Plant biomass and macro‐ and micronutrients jointly drive top‐heavy arthropod trophic pyramids in grassland monocultures

Abstract

Identifying the key drivers of changes in species biomass, abundance, and distribution across trophic levels in terrestrial ecosystems represents a fundamental challenge in ecology. The shape of trophic pyramids—reflecting the relative biomass or abundance of various trophic groups, such as herbivores, omnivores, predators, and parasitoids—is anticipated to vary with both the quantity and nutritional quality of plants. However, a comprehensive understanding of how these factors influence the shape of trophic pyramids in grassland ecosystems remains elusive. In this study, we utilize replicated monocultures of 15 herbaceous species in the Inner Mongolian grassland to investigate the effects of plant biomass, macronutrients (including protein, carbohydrate, and phosphorus), and micronutrients (such as sodium) on the structure of arthropod trophic pyramids and the abundances within different trophic levels. Our results demonstrated that plant biomass, leaf protein‐to‐carbohydrate ratio, and sodium content collectively contributed to a top‐heavy structure in arthropod trophic pyramids, characterized by a relatively higher proportion of predators and parasitoids. Specifically, plant biomass enhanced this top‐heaviness both directly, by increasing the abundance of predators and parasitoids, and indirectly, by increasing chewing and sucking herbivores, which in turn bolstered the abundance of predators and parasitoids. Furthermore, leaf sodium and the protein‐to‐carbohydrate ratio positively influenced parasitoid abundance through increasing the abundance of sucking herbivores and endophytes. Given that climate change and human activities, such as nitrogen fertilization and saline water irrigation, are altering plant biomass, nutrient composition, and sodium availability globally, our findings suggest that these changes will have significant cascading effects on arthropod trophic structures and overall ecosystem functioning.

INTRODUCTION

Understanding the mechanisms that control the species abundance and distribution across trophic levels is a fundamental goal in ecology (Oksanen et al., 1981; Power et al., 1992). Classical views suggest that biomass and abundance distributions across trophic levels are arranged in a pyramid, in which the proportion of biomass and abundance decreases with increasing trophic level (Elton & Nicholson, 1942; Lindeman, 1942). However, diverse biotic communities (e.g., soil nematodes, terrestrial arthropods, and marine fishes) can exhibit highly variable trophic distribution patterns, in part, as a result of variation in biomass and nutritional quality of primary producers, which determine nutrient cycling rates and the overall flow of energy through the ecosystem (DuPont et al., 2009; McCauley et al., 2018; Reuman et al., 2008). A primary driver of abundance across consumer trophic levels is the variation in the biomass of primary producers, which controls the quantity of available food and habitat structure (Chase et al., 2000; Wimp et al., 2010). In addition, trophic pyramid shape may vary with the concentrations of macronutrients (e.g., protein, carbohydrate, and phosphorus) and micronutrients (e.g., sodium) provided by primary producers (Behmer, 2009; Kaspari, 2021). A comprehensive understanding of how primary producer biomass, macro‐ and micronutrients affect the shape of trophic pyramid is lacking, particularly in grassland ecosystems. In grasslands, arthropods, the important components of biodiversity at multiple trophic levels (e.g., herbivores, omnivores, predators, and parasitoids), are an ideal group to quantify the response of trophic pyramids to plant production and tissue quality (Post, 2002; Welti et al., 2020).

Plant biomass is a key regulator of arthropod communities (Siemann, 1998; Wimp et al., 2010). Two opposing hypotheses have been posited regarding the effects of plant biomass on the shape of arthropod trophic pyramids (Figure 1B, Table 1). Hypothesis I‐1 predicts that populations at high trophic levels, such as predators and parasitoids, can benefit from an increase in habitat volume and/or herbivore abundance under high plant biomass conditions, leading to top‐heavy arthropod trophic pyramids (Post et al., 2000; Prather et al., 2021). Higher grassland plant biomass provides more structure for spiders to hang their webs and more flowers to feed predacious and parasitoid wasps (Figure 1C; pathways a and c), increasing the likelihood that herbivore abundance is constrained by top‐down effects (Gómez et al., 2016; Oksanen et al., 1981). Increasing plant biomass can increase food availability for herbivores and omnivores, which indirectly supports more predators and parasitoids (Figure 1C; pathways b, d, i, j, k, l, and m) due to increases in prey for high trophic levels of arthropods (Chase et al., 2000; Siemann, 1998; Wimp et al., 2010). Hypothesis I‐2, in contrast, predicts that increasing plant biomass will support higher proportions of herbivores (Figure 1C; pathway b), such as sucking herbivores and endophytes with high plant food demands (Awmack & Leather, 2002; Lu et al., 2021), increasing the bottom‐heaviness of arthropod trophic pyramids (Hatton et al., 2015; Welti et al., 2020).

FIGURE 1.

Study design and conceptual flow diagrams. (A) The common garden experiment that contains monocultures of 20 plant species in the inner Mongolian grassland, with a breadth of natural variation in plant biomass, plant macronutrients (leaf carbohydrate, protein, phosphorus, and leaf protein:carbohydrate ratio), and leaf sodium concentrations. (B) Conceptual diagram illustrating multiple hypotheses associated with effects of plant biomass, plant macronutrients, and plant sodium on abundance pyramids of arthropod communities (Table 1). On the left side, herbivores and omnivores (light green and purple boxes) dominate the system, while on the right side, predators and parasitoids (light orange and red boxes) dominate the system. Increases in plant biomass, leaf macronutrients, and leaf sodium can drive bottom‐heavy to top‐heavy (Hypothesis I‐1, Hypothesis II‐1, and Hypothesis III‐1), or top‐heavy to bottom‐heavy abundance pyramids of arthropod communities (Hypothesis I‐2, Hypothesis II‐2, and Hypothesis III‐2). (C) A priori conceptual model of links between plant biomass, plant macronutrients and micronutrients and arthropod abundance of multiple trophic levels. Different hypothesized pathways are indicated in diagram by different letters, and are discussed in the main text. Photo in panel A and arthropod illustrations in panel B and C credits: Xiaoming Lu.

