Abstract

Species abundance distributions (SADs) are increasingly used to investigate how species community structure changes in response to environmental variations. SAD models depict the relative abundance of species recorded in a community and express fundamental aspects of the community structure, namely patterns of commonness and rarity. However, the influence of differences in environmental conditions on SAD characteristics is still poorly understood. In this study we used SAD models of carabid beetles (Coleoptera: Carabidae) in three grassland ecosystems (desert, typical, and meadow steppes) in China. These ecosystems are characterized by different aridity conditions, thus offering an opportunity to investigate how SADs are influenced by differences in environmental conditions (mainly aridity and vegetation cover, and hence productivity). We used various SAD models, including the meta-community zero sum multinomial (mZSM), the lognormal (PLN) and Fisher’s logseries (LS), and uni- and multimodal gambin models. Analyses were done at the level of steppe type (coarse scale) and for different sectors within the same steppe (fine scale). We found that the mZSM model provided, in general, the best fit at both analysis scales. Model parameters were influenced by the scale of analysis. Moreover, the LS was the best fit in desert steppe SAD. If abundances are rarefied to the smallest sample, results are similar to those without rarefaction, but differences in models estimates become more evident. Gambin unimodal provided the best fit with the lowest α-value observed in desert steppe and higher values in typical and meadow steppes, with results which were strongly affected by the scale of analysis and the use of rarefaction. Our results indicate that all investigated communities are adequately modeled by two similar distributions, the mZSM and the LS, at both scales of analyses. This indicates (1) that all communities are characterized by a relatively small number of species, most of which are rare, and (2) that the meta-communities at the large scale maintain the basic SAD shape of the local communities. The gambin multimodal models produced exaggerated α-values, which indicates that they overfit simple communities. Overall, Fisher’s α, mZSM θ, and gambin α-values were substantially lower in the desert steppe and higher in the typical and meadow steppes, which implies a decreasing influence of environmental harshness (aridity) from the desert steppe to the typical and meadow steppes.

Highlights

  • Biodiversity loss is the most common consequence of the increasing environmental degradation due to anthropogenic changes (Barnosky et al, 2011; Redford et al, 2015; Kehoe et al, 2017)

  • We collected 6,873 individuals belonging to 25 carabid species: 19 species were collected in the meadow steppe, of which 12 in the first sector and 19 in the second sector; 18 species in the typical steppe, of which 16 in the first sector, 15 in the second sector, and 12 in the third sector; and, 12 species in the sole sector of the desert steppe

  • We found that the basic results (i.e., species abundance distributions (SADs) shape) of the analyses conducted at the smaller scale mirrored those obtained at the larger scale

Read more

Summary

Introduction

Biodiversity loss is the most common consequence of the increasing environmental degradation due to anthropogenic changes (Barnosky et al, 2011; Redford et al, 2015; Kehoe et al, 2017). Local extinctions or even changes in species’ relative abundance lead to alterations in community structure and in ecosystem functioning (McCann, 2000; Balvanera et al, 2006). Studies that model variations in community structure in response to changes in environmental characteristics may provide important information to predict how biodiversity will be affected by alterations in the balance of rare versus dominant species (Tsafack et al, 2019a; Ibanez et al, 2020). SAD models may be used in conservation studies to predict species extinction risk (Kitzes and Harte, 2015) and ecosystem health (related to disturbance) (Dornelas et al, 2009) and, to inform management actions (Kim et al, 2013; Milicicet al., 2017)

Methods
Results
Discussion
Conclusion
Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call