitation and elevation being highest in the south-west (see [44] for more details on regional differences). All 16 grasslands were commercially managed, spanning a gradient of background land-use intensity (LUI) which is a composite measure of mowing frequency, livestock units and amount of fertilization (see S17 Table). By including multiple regions and spanning a gradient in background grassland land-use intensity (rather than conducting a more conventional experiment in highly controlled settings), we aimed to investigate how the de-intensification of different grassland land-use drivers (fertilization vs mowing/grazing) interactively drive plant biodiversity in semi-natural systems. Across all studied grasslands, the “most 10 abundant” species were: Poa pratensis L., Lolium perenne L., Dactylis glomerata L., Taraxacum sp. F.H. Wigg., Alopecurus pratensis L., Phleum pratense L., Trisetum flavescens L., Festuca pratensis Huds., Achillea millefolium L., Bromus hordeaceus L. 2.2 Experimental design In all three regions, we investigated how different land-use drivers (fertilization and mowing/ grazing) influence plant biodiversity via direct or indirect pathways. In this study we focus on the grassland management components fertilization and mowing/grazing, as in commercially managed grasslands within central Europe, those are the most dominant land-use components representing important predictors for both plant biodiversity and biomass production [5, 6, 45]. To do so, we selected 16 managed grassland sites (6 in the Schwäbische Alb and Hainich-Dün, and 4 in Schorfheide-Chorin) and set-up a full factorial experimental design, by reducing the intensity of biomass removal and fertilization in marked treatments from autumn 2019 onwards. Thus, we established four 7×7 m plots per grassland plot, containing one of four combinations: fertilized & biomass removal (+F+R), unfertilized & biomass removal (-F+R), unfertilized & reduced biomass removal (-F-R), fertilized & reduced biomass removal (+F-R) (S1 Fig). In the +F+R treatment the amount of fertilization (0.57–7.43 kg N m-3 year-1), grazing (number of livestock units [LU] per grazing duration, 0.02–5.69 LU * day ha-1 year-1) and mowing (1.02–2.73 number of cuts per year) differed along a land-use gradient (S16 Table). In the -F+R treatment, fertilization was stopped completely, while mowing and grazing continued at the same intensity and frequency as in +F+R plots. In the -F-R treatment, land-use intensity was reduced to mowing only once a year in August or September, while stopping grazing and fertilization completely. This disturbance (here by reduced mowing regime) was done, because some level of minimal disturbance is required in central European grasslands, to prevent shrub encroachment and forest succession [46]. As in the reduced land-use treatment, land-use intensity was reduced in the +F-R treatment, while fertilization (organic or inorganic fertilizer depending on fertilizer type used by farmer) was manually applied to avoid cross-contamination with fertilizer between adjacent treatments. The amount and frequency of fertilizer application was comparable to the amounts of fertilizer applied in the respective fertilized & biomass removal plot. For each treatment (i.e. +F+R, +F-R, -F+R, -F-R), there was one plot in 16 different grasslands (with one subplot located in each plot), yielding to 64 studied plots in total. 2.3 Data collection All data was collected in the spring and summer season, at peak biomass (April to May and July to August respectively), in 2020 and 2021. In each season, we estimated the vascular plant species specific cover in each subplot. Additionally, in each subplot we assessed standing biomass by using a rising plate meter (Jenquip Manual Plate Meter), as the mean of four measurements per subplot. By measuring vegetation height and density, derived from the height above ground level at which the disc (Ø 35.5 cm, 384 g) is supported by the vegetation, the rising plate meter can be used to estimate standing biomass non-destructively [47]. Biomass estimates derived from the rising plate meter were calibrated using data from two additional 1×1 m subplots (100 calibration subplots per region, 300 in total) in each grassland plots (containing +F+R subplots) and in plots within comparable managed grasslands differing in land-use intensity (see BExIS dataset ID 31180). In these calibration plots, during spring 2021, both biomass estimations using the rising plate meter and actual biomass measurements (clipping and drying above ground vegetation) were performed. To quantify light competition for plants below the canopy, we measured light availability of photosynthetically active radiation (using PAR Sensor, Skye Instruments, in μmol sec-1 m-2) both at ground level and at 1.5 m height. To test an alternative indirect relationship of fertilization or biomass removal via soil moisture, we additionally measured soil moisture (ML3 ThetaProbe Soil Moisture Sensor, in %) with three replicates per subplot. In order to account for biomass that has been removed due to mowing or grazing when calculating biomass production for the summer season, we further used data on grazing duration, livestock units per area and hay yield of mowing derived from yearly standardized questionnaires of land-owners and farmers (for more information see 48), which we obtained from the database BExIS (see BExIS dataset ID 26487, http://doi.org/10.17616/R32P9Q). In spring, biomass production was considered equal to standing biomass in spring, as most grasslands were hardly grazed or not mown before our biomass measurements took place (except for one grassland). Additionally, we quantified background fertilization intensity (kg N ha-1 year-1), background grazing intensity (Livestock units * day ha-1 year-1) and background mowing intensity (cuts year-1), for each of the years 2017–2019 (before our experiment was set up) also derived from yearly std. questionnaires [48]. Background intensity of all land-use components were calculated via the LUI calculation tool [49] implemented in BExIS (http://doi.org/10.17616/R32P9Q). 2.4 Data processing and calculations All species names were taxonomically standardized according to the accepted species names in The Plant List (www.theplantlist.org, using the TAXONSTAND R package [50]. For each subplot, we quantified species richness and the effective Shannon Diversity [51, 52]. Effective Shannon Diversity was calculated as: H′=exp(−∑ipilogpi) (1) Standing biomass was quantified using biomass estimates which were converted into standing biomass using a calibration formula. This calibration formula was based on calibration data collected in spring 2021 from 2 x 150 subplots (1×1 m each) in additional grassland sites in all three regions (2 x 50 subplots per regions) including all +F+R plots, containing biomass estimates measured via a rising plate-meter (average per subplot) and actual standing biomass measurements from the same subplot. The calibration formula was then derived by performing a linear model with actual standing biomass (in g m-2) as response variable and rising-plate meter measurements (in 0.5 cm increments) as predictor variable. Prior to calculating the calibration model, we excluded data from five subplots due to measurement errors. Following the calibration model, biomass estimates explained 83.22% of the variance in actual standing biomass (t = 37.94, p > 0.01, RSD = 38.85 g m-2, S18 Table). Using the intercept and slope of the calibration model, we converted the biomass estimates into actual standing biomass following the equation: dry biomass (g m-2) = -38.66 + 8.68 × Plate meter measurement. In addition to standing biomass, we also quantified biomass production in all subplots which was defined as all biomass produced in a given time (Spring until Summer) including biomass removed by mowing or grazing. Only biomass removed by higher trophic levels other than livestock (e.g. herbivorous arthropods) could not be included in our biomass production estimate. We quantified biomass production following Riehl [53]: BiomassproductionSummer=(StandingbiomassSummer*LivestockunitsSummer*GrazingdurationSummer*14700+YieldmowingSummer)−(StandingbiomassSpring*LivestockunitsSpring*GrazingdurationSpring*14700+YieldmowingSpring) (2) with livestock units in unit m-2, grazing duration in number of grazing days, ‘14700’ referring to the mean fodder consumption of an average livestock unit in g per day [53], and yield mowing in g m-2. In particular, yield mowing is defined for spring, as the sum between all mowing events between the start of the growing season (1st of April) and the end of the first grassland survey (15th of May), and in summer, as the sum for all mowing events between the end of the first grassland survey and the end of the second grassland survey (14th of August). These periods were also applied to calculate the grazing duration and livestock units per subplot. We only calculated biomass production in summer as in this season we expect a mismatch between standing biomass and biomass production due to land-use (e.g. removal of biomass due to mowing or grazing), while in spring biomass measurements were in most plots done before mowing and grazing events took place, and hence standing biomass was almost equivalent to biomass production of the early spring season. To quantify light competition in each subplot we calculated the inverse ratio between the average light availability of photosynthetically active radiation above canopy height and ground level (light availability hereafter, light availability increases from 0 to 1) per subplot. Additionally, to obtain a mean background fertilization intensity value over the time before our experiment was set up (2017–2019), background fertilization intensity was averaged across time. 2.5 Data analyses All data analyses were performed using the statistical software R v. 4.1.1 [54]. Statistical analyses were performed using the packages lme4 [55], lmerTest [56], stats4 [54], multcomp [57], car [58], MuMIn [59] and PIECEWISESEM [60]. 2.5.1 Overall effect of treatments on plant species richness and diversity To study how the different treatments (i.e. +F+R, +F-R, -F+R, -F-R) affected species richness, Shannon diversity, standing biomass or biomass production, we performed linear mixed models including treatment and region as fixed effects and grassland as a random effect. To test for pairwise comparisons between the different treatments we estimated least-square means and computed contrasts using the function emmeans from the package emmeans [61]. For all models we performed a forward model selection procedure. Specifically, we started with a model without fixed factors and we then stepwise added treatment and region as additional predictors if these i) did not increase the AIC and ii) did not exceed a variance inflation factor (VIF) of 3 [62], using the vif function from the package car. That model selection procedure allowed us to select the most parsimonious model while avoiding multicollinearity between predictors. Importantly, even when not being part of the most parsimonious model, we always included treatment in our final model. Furthermore, if both treatment and region were part of the most parsimonious model, we tested for an interaction between these, by comparing the AIC of the final model and the model with an interaction term. 2.5.2 Treatment-mediated changes in community composition To test how the different treatments induced changes in plant community composition, separately for each region and season in 2021 (for results on 2020 see S10 Fig), we used a non-metric multidimensional scaling (NMDS) analysis, based on Bray–Curtis similarity index as an ordination technique using the function metaMDS from the package vegan [63]. Furthermore, to statistically test the dissimilarities between the species compositions of the different treatments after two years since the experiment was set up (in 2021), we performed a permutational multivariate analysis of variance (PERMANOVA), using the package pairwiseAdonis [64]. To visualize whether certain environmental factors explain shifts in plant community composition we further fitted environmental vectors (light availability, soil moisture and standing biomass) onto the NMDS ordination for all regions combined (for results on 2020 see S11 Fig). Correlation of environmental variables with the NMDS axes was tested using the envfit() function from the package vegan [63]. 2.5.3 Direct and indirect effects of land-use on species richness and diversity We expected that fertilization can both directly, but also indirectly (mediated via changes in biomass production, standing biomass and light availability) affect species richness and Shannon diversity. Furthermore, we expected biomass removal to mediate the indirect effect of fertilization on species richness/Shannon diversity by removing biomass and thereby decrease light competition. We additionally tested for an indirect relationship between fertilization and biomass removal via soil moisture. To guarantee comparability between spring and summer data, we used soil moisture as an additional indirect pathway in both spring and summer based models. To test these assumptions, we constructed a hypothesis driven causal model using linear mixed-effect models within a PiecewiseSEM [60]. In our models, we converted the treatment ID into the dummy variables “fertilization”, and “biomass removal”, with “0” indicating absence and “1” presence of the respective land-use component. We further log-transformed light availability, to meet model assumptions regarding linearity. To statistically correct for the influence of year and sampling date we further constructed a linear model with each continuous variable (species richness, species diversity, biomass production, standing biomass and log light availability) as response, with year and sampling date as fixed effect and with an interaction term. Finally, we extracted the model residuals (without the explained variance by year and sampling date) which were used as input for all further SEM models. We constructed four separate, a priori hypothesized models, testing the causal relationship between the land-use drivers fertilization and biomass removal with species richness or Shannon diversity in spring and summer separately. First, we ran the initial SEM models as a list of causal relationships. Secondly, we inspected all initial model results, goodness-of-fit tests and Fisher’s C statistics, and, only if necessary, added predictors that significantly improved the AIC and the model fit with p-values higher than 0.05. During this process, for both summer models we compared competing models ex- or including an error correlation structure between biomass production and soil moisture (species richness model: AIC without correlated error = 93.3, with correlated error = 92.1; Shannon diversity model: AIC without correlated error = 95.1, with correlated error = 94.0). For spring models, no error correlation structure between biomass production and soil moisture could be tested, as these models did not include a biomass production variable. This was done in order to increase the goodness-of-fit of the models, while we deliberately did not include any causal relationship between biomass production and soil moisture since it was difficult to distinguish between cause and effect. To statistically correct for the confounding effects of covarying factors we included background fertilization intensity of the time before the experiment was set up (mean of 2017–2019) as covariate for models predicting species richness/diversity, standing biomass and biomass production. In all models, grassland was treated as a random factor nested within study region by using the lme function from the package nlme [65]. As the residual variance of light availability increased with increasing standing biomass, we included a power variance structure for the variance covariate standing biomass in all models predicting light availability, to account for heterogeneity in the residuals. We inspected the assumptions of normality of model residuals visually. PiecewiseSEMs were performed using the R package PIECEWISESEM [60]. 