TABLE 1.

Proposed hypotheses predicting the regulation of the trophic pyramids of arthropod abundance in grasslands.

Hypothesis	No.	Prediction
Plant biomass	I‐1	Plant biomass can promote abundance of predators and parasitoids by increasing herbivore abundance or habitat volume, potentially leading to top‐heavy arthropod trophic pyramids1,2
	I‐2	Increases in plant biomass, through increasing the resource and energy base, benefit dominant herbivorous feeding groups and thus promote bottom‐heavy arthropod trophic pyramids3,4
Plant macronutrient	II‐1	Plant macronutrient concentrations and stoichiometry may increase herbivore and thus predator abundance through trophic interactions, increasing top‐heavy abundance pyramids5–7
	II‐2	Plant macronutrients concentrations increase in abundance of herbivores than predators and parasitoids, promoting bottom‐heaviness of trophic pyramid8,9
Plant sodium	III‐1	Increases in leaf sodium promote herbivores abundance and thus support more parasitoids, thereby increasing the trophic pyramid's top‐heaviness10,11
	III‐2	Increases in leaf sodium are more effective at attracting plant consumers (herbivores and omnivores) than predators and parasitoids, leading to bottom‐heavy trophic pyramids12,13

Note: References are given by numbers in table: ¹Siemann (1998), ²Post et al. (2000), ³Prather et al. (2020), ⁴Hatton et al. (2015), ⁵Awmack and Leather (2002), ⁶Lind et al. (2017), ⁷McCauley et al. (2018), ⁸Joern et al. (2012), ⁹Prather et al. (2021), ¹⁰Welti et al. (2020), ¹¹Kaspari et al. (2017), ¹²Simpson et al. (2006) and ¹³Clay et al. (2014).

Plant macronutrients (e.g., carbohydrates, proteins, and phosphorus) are essential for the growth, reproduction, and survival of arthropods, with strong effects on the abundance and structure of arthropod communities (Joern et al., 2012; Kaspari, 2021). As most arthropods need to maintain homeostasis with respect to their elemental composition, macronutrients are needed in optimal levels and ratios for arthropod growth and reproduction (Boersma & Elser, 2006; Cease et al., 2012). Two hypotheses predict opposing effects of plant macronutrients on the trophic pyramid shape of arthropod communities through bottom‐up controls (Figure 1B, Table 1). Hypothesis II‐1 predicts that macronutrient concentrations and stoichiometry (e.g., protein‐to‐carbohydrate ratios) can increase herbivore abundance, which in turn supports more predators and parasitoids through food webs (Figure 1C; pathways e, j, k, m), increasing the top‐heaviness of trophic pyramids (Lind et al., 2017; McCauley et al., 2018). However, these bottom‐up effects may depend on specific interactions between different herbivorous feeding groups and higher trophic levels (Awmack & Leather, 2002; Behmer, 2009). For instance, an increase in foliar protein concentrations can increase the abundance of certain parasitoid species indirectly by supporting sucking herbivores, which can quickly access nutrients from leaf and phloem sap using their piercing‐sucking mouthparts (Awmack & Leather, 2002). In contrast, Hypothesis II‐2 predicts that increases in plant macronutrients (e.g., proteins and phosphorus) tend to yield larger increases in the abundance of herbivores compared to predators (Figure 1C; pathway e), hence increasing the bottom‐heaviness of the trophic pyramid (Prather et al., 2021; Welti et al., 2020). The amount of macronutrients found in plant tissue is highly variable across different plant species with expected strong effects on herbivores, but not necessarily on predators as their food (meat) is less variable in quality (Awmack & Leather, 2002, Behmer, 2009).

Plant sodium (Na) is a critical micronutrient for arthropods to maintain neural function, osmotic balance, reproduction, and development (Kaspari, 2021; Welti et al., 2020). Variation in sodium concentrations among different plant species can shift the abundance of different arthropod groups and modify the shape of trophic pyramids (Kaspari et al., 2021). Two hypotheses predict the contrasting roles of plant sodium in shaping trophic pyramids (Figure 1B, Table 1). Hypothesis III‐1 predicts that plant sodium promotes the abundance of plant consumers, which in turn supports more trophic levels in arthropod food webs (Figure 1C; pathway g, h, i, j, k, l, and m), leading to increased top‐heaviness of trophic pyramids (Kaspari et al., 2017; Welti et al., 2020). Hypothesis III‐2, in contrast, predicts that plant sodium is more effective at attracting plant consumers than predators and parasitoids (Figure 1C; pathway g and h), leading to high proportions of herbivores and omnivores and thus bottom‐heavy trophic pyramids (Clay et al., 2014; Simpson et al., 2006). Herbivores are vulnerable to sodium limitation as a result of the imbalances between the sodium concentrations required by animals and those found in plants, which generally do not require sodium (Welti & Kaspari, 2021). Predators likely do not suffer from sodium deficits as their diet is already rich in sodium (Kaspari, 2021).

