Species inventories based on various data sources have been widely used in biodiversity research, conservation policy formulation, reserve designation and biodiversity resource management. In this paper, we explored the relationships of species inventories obtained from different sources and whether they would affect the inference of biodiversity patterns and their environmental drivers.
We compiled the species inventories from different data sources (observational data including large amounts of citizen-based observational records and digitalized specimens, and avifauna data extracted from avifaunas which mainly integrated professional-based species surveys, expert knowledge and documentary records) at the prefectural level in China. Then we explored the relationships of different inventories and compared the correlations between the taxonomic, phylogenetic, functional diversity calculated from different datasets and the environmental factors.
The results showed that the avifauna datasets contributed more additional species to the combined species inventories when the species richness was relatively low and vice versa. Species inventories integrated from two different data sources formed complementary relationship rather than nested or totally different relationships. In addition, the species inventories based on observational data had no obvious disadvantage or were even better at inferring the biodiversity patterns than those based on avifauna data. The stepwise multiple regression analyses showed that the best models were the ones using the species inventories combined by observational and avifauna dataset, and the best models built with different datasets included inconsistent environmental variables. Thus, the species inventories from different data sources will indeed affect the inference of the correlations between taxonomic diversity, phylogenetic diversity, functional diversity and environmental factors. Moreover, although it may be more reliable to use a combined species inventory to analyze the relationship between diversity indices and environmental factors, individualized improvement schemes should be proposed for different data sources to fill the data gaps.