3 Results 3.1 Overall effects of fertilization and biomass removal on plant species richness and species composition in different seasons Overall, we did not detect a consistent effect of treatments on species richness (Table 1, S2 Fig). Similar, subplots with +F+R did not significantly differ from any other treatment (see results of pairwise differences in S2 Table). However, we generally observed a lower species richness in +F-R subplots compared to all other treatments (although not statistically significant). In the Hainich-Dün in spring 2021, species richness was found to be marginally significantly, lower in subplots which were fertilized but had reduced biomass removal compared to subplots with no fertilization but biomass removal (S2 Table). Our NMDS results showed no significant differences in community composition between the different treatments in neither of the regions, nor in any of the different seasons of the year 2021 (Fig 2, S7 Table). 10.1371/journal.pone.0287039.g002 Fig 2 Changes in community composition. Non-metric multidimensional scaling (NMDS) based on Bray–Curtis similarity index for the plant communities for each treatment combination in spring (A-C) and summer (D-F) for 2021 (A,D) Schwäbische Alb, (B,E) Hainich-Dün, (C,F) Schorfheide Chorin. Hull volumes represent clusters of plant communities within a given treatment. Treatments are colour coded as fertilized & biomass removal (+F+R): purple; fertilized & reduced biomass removal (+F-R): bright blue, unfertilized & biomass removal (-F+R): green; unfertilized & reduced biomass removal (-F-R): yellow. Detailed model summaries are shown in the Supporting Information (S7 Table). 10.1371/journal.pone.0287039.t001 Table 1 Effects of unfertilized & biomass removal (-F+R), fertilized & reduced biomass removal (+F-R) and unfertilized & reduced biomass removal (-F-R) treatment on species richness. Treatment effects indicate standardized effect sizes, with associated confidence intervals in squared brackets. P-values (P) indicate overall significant effect of treatments on species richness. If most parsimonious model included an interaction between treatment and region, we also show regional specific effect sizes. For more detailed information on statistical results see S1 Table. Note that rows are relative to the intercept (Alb and +F+R treatment). Std. effect sizes of treatments on species richness P all Alb Hai Sch -F+R SP ‘20 -0.04 [-0.22, 0.14] - - - 0.46 SU ‘20 - 0.14 [-0.18, 0.47] 0.00 [-0.45, 0.17] -0.18 [-0.49, 0.09] 0.11 SP ‘21 - 0.10 [-0.24, 0.44] -0.19 [-0.62, 0.04] -0.04 [-0.44, 0.17] 0.74 SU ‘21 - 0.25 [-0.05, 0.56] 0.20 [-0.34, 0.24] 0.03 [-0.49, 0.05] 0.12 +F-R SP ‘20 -0.14 [-0.33, 0.04] - - - 0.46 SU ‘20 - -0.14 [-0.47, 0.18] -0.15 [-0.32, 0.30] -0.27 [-0.41, 0.16] 0.11 SP ‘21 - 0.08 [-0.26, 0.42] 0.00 [-0.41, 0.25] -0.08 [-0.46, 0.14] 0.74 SU ‘21 - -0.04 [-0.34, 0.27] -0.14 [-0.39, 0.19] -0.13 [-0.35, 0.18] 0.12 -F-R SP ‘20 -0.07 [-0.25, 0.12] - - - 0.46 SU ‘20 - 0.02 [-0.30, 0.35] -0.13 [-0.46, 0.16] -0.18 [-0.49, 0.09] 0.11 SP ‘21 - 0.10 [-0.24, 0.44] 0.07 [-0.35, 0.30] 0.06 [-0.34, 0.26] 0.74 SU ‘21 - 0.07 [-0.23, 0.38] 0.07 [-0.29, 0.29] -0.06 [-0.40, 0.14] 0.12 3.2 Direct and indirect effects of land-use and environmental factors on plant species richness and Shannon diversity in different seasons We tested the direct and indirect effects (via light availability and alternatively via soil moisture) of fertilization and biomass removal on species richness by using a piecewise SEM for both spring and summer separately (Table 2). In spring, the SEM fitted species richness (Fisher’s C = 15.135, df = 14, p = 0.369, n = 105, S2 Table) and Shannon diversity (Fisher’s C = 17.263, df = 14, p = 0.242, n = 105, S3 Table) well (Fig 3A and 3B). For the SEM model on the spring data, we found no evidence for direct or indirect (via light availability or soil moisture) causal pathways between fertilization or biomass removal and species richness/Shannon diversity. We found a strong partial effect between standing biomass (hereafter simply ‘biomass’) and light availability in the understory of the vegetation (direct path coeff. = −0.75). However, we did not find any significant relationship between light availability and species richness/Shannon diversity. In contrast, soil moisture was found to be directly positively related to species richness (direct path coeff. = 0.37) and Shannon diversity (direct path coeff. = 0.34). Additionally, we found that biomass removal strongly reduced the accumulation of biomass (direct path coeff. = −0.46), although no significant direct or indirect pathways between biomass removal and species richness or Shannon diversity were found. Overall, the SEM models in spring explained only 15% of the variance observed in species richness and 14% of the variance in species diversity. The most important predictor for species richness in spring was soil moisture (partial r2 = 0.16), followed by fertilization (partial r2 = 0.005), light availability (partial r2 = 0.005) and biomass removal (partial r2 = 0.004). For Shannon diversity, the most important predictor was soil moisture (partial r2 = 0.142), followed by light availability (partial r2 = 0.008), biomass removal (partial r2 = 0.008) and fertilization (partial r2 = 0.008). An NMDS analysis of the spring data showed that the effects of soil moisture on species richness and Shannon diversity cannot be explained by changes in species composition (Fig 4, S9 Table). 10.1371/journal.pone.0287039.g003 Fig 3 Structural equation models with richness and diversity as main response. Structural equation models for testing the direct and indirect effects of different land-use drivers (fertilization and biomass removal) on species richness in (A) spring (Fischer’s C = 15.135, p = 0.369) and (C) summer (Fischer’s C = 24.063, p = 0.344), and on Shannon diversity in (B) spring (Fischer’s C = 17.263, p = 0.242), and (D) summer (Fischer’s C = 26.006, p = 0.251) of both 2020 and 2021. Solid lines represent positive relationships and dotted lines negative relationships. Each arrow connection among variables indicates standardized path coefficients and significant levels (***p < 0.001, **p < 0.01, *p < 0.05). Detailed model summaries with unstandardized path coefficients are shown in the Supporting Information (S3 and S6 Tables). 10.1371/journal.pone.0287039.g004 Fig 4 Community composition and environmental factors. Non-metric multidimensional scaling (NMDS) based on Bray–Curtis similarity index for the plant communities for each treatment combination and all regions combined in spring (A) and summer (B) for 2021. R2-values are shown for each axis. Arrows printed in grey represent environmental factors: standing biomass, light availability and soil moisture, while arrows printed in red represent species richness and Shannon diversity. Species names represent the most extreme species according to the NMDS axes. Treatments are colour coded as fertilized & biomass removal (+F+R): purple; fertilized & reduced biomass removal (+F-R): bright blue, unfertilized & biomass removal (-F+R): green; unfertilized & reduced biomass removal (-F-R): yellow. Detailed model summaries are shown in the Supporting Information (S9 Table). 10.1371/journal.pone.0287039.t002 Table 2 Selected standardized partial effect sizes of direct and indirect effects via light availability and soil moisture of fertilization or biomass removal on species richness and Shannon diversity with proposed interpretations. Indirect effects were calculated by multiplying respective direct path coefficients. Direction of predicted (Pred) and observed (Obs) relationships are indicated as “pos.” if positive, “neu.” if neutral (if effect is smaller than +/- 0.005), and “neg.” if negative. Bold text indicates whether effect was found to be significant. For further information see S3 and S6 Tables. Effect Mediator Season Magnitude Pred Obs Proposed interpretation Rich Div Fertilization Direct - Spring -0.061 -0.075 neg. neg. Contrary to our hypothesis (1), fertilization did not significantly affect plant species richness or Shannon diversity, neither directly nor indirectly, potentially because species gains as a consequence of a cessation in fertilization may need more time to emerge. Summer -0.125 -0.141 neg. Indirect Light availability Spring -0.006 -0.008 neg. neg. Summer 0.001 0.002 neu. Soil moisture Spring -0.007 -0.006 neg. Summer -0.001 -0.001 neu. Biomass removal Direct - Spring -0.060 -0.082 pos. neg. In line with our hypothesis (2), biomass removal enhanced plant species richness/Shannon diversity directly, supposedly via reducing soil nitrogen pools and preventing litter accumulation. Effects of biomass removal are largest in summer, potentially due to the accumulation of litter in the course of the growing season. Summer 0.205 0.276 pos. pos. Indirect Light availability Spring 0.022 0.029 pos. pos. Contrary to our hypothesis (2), biomass removal decreases Shannon diversity indirectly, by increasing light availability especially in summer, potentially due to lower drought-induced plant survival. However, as alternatively hypothesized, due to increases in soil moisture, mediated through changes in standing biomass, biomass removal positively affected species richness and Shannon diversity, likely due to decreased water stress. Summer -0.066 -0.110 neg.a Soil moisture Spring 0.024 0.023 pos. Summer 0.029 0.028 pos.a a mixed effects found for species richness and Shannon diversity. Here only significant effects are reported, but see S3 and S6 Tables. In summer, the SEM fitted species richness (Fisher’s C = 24.063, df = 22, p = 0.344, n = 117, S5 Table) and Shannon diversity (Fisher’s C = 26.006, df = 22, p = 0.251, n = 117, S6 Table) well (Fig 3C and 3D), including an additional error correlation between biomass production and soil moisture (species richness model: AIC without correlated error = 93.3, with correlated error = 92.1; Shannon diversity model: AIC without correlated error = 95.1, with correlated error = 94.0). For the SEM considering species richness and Shannon diversity, we found clear evidence for a positive, direct causal pathway of biomass removal on species richness/Shannon diversity (direct path coeff. = 0.21, direct path coeff. = 0.28, respectively, Table 1). We found only evidence for an indirect causal pathway between biomass removal and Shannon diversity, not species richness, mediated via changes in light availability (Fig 3, S5 Table). However, we did find evidence for an indirect causal pathway between biomass removal and species richness as well as Shannon diversity via soil moisture (Fig 3, S5 Table). For both models considering species richness or Shannon diversity, biomass removal led to an increase in biomass production (direct path coeff. = 0.57), while at the same time reducing standing biomass (direct path coeff. = -0.74). Furthermore, the accumulation of standing biomass led to a strong decrease in light availability (direct path coeff. = −0.82) and soil moisture (direct path coeff. = −0.24). We also found strong partial effects of biomass production on standing biomass (direct path coeff. = 0.23). However, we did not detect any strong direct or indirect effects of fertilization or background fertilization intensity on neither species richness, Shannon diversity (Table 1) nor biomass production. Overall, the SEM models in summer explained 11% of the variance observed in species richness and 13% of the variance in Shannon diversity. The most important predictor for species richness in summer was soil moisture (partial r2 = 0.104), followed by biomass removal (partial r2 = 0.043), fertilization (partial r2 = 0.025), light availability (partial r2 = 0.017) and background fertilization intensity (partial r2 = 0.008). For Shannon diversity, the most important predictor was soil moisture (partial r2 = 0.075), followed by biomass removal (partial r2 = 0.057), background fertilization intensity (partial r2 = 0.042), light availability (partial r2 = 0.035) and fertilization (partial r2 = 0.024). The NMDS analysis of the summer data showed that the positive effect of soil moisture on species richness in the summer season can be partly explained by changes in species composition (Fig 4, S9 Table), with moist plots being positively associated with both species richness, diversity and the first two NMDS axes, which represented a shift from communities showing higher abundances of Rumex acetosella in dry plots towards a higher abundance of Potentilla erecta in plots with a relatively high soil moisture. 4 Discussion The present study aimed to advance our understanding of the complex relationship between land-use and plant biodiversity, by statistically and experimentally separating the effect of land-use drivers on plant biodiversity. Across three regions in Germany we found varying direct and indirect effects of fertilization and biomass removal on plant biodiversity depending on seasons and biodiversity facet (species richness and Shannon diversity). Although effects were generally weak in spring, in summer, biomass removal enhanced species richness and Shannon diversity, due to direct and partly indirect positive effects which compensated for negative effects (direct and indirect) of fertilization. Thus, our results support the hypothesis that biomass removal, such as mowing and grazing, can have compensatory effects on the generally negative effects of fertilization on plant biodiversity in managed grasslands. To better understand the multivariate links between land-use drivers, observed plant species, as well as plant species richness and Shannon diversity in managed grasslands, we used structural equation modelling to test an integrated causal hypothesis (Fig 1). Although we did not observe strong changes in plant community composition in the different land-use treatments, our structural equation model revealed both direct and indirect effects of land-use on both species richness and Shannon diversity. We hypothesized that fertilization controls species richness and Shannon diversity both directly (e.g. by eutrophication, see [66]) and indirectly, via increased light competition resulting from increased standing biomass, [8, 12] or alternatively, through changes in soil moisture [14]. However, we did not detect strong direct or indirect effects of fertilization on species richness or Shannon diversity in spring and summer, while (non-significant) effects were mainly negative (albeit weakly), as expected. One possible explanation for the lack of strong direct or indirect effects of fertilization on species richness or Shannon diversity might be that species gains in response to cessation of fertilization need more time to emerge, so that only small differences between fertilized and unfertilized treatments are visible in the short term [67, 68]. In general, light limitation as a consequence of fertilization-induced increases in standing biomass is suggested to negatively affect plant species richness and Shannon diversity, by changing competitive abilities of plants, consequentially favouring fast-growing and tall species [12, 19, 20]. However, in the present study we did not observe strong changes in the plant community composition. Only the plant community composition in Schorfheide-Chorin