In spite of the recognized importance of plant biomass, macro‐ and micronutrients effects on arthropod trophic structure, few studies have systematically tested the contrasting hypotheses and predictions (Welti et al., 2020). First, while strong effects of protein, carbohydrates, and phosphorus on arthropod abundance is a common finding (Awmack & Leather, 2002; Joern et al., 2012; Le Gall et al., 2020), the roles of plant micronutrients, such as sodium, can also drive arthropod abundance but are often ignored (Kaspari et al., 2017; Welti & Kaspari, 2021). Second, studies examining the effects of plant macro‐ and micronutrients generally include only a subset of trophic levels, such as herbivores or predators (Joern et al., 2012; Le Gall et al., 2020; Peterson et al., 2021). However, arthropod trophic groups (e.g., herbivores, omnivores, predators, and parasitoids) may show different responses to changes in plant macro‐ and micronutrients availabilities (Clay et al., 2014; Seibold et al., 2018; Welti & Kaspari, 2021). Third, although a few studies have attempted to elucidate the effects of plant biomass and nutrients on the abundance of trophic groups, they did not break herbivores into different guilds according to their feeding modes (Clay et al., 2014; Welti et al., 2020). The specific interactions between plants and herbivorous feeding guilds and between herbivorous feeding guilds and predators or parasitoids could underlie variation in plant‐arthropod trophic pyramid shape.

To test our hypotheses regarding the effects of plant biomass and plant macro‐ and micronutrients on the shape of arthropod trophic pyramids (Figure 1B, Table 1), we used replicated monocultures of 15 herbaceous species in the Inner Mongolian grassland, with a wide range of natural variations in plant biomass and macro‐ and micronutrients under similar environmental contexts (Lu et al., 2021). We sampled arthropods in all plots and determined the taxon abundance within trophic levels and functional feeding guilds. We calculated community trophic mean (CTM) of trophic level as an indicator of arthropod trophic pyramid shape by using arthropod trophic level weighted by abundance of arthropods within each trophic level (Welti et al., 2020). Specifically, we address two questions: first, how do plant biomass, macronutrients (e.g., protein, carbohydrates, phosphorus, and protein‐to‐carbohydrates ratio), and sodium affect the CTM of arthropod trophic level and abundance within different arthropod trophic groups across monocultures of different plant species? Second, what are the mechanisms underlying the effects of plant biomass, macronutrients, and sodium on CTM of arthropod trophic level?

MATERIALS AND METHODS

Study site

The monoculture experiment was established in June 2014 at the Inner Mongolia Grassland Ecosystem Research Station (IMGERS, 116°42′ E, 43°38′ N) of the Chinese Academy of Sciences, located in the Xilin River Basin of Inner Mongolia, China (Bai et al., 2004). The mean annual temperature of the study area is 0.3°C, with the lowest and highest monthly mean temperatures occurring in January and July, respectively. The mean annual precipitation is 346 mm, with more than 80% of the annual rainfall falling in the growing season from April to September. The plant community is dominated by Leymus chinensis (perennial rhizome grass) and Stipa grandis (perennial bunchgrass), both of which are widely distributed in the Eurasian steppe (Bai et al., 2004). The soil texture of the study area is loamy sand (Calcic Chernozem according to the ISSS Working Group RB, 1998).

To remove plants and seed prior to experiment establishment, a 40 × 40 m area was bulldozed and manually plowed, removing 10 cm of top soil. The area was then divided into four blocks, with each block containing 20 plots (each 1.2 × 1.2 m), and the entire area was fenced (Figure 1A). Within each block, the 20 most common native species in the regional plant community were established in monocultures within randomly assigned plots. Seeds were sourced locally, being collected by hand in 2013. Of the initial 20 plant species, only 15 established within plots (Appendix S1: Table S1; 60 plots). For more information on seeding and experimental design parameters, see Lu et al. (2021).

Plant biomass and leaf nutrients measurements

In mid‐August 2018 and 2019, plant biomass was sampled by clipping all plants in a 50 × 50 cm quadrat at the soil surface within each plot. All plants were oven‐dried at 65°C for 48 h and weighed to measure dry plant biomass (in grams per square meter) of each plot (Bai et al., 2004). Leaf concentrations of nonstructural carbohydrates, protein, phosphorus, and sodium were measured from five leaves each from three individual plants/plot, as these are important determinants of plant nutritional quality for arthropods (Awmack & Leather, 2002; Joern et al., 2012; Kaspari, 2021). Leaves were dried for 24 h at 65°C, ground using a ball mill prior to chemical analysis. Leaf phosphorus and sodium concentrations were measured using an elemental analyzer (VarioEL Element Analyzer; Hanau, Germany), protein content was measured in the Bradford assay, and nonstructural carbohydrates (used as carbohydrates in the leaf protein‐to‐carbohydrate ratio) were measured using the phenol–sulfuric acid method (Clissold et al., 2006). Leaf nutrient concentrations were averaged within a plot and the plot was used as the unit of replication prior to data analysis. Monocultures of different plant species covered a wide range of plant biomass, macro‐ and micronutrient variations (Appendix S1: Table S1).

