Although an enormous number of plant and animal species have been directly observed and recorded in online databases, the spatial variation in the composition of the microbiome remains relatively largely unknown. In this study, for the first time, we demonstrate mapping of the relative abundance of the soil microbiome for three terrestrial ecosystems across North America (savanna, boreal and tundra) using airborne image spectroscopy and environmental DNA (eDNA) data. We identified field plots of publicly available eDNA data co-occurring with AVIRIS-NG hyperspectral imagery. An eDNA processing pipeline was developed to generate a consistent profile of the relative abundance for thousands of microbiome operational taxonomic units (OTU) and 225 microbiome families. Using Linear Discriminate Analysis (LDA) scores for the eDNA data, we identified 81 families with the greatest explanatory power based on the community composition between the three ecosystems. A case study example demonstrates our conceptual approach by selecting a dominant and functionally important bacterial family for each ecosystem, with each family representing a specific biomarker. A partial least square regression (PLSR) was applied to spatially predict the relative abundance of each bacteria family (in the boreal, tundra, and savanna ecosystems) from the hyperspectral imagery. For the boreal, Pseudomonadaceae is shown to be a dominant family taxon, as it is a saprophytic family specialised in decomposing the dense organic matter of boreal forest soils. Members of an understudied family of Acidobacteria, so far indicated as AKIW659, are abundant in acidic Arctic soils and peat bogs. Finally, the Micromonosporaceae are dominant and functionally important in drier regions with grass-tree dominated woodlands, being a member of Actinobacteria with a high relative abundance in soils with high carbon content and nitrate leaching. We demonstrate for the first time how the spatial prediction of relative abundance of these bacteria taxa based on remote sensing, showing patterns of the soil microbiome biodiversity and ecosystem function within and across the three ecosystems.