showed a tendency to differ between the land-use treatments. A possible explanation for the weak changes in plant community composition in response to the different land-use treatments is that even when reducing the intensity of certain land-use drivers, residual land-use effects (e.g. from former fertilization) might still be present in the system and disappear only in the long term [69]. Contrary to the weak negative effects of fertilization that we observed, biomass removal (by mowing or grazing) can promote species richness and Shannon diversity under certain conditions [25, 26, 30, 31]. We hypothesized that positive effects of biomass removal on species richness or Shannon diversity potentially compensate for the negative direct and indirect effects of fertilization. In general, and consistent with findings from previous studies [13, 70], biomass removal strongly decreased the accumulation of standing biomass in both seasons, having its greatest effect in summer. Furthermore, we observed that biomass removal increased biomass production, which has also been shown in previous studies [71], likely due to compensatory regrowth. However, direct effects of biomass removal on species richness or diversity differed strongly between seasons. In spring, biomass removal did not significantly affect species richness and Shannon diversity (although non-significant, weak negative effects were observed), while direct effects in summer were found to be significantly positive. Previous studies have found that early-season biomass removal events reduce species richness [6], likely explaining the negative (although weak and non-significant) relationships between biomass removal and both richness and Shannon diversity observed in spring. A possible explanation for the strong direct effects of biomass removal in summer might be that the removal of biomass prevented the accumulation of litter over the growing season [42]. Specifically, previous studies found that litter accumulation strongly affects community composition. This happens not only via changes in light availability but likely also via changes in soil acidity, nutrient mineralization or increased activity of pathogens, consequently having negative effects on species richness and diversity [15, 16]. Another possible mechanism by which biomass removal may have increased plant diversity is through the removal of nutrients, although this should only play a large role in mown (as opposed to grazed) grasslands [42, 43]. Similar to direct effects of biomass removal, indirect effects differed between seasons, with larger effects being generally found in summer compared to spring, suggesting varying importance of environmental variables across the growing season. We hypothesized that increased light availability would promote plant species richness and Shannon diversity, for example due to increased germination and seedling establishment [33, 72, 73]. However, in spring, contrary to our expectations, we found that such indirect effects were absent. Nevertheless, we did find that species richness and Shannon diversity were positively affected by soil moisture in spring, although soil moisture was not driven by biomass removal. A possible explanation for the lack of light mediated effects of land-use on both plant biodiversity in spring, might be that during the start of the growing season, the vegetation canopy is still relatively open and thus light is no limited resource. Another explanation for the general weak relationships between altered light availability (due to experimental changes in land-use drivers) and plant biodiversity observed in spring might be associated with the delayed impact of competitive release on germination, which likely requires longer term observations to be detected [73]. In general, biodiversity change following a forcing event (such as land-use change), may only emerge in the long term, because of delayed immigration events of new species (immigration credit) or delayed extinction events (extinction dept) [74]. Hence, over time, we could expect an increasing importance of indirect effects of land-use on plant biodiversity in plots with reduced biomass removal due to an extinction debt (due to delayed extinction). Similarly, we could expect increased plant biodiversity only to be visible at the longer term, in plots with reduced fertilization intensity, due to an immigration credit. In contrast to the observed relationships in spring, we found significant indirect effects of biomass removal on Shannon diversity in summer mediated through changes in light availability (negatively affecting diversity) as well as soil moisture (positively affecting diversity). Species richness though, was only significantly positively affected through biomass removal-induced changes in soil moisture. Contrary to our results, previous studies observed a positive relationship between light availability and species richness [8, 13]. A possible explanation for our contrasting results might be that contrary to our original hypothesis, in summer, soil moisture is more limiting for plant growth than light availability. Furthermore, the unexpected negative relationships between light availability and Shannon diversity could be explained by an increase in micro-temperatures, that might lead to increasing levels of transpiration especially in the summer season, potentially decreasing plant diversity [36, 37]. The positive indirect effect of biomass removal on species richness and Shannon diversity in summer mediated through changes in soil moisture, might be explained by short term effects of biomass removal. Thus, biomass removal may promote soil moisture by increasing rainfall recharge in upper soil layers due to reduced vegetation canopy coverage and root water consumption [75], although increased evaporation due to soil exposure to air can also have negative effects on soil moisture under certain conditions [76]. In general, disturbance is suggested to release resources, not previously available in the system [77]. Thus, in line with previous findings [78], our results suggest that biomass removal may indirectly promote plant species richness and Shannon diversity due to increased soil moisture in managed grasslands. Further, we found that the positive effect of biomass removal mediated through soil moisture on species richness and Shannon diversity in summer, could also be partly explained by shifts in the community composition. For example Potentilla erecta (L.) Raeusch., was slightly associated with moist plots, while dryer plots were slightly associated with Rumex acetosella L. However, as our experiment only aimed to manipulate the intensity of land-use drivers, we cannot fully unravel the relative importance of land-use mediated changes in light availability in comparison to soil moisture and how that affects plant biodiversity. In the present study, the effects of both mowing and grazing were summarized mainly as effects of biomass removal. However, although both mowing and grazing have been found to promote plant biodiversity in grasslands [25, 26, 30, 31], the underlying mechanisms may differ, with grazing causing more patchy biomass removal, while mowing having more uniform effects [79]. As most of the grassland sites studied here were managed as meadows, it is possible that direct and indirect effects of biomass removal on plant biodiversity observed in this study were mostly driven by effects of mowing. Thus, to truly understand the complex interaction between different land-use drivers on plant biodiversity in managed grasslands it is necessary to not only disentangle the effects of fertilization and biomass removal, but also account for the type of biomass removal. This study builds on previous studies, aiming to statistically and experimentally separate the effects of different drivers of land-use on plant biodiversity in managed grasslands [6, 12, 80–83]. However, contrary to these previous studies, we used high land-use intensity as a default in comparison to treatments in which the intensity of certain land-use drivers was reduced (i.e. cessation in fertilization, and a reduction in biomass removal). This enabled us to disentangle the short-term direct and indirect effects of reduced fertilization and biomass removal in managed grasslands. We found that indirect effects of land-use on plant biodiversity were mainly driven by biomass removal, in particular in summer, mediated through both light availability and soil moisture. Specifically, the importance of soil moisture for plant biodiversity found in this study supports findings from previous studies [14, 38], suggesting that indirect effects of land-use mediated through light availability may depend on other environmental factors (such as soil moisture) that limit plant growth. Importantly, as climate change scenarios predict increasing temperatures in the future [84], the relative importance of indirect pathways of land-use mediated through changes in soil moisture might even increase. Overall, we did not find that indirect effects through changes in standing biomass were of greater importance than direct effects of land-use, in particular biomass removal. While in the present short-term study, we could not prove that a cessation in fertilization improves plant biodiversity, we did show that in these studied managed grasslands with likely high soil nutrient loads, decreasing biomass removal through a reduction in grazing intensity or mowing frequency negatively affects plant diversity. However, the relatively small number of sites and years investigated in this study, still limits insights into potential lag effects of land-use drivers on plant biodiversity. Thus, further research is needed to disentangle the complex interaction of different land-use drivers in managed grasslands. Additionally, as our experiment covered a relatively small gradient of land-use intensity, future studies should investigate the interactive effects of different land-use drivers on plant biodiversity along a wider gradient of land-use, while further comparing direct and indirect short vs long term land-use mediated effects. Nevertheless, this study may help us to understand how different land-use drivers interactively affect and thereby control biodiversity, which is crucial for informing biodiversity conservation. Supporting information S1 Fig Overview showing the full factorial design. In all three regions, four grasslands in the Schorfheide-Chorin, six grasslands in the Schwäbische Alb, and 6 grasslands in the Hainich-Dün (16 in total) were selected, where background land-use varied in mowing frequency, grazing intensity and fertilizer input. Within those grasslands, we established four 7×7 m treatments as a full factorial design with: fertilized & biomass removal, fertilized & reduced biomass removal (grazing stopped, but one late cut), unfertilized & biomass removal, unfertilized & reduced biomass removal. In all plots a 1×1 m subplot was established, where we measured plant biomass and performed vegetation surveys. Plot and subplot location was partly randomized in the grassland. (DOCX) Click here for additional data file. S2 Fig Species richness across treatments. Response of species richness (m-2) on the fertilized & biomass removal (+F+R), unfertilized & biomass removal (-F+R), unfertilized & reduced biomass removal (-F-R) and fertilized & reduced biomass removal (+F-R) treatment for all three regions (Alb: Schwäbische Alb; Sch: Schorfheide-Chorin; Hai: Hainich-Dün). Whiskers correspond to the first and third quartiles. Treatments are colour coded as fertilization & biomass removal: purple; fertilized & reduced biomass removal: bright blue, unfertilized & biomass removal: green; unfertilized & reduced biomass removal: yellow. We did not observe any clear significant differences between any of the treatments (S2 Table). (DOCX) Click here for additional data file. S3 Fig Shannon diversity per treatment. Response of Shannon diversity (m-2) on the fertilized & biomass removal (+F+R), unfertilized & biomass removal (-F+R), unfertilized & reduced biomass removal (-F-R), and fertilized & reduced biomass removal (+F-R) treatment for all three regions (Alb: Schwäbische Alb; Sch: Schorfheide-Chorin; Hai: Hainich-Dün). Whiskers correspond to the first and third quartiles. Treatments are colour coded as fertilization & biomass removal: purple; fertilized & reduced biomass removal: bright blue, unfertilized & biomass removal: green; unfertilized & reduced biomass removal: yellow. Bars sharing a letter (a,b) do not differ significantly (p<0.05, S10 Table). We only detected significant differences between treatments in summer 2020 within the Schwäbische Alb, thus no letters are shown for other years, seasons or regions. (DOCX) Click here for additional data file. S4 Fig Standing biomass per treatment. Response of standing biomass (g m-2) on the fertilized & biomass removal (+F+R), unfertilized & biomass removal (-F+R), unfertilized & reduced biomass removal (-F-R), and fertilized & reduced biomass removal (+F-R) treatment for all three regions (Alb: Schwäbische Alb; Sch: Schorfheide-Chorin; Hai: Hainich-Dün). Whiskers correspond to the first and third quartiles. Treatments are colour coded as fertilization & biomass removal: purple; fertilized & reduced biomass removal: bright blue, unfertilized & biomass removal: green; unfertilized & reduced biomass removal: yellow. Bars sharing a letter (a,b) do not differ significantly (p<0.05, S12 Table). (DOCX) Click here for additional data file. S5 Fig Biomass production per treatment. Response of biomass production (g m-2) on the fertilization & biomass removal (+F+R), unfertilized & biomass removal (-F+R), unfertilized & reduced biomass removal (-F-R), and fertilized & reduced biomass removal (+F-R) treatment for all three regions (Alb: Schwäbische Alb; Sch: Schorfheide-Chorin; Hai: Hainich-Dün). Treatments are colour coded as fertilization & biomass removal: purple; fertilized & reduced biomass removal: bright blue, unfertilized & biomass removal: green; unfertilized & reduced biomass removal: yellow. Bars sharing a letter (a,b) do not differ significantly (p<0.05, S14 Table). (DOCX) Click here for additional data file. S6 Fig Correlation between land-use components. Correlation between background LUI Index components, background fertilization intensity (kg N m-3 year-1), background mowing intensity (cuts year-1) and background grazing intensity (Livestock units * d ha-1 year-1) averaged across the years 2017–2019 and combined for all regions. (DOCX) Click here for additional data file. S7 Fig Correlation between variables of SEM in spring. Correlation matrix plot for the relationships between richness and Shannon diversity as well as the response variables standing biomass, light availability, and background fertilization intensity for spring of both 2020 and 2021, and colour coded for each region separately (Alb: Schwäbische Alb (red); Sch: Schorfheide-Chorin (blue); Hai: Hainich-Dün (green)). (DOCX) Click here for additional data file. S8 Fig Correlation between variables of SEM in summer. Correlation matrix plot for the relationships between richness and Shannon diversity as well as the response variables standing biomass, biomass production, light availability, and background fertilization intensity for summer of both 2020 and 2021, and colour coded for each region separately (Alb: Schwäbische Alb (red); Sch: Schorfheide-Chorin (blue); Hai: Hainich-Dün (green)). (DOCX) Click here for additional data file. S9 Fig Partial residual plots, showing the effect of light availability and soil moisture on richness in spring (A,B) and summer (C,D), as well as on Shannon diversity in spring (E,F) and summer (G,H). Partial residuals were extracted