Arthropod sampling and identification

Arthropods were collected from the monoculture plots using a sweep net (diameter 32.0 cm) in mid‐August 2018 and 2019. Sampling took place between 10:00 am and 4:00 pm on rain‐free days. We conducted 50 sweeps per plot (25 sweeps per sampling period, with two sampling periods in August). Each sweep involved a 180° arc through the vegetation canopy, followed by a quick turn and reverse to capture vegetation‐dwelling arthropods. We varied the sweep direction to ensure comprehensive coverage (Appendix S1: Figure S1). Directions A and B consisted of eight sweeps each, as the 32.0 cm‐diameter net effectively covered the plot width. Direction C involved nine sweeps to account for the longer diagonal distance. The contents of the sweep net were put in bottles containing ethyl acetate until sorting. In the laboratory, all arthropod individuals were identified to species by optical microscopy, or to morphospecies identified to genus or family. Based on personal observations and a literature review (Carmona et al., 2011; Lu et al., 2022), each morphospecies was assigned to one of four trophic categories: herbivores, omnivores, predators, and parasitoids (Appendix S1: Table S2). Herbivores were further assigned to one of the three feeding guilds: sucking/piercing herbivore, chewing herbivore, and endophytes (Carmona et al., 2011, Lu et al., 2022), based on whether specimens had chewing mouthparts, sucking mouthparts, or feed within tissues, respectively (Appendix S1: Table S2). However, sweep net sampling may be less effective for capturing ground‐active omnivores and predators, such as ants and beetles, as well as endophytic insects that inhabit plant tissues. We further quantified the abundance of herbivores, omnivores, predators, and parasitoids across the monoculture plots of different plant species (Appendix S1: Table S3).

Community trophic mean

We calculated the CTM of arthropod trophic positions for each monoculture plot using the following equation:

$$\mathrm{CTW}=\frac{1}{P}\sum _{i=1}^N{p}_i\times t,$$ (1)

where P represents the total number of arthropod individuals, N is the number of trophic levels, p _i is the number of arthropods of trophic level i, and t is the trophic level value (Ricotta & Moretti, 2011; Welti et al., 2020).

We assigned a trophic level value of 2 to herbivores and of 2.5 to omnivores. Omnivores, such as ants, occupy a trophic position that is intermediate between herbivores and predators. This classification is supported by N isotopic analyses across a diverse range of grasslands, which indicate an average trophic position of 2.5 for omnivores (Welti et al., 2020). Predators and parasitoids were both assigned a trophic level of 3. While predatory arthropods may serve as hosts for parasitoid species (Fei et al., 2023), this interaction could suggest a higher the trophic level for parasitoids as a group. However, most parasitoids are specialized, primarily targeting herbivorous insects as their hosts (Frago & Zytynska, 2023; Kraaijeveld et al., 1998). In contrast, predators exhibit a more polyphagous feeding, strategy, preying on a wider variety of organisms, including herbivores, other predators, and parasitoids (Colfer & Rosenheim, 2001; Hurd & Eisenberg, 1990; Rosenheim, 1998; Rosenheim et al., 1995). Therefore, it is not justified to assign a higher trophic level to parasitoids than to predators. CTM of trophic level simplifies biological complexity and is a straightforward indicator of community trophic structure (Welti et al., 2020). A high CTM of trophic level indicates a top‐heavy trophic pyramid (e.g., relatively higher proportions of predators and parasitoids) while a low CTM of trophic level indicates a bottom‐heavy trophic pyramid (e.g., relatively higher proportions of herbivores and omnivores; Welti et al., 2020). Because the high mobility and feeding activity of chewing insects, such as grasshoppers, across different monoculture plots may influence CTM estimates, we also recalculated the CTM after excluding these insects.

Data analysis

We used linear mixed‐effects models (LMMs) to determine the responses of CTM values to changes in plant biomass, leaf macronutrients, and leaf sodium concentrations across the two sampling years. Prior to data analysis, plant biomass was log₁₀‐transformed and abundance of different trophic levels was natural log transformed to improve normality and homogeneity of variance. Following transformation, normality was tested using the Shapiro–Wilk goodness of fit test. In the LMMs, plant biomass, leaf carbohydrate concentrations, leaf protein concentrations, protein‐to‐carbohydrate ratio, leaf phosphorus concentrations, and leaf sodium concentrations were treated as fixed factors, and block nested within year was treated as a random factor. LMMs were conducted using the Program R package lme4 (Douglas Bates et al., 2015; R Core Team, 2023). We used the first‐ and second‐order polynomials to capture linear and nonlinear responses of CTM values to variations of plant biomass and macro‐ and micronutrients (Boersma & Elser, 2006). We compared the fit of the two models by using small sample size corrected Akaike information criterion (AICc) (Anderson & Burnham, 2002). When the AICc difference (ΔAICc) was <2 between two significant functions, the simpler model was chosen, where ‘simplicity’ increased from linear to quadratic. We further calculated the marginal and conditional R ² values for each fitted model using the MuMIn package (Barton, 2020), representing the variance explained by fixed factors and the variance explained by both fixed and random factors, respectively (Nakagawa & Schielzeth, 2013). To test the robustness of our findings, we conducted a sensitivity analysis by recalculating CTM after excluding chewing insects. The results were consistent with those from the full dataset, showing similar effects of plant variables on CTM in analyses with and without chewing insects (Appendix S1: Figure S2, Table S4). Thus, the inclusion of chewing insects does not change our conclusions about the effects of plant biomass and nutrient content on CTM.

