Abstract To predict how biodiversity will respond to global change, it is crucial to understand the relative roles of abiotic drivers and biotic interactions in driving associations between the biodiversity of disparate taxa. It is particularly challenging to understand diversity–diversity links across domains and habitats, because data are rarely available for multiple above‐ and below‐ground taxa across multiple sites. Here, we analyse data from a unique biodiversity data set gathered across a variety of oceanic temperate terrestrial habitats in Wales, comprising 300 sites with co‐located soil microbial, plant, bird and pollinator surveys along with climate and soil physicochemical information. Soil groups are analysed using metabarcoding of the 16S, ITS1 and 18S DNA regions, allowing in‐depth characterisation of microbial and soil animal biodiversity. We explore biodiversity relationships along three aspects of community composition: First, we assess correlation between the alpha diversity of different groups. Second, we assess whether biotic turnover between sites is correlated across different groups. Finally, we investigate the co‐occurrence of individual taxa across sites. In each analysis, we assess the contribution of linear or nonlinear environmental effects. We find that a positive correlation between alpha diversity of plants, soil bacteria, soil fungi, soil heterotrophic protists, bees and butterflies is in fact driven by complex nonlinear responses to abiotic drivers. In contrast, environmental variation did not account for positive associations between the diversity of plants and both birds and AM fungi, suggesting a role for biotic interactions. Both the diversity and taxon‐level associations between the differing soil groups remained even after accounting for nonlinear environmental gradients. Above‐ground, spatial factors played larger roles in driving biotic communities, while linear environmental gradients were sufficient to explain many group‐ and taxon‐level relationships. Synthesis. Our results show how nonlinear responses to environmental gradients drive many of the relationships between plant biodiversity and the biodiversity of above‐ and below‐ground biological communities. Our work shows how different aspects of biodiversity might respond nonlinearly to changing environments and identifies cases where management‐induced changes in one community could either influence other taxa or lead to loss of apparent biological associations.