from model predicting species richness or Shannon diversity of respective SEM model. For more details on model estimates and significance see S3–S6 Tables. (DOCX) Click here for additional data file. S10 Fig Changes in community composition in 2020. Non-metric multidimensional scaling (NMDS) based on Bray–Curtis similarity index for the plant communities for each treatment combination in spring (A-C) and summer (D-F) 2020 (A,D) Schwäbische Alb, (B,E) Hainich-Dün, (C,F) Schorfheide Chorin. Hull volumes represent clusters of plant communities within a given treatment. Treatments are colour coded as fertilized & biomass removal (+F+R): purple; fertilized & reduced biomass removal (+F-R): bright blue, unfertilized & biomass removal (-F+R): green; unfertilized & reduced biomass removal (-F-R): yellow. Detailed model summaries are shown in the Supporting Information (S15 Table). (DOCX) Click here for additional data file. S11 Fig Community composition and environmental factors in 2020. Non-metric multidimensional scaling (NMDS) based on Bray–Curtis similarity index for the plant communities for each treatment combination and all regions combined in spring (A) and summer (B) for 2020. R2-values are shown for each axis. Arrows printed in grey represent environmental factors: standing biomass, light availability and soil moisture, while arrows printed in red represent species richness and Shannon diversity. Species names represent the most extreme species according to the NMDS axes. Treatments are colour coded as fertilized & biomass removal (+F+R): purple; fertilized & reduced biomass removal (+F-R): bright blue, unfertilized & biomass removal (-F+R): green; unfertilized & reduced biomass removal (-F-R): yellow. Detailed model summaries are shown in the Supporting Information (S16 Table). (DOCX) Click here for additional data file. S1 Table Species richness in response to treatments. Linear mixed effect model showing the effect of the unfertilized & reduced biomass removal (-F-R), fertilized & reduced biomass removal (+F-R), unfertilized & biomass removal (-F+R) on species richness in comparison with the fertilized & biomass removal treatment for each regions (Alb: Schwäbische Alb; Sch: Schorfheide-Chorin; Hai: Hainich-Dün), as well as for different years and seasons. (DOCX) Click here for additional data file. S2 Table Pairwise comparison of species richness across treatments. Pairwise comparisons of the species richness in the fertilization & biomass removal (+F+R), unfertilized & reduced biomass removal (-F-R), unfertilized & biomass removal (-F+R) and fertilized & reduced biomass removal (+F-R) treatment, for each region (Alb: Schwäbische Alb; Sch: Schorfheide-Chorin; Hai: Hainich-Dün), as well as for different years and seasons. Significant (< 0.05) contrasts are written in bold. (DOCX) Click here for additional data file. S3 Table PiecewiseSEM model fit for model with main response species richness in spring of both 2020 and 2021 with the main responses biomass removal (unfertilized & biomass removal treatment), fertilization (fertilized & reduced biomass removal treatment), standing biomass, log(light availability), background fertilization, region, sampling date and year. Explained variances for species richness: mar. R2 = 0.15 (adj. R2 = 0.40), standing biomass: mar. R2 = 0.23 (adj. R2 = 0.31); log(light availability): mar. R2 = 0.69 (adj. R2 = 0.69); soil moisture: mar. R2 = 0.02 (adj. R2 = 0.61). Alb: Schwäbische Alb; Sch: Schorfheide-Chorin; Hai: Hainich-Dün. (DOCX) Click here for additional data file. S4 Table PiecewiseSEM model fit for model with main response Shannon diversity in spring of both 2020 and 2021 with the main responses biomass removal (unfertilized & biomass removal treatment), fertilization (fertilized & reduced biomass removal treatment), standing biomass, log(light availability), background fertilization region, sampling date and year. Explained variances for Shannon diversity: mar. R2 = 0.14 (adj. R2 = 0.35), standing biomass: mar. R2 = 0.23 (adj. R2 = 0.31); log(light availability): mar. R2 = 0.69 (adj. R2 = 0.69); soil moisture: mar. R2 = 0.02 (adj. R2 = 0.61). Alb: Schwäbische Alb; Sch: Schorfheide-Chorin; Hai: Hainich-Dün. (DOCX) Click here for additional data file. S5 Table PiecewiseSEM model fit for model with main response species richness in summer of both 2020 and 2021 with the main responses biomass removal (unfertilized & biomass removal treatment), fertilization (fertilized & reduced biomass removal treatment), standing biomass, biomass production, log(light availability), background fertilization region, sampling date and year. Explained variances for species richness: mar. R2 = 0.11 (adj. R2 = 0.56), standing biomass: mar. R2 = 0.35 (adj. R2 = 0.69); biomass production: mar. R2 = 0.32 (adj. R2 = 0.43); log(light availability): mar. R2 = 0.66 (adj. R2 = 0.67); soil moisture: mar. R2 = 0.02 (adj. R2 = 0.74). Alb: Schwäbische Alb; Sch: Schorfheide-Chorin; Hai: Hainich-Dün. (DOCX) Click here for additional data file. S6 Table PiecewiseSEM model fit for model with main response Shannon diversity in summer of both 2020 and 2021 with the main responses biomass removal (unfertilized & biomass removal treatment), fertilization (fertilized & reduced biomass removal treatment), standing biomass, biomass production, log(light availability), background fertilization region, sampling date and year. Explained variances for Shannon diversity: mar. R2 = 0.10 (adj. R2 = 0.36), standing biomass: mar. R2 = 0.35 (adj. R2 = 0.69); biomass production: mar. R2 = 0.32 (adj. R2 = 0.43); log(light availability): mar. R2 = 0.66 (adj. R2 = 0.67); soil moisture: mar. R2 = 0.02 (adj. R2 = 0.74). Alb: Schwäbische Alb; Sch: Schorfheide-Chorin; Hai: Hainich-Dün. (DOCX) Click here for additional data file. S7 Table Pairwise comparison (PERMANOVA) of plant community composition per treatment in 2021 between the fertilized & biomass removal (+F+R), unfertilized & reduced biomass removal (-F-R), fertilized & reduced biomass removal (+F-R), unfertilized & biomass removal (-F+R) treatments for all three regions (Alb: Schwäbische Alb; Sch: Schorfheide-Chorin; Hai: Hainich-Dün) and for both spring and summer 2021 (see Fig 3). (DOCX) Click here for additional data file. S8 Table Permutation test of fitted vectors of the environmental variables (standing biomass, light availability and soil moisture), plant species richness and Shannon diversity in 2021 on the NMDS ordination (NMDS1 and NMDS2) for spring and summer of 2021 across all regions (Schwäbische Alb, Hainich-Dün, Schorfheide-Chorin) (S5 Fig). (DOCX) Click here for additional data file. S9 Table Shannon diversity in response to treatments. Linear mixed effect model showing the effect of the unfertilized & reduced biomass removal (-F-R), fertilized & reduced biomass removal (+F-R), unfertilized & biomass removal (-F+R) on Shannon diversity in comparison with the fertilized & biomass removal (+F+R) treatment for each regions (Alb: Schwäbische Alb; Sch: Schorfheide-Chorin; Hai: Hainich-Dün), as well as for different years and seasons. (DOCX) Click here for additional data file. S10 Table Pairwise comparison of Shannon diversity across treatments. Pairwise comparisons of the Shannon diversity in the fertilization & biomass removal (+F+R), unfertilized & reduced biomass removal (-F-R), unfertilized & biomass removal (-F+R) and fertilized & reduced biomass removal (+F-R) treatment, for each region (Alb: Schwäbische Alb; Sch: Schorfheide-Chorin; Hai: Hainich-Dün), as well as for different years and seasons. No pairwise comparison shown for spring 2021, as the predictor ‘treatment’ was not part of the most parsimonious model. Significant (< 0.05) contrasts are written in bold. (DOCX) Click here for additional data file. S11 Table Standing biomass in response to treatments. Linear mixed effect model showing the effect of the unfertilized & reduced biomass removal (-F-R), fertilized & reduced biomass removal (+F-R), unfertilized & biomass removal (-F+R) on standing biomass in comparison with the fertilized & biomass removal treatment for each regions (Alb: Schwäbische Alb; Sch: Schorfheide-Chorin; Hai: Hainich-Dün), as well as for different years and seasons. (DOCX) Click here for additional data file. S12 Table Pairwise comparison of standing biomass across treatments. Pairwise comparisons of the standing biomass in the fertilization & biomass removal, unfertilized & reduced biomass removal, unfertilized & biomass removal and fertilized & reduced biomass removal treatment, for each region (Alb: Schwäbische Alb; Sch: Schorfheide-Chorin; Hai: Hainich-Dün), as well as for different years and seasons. Significant (< 0.05) contrasts are written in bold. Due to missing data on fertilized & biomass removal treatments in spring 2020 for the Schorfheide-Chorin no pairwise contrasts shown (*). (DOCX) Click here for additional data file. S13 Table Biomass production in response to treatments. Linear mixed effect model showing the effect of the unfertilized & reduced biomass removal (-F-R), fertilized & reduced biomass removal (+F-R), unfertilized & biomass removal (-F+R) on biomass production in comparison with the fertilized & biomass removal treatment for each regions (Alb: Schwäbische Alb; Sch: Schorfheide-Chorin; Hai: Hainich-Dün), in summer for all different years. (DOCX) Click here for additional data file. S14 Table Pairwise comparison of biomass production across treatments. Pairwise comparisons of biomass production in the fertilization & biomass removal (+F+R), unfertilized & reduced biomass removal (-F-R), unfertilized & biomass removal (-F+R) and fertilized & reduced biomass removal (+F-R) treatment, for each region (Alb: Schwäbische Alb; Sch: Schorfheide-Chorin; Hai: Hainich-Dün), as well as for different years and seasons. Significant (< 0.05) contrasts are written in bold. Due to missing data on fertilized & biomass removal treatments in spring 2020 for the Schorfheide-Chorin no pairwise contrasts shown (*). (DOCX) Click here for additional data file. S15 Table Pairwise comparison (PERMANOVA) of plant community composition per treatment in 2020 between the fertilized & biomass removal (+F+R), unfertilized & reduced biomass removal (-F-R), fertilized & reduced biomass removal (+F-R), unfertilized & biomass removal (-F+R) treatments for all three regions (Alb: Schwäbische Alb; Sch: Schorfheide-Chorin; Hai: Hainich-Dün) and for both spring and summer 202O (see S9 Fig). (DOCX) Click here for additional data file. S16 Table Permutation test of fitted vectors of the environmental variables (standing biomass, light availability and soil moisture), plant species richness and Shannon diversity in 2020 on the NMDS ordination (NMDS1 and NMDS2) for spring and summer of 2020 across all regions (Schwäbische Alb, Hainich-Dün, Schorfheide-Chorin) (S10 Fig). (DOCX) Click here for additional data file. S17 Table Mean LUI drivers (grazing, mowing and fertilization) averaged across 2017 to 2019 for each region separately (Alb: Schwäbische Alb, Hai: Hainich-Dün, Sch: Schorfheide-Chorin). (DOCX) Click here for additional data file. S18 Table Calibration model of biomass estimates based on a linear regression between actual standing biomass measurements (dry matter in g m-2) and rising plate meter measurement (1/2 cm increments), combined for all regions. Calibration was performed for data measured in spring 2021. (DOCX) Click here for additional data file. S1 File (TXT) Click here for additional data file. S2 File (RMD) Click here for additional data file. We thank Svenja Kunze, Ralph Bolliger, Uta Schumacher, Jörg Hailer, Christin Schreiber, Victoria Henning and many students for helping us to collect data in the field. We also thank Christian Wirth and Christiane Roscher for their input during the conceptual phase of this paper. Also many thanks to the managers of the three Exploratories, Max Müller, Robert Künast, Franca Marian, and all former managers for their work in maintaining the plot and project infrastructure; Victoria Grießmeier for giving support through the central office, Andreas Ostrowski for managing the central database, and Markus Fischer, Eduard Linsenmair, Dominik Hessenmöller, Ingo Schöning, François Buscot, Ernst-Detlef Schulze, Wolfgang W. Weisser and the late Elisabeth Kalko for their role in setting up the Biodiversity Exploratories project. We thank the administration of the Hainich national park, the UNESCO Biosphere Reserve Swabian Alb and the UNESCO Biosphere Reserve Schorfheide-Chorin as well as all land owners for the excellent collaboration. Field work permits were issued by the responsible state environmental offices of Baden-Württemberg, Thüringen, and Brandenburg (according to § 72 BbgNatSchG). 10.1371/journal.pone.0287039.r001 Decision Letter 0 Jenkins David G. Academic Editor © 2023 David G. Jenkins 2023 David G. Jenkins https://creativecommons.org/licenses/by/4.0/ This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Submission Version0 23 Jan 2023 PONE-D-22-34453Biomass removal promotes plant diversity after short-term extensification of managed grasslandsPLOS ONE Dear Dr. Andraczek, Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process. And thanks for you patience with the review process - I got sick with covid (still positive - yay) - which explains some of the delay. But more importantly, multiple reviewers declined to review the paper and rather than continuing to chase down more at this time, I suggest that figures be processed to make them more legible. If I received this as a reviewer, I would say "nope - can't read figures." I also suggest some minor edits that may help. So "Major Revision" here is intended to flag that it will go to reviewers (vs reject [a go away] or minor revision [editor could handle it]). Please see below for details, and then resubmit with me suggested as Academic Editor again, including a quick summary in your cover letter of how you approached the list below. I will then try again on the reviewer-hunting process, with the hope that edits will make it more amenable to reviewers. My comments: 1. What is the difference between intensification and extensification? If different, please explain. If actually the same, then I suggest using “intensification” consistently. 2. Must the experiment be justified in a BEF context? Why not simply a matter of understanding land use intensification effects on biodiversity? In other words, the BEF context may not be needed and only adds an extra layer of inference here, whereas biodiversity seems worthy of preserving regardless. Think about it. 3. Shannon diversity: please define with its equation. It could be the original H’, but as Jost (2006) nicely demonstrated, that is an entropy that requires exp(H’) to obtain effective diversity. Thus “Shannon diversity” has taken on several meanings of late. I suggest calculating and using Jost’s effective diversity, which corresponds to Hill number 1 (1D). 4. Fig. 1 Are the numbers related to a sequence (implied by numbering)? If not, perhaps they can be removed. Also, I suggest more obviously dashed lines for negative effects: on the pdf the resolution is too low, meaning that dashes are not clearly different from the solid lines – especially when zoomed out to be comparable to the size of a figure in a printed journal page (imagine figures squeezed into a column or ½ width of paper). 