We used generalized additive mixed models (GAMMs) to evaluate the partial effects of plant variables on arthropod abundance of different trophic levels as the results of a Shapiro–Wilk normality test showed that arthropod abundance had an abnormal distribution (Appendix S1: Table S5). GAMMs estimate response curves with a nonparametric smoothing function, allowing test nonlinear and linear responses of data to explanatory variables (Wood et al., 2016). In the GAMMs, all the plant variables were treated as fixed factors, and block nested within year was treated as a random factor. A negative binomial distribution was assumed and a log link function was used in these models due to a high proportion of zero values (Clark et al., 2016; Long, 1997). For each response variable, we set the basis dimension (i.e., k) of the smoothing parameter to three, equaling a polynomial function of maximum third order to avoid overfitting GAMMs. GAMMs were fit using mgcv packages (Wood, 2022).

Structural equation modeling (SEM) was performed to analyze hypothetical pathways identifying the effects of plant biomass, leaf macronutrients, and sodium directly and indirectly on the abundance of arthropods within trophic levels (herbivores, omnivores, predators, and parasitoids) across two sampling years (Figure 1C). To avoid high collinearity among predictor variables, we ensured that the correlation coefficients among all predictor variables retained in the final model were less than the standard threshold of |r| = 0.7 (Craney & Surles, 2002). Consequently, leaf carbohydrate concentration was excluded from SEM analyses (Appendix S1: Figure S3). To avoid overfitting and increase degrees of freedom, we further simplified the SEM model by removing any path with a coefficient >0.1 when it was not significant (Clark et al., 2016). The SEM was fit using the “lavaan” package (Rosseel, 2012). We used nonparametric Bollen–Stine bootstrapping estimations to account for linear and nonlinear relationships, with good model fit specified by a Bollen–Stein bootstrap p > 0.10 (Bollen & Stine, 1992).

RESULTS

Plant biomass

Consistent with the hypothesis that predators/parasitoids are limited by plant biomass, CTM of trophic level increased nonlinearly with rising plant biomass (Figure 2A; Appendix S1: Table S6). Plant biomass played a dual role in promoting predator abundance: It decreased omnivore abundance and increased the abundance of chewing herbivores (Figure 3; Appendix S1: Table S7). Additionally, plant biomass directly and indirectly enhanced parasitoid abundance by increasing populations of sucking herbivores and endophytes. After controlling for the effects of leaf macronutrients and sodium concentrations on arthropods, we observed that the partial residuals of the herbivore and parasitoid abundance increased nonlinearly, while the partial residual of the predator abundance increased linearly with increasing plant biomass (Figure 4; Appendix S1: Table S8). In contrast, the partial residual of omnivore abundance decreased linearly with increasing plant biomass. Among the herbivorous feeding groups, the partial residual of the abundance of sucking and chewing herbivores increased linearly, whereas the partial residuals of endophyte abundance increased nonlinearly with increasing plant biomass (Appendix S1: Table S9 and Figure S4).

FIGURE 2.

Effects of plant biomass (A), leaf carbohydrate concentrations (B), leaf protein concentrations (C), protein:carbohydrate (D), leaf phosphorus concentrations (E), and leaf sodium concentrations (F) on community trophic mean across monocultures of different plant species. Solid lines indicate fits of bivariate linear and nonlinear mixed effects models with block nested within year as a random factor. We used first‐ and second‐order polynomials to capture linear and nonlinear responses of CTM values to variations of plant biomass and macro‐and micronutrients. When the Akaike information criterion (AICc) difference (ΔAICc) was <2 between two significant functions, the simpler function was chosen, where ‘simplicity’ increased from linear to quadratic. The shaded gray areas indicate 95% CI for the regressions. Marginal $R_m^2$ (variance explained by fixed effects), df, and p values are also shown. Details of Model selection and Statistical test results are shown in Appendix S1: Table S6. Arthropod illustrations credits: Xiaoming Lu.

FIGURE 3.

Structural equation model of the direct and indirect effects of plant biomass, leaf protein:carbohydrate, and leaf sodium across the monocultures of different plant species on arthropod abundance of different trophic levels (herbivore, omnivore, predator, and parasitoid). Solid arrows represent significant paths (p < 0.10), and dashed arrows designate nonsignificant paths. Numbers near the lines show standardized regression weights. Red (blue) arrows indicate positive (negative) effects. Model fit metrics (χ² and Bollen–Stein bootstrap p value) are shown. R ² values for component models are given below the boxes of endogenous variables. Statistical test for standardized path coefficients of structural equation model results is shown in Appendix S1: Table S7. Plant and arthropod illustrations credits: Xiaoming Lu.

FIGURE 4.

Partial effects of plant biomass, leaf carbohydrate concentrations, leaf protein concentrations, protein:carbohydrate ratio, leaf phosphorus concentrations, and leaf sodium concentrations on abundance of herbivore (A–F), omnivore (G–L), predator (M–R), and parasitoid (S–X) quantified using generalized additive mixed models (GAMMs). The y‐axis values are based on partial residues and indicate the relative influence of each explanatory variable on the prediction. The shaded gray areas indicate an approximation of the 95% CI of the smoothed curve. Solid lines represent significant effect of plant variables on arthropod abundance; dashed lines were nonsignificant. See Appendix S1: Table S8 for full model results. Arthropod illustrations by Xiaoming Lu.

