Competition model explains trends of long-term fertilization in plant communities

Abstract

Over 40 years ago, Kempton (Biometrics, 35, 1979, 307) reported significant modification to plant community structure following a long-term fertilization experiment. Many researchers have investigated this phenomenon in the years since. Collectively, these studies have shown consistent shifts in rank abundance relationships among species in communities following fertilization. The previous studies indicated that fertilization affects community structure through several critical processes, including trait-based functional response, reordering of species in rank abundance diagram (RAD), and niche dimensionality, although some questions have remained. How does the species reordering driven by the plant responses cause characteristic trends in temporal changes of RAD? Why are those trends ubiquitous in various systems? To answer those questions, we theoretically investigated the effects of fertilization on community structure based on a colonization model (or Levins model) with competition-fecundity trade-offs, which can result in the coexistence of multiple species under competition. The model represents characteristic RAD, which can be an adequate tool to study community composition. Our theoretical model comprehensively represents observed trends in rank abundance relationships following long-term fertilization and suggests that competitive interactions among species are a critical factor in structuring species diversity in plant communities.

Keywords: community structure; competition–fecundity trade‐off; nutrient; rank abundance diagram.

FIGURE 1

(a) Fecundity functions used in analyses representing competition–fecundity trade‐off. The original condition was expressed as f _0,i  = 2(1−Exp[−4i/80])/(1 + Exp[−4i/80]), and fertilization was expressed as f _1,i  = 3(1−Exp[−16i/80])/(1 + Exp[−16i/80]) or f _2,i  = 3(1−Exp[−18i/80])/(1 + Exp[−18i/80]). (b) Relative improvements in fecundities following fertilization. It was assumed that 80 species could exist within the simulation.

FIGURE 2

Equilibrium frequency distributions of sites with each species and RADs under fecundities illustrated in Figure 1. We excluded species with a relative abundance <10⁻⁵ from the RADs. Blue points represent the eight most abundant species in Figure 1a, and red circles represent species that were absent in Figure 1a. Fraction numbers on the RADs represent species compositions, where the denominator and numerator indicate the total number of species and that of lost species by fertilization, respectively. Parameters were n = 80, q = 0.3, and m = 0.2.

FIGURE 3

Dynamic time series of frequencies of sites with each species. Fertilization was initiated at t = 10,000. Left panels show cumulative frequency plots, where a space between two curves represents a specific frequency. Species are arranged by competitive ability from left to right. Right panels represent density plots of the frequencies of sites with each species on competitive ability. An immigration term was included at a rate of 10⁻¹⁰. Parameters were n = 80, q = 0.3, and m = 0.2.

FIGURE 4

Temporal change in an RAD following fertilization, in which k means 1,000 timesteps after the initiation of fertilization. Fraction values on each plot represent species composition, where the denominator and numerator indicate the total number of species and that of lost species by fertilization, respectively. Parameters were n = 80, q = 0.3 and m = 0.2.

FIGURE 5

Species richness (a) and biomass (b) at equilibrium under competition–fecundity trade‐offs with varying combinations of intensity of saturation, α, and maximum fecundity, β. Parameters were n = 80, q = 0.3, and m = 0.2.