5. MOST IMPORTANTLY: Resolution of all Figures in the received pdf (and what reviewers see) is too low, so that text in figures is illegible and shapes are very indistinct. I expect negative reviews based only on this technical problems rather than merits of the work. According to the PLOS One office, "It appears the reason this error is occurring is because the authors provided figure files that are incompatible with Editorial Manager's PDF compilation. We encourage authors to use the the PACE tool (https://pacev2.apexcovantage.com/) on all of their Fig files to ensure the images will be of the highest quality in PDF form." Thus I strongly suggest following those PLOS figure guidelines for a revised submission. Please submit your revised manuscript by Mar 09 2023 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file. Please include the following items when submitting your revised manuscript: A letter that responds to each point raised above. You should upload this letter as a separate file labeled 'Response to Reviewers'. A new copy of your manuscript with improved figures and edits for my comments (as you see fit). You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'. An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'. This will go to reviewers. Guidelines for resubmitting your figure files are available below, at the end of this letter. We look forward to receiving your revised manuscript. Kind regards, David G. Jenkins, PhD Academic Editor PLOS ONE Journal Requirements: When submitting your revision, we need you to address these additional requirements. 1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at  https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and  https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf 2. We note that the grant information you provided in the ‘Funding Information’ and ‘Financial Disclosure’ sections do not match.  When you resubmit, please ensure that you provide the correct grant numbers for the awards you received for your study in the ‘Funding Information’ section. 3. Thank you for stating the following in the Acknowledgments Section of your manuscript:  "The work has been funded by the German Research Foundation (DFG) Priority Program 1374 ‘Biodiversity-Exploratories’ (BEF-Loops Nr.433266560, Exploratories project phase Nr. PL 891/3-1)." We note that you have provided funding information that is not currently declared in your Funding Statement. However, funding information should not appear in the Acknowledgments section or other areas of your manuscript. We will only publish funding information present in the Funding Statement section of the online submission form.  Please remove any funding-related text from the manuscript and let us know how you would like to update your Funding Statement. Currently, your Funding Statement reads as follows:  "The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript." Please include your amended statements within your cover letter; we will change the online submission form on your behalf. 4. In your Data Availability statement, you have not specified where the minimal data set underlying the results described in your manuscript can be found. PLOS defines a study's minimal data set as the underlying data used to reach the conclusions drawn in the manuscript and any additional data required to replicate the reported study findings in their entirety. All PLOS journals require that the minimal data set be made fully available. For more information about our data policy, please see http://journals.plos.org/plosone/s/data-availability. Upon re-submitting your revised manuscript, please upload your study’s minimal underlying data set as either Supporting Information files or to a stable, public repository and include the relevant URLs, DOIs, or accession numbers within your revised cover letter. For a list of acceptable repositories, please see http://journals.plos.org/plosone/s/data-availability#loc-recommended-repositories. Any potentially identifying patient information must be fully anonymized. Important: If there are ethical or legal restrictions to sharing your data publicly, please explain these restrictions in detail. Please see our guidelines for more information on what we consider unacceptable restrictions to publicly sharing data: http://journals.plos.org/plosone/s/data-availability#loc-unacceptable-data-access-restrictions. Note that it is not acceptable for the authors to be the sole named individuals responsible for ensuring data access. We will update your Data Availability statement to reflect the information you provide in your cover letter. 5. Please include a caption for figure 2.  [NOTE: The submitted manuscript is attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.] While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step. Attachment Submitted filename: PONE-D-22-34453.pdf Click here for additional data file. 10.1371/journal.pone.0287039.r002 Author response to Decision Letter 0 Submission Version1 10 Feb 2023 Editor’s comments We very much appreciate that the editor finds our study interesting and for giving us a chance to resubmit the manuscript. Regarding the specific comments by the editor, we refer to our responses on the more detailed comments below. #1. What is the difference between intensification and extensification? If different, please explain. If actually the same, then I suggest using “intensification” consistently. We thank the editor for this useful suggestion on clarifying the difference between ‘intensification’ and ‘extensification’. As a short answer here: with land-use extensification we are referring to the opposite of land-use intensification, hence a reduction in land-use intensity (i.e. reduction in mowing frequency, grazing or fertilization intensity). We now provided a definition in the introduction to improve clarity about the terminologies used (see changes in lines 56-57). #2. Must the experiment be justified in a BEF context? Why not simply a matter of understanding land use intensification effects on biodiversity? In other words, the BEF context may not be needed and only adds an extra layer of inference here, whereas biodiversity seems worthy of preserving regardless. Think about it. We originally included the BEF context to stress the importance of why we need to advance our understanding on how land-use drives biodiversity loss, as this loss may have detrimental consequences for ecosystem functioning. However, as the editor correctly points out, that rather adds an additional layer of complexity to our study, while biodiversity in itself is worthy of preserving regardless. Thus, we agree to the editor’s suggestion that our introduction would benefit from refining our focus and now excluded the link to the BEF context (the sentences on lines 51-53 of our original manuscript has been removed). #3. Shannon diversity: please define with its equation. It could be the original H’, but as Jost (2006) nicely demonstrated, that is an entropy that requires exp(H’) to obtain effective diversity. Thus “Shannon diversity” has taken on several meanings of late. I suggest calculating and using Jost’s effective diversity, which corresponds to Hill number 1 (1D). Following this suggestion, we now adjusted all our analyses by using exp(H’) instead of just H’. We now also provided the specific formula used to calculate effective diversity in the methods section (see changes in lines: 210-212). While most of the results did not change after using the exp(H’), as the only qualitative difference we now find that indirect effects of biomass removal on diversity were not only mediated via changes in light availability, but also via soil moisture (Fig. 4, Table 1; lines 369-371, 489-493). Hence, using effective diversity revealed the relative importance of soil moisture in summer as an indirect pathway mediated by biomass removal, similar to what we observed for species richness, making our findings more consistent. #4. Fig. 1 Are the numbers related to a sequence (implied by numbering)? If not, perhaps they can be removed. Also, I suggest more obviously dashed lines for negative effects: on the pdf the resolution is too low, meaning that dashes are not clearly different from the solid lines – especially when zoomed out to be comparable to the size of a figure in a printed journal page (imagine figures squeezed into a column or ½ width of paper). Yes, the numbers within the figure are related to a sequence of supporting references for each path within the causal framework and hence, we would prefer to keep the numbers within the figure. We now clarified this issue in line 108. The specific references are included in the figure caption. #5. Also, I suggest more obviously dashed lines for negative effects: on the pdf the resolution is too low, meaning that dashes are not clearly different from the solid lines – especially when zoomed out to be comparable to the size of a figure in a printed journal page (imagine figures squeezed into a column or ½ width of paper). MOST IMPORTANTLY: Resolution of all Figures in the received pdf (and what reviewers see) is too low, so that text in figures is illegible and shapes are very indistinct. I expect negative reviews based only on this technical problems rather than merits of the work. According to the PLOS One office, "It appears the reason this error is occurring is because the authors provided figure files that are incompatible with Editorial Manager's PDF compilation. We encourage authors to use the the PACE tool (https://pacev2.apexcovantage.com/) on all of their Fig files to ensure the images will be of the highest quality in PDF form." Thus I strongly suggest following those PLOS figure guidelines for a revised submission. We agree, and apologize for the low figure quality of the figures we provided. We now increased the quality of all figures (and ensured journal requirements for all figures using the PACE tool) within the document and, to better visualize negative of positive hypothesised relationships, also selected a different line-type for the dashed lines. Attachment Submitted filename: Response to Reviewers.docx Click here for additional data file. 10.1371/journal.pone.0287039.r003 Decision Letter 1 Jenkins David G. Academic Editor © 2023 David G. Jenkins 2023 David G. Jenkins https://creativecommons.org/licenses/by/4.0/ This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Submission Version1 29 Mar 2023 PONE-D-22-34453R1Biomass removal promotes plant diversity after short-term extensification of managed grasslandsPLOS ONE Dear Dr. Andraczek, Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process. You will see that we received commens from one outside expert reviewer, which I supplemented with my own detailed review, Both are written in the spirit of improving the manuscript toward publication, and I encourage you to read and respond to each comment with your revisions so that a decision can be made promptly on your revised manuscript. I anticipate handling that phase without sending it out for external review again, given that my comments may lead to more revisions than the reviewer's, and that the next decision should not be too difficult. Please submit your revised manuscript by 30 May 2023. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file. Please include the following items when submitting your revised manuscript:A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'. A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'. An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'. If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter. If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols. We look forward to receiving your revised manuscript. Kind regards, David G. Jenkins, PhD Academic Editor PLOS ONE Additional Editor Comments: Dear Dr. Andraczek and co-authiors, Below you will find I sought at least two reviewers but in the end obtained only one. Thus I also wrote comments aimed to improve the manuscript – in the way a reviewer should. Please do not assign any more weight to my comments below in your responses and revisions than you would to an anonymous reviewer (i.e., feel free to rebut my thoughts with good arguments). I find Reviewer 1 is generally complementary about the manuscript and provides good suggestions. As usual, the reviewer and I see different features to emphasize – a good thing! I especially agree with the reviewer in the suggestion that the short-term effects should be emphasized, which would be consistent with my comments below about potential, and acknowledged, lag effects. Reviewer 1 suggested Minor Revisions, but I think the aggregate of their comments and mine add up to Major Revision (for what that’s worth). I also think that having resolved or dispelled comments here, the manuscript should be ready to accept. Please see the details on Reviewer 1’s comments via the information provided by PLOS One. Below are my comments on the manuscript. Finally, thanks for your patience in this review process. Sincerely, David G. Jenkins Title and lines 56-57 and elsewhere. The word “extensification” is not the opposite of “intensification” and is not used by the cited papers [9, 10]. If I look it up in online dictionaries I see definitions like this: 1. the process of making something (more) extensive 2. the geographic spread and distribution of any technology, especially agriculture. Thus I suggest it be replaced with an antonym, such as one of those listed at https://www.thesaurus.com/browse/intensification? Or perhaps "mitigation"? Or more clumsy: "de-intensification"? Line 247. I expected SEM software etc. to be listed in this section. It may help to re-organize subheadings? Line 255. I strongly encourage authors to move beyond repeated pairwise Tukey tests and rely more on GLMs and SEMs for interpreting statistical trends. Paragraph starting at line 292. I strongly recommend the authors use AIC.psem in the piecewise SEM package to compare models for “efficiency” or “plausibility.” This will provide a more robust test of which model to spend more time with than other tests, and then leaves goodness of fit measures as a way to criticize the most efficient model selected by AIC. See extensive literature by Burnham and Anderson for more details on AICs. The AIC results should then be a table in the manuscript, followed by results for the "winning" model. Fig. 2 revealed little to me. I recommend the authors instead show and discuss a table of GLM output from lme4: are coefficients for biomass removal and fertilizer addition clearly non-zero? What are the signs? What is the overall fit? Speaking of regression output, S1 Table is close to what I sought, but I think it reveals errors that matter: • Why is is Spring 2020 different from other time steps in its model structure? • The intercept (default by R) is Alb but also +F + R. It may make more sense to choose (using relevel in R) to use Alb but -F -R for a control condition as the intercept. Also I suggest converting SE to 95% confidence intervals (1.96 X SE) and omitting the DF and t value columns. And note that rows are relative to the intercept. Finally, the study was designed as if site and time could be random effects, though with too few levels of each to really help as random effects. That makes all the sites and times show as separate estimates. The minimum number of levels needed for random effects in a mixed effect model may not be met here (that threshold is not clear), but I suggest trying a GLMM instead of the GLM. That would potentially move sites and time into random effects, leaving main effects of interest as fixed effects and making interpretation more straight-forward. I may be wrong - it may not work because sample size is limiting. If so, please just say so in your response to review comments. But - if it works, it may be more efficient and effective as a test of treatment effects. Line 313. This sentence could be edited to be more direct. Beyond that, this section 4.1 based on regression could be argued (above?) as seeking overall (i.e., direct and indirect) effects of treatments on diversity, whereas SEM seeks to tease apart the direct from the indirect. Thus