Plant macronutrients

Consistent with the hypothesis that predators and parasitoids are constrained by macronutrient stoichiometry, CTMs exhibited a quadratic increase with increasing leaf protein‐to‐carbohydrate ratios (Figure 2D; Appendix S1: Table S6). In contrast, no significant responses were observed in CTMs when leaf nonstructural carbohydrates, protein, or phosphorus concentrations were increased independently (Figure 2; Appendix S1: Table S6). Notably, elevated leaf protein‐to‐carbohydrate ratios were associated with an increased abundance of sucking herbivores (Figure 3; Appendix S1: Table S7), which subsequently led to a rise in the abundance of parasitoids. After controlling for the effects of plant biomass and leaf sodium concentrations on arthropods, the partial residual of the abundance of herbivores showed a linear increase with rising leaf protein‐to‐carbohydrate ratios (Figure 4; Appendix S1: Table S8). Additionally, the abundance of sucking herbivores decreased linearly with increasing leaf carbohydrate concentrations, while it increased linearly with increasing leaf protein concentrations and leaf protein‐to‐carbohydrate ratios (Appendix S1: Table S9 and Figure S4).

Plant sodium

In accordance with the hypothesis that predators and parasitoids are influenced by plant sodium levels, we observed a quadratic increase in consumer CTM as leaf sodium concentrations increased (Figure 2F; Appendix S1: Table S6). Importantly, elevated leaf sodium levels directly enhanced the abundance of sucking and chewing herbivores (Figure 3; Appendix S1: Table S7). This increase in herbivore abundance subsequently promoted abundances of parasitoids and predators. After controlling for the effects of plant biomass and leaf macronutrient contents on arthropods, the partial residuals of the abundance of herbivore increased nonlinearly, while the partial residual of abundance of predator and parasitoid increased linearly with rising leaf sodium concentrations (Figure 4; Appendix S1: Table S8). Additionally, the partial residual abundance of both sucking and chewing herbivores exhibited a nonlinear increase with increasing leaf sodium concentrations (Appendix S1: Table S9 and Figure S4).

DISCUSSION

Plant biomass promotes top‐heavy arthropod trophic pyramids

The initial removal of topsoil may have disrupted soil structure and altered soil properties, potentially homogenizing soil arthropod communities (Rowen et al., 2020). As a result, variation in arthropod abundance across trophic levels in our monoculture experiment may be driven more by plant species identity than by soil heterogeneity. Against this background, we found that plant biomass was positively associated with CTM, supporting Hypothesis I‐1 that increasing plant biomass increased the proportions of arthropods at higher trophic levels (Siemann, 1998; Wimp et al., 2010). This finding contrasts with earlier studies reporting that greater plant biomass leads to more bottom‐heavy trophic pyramids (Hatton et al., 2015; Welti et al., 2020). Such discrepancies may reflect both ecological and methodological differences among studies. First, many previous studies were conducted in aquatic systems, where highly mobile predators can move across large spatial scales, weakening the influence of local primary production on upper trophic levels (McCauley et al., 2018). Second, differences may arise from whether trophic structure is evaluated using abundance‐ or biomass‐based metrics (Cohen et al., 2003). In many biomass‐based analyses, trophic structure is derived by estimating biomass across trophic levels from the relative abundance of taxa (Elton & Nicholson, 1942; McCauley et al., 2018). In grasslands, greater plant biomass often disproportionately benefits large herbivores such as locusts, whereas predator and parasitoid populations may respond less strongly (Prather & Kaspari, 2019; Wimp et al., 2010). As a result, biomass‐based pyramids may overemphasize the contribution of herbivores and thus yield a more bottom‐heavy trophic structure.

Three nonmutually exclusive mechanisms may explain the positive association between plant biomass and CTM. First, greater plant biomass may directly increase the abundance of predators and parasitoids. Plant species with higher biomass typically provide more habitat volume and greater structural complexity, which can support larger populations of spiders and parasitoid wasps. In turn, this may increase the representation of arthropods at higher trophic levels and promote more top‐heavy trophic pyramids through stronger top‐down regulation of herbivores and omnivores (Oksanen et al., 1981; Prather et al., 2021). Second, plant biomass may exhibit an indirect positive effect on predator abundance through decreasing omnivore abundance and increasing chewer abundance. In grasslands, chewing herbivores can provide substantial prey resources for predatory arthropods, whereas some omnivores, such as members of Sarcophagidae and Calliphoridae, may also consume animal‐based resources and thereby modify trophic pathways (Sivinski et al., 1999). Plant species with high biomass may create a complex structure that hinders omnivores from locating their hosts, thereby alleviating the disturbance effects of these omnivores on predators (Prather et al., 2021; Wimp et al., 2010). Additionally, active attacks on grasshoppers have been observed in many spider species in more productive habitats, which is a major driving factor maintaining spider abundance (Yip et al., 2008). Third, plant biomass increases parasitoid abundance through promoting sucking herbivores, such as aphids, and endophytes, like leaf miners, both of which have high dietary requirements for plant material. An increase in the abundance of these host insects is likely to lead to a simultaneous increase in the survival and reproduction rates of their specialist parasitoids (Hawkins et al., 1997).

