Community biomass and bottom up multivariate nutrient complementarity mediate the effects of bioturbator diversity on pelagic production

Abstract

Tests of the biodiversity and ecosystem functioning (BEF) relationship have focused little attention on the importance of interactions between species diversity and other attributes of ecological communities such as community biomass. Moreover, BEF research has been mainly derived from studies measuring a single ecosystem process that often represents resource consumption within a given habitat. Focus on single processes has prevented us from exploring the characteristics of ecosystem processes that can be critical in helping us to identify how novel pathways throughout BEF mechanisms may operate. Here, we investigated whether and how the effects of biodiversity mediated by non-trophic interactions among benthic bioturbator species vary according to community biomass and ecosystem processes. We hypothesized that (1) bioturbator biomass and species richness interact to affect the rates of benthic nutrient regeneration [dissolved inorganic nitrogen (DIN) and total dissolved phosphorus (TDP)] and consequently bacterioplankton production (BP) and that (2) the complementarity effects of diversity will be stronger on BP than on nutrient regeneration because the former represents a more integrative process that can be mediated by multivariate nutrient complementarity. We show that the effects of bioturbator diversity on nutrient regeneration increased BP via multivariate nutrient complementarity. Consistent with our prediction, the complementarity effects were significantly stronger on BP than on DIN and TDP. The effects of the biomass-species richness interaction on complementarity varied among the individual processes, but the aggregated measures of complementarity over all ecosystem processes were significantly higher at the highest community biomass level. Our results suggest that the complementarity effects of biodiversity can be stronger on more integrative ecosystem processes, which integrate subsidiary "simpler" processes, via multivariate complementarity. In addition, reductions in community biomass may decrease the strength of interspecific interactions so that the enhanced effects of biodiversity on ecosystem processes can disappear well before species become extinct.

Figure 1. Effects (mean ±1 SE) of invertebrate biomass, species richness and taxonomic composition on (a, d) dissolved inorganic nitrogen (DIN) flux, (b, e) total dissolved phosphorus (TDP) flux and (c, f) bacterioplankton production (BP).

Left-hand panels show the interactive effects of invertebrate biomass and species richness on ecosystem processes. Right-hand panels show the overall effects (irrespective of biomass) of species richness and composition (nested factor) on the magnitude of ecosystem processes analyzed. The overall linear effect of species richness was calculated by regressing ecosystem process data from all individual microcosms (n = 84, controls not included) as a function of species richness across all biomass levels. The overall effects of taxonomic composition are shown by nested comparisons among mean values across all biomass levels for each individual species and 2-species mixtures. Treatments marked with different letters within the same species richness level differ significantly from one another (Tukey test, P<0.05). Csp. = Chironomus sp., Hs = Heteromastus similis, Ha = Heleobia australis.

Figure 2. Transgressive overyielding for the ecosystem processes analyzed as a function of invertebrate biomass.

Diversity ‘effect sizes’ (standardized ln response ratios) were estimated for individual ecosystem processes for each biomass level by comparing the proportional response of 2- (open symbols) and 3-species mixtures (filled symbols) to their respective best constituent monoculture. Overall cumulative effect sizes (gray diamonds) and their ±95% bootstrapped CI’s were calculated from the weighted integration of the individual effect sizes calculated for all combined ecosystem processes and species richness treatments throughout 9999 iterations. Significant overall transgressive overyielding occurs if the value of LR_trans and its confidence interval are greater than zero (dashed line). Numbers in parentheses represent the proportions of treatments with LR_trans >0 for the respective invertebrate biomass and species richness levels. Abbreviations for ecosystem processes are as in figure 2.

Figure 3. Transgressive overyielding for the ecosystem processes analyzed.

Diversity effect sizes (ln response ratios) and their ±95% bootstrapped CI were estimated from the weighted integration, throughout 9999 iterations, of the effect sizes calculated from the proportional response of 2- and 3-species mixtures to their respective best constituent monoculture for each biomass level. Significant overall transgressive overyielding occurs if the value of LR_trans and its confidence interval are greater than zero (dashed line). Abbreviations for ecosystem processes are as in figure 2.

Figure 4. Isopleths showing changes in bacterial production (BP) as a function of dissolved inorganic nitrogen (DIN) and total dissolved phosphorus (TDP) fluxes in microcosms inhabited by benthic bioturbators.