regression evaluates overall / net effects while SEM evaluates details. Perhaps that can help reconcile the greater information in Fig. 4 than in Fig. 2. And by the way, the use of SEM is an advance to make clear for this research subject – much research has not used it but work here shows it helps bring clarity. Line 417. I don’t know what is meant by “land use reductions.” Paragraph starting at Line 425 and subsequent paragraphs. It is usually expected that Figures are not cited in Discussion, to ensure that Results are clearly separate from Discussion. If it feels like Figures need to be used, then the sentence may belong better in Results. Line 438. Yes – I agree; time lags are often observed, and some citations to support this would be useful here. Paragraph starting at Line 448. This paragraph covers almost 3 pages! Organize and subdivide. Line 522. Careful! Not only is “extensification”used here again, but I think the authors need to be careful to avoid claiming a first. Consult the rich literature by Tilman and a small army of others at Cedar Creek. In addition, other studies that used a broader range of fertilizer and biomass removal have explored hump-shaped patterns related to the Intermediate Disturbance Hypothesis, Intermediate Productivity Hypothesis, and Huston’s combination of those two ideas (his Dynamic Equilibrium Model). There is a long history of inquiry here that includes factors here as part of the story, as the co-authors well know (they have been involved in some). Among others, a few that may be relevant include: • Feng-Wei Xu, Jian-Jun Li, Li-Ji Wu, Xiao-Ming Lu, Wen Xing, Di-Ma Chen, Biao Zhu, Shao-Peng Wang, Lin Jiang, Yong-Fei Bai, Resource enrichment combined with biomass removal maintains plant diversity and community stability in a long-term grazed grassland, Journal of Plant Ecology, 113(5), October 2020, Pages 611–620, https://doi.org/10.1093/jpe/rtaa046 • Band, N., Kadmon, R., Mandel, M. and DeMalach, N., 2022. Assessing the roles of nitrogen, biomass, and niche dimensionality as drivers of species loss in grassland communities. Proceedings of the National Academy of Sciences, 119(10), p.e2112010119. • Carnus, T., Connolly, J., Kirwan, L. and Finn, J.A., 2013. Plant diversity effects are robust to cutting severity and nitrogen application in productive grasslands. In The role of grasslands in a green future: threats and perspectives in less favoured areas. Proceedings of the 17th Symposium of the European Grassland Federation, Akureyri, Iceland, 23-26 June 2013 (pp. 195-197). Agricultural University of Iceland. I think the concluding paragraph could benefit from a sentence or two on limits of the study – as difficult as it is to get this much data, it still has limits in the number places and times, for strong analyses and to capture acknowledged potential lag effects. [Note: HTML markup is below. Please do not edit.] Reviewers' comments: Reviewer's Responses to Questions Comments to the Author 1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation. Reviewer #1: (No Response) ********** 2. Is the manuscript technically sound, and do the data support the conclusions? The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented. Reviewer #1: Yes ********** 3. Has the statistical analysis been performed appropriately and rigorously? Reviewer #1: Yes ********** 4. Have the authors made all data underlying the findings in their manuscript fully available? The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified. Reviewer #1: Yes ********** 5. Is the manuscript presented in an intelligible fashion and written in standard English? PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here. Reviewer #1: Yes ********** 6. Review Comments to the Author Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters) Reviewer #1: The authors present results on plant biodiversity, community composition and plant biomass from an in-field experiment in which the land-use components fertilization and biomass removal have been reduced in a full-factorial design. The results represent the short-term effects one and two years after the experiments have been implemented. Although effects on the plant community are very small overall, the study does provide interesting insights into possible mechanistic pathways as the authors also included light availability and soil moisture in their models. Both the experimental design and the statistical analyses are done rigorously and correctly, hence the conclusions are supported by the data. * The main strength of the experiment compared to previous studies is the implementation of all treatments within the same grassland sites. This could be emphasised more in the introduction. * The introduction lacks an explanation of the expected "seasonal variability of relationships" mentioned in the legend of Figure 1. As spring and summer data are analysed and discussed separately, a section on seasonal differences should be added in the introduction. * The (indirect) effect of biomass production on diversity is mentioned in the hypotheses, but not explained and motivated in the Introduction. As biomass production is an important component in the model structure, the expected effects and mechanisms should be added to the introduction. * The experiments were set up in grasslands of different land-use intensity. While the average intensity values are given in the Supplement, the description in the Methods section is inconsistent: In line 142f "a gradient in land-use intensity" is mentioned, while in line 163 management is described as "relatively high" and in line 163 as "medium to high". In addition to making the statements consistent, I recommend describing the range of land-use components on the grasslands by the actual number and time of cuts, amount of fertilizer and/or number of grazing animals (and not just in terms of the standardized land-use intensity). * The description of the treatments is in some parts difficult to read, especially if several treatments are mentioned in one sentence. I would prefer the use of abbreviations as done for the supplementary tables. The authors could also consider calling the treatment with fertilizer and biomass removal the control. * The terms subplot, plot and field are used inconsistently in the first part of the manuscript. See for instance line 176f. Please use consistent terminology and I would recommend not using "field" but rather "grassland". Minor comments: Line 146: "the 10 most abundant" Line 171: Why was the fertilizer manually applied in this treatment? Why was is not fertilized together with the "control" treatment and the remaining grassland? Line 173f: This information can be moved to the next section on field data collection Line 158: What does "setting-up" mean exactly? Does this mean that the plots were marked or was e.g. fertilization in autumn 2019 already reduced? Please clarify in which year/season the treatments were applied for the first time. Line 243 ff: This sentence contains the same information twice. Line 254: remove "with" Line 257: which additional predictors are those? Line 313 ff: there is no need to mention the type of treatment again in the parentheses Line 350f (and later): it is not quite clear what the "respectively" refers to. The path coefficients should be given directly after richness and diversity. Figure 2: the y-axis label seems incorrect (remove the minus) ********** 7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files. If you choose “no”, your identity will remain anonymous but your review may still be made public. Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy. Reviewer #1: No ********** [NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.] While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step. 10.1371/journal.pone.0287039.r004 Author response to Decision Letter 1 Submission Version2 8 May 2023 Editor’s comments I find Reviewer 1 is generally complementary about the manuscript and provides good suggestions. As usual, the reviewer and I see different features to emphasize – a good thing! I especially agree with the reviewer in the suggestion that the short-term effects should be emphasized, which would be consistent with my comments below about potential, and acknowledged, lag effects. Reviewer 1 suggested Minor Revisions, but I think the aggregate of their comments and mine add up to Major Revision (for what that’s worth). I also think that having resolved or dispelled comments here, the manuscript should be ready to accept. Please see the details on Reviewer 1’s comments via the information provided by PLOS One. Below are my comments on the manuscript. Finally, thanks for your patience in this review process. We very much appreciate the complementary comments provided by both the editor and the reviewer. We fully agree with the general recommendation to emphasize short-term and lag effects and adjusted the respective sections in our discussion (see changes in lines: 39, 126-128, 451-454, 541, 557-558). Regarding the specific comments by the editor and the reviewers, we refer to our responses to the more detailed comments below. Title and lines 56-57 and elsewhere. The word “extensification” is not the opposite of “intensification” and is not used by the cited papers [9, 10]. If I look it up in online dictionaries I see definitions like this: 1. the process of making something (more) extensive 2. the geographic spread and distribution of any technology, especially agriculture. Thus I suggest it be replaced with an antonym, such as one of those listed at https://www.thesaurus.com/browse/intensification? Or perhaps "mitigation"? Or more clumsy: "de-intensification"? We thank the editor for this comment and now used the term “de-intensification” when referring to reduction of land-use intensity throughout the manuscript and within our title (“Biomass removal promotes plant diversity after short-term de-intensification of managed grasslands”). See also lines 1-2, 56, 110, 127 and 154. Line 247. I expected SEM software etc. to be listed in this section. We agree and now added the SEM software in line 257. It may help to re-organize subheadings? We agree and re-organized and re-named the subheadings in the methods section to allow a better distinction between “conventional” statistical methods (linear mixed effect models and pairwise contrasts) and more in-depth analyses using multivariate analyses (NMDS) as well as piecewiseSEM. Specifically, we refined subheadings in the methods (see lines: 258, 272, 284) and result section (see lines: 324). Line 255. I strongly encourage authors to move beyond repeated pairwise Tukey tests and rely more on GLMs and SEMs for interpreting statistical trends. We fully agree with the editor that GLMs and SEMs provide more in-depth statistical results compared to analyses based on pairwise Tukey tests. We noticed that we communicated this wrongly in our original manuscript and apologize for the confusion: In our original manuscript we did not use the Tukey HSD test (as incorrectly mentioned in our methods section) to compute pairwise differences. Rather, after testing for overall effects of treatments within our Linear Mixed Model, we applied the more advanced emmeans function from the package “emmeans”, which computes contrasts between treatments based on estimated least-square means. To clarify this information we corrected all respective description about pairwise tests throughout our manuscript (see lines: 261-263, 327-328). In our manuscript we deliberately analysed differences between treatments using “conventional” Linear Mixed Models as a first step, to test for treatment effects. However, we also agree with the editor that more advanced statistical methods are needed to provide more insights into the mechanisms (e.g. through changes in vegetation biomass [and hence light competition], or changes in soil moisture) by which these treatments influence plant biodiversity indicators. Hence, in our manuscript we used piecewiseSEM for this purpose (methods: section 3.5.3, results: section 4.2). Paragraph starting at line 292. I strongly recommend the authors use AIC.psem in the piecewise SEM package to compare models for “efficiency” or “plausibility.” This will provide a more robust test of which model to spend more time with than other tests, and then leaves goodness of fit measures as a way to criticize the most efficient model selected by AIC. See extensive literature by Burnham and Anderson for more details on AICs. The AIC results should then be a table in the manuscript, followed by results for the "winning" model. We thank the editor for this useful suggestion and now compared the respective models using AIC.psem from the piecewise SEM package. Accordingly, in the methods we now explained that we compared the AICs of each model when alternative model structures were tested (see lines: 304-311, 375-379). Specifically, we started with constructing a priori models (separately for richness and diversity in spring and summer) based on our hypothetical framework introduced in our paper. If alternative pathways improved model fit, we now additionally compared competing models based on the AIC. This was only done in summer, as there, an error correlation between biomass production and soil moisture improved overall model fit. Hence, we reported these AIC results in our manuscript. However, as model fit of our a priori developed spring models was already good, we did not compare any competing models with the spring models, in line with the philosophy that ideally, only a priori developed models in SEMs are shown, unless model fit suggests that initial models are not adequate yet (see lines: 306-311). Fig. 2 revealed little to me. I recommend the authors instead show and discuss a table of GLM output from lme4: are coefficients for biomass removal and fertilizer addition clearly non-zero? What are the signs? What is the overall fit? We agree, and now put Figure 2 of the original manuscript into our supplement, and additionally provided a table showing the standardized effect sizes of the treatment effects on species richness in the main manuscript (see Table 1, lines 335-341). In line with the suggestions from the editor, this table now more clearly shows the sign, confidence intervals and overall fit for each model. Speaking of regression output, S1 Table is close to what I sought, but I think it reveals errors that matter: • Why is is Spring 2020 different from other time steps in its model structure? The statistical output for the model testing the treatment effect on species richness in spring 2020 is different compared to the other seasons or years, as the most parsimonious model (based on AIC) did not contain an interaction between treatment and the respective regions. To emphasise this fact, we now provided some additional information in the table caption (see lines in supplement: 113-114). • The intercept (default by R) is Alb but also +F + R. It may make more sense to choose (using relevel in R) to use Alb but -F -R for a control condition as the intercept. While we agree that from a theoretical point of view, -F-R might be a more reasonable default, we deliberately choose +F+R as default, as in our experimental design the “unmanipulated” plots were +F+R (where land-use intensity reflected the local management by the respective farmer). Hence, we would prefer to keep +F+R as default in our models, unless the editor wishes otherwise. Also I suggest converting SE to 95% confidence intervals (1.96 X SE) and omitting the DF and t value columns. And note that rows are relative to the intercept. We agree and now provided the 95% CI in all our tables, omitted the DF and t-values, and added the note that rows are relative to the intercept. Finally, the study was designed