Plant protein: Carbohydrate promotes top‐heavy arthropod trophic pyramids

Protein, carbohydrate, and phosphorus are important macronutrients for maintaining population sizes of many herbivorous feeding groups, indirectly increasing proportions of predators and parasitoid abundances and increasing CTMs of trophic level (Awmack & Leather, 2002; Hawkins et al., 1997; Joern et al., 2012). However, we found that increases in protein‐to‐carbohydrate ratios, rather than macronutrients amounts, significantly promoted CTM across monocultures of different plant species, partially supporting Hypothesis II‐1. This may be because all animals need to maintain elemental homeostasis, and herbivores have evolved the capacity to regulate their macronutrients intake to a specific ratio (Boersma & Elser, 2006). Nevertheless, ratio‐based transformations should be interpreted cautiously, because they can obscure the independent effects of individual nutrients; when possible, covariance‐based approaches provide a more informative alternative (Raubenheimer & Simpson, 1992; Raubenheimer & Simpson, 1994). Our findings further demonstrate that plant protein and carbohydrate content play a significant role in shaping arthropod communities. Even after accounting for the influence of plant biomass and leaf sodium concentrations, we observed that increases in the ratio of protein to carbohydrates across plant species led to a higher abundance of sucking herbivores and endophytes. Sucking herbivores (e.g., aphids) and endophytes (e.g., leaf miners) likely need more protein than carbohydrates, relative to environmental availability, for their growth and reproduction, and benefit from a high protein‐to‐carbohydrate ratio in their plant food (Awmack & Leather, 2002; Lu et al., 2022; Prather et al., 2021). In contrast, plants with high protein‐to‐carbohydrate ratios have been shown to exert strong negative effects on abundance of some locusts (chewing herbivores) due to carbohydrate limitations for their reproduction and migration (Cease et al., 2012; Le Gall et al., 2020). In our study, the low abundance of migratory locusts may explain why protein and carbohydrate ratios had no detectable effect on the abundance of chewing herbivores. Likewise, neither the absolute amounts nor the ratio of foliar protein and carbohydrates affected omnivore abundance. This weak response may reflect the broad dietary flexibility of omnivores. For example, some taxa, such as lauxaniid flies, can obtain protein from animal tissues, whereas members of Muscidae can acquire carbohydrates from floral nectar (Bänziger & Pape, 2004). Such alternative nutrient sources may reduce their dependence on foliar macronutrients and thereby weaken their responses to variation in leaf nutritional composition.

More importantly, increases in the plant protein‐to‐carbohydrate ratios increased the abundance of parasitoids by increasing the abundance of sucking herbivores. Parasitoid wasps often attack sucking herbivores (e.g., aphids) by laying a single egg into the body of their host, and using most of the host's macronutrients for development and growth of parasitoid larva until pupation (Hawkins et al., 1997; Petermann et al., 2010). Consequently, increasing plant protein‐to‐carbohydrate ratios across different plant species could increase the abundance of associated aphids, leading to an increase in the abundance of associated parasitoid species, promoting top‐heavy arthropod trophic pyramids.

Plant sodium promotes top‐heavy arthropod trophic pyramids

In line with Hypothesis III‐1, we found that increases in leaf sodium concentrations across monocultures of different plant species promote higher CTM of trophic levels. Our results support a previous study finding that higher plant sodium availability promotes more top‐heavy food webs across a diverse set of grasslands (Welti et al., 2020). Our study revealed a distinct relationship between leaf sodium levels and herbivore abundance. When controlling for the influence of plant biomass and leaf macronutrient content, we found that higher leaf sodium concentrations across plant species were linked to a greater abundance of both sucking and chewing herbivores. Sucking and chewing herbivores often have sodium tissue levels 100‐ to 1000‐fold higher than their food, and they accumulate sodium from sodium‐poor plant tissue to maintain reproduction and development (Clay et al., 2014; Kaspari, 2021). In contrast, leaf sodium did not significantly affect the abundance of endophytic herbivores or omnivores. For endophytic herbivores, this weak pattern may partly reflect methodological limitations, as sweep‐net sampling is less effective at capturing insects that feed within plant tissues and may therefore underestimate their responses to foliar sodium (Welti & Kaspari, 2021). For omnivores, the lack of response may instead reflect dietary flexibility, because some taxa, such as lauxaniid flies, can obtain sufficient sodium from animal tissues and thus may depend less on leaf sodium (Bänziger & Pape, 2004).

High leaf sodium promotes the abundance of parasitoids and predators via increasing the abundance of sucking and chewing herbivores, leading to top‐heavy trophic pyramids. These results contradict the findings of sodium addition experiments showing that increases in plant sodium stimulated higher proportions of herbivores and omnivores than predators and parasitoids, which could decrease the likelihood that abundance of plant consumers is constrained by top‐down effects and lead to bottom‐heavy arthropod trophic pyramids (Clay et al., 2014; Welti et al., 2019). However, we found that high plant sodium concentrations across different plant species promoted the abundance of parasitoids through increasing their specialist prey (e.g., aphids) and promoted predator abundance through increasing chewers (e.g., grasshopper), which in turn stabilized arthropod food webs and maintained top‐heavy arthropod trophic pyramids (Hawkins et al., 1997, Petermann et al., 2010). These results suggest that plant sodium has a strong bottom‐up effect on arthropod abundance across multiple trophic levels within natural arthropod communities (Kaspari, 2021; Welti et al., 2020). Conversely, the low sensitivity of predator abundance to the increased sucking herbivores under high level of plant sodium could be due to the more generalist feeding choices of many predators, such as spiders, obtaining sodium by feeding on other insects or by cannibalism (Welti et al., 2020; Yip et al., 2008).

