Multidimensional biases, gaps and uncertainties in global plant occurrence information

Abstract

Plants are a hyperdiverse clade that plays a key role in maintaining ecological and evolutionary processes as well as human livelihoods. Glaring biases, gaps, and uncertainties in plant occurrence information remain a central problem in ecology and conservation, but these limitations have never been assessed globally. In this synthesis, we propose a conceptual framework for analyzing information biases, gaps and uncertainties along taxonomic, geographical, and temporal dimensions and apply it to all c. 370,000 species of land plants. To this end, we integrated 120 million point-occurrence records with independent databases on plant taxonomy, distributions, and conservation status. We find that different data limitations are prevalent in each dimension. Different information metrics are largely uncorrelated, and filtering out specific limitations would usually lead to extreme trade-offs for other information metrics. In light of these multidimensional data limitations, we critically discuss prospects for global plant ecological and biogeographical research, monitoring and conservation, and outline critical next steps towards more effective information usage and mobilization. We provide an empirical baseline for evaluating and improving global floristic knowledge and our conceptual framework can be applied to the study of other hyperdiverse clades. AbstractPlants are a hyperdiverse clade that plays a key role in maintaining ecological and evolutionary processes as well as human livelihoods. Glaring biases, gaps, and uncertainties in plant occurrence information remain a central problem in ecology and conservation, but these limitations have never been assessed globally. In this synthesis, we propose a conceptual framework for analyzing information biases, gaps and uncertainties along taxonomic, geographical, and temporal dimensions and apply it to all c. 370,000 species of land plants. To this end, we integrated 120 million point-occurrence records with independent databases on plant taxonomy, distributions, and conservation status. We find that different data limitations are prevalent in each dimension. Different information metrics are largely uncorrelated, and filtering out specific limitations would usually lead to extreme trade-offs for other information metrics. In light of these multidimensional data limitations, we critically discuss prospects for global plant ecological and biogeographical research, monitoring and conservation, and outline critical next steps towards more effective information usage and mobilization. We provide an empirical baseline for evaluating and improving global floristic knowledge and our conceptual framework can be applied to the study of other hyperdiverse clades.PeerJ PrePrints | https://dx.doi.org/10.7287/peerj.preprints.1326v2 | CC-BY 4.0 Open Access |

Finding needles in the haystack: where to look for rare species in the American tropics

“…3) are there general patterns of geographic disjunctions across taxa in rare species? Since data available from GBIF are known to contain taxonomic and geographic errors and biases (Maldonado et al 2015, Meyer et al 2016, we validate our results with a nearly independent dataset from vegetation plots of the Amazon Tree Diversity Network (ATDN) based on ter Steege et al (2013). We use this comparison to discuss the suitability of collection data to assess species rarity.…”

“…The data downloaded from GBIF are compiled from a large variety of data collectors and data contributors ( 1900 in this case; < doi:10.15468/dl.om11gi > for a detailed list of data contributors), mainly from herbarium specimens and human observations. The advantages of using GBIF for biodiversity analyses are unprecedented amounts of occurrences and an exhaustive spatial coverage, but concerns on data quality and the impact of sampling biases have been raised (Maldonado et al 2015, Meyer et al 2016. We addressed the issues of geographic and taxonomic data quality by intensive, automated cleaning.…”

A balance of winners and losers in the Anthropocene

“…Hence, to tackle questions about biodiversity change across the globe, we must rely on data from compilations of individual ecological studies and monitoring efforts. Although we recognise that ecological research effort has been geographically biased (Martin et al 2012;Gonzalez et al 2016;Meyer et al 2016;Vellend et al 2017), this bias affects both population-and assemblage-level studies, and is ultimately driven by the geographic distribution of ecological effort.…”

Anthropogenic threats can have cascading homogenizing effects on the phylogenetic and functional diversity of tropical ecosystems

“…However, fine-scale, comprehensive presence-absence data are unavailable for many of the plants, mammals, and birds of Madagascar. Furthermore, point occurrence (presence) data are often subject to numerous biases and uncertainties (Meyer et al 2016, Park and Davis 2017, Daru et al 2018. Thus, to accurately infer patterns of biodiversity and community structure, we generated presence-absence matrices at the scale of administrative provinces (n = 6) and regions (n = 22), using species checklists, documented observations, and verified range distributions.…”