as if site and time could be random effects, though with too few levels of each to really help as random effects. That makes all the sites and times show as separate estimates. The minimum number of levels needed for random effects in a mixed effect model may not be met here (that threshold is not clear), but I suggest trying a GLMM instead of the GLM. That would potentially move sites and time into random effects, leaving main effects of interest as fixed effects and making interpretation more straight-forward. I may be wrong - it may not work because sample size is limiting. If so, please just say so in your response to review comments. But - if it works, it may be more efficient and effective as a test of treatment effects. We fully agree that a GLMM with regions and year as random effects would help to correct our estimates. However, unfortunately, our study design lacks the required number of levels (region only 3, year only 2) for these variables to be used as random effects (rough threshold of minimal 5 levels, see Chapter 11, Section 5 in “Data Analysis Using Regression and Mulitilevel/Hierarchical models” by Gelman and Hill, 2007). Hence, while appreciating the suggestion by the editor, a GLMM with random effects would unfortunately not improve our current modelling approach. Line 313. This sentence could be edited to be more direct. We agree, and reformulated and restructured this sentence to be more direct (see line: 326-329). Beyond that, this section 4.1 based on regression could be argued (above?) as seeking overall (i.e., direct and indirect) effects of treatments on diversity, whereas SEM seeks to tease apart the direct from the indirect. Thus regression evaluates overall / net effects while SEM evaluates details. Perhaps that can help reconcile the greater information in Fig. 4 than in Fig. 2. And by the way, the use of SEM is an advance to make clear for this research subject – much research has not used it but work here shows it helps bring clarity. We fully agree and appreciate the editors recognition of our use on more “advanced” statistical methods, such as SEM, in addition to “conventional” statistics, to distinguish between direct and indirect treatment effects. According to the editors suggestions, we now renamed subheadings in the methods section and in the results section, to clarify how the different statistical approaches used in our study answer different questions. We hope that this will also help the reader to reconcile the greater information in Fig. 2. Line 417. I don’t know what is meant by “land use reductions.” We reformulated the sentence to increase clarity (see line: 431-433). Paragraph starting at Line 425 and subsequent paragraphs. It is usually expected that Figures are not cited in Discussion, to ensure that Results are clearly separate from Discussion. If it feels like Figures need to be used, then the sentence may belong better in Results. We agree with the suggestion of the editor and removed all references to Figures or Tables from the discussion. Line 438. Yes – I agree; time lags are often observed, and some citations to support this would be useful here. We agree and now provided additional references on lag effects found in previous studies (see line: 454). Paragraph starting at Line 448. This paragraph covers almost 3 pages! Organize and subdivide. We apologize for the inconvenient format of this paragraph and re-organized the respective section into three separate paragraphs (see re-organized paragraphs in lines: 464-538). Line 522. Careful! Not only is “extensification” used here again, but I think the authors need to be careful to avoid claiming a first. Consult the rich literature by Tilman and a small army of others at Cedar Creek. In addition, other studies that used a broader range of fertilizer and biomass removal have explored hump-shaped patterns related to the Intermediate Disturbance Hypothesis, Intermediate Productivity Hypothesis, and Huston’s combination of those two ideas (his Dynamic Equilibrium Model). There is a long history of inquiry here that includes factors here as part of the story, as the co-authors well know (they have been involved in some). Among others, a few that may be relevant include: • Feng-Wei Xu, Jian-Jun Li, Li-Ji Wu, Xiao-Ming Lu, Wen Xing, Di-Ma Chen, Biao Zhu, Shao-Peng Wang, Lin Jiang, Yong-Fei Bai, Resource enrichment combined with biomass removal maintains plant diversity and community stability in a long-term grazed grassland, Journal of Plant Ecology, 113(5), October 2020, Pages 611–620, https://doi.org/10.1093/jpe/rtaa046 • Band, N., Kadmon, R., Mandel, M. and DeMalach, N., 2022. Assessing the roles of nitrogen, biomass, and niche dimensionality as drivers of species loss in grassland communities. Proceedings of the National Academy of Sciences, 119(10), p.e2112010119. • Carnus, T., Connolly, J., Kirwan, L. and Finn, J.A., 2013. Plant diversity effects are robust to cutting severity and nitrogen application in productive grasslands. In The role of grasslands in a green future: threats and perspectives in less favoured areas. Proceedings of the 17th Symposium of the European Grassland Federation, Akureyri, Iceland, 23-26 June 2013 (pp. 195-197). Agricultural University of Iceland. We thank the editor for this important comment and apologize for the confusion. We fully agree, that there are many previous studies, investigating the effect of different land-use drivers on plant biodiversity. To clarify this, we now rephrased the respective sections in our discussion (see lines: 539-543), emphasizing that we build on knowledge of previous studies. Additionally, we now specifically refer to our experimental design (high land-use intensity as a default in comparison to treatments in which the intensity of certain land-use drivers was reduced) to emphasize the novelty of our study compared to previous research (see lines: 541-543). Finally, we very much appreciated the useful references provided and referred to them in our manuscript. I think the concluding paragraph could benefit from a sentence or two on limits of the study – as difficult as it is to get this much data, it still has limits in the number places and times, for strong analyses and to capture acknowledged potential lag effects. We fully agree, and now provided some additional information on the limitations of our study in the discussion section and highlighted avenues for future research (see lines: 557-558, 560-563). Reviewer #1: Reviewer #1: The authors present results on plant biodiversity, community composition and plant biomass from an in-field experiment in which the land-use components fertilization and biomass removal have been reduced in a full-factorial design. The results represent the short-term effects one and two years after the experiments have been implemented. Although effects on the plant community are very small overall, the study does provide interesting insights into possible mechanistic pathways as the authors also included light availability and soil moisture in their models. Both the experimental design and the statistical analyses are done rigorously and correctly, hence the conclusions are supported by the data. We thank the reviewer for the positive feedback on our study. Regarding the specific comments by the reviewer, we refer to our responses to the more detailed comments below. * The main strength of the experiment compared to previous studies is the implementation of all treatments within the same grassland sites. This could be emphasised more in the introduction. We thank the reviewer for this suggestion and now emphasized our experimental treatment within the introduction (see lines: 102-105). * The introduction lacks an explanation of the expected "seasonal variability of relationships" mentioned in the legend of Figure 1. As spring and summer data are analysed and discussed separately, a section on seasonal differences should be added in the introduction. We agree and now added two sections providing further rational on why we expect “seasonal variability of relationships” in our introduction (see lines: 66-69, 88-90). * The (indirect) effect of biomass production on diversity is mentioned in the hypotheses, but not explained and motivated in the Introduction. As biomass production is an important component in the model structure, the expected effects and mechanisms should be added to the introduction. We now clarified our motivation why indirect effects of fertilization are mediated by changes in biomass production in our introduction (see lines: 63-66). * The experiments were set up in grasslands of different land-use intensity. While the average intensity values are given in the Supplement, the description in the Methods section is inconsistent: In line 142f "a gradient in land-use intensity" is mentioned, while in line 163 management is described as "relatively high" and in line 163 as "medium to high". In addition to making the statements consistent, I recommend describing the range of land-use components on the grasslands by the actual number and time of cuts, amount of fertilizer and/or number of grazing animals (and not just in terms of the standardized land-use intensity). We apologize for the inconsistency when describing the land-use intensity gradient at the given grasslands and now provided information on the range of the intensity of each land-use driver in our methods section (see lines: 171-173). * The description of the treatments is in some parts difficult to read, especially if several treatments are mentioned in one sentence. I would prefer the use of abbreviations as done for the supplementary tables. The authors could also consider calling the treatment with fertilizer and biomass removal the control. We appreciate the useful suggestion to use the treatment abbreviations and adjusted the treatment names throughout the manuscript. Within the figures, we only provided the full treatment names, while in the caption, we referenced both the full and the abbreviated treatment name. This was done to guarantee the connection between abbreviated and full treatment names, to make sure readers understand our figures even if they did not read the full manuscript. * The terms subplot, plot and field are used inconsistently in the first part of the manuscript. See for instance line 176f. Please use consistent terminology and I would recommend not using "field" but rather "grassland". We fully agree and now clarified terminologies throughout the manuscript when referring to “plots” or “subplots”. Additionally, we now use “grassland” instead of “field” throughout the manuscript. Minor comments: Line 146: "the 10 most abundant" We agree and added quotation marks in line 156. Line 171: Why was the fertilizer manually applied in this treatment? Why was is not fertilized together with the "control" treatment and the remaining grassland? The +F-R and the -F-R subplots were both located in a plot where land-use was reduced. While in the +F+R, the farmer applied fertilizer using heavy machinery (fertilizer is sprayed over relatively large area), fertilizer was manually applied in the +F-R subplots, to avoid cross-contamination between treatments. We clarified this information in the methods section, see lines: 179-182. Line 173f: This information can be moved to the next section on field data collection We agree and deleted the respective text from the “Experimental design” section. Line 158: What does "setting-up" mean exactly? Does this mean that the plots were marked or was e.g. fertilization in autumn 2019 already reduced? Please clarify in which year/season the treatments were applied for the first time. With “setting-up” we refer to plots, that were marked and, in some cases (+F-R and -F-R) fenced in autumn 2019. Reduction in both biomass removal or fertilization was started from autumn 2019 onwards. We specified this information and provided further clarification in our method section (see lines: 167-168). Line 243 ff: This sentence contains the same information twice. We now rephrased the respective sentence (see lines: 248-251). Line 254: remove "with" We adjusted respective sentence. Line 257: which additional predictors are those? To test how species richness differed between treatments, we stepwise added “treatment” and “region”. The predictor “treatment” was always kept in the final model, irrespective of whether it was part of the most parsimonious model or not. We clarified this information in the methods section in line: 264-267, 269-271. Line 313 ff: there is no need to mention the type of treatment again in the parentheses# We agree and adjusted the respective sentences. Line 350f (and later): it is not quite clear what the "respectively" refers to. The path coefficients should be given directly after richness and diversity. We agree and adjusted the respective sentences (see line: 362-364). Figure 2: the y-axis label seems incorrect (remove the minus) Species richness was quantified as the number of vascular plant species within a 1x1 m subplot. Hence, we used the label “species richness (m-2)” to indicated that species richness is given per square meter. Attachment Submitted filename: Response to Reviewers.docx Click here for additional data file. 10.1371/journal.pone.0287039.r005 Decision Letter 2 Jenkins David G. Academic Editor © 2023 David G. Jenkins 2023 David G. Jenkins https://creativecommons.org/licenses/by/4.0/ This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Submission Version2 29 May 2023 Biomass removal promotes plant diversity after short-term de-intensification of managed grasslands PONE-D-22-34453R2 Dear Dr. Andraczek, We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements. Thanks again for your patience in the review process, and thank you for handling review comments so well. Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication. An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org. If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org. Kind regards, David G. Jenkins, PhD Academic Editor PLOS ONE 10.1371/journal.pone.0287039.r006 Acceptance letter Jenkins David G. Academic Editor © 2023 David G. Jenkins 2023 David G. Jenkins https://creativecommons.org/licenses/by/4.0/ This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. 8 Jun 2023 PONE-D-22-34453R2 Biomass removal promotes plant diversity after short-term de-intensification of managed grasslands Dear Dr. Andraczek: I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department. If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org. If we can help with anything else, please email us at plosone@plos.org. Thank you for submitting your work to PLOS ONE and supporting open access. Kind regards, PLOS ONE Editorial Office Staff on behalf of Dr. David G. Jenkins Academic Editor PLOS ONE ==== Refs References 1 Bengtsson J , Bullock JM , Egoh B , Everson C , Everson T , O’Connor T , et al . Grasslands-more important for ecosystem services than you might think. Ecosphere. 2019;10 :e02582. doi: 10.1002/ecs2.2582 2 Sala OE , Chapin FS , Armesto JJ , Berlow E , Bloomfield J , Dirzo R , et al . Global Biodiversity Scenarios for the Year 2100. American Association for the Advancement of Science. 2000:1770–4. doi: 10.1126/science.287.5459.1770 10710299 3 Laliberté E , Wells JA , Declerck F , Metcalfe DJ , Catterall CP , Queiroz C , et al . Land-use intensification reduces functional redundancy and response diversity in plant communities. Ecol Lett. 2010;13 :76–86. doi: 10.1111/j.1461-0248.2009.01403.x 19917052 4 IPBES-7. Summary for policymakers of the global assessment report on biodiversity and ecosystem services of the Intergovernmental Science-Policy Platform on Biodiversity and Ecosystem Services: Advanced Working Copy; 2019. 