Blind spots in global soil biodiversity and ecosystem function research

The study addresses how well current macroecological research represents global soil biodiversity and ecosystem functions, and identifies blind spots that hinder understanding of biodiversity–ecosystem function (BEF) relationships in soils. While soils host vast biodiversity that underpins key ecosystem services such as climate regulation, nutrient cycling, and food production, global-scale analyses of soil BEF relationships are scarce compared to aboveground systems. Existing soil data suffer from spatial, taxonomic, functional, and temporal gaps, with strong biases toward temperate regions and certain taxa (bacteria and fungi). The purpose is to compile and assess global macroecological studies of soil taxa and functions, evaluate their coverage of environmental gradients (soil, climate, topography, land cover), and prioritize actions to close knowledge gaps critical for conservation and policy.

Prior work has demonstrated the importance of soil biodiversity for ecosystem multifunctionality based on local and biome-specific studies. In contrast, aboveground biodiversity has extensive global datasets and BEF analyses, albeit with known biases. Soil biodiversity information is fragmented across literature, collections, and non-interoperable databases (e.g., EDAPHOBASE, global ants databases, Earth Microbiome Project), with limited representation on global platforms like GBIF. Even for bacteria and fungi, taxonomic depth is insufficient. Temporal data are particularly lacking, impeding assessments of trends and responses to global change. Recent syntheses for specific taxa (nematodes, earthworms) show progress but retain methodological heterogeneity and temperate-region biases. These shortcomings limit capacity to generalize patterns, model distributions, and inform governance and land management.

The authors conducted a comprehensive literature search (Web of Knowledge, Nov 2018; 1945–2018) using broad keywords for global/continental soil biodiversity and function. Inclusion criteria required studies on soil taxa and/or soil ecosystem functions spanning more than one continent or across continental scales, with georeferenced sampling locations. From 1203 initial studies, 62 global-scale studies (1995–2018) were retained and classified by nine soil taxonomic groups and five ecosystem functions; 45 provided usable georeferenced site data (~72.6% of suitable studies). They collated 17,186 unique sampling sites (WGS84), excluding Antarctica and Greenland. To assess environmental representativeness, they compared sample-based distributions of 15 environmental and diversity variables (soil properties: carbon, pH, sand/silt/clay, soil type; climate: mean temperature and precipitation, seasonality, aridity, potential evapotranspiration; geomorphology: elevation; land cover; vascular plant richness) against global distributions. Continuous variables were binned using Jenks natural breaks; categorical variables retained original classes. They also combined variables into three groups (land cover: land cover, plant diversity, elevation; soils: carbon, sand, pH; climate: mean temperature and precipitation plus their seasonality) to detect redundancy and gaps. Representation was quantified across global biomes and IPBES regions. A multivariate coverage analysis used Mahalanobis distance (with chi-squared quantiles) to map how environmental space around each study’s centroid was covered, identifying outliers and underrepresented conditions. The analysis focused on coverage rather than true diversity completeness due to limited access to raw biodiversity/functional data and standardization issues.

- Dataset compiled 17,186 sampling sites: 12,915 for biodiversity, 3,318 for ecosystem functions, and 977 for biomass; aboveground databases (e.g., PREDICTS ~29,000 sites) are larger. - Taxa coverage: Bacteria (N=3,453), fungi (N=1,687), and Formicoidea (ants; N=3,024) together comprise 48.8% of soil biodiversity records; Rotifera (N=41), Collembola (N=27), and Acari (N=10) are severely underrepresented. - Functions coverage: Soil respiration dominates with 78.8% (N=2,616) of function records; nutrient cycling and secondary productivity functions are weakly represented. - Overlap of biodiversity and function is minimal: only ~0.3% of all sites (≈63–67 sites) have both biodiversity and function data, often with taxonomic or temporal mismatches. - Strong spatial bias: Temperate biomes are overrepresented; 62% of function sites are in temperate systems. Tropical/subtropical, montane grasslands, mangroves, flooded grasslands and savannas, hyper-arid areas are sparsely sampled. - Environmental blind spots: Extreme soil carbon (very high/low) and several soil types (e.g., durisols, stagnosols, umbrisols) are underrepresented. Soil texture and pH are better represented except at extremes. - Climate space coverage is poor: low/high potential evapotranspiration/aridity, high seasonality, low precipitation, and thermal extremes (very hot/cold) are largely missing; 59.6% of global climate condition combinations are not covered by any study. - Land cover bias: Sites near urban areas are overrepresented. Lichens, mosses, bare areas, and shrublands are neglected, particularly in function studies. - Regional environmental coverage: Most studies cover less than 50% of environmental space in most regions; relatively higher coverage in Central and West Europe and the Caribbean (both biodiversity and function), and Central/Northeast Asia plus North/South America for functions; lowest in North Africa and West Asia. - Temporal data are almost absent: most studies are single sampling events without resampling, hindering trend analysis and attribution to global change drivers. - Data sharing and methodological heterogeneity limit synthesis: poor interoperability and diverse sampling/analytical methods constrain macroecological analyses and cross-study integration.

The assessment reveals that current global soil macroecological knowledge is constrained by substantial spatial, environmental, taxonomic, functional, and temporal blind spots. These biases impede robust inference of soil biodiversity patterns, modeling across environmental gradients, and evaluation of biodiversity–ecosystem function relationships at macroecological scales. The overrepresentation of temperate regions and specific taxa/functions, together with missing climate extremes and soil types, undermines generality and predictive capacity, especially under accelerating global change. Minimal co-location of biodiversity and function data prevents comprehensive BEF analyses. The findings underscore the need for standardized protocols, improved data mobilization and interoperability, and coordinated monitoring that jointly captures biodiversity and functions across diverse environmental spaces, including neglected regions and conditions. Addressing legal and logistical barriers (e.g., ABS/Nagoya constraints) and fostering international, multilateral frameworks will be critical to enable global soil ecology efforts and to inform policy and land management with reliable, globally representative evidence.

This study synthesizes global soil macroecological datasets to identify major blind spots across space, environment, taxa, functions, and time, and quantifies the limited overlap between biodiversity and function measurements. It provides priority targets for future research: expand sampling in underrepresented regions (tropics, drylands, high latitudes, hyper-arid and montane systems), environmental extremes (soil carbon, climatic extremes), and neglected taxa/functions; increase co-located biodiversity–function sampling; and develop temporally explicit monitoring through standardized global frameworks. The authors advocate two complementary pathways: systematic data mobilization via open platforms (e.g., GBIF) with harmonized metadata and standards, and coordinated, standardized global sampling campaigns. They also call for multilateral solutions to ABS constraints and capacity building to empower global participation. These actions will enhance predictive modeling of soil ecosystems and support national and global conservation and sustainability goals.

- Strong geographic bias toward temperate regions; tropical/subtropical and arid regions are under-sampled, limiting global generalization. - Severe taxonomic and functional biases (focus on bacteria, fungi, soil respiration) and sparse coverage for many soil taxa and functions. - Minimal co-location of biodiversity and function data (~0.3% overlap), often with thematic/temporal mismatches. - Predominantly single-time sampling with scarce temporal replication, preventing trend analyses and causal attribution under global change. - Methodological heterogeneity (sampling, extraction, molecular protocols) and poor data interoperability constrain cross-study synthesis and certain analyses (e.g., compositional turnover). - Potential analytical bias from treating coverage globally (not continent-partitioned), which may overrepresent some conditions while underrepresenting others. - Legal and logistical barriers (ABS/Nagoya Protocol) impede international sample and data sharing, particularly affecting global initiatives.