Estimating global species richness using symbolic data meta-analysis

Huan Lin, M. Julian Caley, Scott A. SissonEcography 2017

Abstract

Global species richness is a key biodiversity metric. Despite recent efforts to estimate global species richness, the resulting estimates have been highly uncertain and often logically inconsistent. Estimates lower down either the taxonomic or geographic hierarchies are often larger than those above. Further, these estimates have been represented in a wide variety of forms, including intervals (a, b), point estimates with no uncertainty, and point estimates with either symmetrical or asymmetrical bounds, making it difficult to combine information across different estimates. Here, we develop a Bayesian hierarchical approach to estimate the global species richness from published studies. It allows us to recover interval estimates at each level of the hierarchy, even when data are partially or wholly unobserved, while respecting logical constraints, and to determine the effects of estimation on the whole hierarchy of obtaining future estimates anywhere within it.

Introduction

Determining the total number of species on Earth is a fundamental challenge in ecology. However, current literature presents a fragmented landscape of estimates. Some studies provide single point estimates, while others offer ranges or asymmetric bounds. A significant issue is logical inconsistency: meta-analyses often reveal that the sum of estimates for specific subgroups (e.g., insects) exceeds the total estimate for their parent group (e.g., arthropods). This discrepancy arises because studies often operate in isolation using disparate methodologies.

To address this, we conducted a meta-analysis of 45 previously published estimates, covering seven categories: coral reefs, marine species, terrestrial species, arthropods, insects, beetles, and global eukaryotes. The dataset included 17 point estimates, 19 intervals, and 9 combined forms.

Methods

We developed a Bayesian hierarchical framework to reconcile these diverse data types and enforce taxonomic consistency using two primary innovations:

  1. Symbolic Data Analysis (SDA): We treated intervals $(a, b)$ as bivariate random vectors consisting of a midpoint and a log-range. Point estimates were treated as intervals with missing uncertainty ranges, which the model then estimated.
  2. Hierarchical Modeling: We constructed a taxonomic hierarchy (e.g., Global → Marine/Arthropods → Insects). The model enforces "sum-to-parent" constraints, ensuring that the sum of child taxa estimates equals the parent taxon's total.
  3. Information Sharing: The Bayesian approach allows data-rich groups (like insects) to inform the estimates of data-poor groups (like "other global species") across the hierarchy.

The model was implemented using Markov chain Monte Carlo (MCMC) simulations in the Stan software package.

Key Insights and Results

A primary finding of this study is the relationship between the magnitude of an estimate and its associated uncertainty. We observed a moderately strong positive correlation (0.57) between the midpoint of an estimate and its log-range.

Correlation between standardized midpoints and log-ranges across different taxonomic and geographic categories.

This scaling suggests that as the estimated number of species in a group increases, our certainty regarding that number decreases. This relationship justifies the use of a bivariate model to capture how uncertainty grows with richness.

Global Species Richness

Our model estimates a total of 22.3 million eukaryotic species, with a 95% Highest Posterior Density (HPD) interval of 11.2 to 36.3 million. This figure is significantly higher than the widely cited estimate of ~8.7 million (Mora et al., 2011). We argue that many previous estimates are likely too low because they do not account for the logical constraints of the taxonomic hierarchy.

Taxonomic Findings

  • Arthropod Dominance: Terrestrial arthropods are estimated at ~10.7 million species, representing nearly half of all global eukaryotes.
  • Correction of Estimates: The hierarchical model "corrects" individual studies. For instance, credible intervals for coral reef species were significantly revised when forced into consistency with broader marine and global data.

Discussion and Limitations

While this framework provides a more consistent global estimate, several limitations remain. The model assumes each published estimate is independent, whereas many studies may share underlying data or expert biases. Additionally, we simplified asymmetric intervals by using geometric midpoints; future work could employ trivariate modeling to capture this nuance.

Furthermore, this analysis was restricted to global-scale estimates to avoid the complexities of species turnover (beta-diversity) inherent in regional data. Future iterations could incorporate regional studies by modeling spatial diversity patterns. Ultimately, the accuracy of this meta-analysis remains dependent on the quality of the underlying primary research.

Summary updated

Abstract

Global species richness is a key biodiversity metric. Despite recent efforts to estimate global species richness, the resulting estimates have been highly uncertain and often logically inconsistent. Estimates lower down either the taxonomic or geographic hierarchies are often larger than those above. Further, these estimates have been represented in a wide variety of forms, including intervals (a, b), point estimates with no uncertainty, and point estimates with either symmetrical or asymmetrical bounds, making it difficult to combine information across different estimates. Here, we develop a Bayesian hierarchical approach to estimate the global species richness from published studies. It allows us to recover interval estimates at each level of the hierarchy, even when data are partially or wholly unobserved, while respecting logical constraints, and to determine the effects of estimation on the whole hierarchy of obtaining future estimates anywhere within it
BibTeX
@inproceedings{lin2017estimating,
  author = {Huan Lin and M. Julian Caley and Scott A. Sisson},
  title = {{Estimating global species richness using symbolic data meta-analysis}},
  booktitle = {Ecography},
  year = {2017},
}
Figure