Caveats and implications for monoculture experiments

We found that plant biomass, leaf protein‐to‐carbohydrate ratio, and sodium are the key regulators of arthropod trophic pyramids. However, whether these relationships generalize to natural communities remains unclear. In natural systems, plant species assemble non‐randomly, and neighboring species with contrasting nutrient profiles may modify or obscure the effects of the macro‐ and micronutrient composition of any single plant species on arthropod trophic structure (Bennett & Gratton, 2013). An important next step, therefore, is to test whether the nutrient‐mediated patterns observed in this monoculture study are maintained in more compositionally complex plant communities (Loranger et al., 2013). In addition, a substantial proportion of the variation in CTM remained unexplained by plant biomass and elemental nutrient traits alone. This suggests that other plant characteristics may also shape arthropod trophic structure. In particular, incorporating chemical defense traits, such as total polyphenols and condensed tannins, and biomechanical traits, such as leaf toughness, may strengthen inference because these traits can interact with nutrient availability to influence herbivore performance and, ultimately, arthropod abundance (Poelman et al., 2008). Nutrient stoichiometry may also be important. For example, the nitrogen‐to‐phosphorus ratio can affect herbivore growth and reproduction and should therefore be considered when evaluating the effects of plant nutrients on arthropod communities (Zhang et al., 2014). A further priority is to account for temporal variation. In grasslands, plant biomass and functional traits change over the course of the growing season, and these shifts may alter arthropod responses through time. Repeated sampling of both plants and arthropods across the season would therefore provide a stronger basis for understanding how seasonal dynamics in plant biomass and nutrient composition influence arthropod diversity and trophic structure (Awmack & Leather, 2002).

This study has several limitations that should be considered when interpreting our results. First, we evaluated the effects of plant biomass and nutrient traits, including both macro‐ and micronutrients, on arthropod trophic groups and the CTM within relatively small experimental plots. Because many arthropods are highly mobile, individuals may move rapidly among plots, potentially weakening the effects of plant quality and quantity on the abundance of different trophic groups across monoculture treatments. Larger plots could help reduce this problem, although such designs often require a trade‐off with replication. Future studies should therefore quantify arthropod trophic groups over broader spatial scales to more fully characterize how arthropod trophic pyramids respond to variation in plant resources. In addition, the direction of causality may not be exclusively bottom‐up. Predators may indirectly influence plant biomass and nutrient content by altering herbivore abundance or behavior (Hawlena & Schmitz, 2010). To disentangle these reciprocal effects, future work should use more rigorous insect enclosure experiments to test how predator‐induced changes in herbivores cascade to plant biomass and nutrient composition in monoculture systems (Schmitz, 2004). Second, our monoculture experiment did not directly manipulate plant quality and quantity, in contrast to many laboratory studies that employ synthetic diets or field fertilization with varying nutrient concentrations (Behmer, 2009; Prather & Kaspari, 2019). For instance, synthetic diets with phosphorus content ranging from 0.5% to 1.0% can enhance the growth of various arthropods (Boersma & Elser, 2006). However, excessively high phosphorus concentrations, such as 3.5%, resulting from nutrient addition, can significantly impair the growth of mayflies (Frost & Elser, 2002). Similarly, in a sodium addition experiment, grasshopper nymphs reared on a plant diet containing 1% sodium exhibited higher survival rates compared to those on a high‐sodium diet (Peterson et al., 2021). These findings suggest that arthropod responses to manipulated nutrient levels are likely nonlinear, as different arthropod species have distinct optimal nutrient concentrations (Lu et al., 2021). Given the substantial natural variation in phosphorus (1.8%–3.7%) and sodium (0.9%–2.6%) content among plant species in our monocultures, our interpretation of the quadratic effect of plant quality on CTM is particularly insightful.

CONCLUSIONS

Understanding what drives variation in organismal abundance and trophic structure is a fundamental goal in ecology. Our previous study showed that plant biomass and nutrient traits independently affect arthropod abundance across taxonomic groups (Lu et al., 2021). Here, we extend that finding by showing that natural variation in plant biomass, leaf protein‐to‐carbohydrate ratio, and leaf sodium concentration across monocultures is associated with more top‐heavy arthropod trophic pyramids. These effects appear to arise through a combination of direct and indirect pathways. Plant biomass increased predator abundance directly and indirectly through lower omnivore abundance and higher chewer abundance, and it increased parasitoid abundance directly and indirectly through higher abundances of sucking herbivores and endophytes. Higher leaf protein‐to‐carbohydrate ratios and leaf sodium concentrations were also linked to greater parasitoid abundance through their positive effects on sucking herbivores. Together, these results indicate that interactions among omnivores, predators, herbivore feeding guilds, and parasitoids are central to understanding how natural variation in plant quantity and quality structures arthropod food webs in grasslands. This study therefore provides a mechanistic basis for linking plant trait variation to trophic dynamics and ecosystem functioning.

AUTHOR CONTRIBUTIONS

Yongfei Bai designed the research. Xiaoming Lu conceived the study idea. Xiaoming Lu, Xuezhen Zhao, Taiki Tachibana, Kei Uchida, and Takehiro Sasaki performed the experiments and collected the data. Xiaoming Lu, Ellen A. R. Welti, and Yongfei Bai conducted the data analysis and wrote the manuscript. Yongfei Bai edited the manuscript. All authors contributed substantially to manuscript revisions.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest.

Supporting information

Appendix S1.

Lu, Xiaoming , Zhao Xuezhen, Tachibana Taiki, Welti Ellen A. R., Uchida Kei, Sasaki Takehiro, and Bai Yongfei. 2026. “Plant Biomass and Macro‐ and Micronutrients Jointly Drive Top‐Heavy Arthropod Trophic Pyramids in Grassland Monocultures.” Ecology 107(9): e70483. 10.1002/ecy.70483

DATA AVAILABILITY STATEMENT

Data and code (Lu et al., 2026) are available in Dryad at https://doi.org/10.5061/dryad.xsj3tx9wn.