5 Blüthgen N , Dormann CF , Prati D , Klaus VH , Kleinebecker T , Hölzel N , et al . A quantitative index of land-use intensity in grasslands: Integrating mowing, grazing and fertilization. Basic and Applied Ecology. 2012;13 :207–20. doi: 10.1016/j.baae.2012.04.001 6 Socher SA , Prati D , Boch S , Müller J , Baumbach H , Gockel S , et al . Interacting effects of fertilization, mowing and grazing on plant species diversity of 1500 grasslands in Germany differ between regions. Basic and Applied Ecology. 2013;14 :126–36. doi: 10.1016/j.baae.2012.12.003 7 Minden V , Scherber C , Cebrián Piqueras MA , Trinogga J , Trenkamp A , Mantilla-Contreras J , et al . Consistent drivers of plant biodiversity across managed ecosystems. Philos Trans R Soc Lond, B, Biol Sci 2016. doi: 10.1098/rstb.2015.0284 27114585 8 Hautier Y , Niklaus PA , Hector A . Competition for light causes plant biodiversity loss after eutrophication. Science. 2009;324 :636–8. doi: 10.1126/science.1169640 19407202 9 Kleijn D , Kohler F , Báldi A , Batáry P , Concepción ED , Clough Y , et al . On the relationship between farmland biodiversity and land-use intensity in Europe. Proc Biol Sci. 2009;276 :903–9. doi: 10.1098/rspb.2008.1509 19019785 10 Tilman D. Resource Competition and Community Structure: Princeton Univ. Press; 1982. 11 Harpole WS , Sullivan LL , Lind EM , Firn J , Adler PB , Borer ET , et al . Addition of multiple limiting resources reduces grassland diversity. Nature. 2016;537 :93–6. doi: 10.1038/nature19324 27556951 12 Borer ET , Seabloom EW , Gruner DS , Harpole WS , Hillebrand H , Lind EM , et al . Herbivores and nutrients control grassland plant diversity via light limitation. Nature. 2014;508 :517–20. doi: 10.1038/nature13144 24670649 13 Grace JB , Anderson TM , Seabloom EW , Borer ET , Adler PB , Harpole WS , et al . Integrative modelling reveals mechanisms linking productivity and plant species richness. Nature. 2016;529 :390–3. doi: 10.1038/nature16524 26760203 14 Dickson TL , Foster BL . Fertilization decreases plant biodiversity even when light is not limiting. Ecol Lett. 2011;14 :380–8. doi: 10.1111/j.1461-0248.2011.01599.x 21332902 15 Foster BL , Gross KL . Species Richness in a Successional Grassland: Effects of Nitrogen Enrichment and Plant Litter. Ecology. 1998;8 :2593–602. 16 Berendse F. Implications of increased litter production for plant biodiversity. Trends in Ecology & Evolution. 1999;14 :4–5. doi: 10.1016/s0169-5347(98)01451-7 10234237 17 Li W , Wen S , Hu W , Du G . Root-shoot competition interactions cause diversity loss after fertilization: a field experiment in an alpine meadow on the Tibetan Plateau. Journal of Plant Ecology. 2011;4 :138–46. doi: 10.1093/jpe/rtq031 18 Kotowski W , Diggelen R . Light as an environmental filter in fen vegetation. Journal of Vegetation Science. 2004;15 :583–94. doi: 10.1111/j.1654-1103.2004.tb02300.x 19 Vojtech E , Turnbull LA , Hector A . Differences in light interception in grass monocultures predict short-term competitive outcomes under productive conditions. PLoS ONE. 2007;2 :e499. doi: 10.1371/journal.pone.0000499 17565366 20 DeMalach N , Zaady E , Kadmon R . Light asymmetry explains the effect of nutrient enrichment on grassland diversity. Ecol Lett. 2017;20 :60–9. doi: 10.1111/ele.12706 27933739 21 Parreño MA , Schmid B , Petchey OL . Comparative study of the most tested hypotheses on relationships between biodiversity, productivity, light and nutrients. Basic and Applied Ecology. 2021;53 :175–90. doi: 10.1016/j.baae.2021.03.012 22 Rose L , Leuschner C . The diversity–productivity relationship in a permanent temperate grassland: negative diversity effect, dominant influence of management regime. Plant Ecology & Diversity. 2012;5 :265–74. doi: 10.1080/17550874.2012.723763 23 Collins , Knapp , riggs , Blair , Steinauer . Modulation of diversity by grazing and mowing in native tallgrass prairie. Science. 1998;280 :745–7. doi: 10.1126/science.280.5364.745 9563952 24 Wang C , Long R , Wang Q , Liu W , Jing Z , Zhang L . Fertilization and litter effects on the functional group biomass, species diversity of plants, microbial biomass, and enzyme activity of two alpine meadow communities. Plant Soil. 2010;331 :377–89. doi: 10.1007/s11104-009-0259-8 25 Yang H , Jiang L , Li L , Li A , Wu M , Wan S . Diversity-dependent stability under mowing and nutrient addition: evidence from a 7-year grassland experiment. Ecol Lett. 2012;15 :619–26. doi: 10.1111/j.1461-0248.2012.01778.x 22487498 26 Hautier Y , Vojtech E , Hector A . The importance of competition for light depends on productivity and disturbance. Ecol Evol. 2018;8 :10655–61. doi: 10.1002/ece3.4403 30519395 27 Barker C , Power S , Bell J , Orme C . Effects of habitat management on heathland response to atmospheric nitrogen deposition. Biological conservation. 2004;120 :41–52. doi: 10.1016/j.biocon.2004.01.024 28 Jones L , Stevens C , Rowe EC , Payne R , Caporn S , Evans CD , et al . Can on-site management mitigate nitrogen deposition impacts in non-wooded habitats? Biological conservation. 2016;212 :464–75. doi: 10.1016/j.biocon.2016.06.012 29 Olff H , Ritchie ME . Effects of herbivores on grassland plant diversity. Trends in Ecology & Evolution. 1998;13 :261–5. doi: 10.1016/s0169-5347(98)01364-0 21238294 30 Adler P , Raff D , Lauenroth W . The effect of grazing on the spatial heterogeneity of vegetation. Oecologia. 2001;128 :465–79. doi: 10.1007/s004420100737 28547391 31 Loucougaray G , Bonis A , Bouzillé J-B . Effects of grazing by horses and/or cattle on the diversity of coastal grasslands in western France. Biological conservation. 2004;116 :59–71. doi: 10.1016/S0006-3207(03)00177-0 32 Dickson TL , Foster BL . The relative importance of the species pool, productivity and disturbance in regulating grassland plant species richness: a field experiment. Journal of Ecology. 2008;96 :937–46. doi: 10.1111/j.1365-2745.2008.01420.x 33 Kladivová A , Münzbergová Z . Interacting effects of grazing and habitat conditions on seedling recruitment and establishment. J Veg Sci. 2016:834–43. 34 Stewart GB , Pullin AS . The relative importance of grazing stock type and grazing intensity for conservation of mesotrophic ‘old meadow’ pasture. Journal for Nature Conservation. 2008;16 :175–85. doi: 10.1016/j.jnc.2008.09.005 35 Teuber LM , Hölzel N , Fraser LH . Livestock grazing in intermountain depressional wetlands—Effects on plant strategies, soil characteristics and biomass. Agriculture, Ecosystems & Environment. 2013;175 :21–8. doi: 10.1016/j.agee.2013.04.017 36 Ryser P. Influence of gaps and neighbouring plants on seedling establishment in limestone grassland: experimental field studies in northern Switzerland: ETH Zurich; 1990. 37 Olff H , Pegtel DM , Van Groenendael JM , Bakker JP . Germination Strategies During Grassland Succession. J Ecol. 1994:69–77. 38 Yan X , Diez J , Huang K , Li S , Luo X , Xu X , et al . Beyond resource limitation: an expanded test of the niche dimension hypothesis for multiple types of niche axes. Oecologia. 2020;193 :689–99. doi: 10.1007/s00442-020-04713-w 32681295 39 Grace JB , Pugesek BH . A Structural Equation Model of Plant Species Richness and Its Application to a Coastal Wetland. The American Naturalist. 1997:436–60. 40 Schmitzberger I , Wrbka T , Steurer B , Aschenbrenner G , Peterseil J , Zechmeister HG . How farming styles influence biodiversity maintenance in Austrian agricultural landscapes. Agriculture, Ecosystems & Environment. 2005;108 :274–90. doi: 10.1016/j.agee.2005.02.009 41 Molina CD , Tognetti PM , Graff P , Chaneton EJ . Mowing does not redress the negative effect of nutrient addition on alpha and beta diversity in a temperate grassland. J Ecol. 2021;109 :1501–10. doi: 10.1111/1365-2745.13573 42 Schaffers AP . Soil, biomass, and management of semi-natural vegetation: Part II. Factors controlling species diversity. Plant Ecol. 2002:247–268. 43 Lamb EG . Direct and indirect control of grassland community structure by litter, resources, and biomass. Ecology. 2008;89 :216–25. doi: 10.1890/07-0393.1 18376563 44 Fischer M , Bossdorf O , Gockel S , Hänsel F , Hemp A , Hessenmöller D , et al . Implementing large-scale and long-term functional biodiversity research: The Biodiversity Exploratories. Basic and Applied Ecology. 2010;11 :473–85. doi: 10.1016/j.baae.2010.07.009 45 Socher SA , Prati D , Boch S , Müller J , Klaus VH , Hölzel N , Fischer M . Direct and productivity-mediated indirect effects of fertilization, mowing and grazing on grassland species richness. J Ecol. 2012;100 :1391–9. doi: 10.1111/j.1365-2745.2012.02020.x 46 Ellenberg H. Vegetation Mitteleuropas mit den Alpen in ökologischer, dynamischer und historischer Sicht. Stuttgart: Ulmer Verlag; 1996. 47 López-Díaz JE , Roca-Fernández AI , González-Rodríguez A . Measuring Herbage Mass by Non-Destructive Methods: A Review. Journal of Agricultural Science and Technology. 2011:303–14. 48 Vogt J , Klaus VH , Both S , Fürstenau C , Gockel S , Gossner MM , et al . Eleven years’ data of grassland management in Germany. Biodivers Data J. 2019;7 :e36387. doi: 10.3897/BDJ.7.e36387 31598068 49 Ostrowski A, Lorenzen K, Petzold E, Schindler S. Land use intensity index (LUI) calculation tool of the Biodiversity Exploratories project for grassland survey data from three different regions in Germany since 2006, BEXIS 2 module. zenodo. 2020. 50 Cayuela L, Macarro I, Stein A, Oksanen J. Taxonstand: Taxonomic Standardization of Plant Species Names. R package version 2.2. 2019. 51 Magurran AE . Measuring Biological Diversity. Oxford: Blackwell Publishing; 2004. 52 Jost L . Entropyand diversity. OIKOS. 2006;113 :363–75. doi: 10.1111/j.2006.0030-1299.14714.x 53 Riehl G. Ermittlung von Erträgen auf dem Grünland. Sächsisches Landesamt für Umwelt, Landwirtschaft und Geologie,. 2001. 54 R Core Team. R: A language and environment for statistical computing: R Foundation for Statistical Computing. Vienna, Austria; 2021. 55 Bates D , Maechler M , Bolker B , Walker S . Fitting Linear Mixed-Effects Models Using lme4. Journal of Statistical Software. 2015;67 :1–48. doi: 10.18637/jss.v067.i01 56 Kuznetsova A , Brockhoff P , Christensen R . lmerTest Package: Tests in Linear Mixed Effects Models. Journal of Statistical Software. 2017;82 :1–26. doi: 10.18637/jss.v082.i13 57 Hothorn T , Bretz F , Westfall P . Simultaneous Inference in General Parametric Models. Biometrical Journal. 2008;50 :346–63. doi: 10.1002/bimj.200810425 18481363 58 Fox J , Weisberg S . An {R} Companion to Applied Regression, Third Edition. Thousand Oaks CA: Sage. 2019. 59 Barton K. MuMIn: Multi-Model Inference. R package version 1.43.17. 2020. 60 Lefcheck JS . piecewiseSEM: Piecewise structural equation modelling in r for ecology, evolution, and systematics. Methods Ecol Evol. 2016;7 :573–9. doi: 10.1111/2041-210X.12512 61 Lenth RV. Estimated Marginal Means, aka Least-Squares Means: R package version 1.7.0. 2021. 62 Zuur AF , Ieno EN , Elphick CS . A protocol for data exploration to avoid common statistical problems. Methods in Ecology and Evolution. 2010;1 :3–14. doi: 10.1111/j.2041-210X.2009.00001.x 63 Oksanen J, Blanchet FG, Friendly M, Kindt R, Legendre P, McGlinn D, et al. vegan: Community Ecology Package. R package version 2.5–7. 2020. 64 Arbizu PM. pairwiseAdonis: Pairwise Multilevel Comparison using Adonis. R package version 0.4. 2017. 65 Pinheiro J, Bates D, DebRoy S, Sarkar D, R Core Team. nlme: Linear and Nonlinear Mixed Effects Models. 2021. 66 Yang G-J , Lü X-T , Stevens CJ , Zhang G-M , Wang H-Y , Wang Z-W , et al . Mowing mitigates the negative impacts of N addition on plant species diversity. Oecologia. 2019;189 :769–79. doi: 10.1007/s00442-019-04353-9 30725373 67 Dullinger S , Essl F , Rabitsch W , Erb K-H , Gingrich S , Haberl H , et al . Europe’s other debt crisis caused by the long legacy of future extinctions. Proc Natl Acad Sci U S A. 2013;110 :7342–7. doi: 10.1073/pnas.1216303110 23589873 68 Faust C , Storm C , Schwabe A . Shifts in plant community structure of a threatened sandy grassland over a 9-yr period under experimentally induced nutrient regimes: is there a lag phase? J Veg Sci. 2012;23 :372–86. doi: 10.1111/j.1654-1103.2011.01355.x 69 Olff H , Bakker JP . Long-Term Dynamics of Standing Crop and Species Composition After the Cessation of Fertilizer Application to Mown Grassland. Journal of applied ecology. 1991;28 :1040–52. 70 Grace JB , Adler PB , Stanley Harpole W , Borer ET , Seabloom EW . Causal networks clarify productivity-richness interrelations, bivariate plots do not. Funct Ecol. 2014;28 :787–98. doi: 10.1111/1365-2435.12269 71 Altesor A , Oesterheld M , Leoni E , Lezama F , Rodríguez C . Effect of grazing on community structure and productivity of a Uruguayan grassland. Plant Ecol. 2005;179 :83–91. doi: 10.1007/s11258-004-5800-5 72 Klaus VH , Schäfer D , Kleinebecker T , Fischer M , Prati D , Hölzel N . Enriching plant diversity in grasslands by large-scale experimental sward disturbance and seed addition along gradients of land-use intensity. JPECOL. 2017:rtw062. doi: 10.1093/jpe/rtw062 73 Jutila H , Grace JB . Effects of disturbance on germination and seedling establishment in a coastal prairie grassland: A test of the competitive release hypothesis. Journal of Ecology. 2002. 74 Jackson ST , Sax DF . Balancing biodiversity in a changing environment: extinction debt, immigration credit and species turnover. Trends in Ecology & Evolution. 2010;25 :153–60. doi: 10.1016/j.tree.2009.10.001 19879014 75 Dai L , Fu R , Guo X , Du Y , Zhang F , Cao G . Soil Moisture Variations in Response to Precipitation Across Different Vegetation Types on the Northeastern Qinghai-Tibet Plateau. Front Plant Sci. 2022;13 :854152. doi: 10.3389/fpls.2022.854152 35463396 76 Wang Le , Xu H , Zhang H , Zhang Y . Grazing and Mowing Affect the Carbon-to-Nitrogen Ratio of Plants by Changing the Soil Available Nitrogen Content and Soil Moisture on the Meadow Steppe, China. Plants (Basel) 2022. doi: 10.3390/plants11030286 35161265 77 Rabinowitz D , Rapp JK . Colonization and Establishment of Missouri Prairie Plants on Artificial Soil Disturbances. II. Detecting Small-Scale Plant-to-Plant Interactions and Separating Disturbance from Resource Provision. Am J Bot. 1985:1629–34. 78 Gibson DJ , Urban J , Baer SG . Mowing and fertilizer effects on seedling establishment in a successional old field. Journal of Plant Ecology. 2011;4 :157–68. doi: 10.1093/jpe/rtr020 79 Tälle M , Deák B , Poschlod P , Valkó O , Westerberg L , Milberg P . Grazing vs. mowing: A meta-analysis of biodiversity benefits for grassland management. Agriculture, Ecosystems & Environment. 2016;222 :200–12. doi: 10.1016/j.agee.2016.02.008 80 Band N , Kadmon R , Mandel M , DeMalach N . Assessing the roles of nitrogen, biomass, and niche dimensionality as drivers of species loss in grassland communities. Proc Natl Acad Sci U S A. 2022;119 :e2112010119. doi: 10.1073/pnas.2112010119 35235460 81 Carnus T, Connolly J, Kirwan L, Finn JA. Plant diversity effects are robust to cutting severity and nitrogen application in productive grasslands: In: The role of grasslands in a green future: threats and perspectives in less favoured areas. Grassland Science in Europe. 2013:195–7. 82 Xu F-W , Li J-J , Wu L-J , Lu X-M , Xing W , Chen D-M , et al . Resource enrichment combined with biomass removal maintains plant diversity and community stability in a long-term grazed grassland. Journal of Plant Ecology. 2020;13 :611–20. doi: 10.1093/jpe/rtaa046 83 Wilson SD , Tilman D . Component of Plant Competition Along an Experimental Gradient of Nitrogen Availability. Ecology. 1991;72 :1050–65. 84 IPCC, editor. Climate change 2014. Impacts, adaption, and vulnerability.: Part A: Global and sectoral aspects. Contribution of working group II to the Fifth assessment report of the intergovernmental panel on climate change.: Cambridge University Press; 2014.