Global meta-analysis shows reduced quality of food crops under inadequate animal pollination

Animal pollination is critical for the reproduction of flowering plants and supports many cultivated food crops, particularly fruits, vegetables, nuts, and spices. While staple crops are often wind- or self-pollinated, many animal-pollinated crops provide essential micronutrients that contribute to healthy diets. Recent declines in pollinator diversity and abundance, combined with expanding cultivation of pollination-dependent crops, raise concerns about pollination deficits that threaten not only crop yields and their stability but also aspects of human health. Prior syntheses have largely quantified pollination effects on yield metrics (e.g., fruit set, seed weight) and yield stability, leaving a gap regarding the influence of pollination on multi-dimensional food quality, including organoleptic and nutritional traits. This study aims to fill that gap by providing a global quantitative assessment of how animal pollination affects food quality and marketability across major crops and geographies, and by evaluating whether current pollination services are sufficient or whether deficits are detectable.

Global studies demonstrate that pollinators (managed and wild) enhance crop yield and stabilize yields across space and time, highlighting their economic and nutritional importance. However, most meta-analyses have focused on yield-related outcomes, with limited comprehensive assessment of quality traits. Isolated studies show that pollinators can improve size, shape, firmness, shelf life, and commercial grade, suggesting hormonal mechanisms linked to fertilization. Evidence on nutrient content is mixed, potentially due to dilution effects as fruits grow larger and due to environmental and agronomic influences. There is also interest in pollination deficits as indicators of declining services, though managed pollinators can buffer deficits in some systems. The literature lacks broad synthesis on quality outcomes across diverse crop types, pollinator groups, environments, and climates, motivating the present global meta-analysis.

The authors conducted a systematic literature survey in Scopus and ISI Web of Science Core Collection, performing two searches using PICO-based strings targeting crops, pollinators (including insects and vertebrates), and quality outcomes. The first search (Jan 28, 2021) covered 1960 (Scopus) and 1985 (WoS) to 2021; the second (Feb 23, 2023) updated through 2023. Grey literature was added via Google Scholar searches (first 600 results by relevance in both 2021 and 2023). Non-English studies with English abstracts were considered. PRISMA and PRISMA-EcoEvo guidelines documented selection steps. Inclusion criteria: manipulative experiments with at least two pollination treatments; at least one edible-part quality trait measured (e.g., physical traits, shelf life, commercial grade, taste, macro-/micronutrients); fruit production under all relevant treatments; primary data only. Exclusions included studies reporting solely fruit set, yield, seed traits without quality measures. Pollination metrics: (1) Pollination service = open pollination vs pollinator exclusion (animals excluded; wind allowed). (2) Pollination deficit = hand pollination vs open pollination, using hand pollination as a proxy for maximal pollen deposition (saturation). Data extraction: Means, standard deviations (derived from SE, CI, or range when needed), and sample sizes were extracted from text, tables, figures, or raw data (when available; missing data requested for papers from 2000 onward). Negative quality measures were transformed to positive counterparts when possible. Effect size was the log response ratio (lnRR = ln(X_treatment/X_control)); percentage change computed as (e^{lnRR} − 1)*100. Only effect sizes with variance were included; weighting by inverse variance. Datasets: 190 studies total. For pollination service: 1197 effect sizes from 153 publications. For pollination deficit: 682 effect sizes from 86 publications. Studies spanned 48 countries and 48 crops (1968–2023), including field and greenhouse experiments. Most studies measured organoleptic traits; nutritional traits were less frequent. Moderators extracted: quality trait (pooled into organoleptic vs nutritional; and into seven categories: size, shape, external appearance and taste, firmness, commercial grade, macronutrients, micronutrients), pollinator group (honeybee Apis, bumblebee Bombus, pollinator community, other single species), crop type (vegetables, fruits, nuts, edible oil/proteinaceous, stimulants, spices/condiments), experimental scale (flower, branch, plant), cropping environment (field, greenhouse), climate (tropical, subtropical, temperate), and publication year. Meta-analytic modeling: Multi-level (mixed-effects) meta-analyses using rma.mv (metafor in R). Random-effects structures identified via AIC: for pollination service, crossed random effects of publication ID and year of experiment plus an effect-size-level residual term; for pollination deficit, publication ID plus residual term. Variance–covariance matrices accounted for non-independence among clustered outcomes (assumed sampling correlation 0.5; sensitivity tested 0.8) and shared controls (adjusted effective sample sizes). Fixed-effect moderators were tested in separate models due to collinearity and incomplete combinations; models compared to null via LRT and AIC; marginal R² reported for significant moderators. Omnibus tests (Q statistics), coefficient CIs, and within-group estimates were presented. Bias and sensitivity: Funnel plots of residuals showed no extreme asymmetry; modified Egger’s test indicated no publication bias in service or deficit datasets. Time-lag bias was not detected. Sensitivity analyses excluding high-contribution studies, influential points (Cook’s distance), alternative correlation assumptions, and alternative clustering for variance–covariance construction showed robust results.

- Overall pollination service effect: Animal pollination improved overall crop quality by 23% (95% CI: 16% to 30%; p < 0.001). - By quality domain: Organoleptic traits increased by 27% (CI: 20% to 34%); nutritional traits increased by 7% (CI: 1% to 14%). - Detailed traits: Largest improvements in size, shape, and commercial grade; firmness and micronutrients improved with lower statistical support (LRT p < 0.001; marginal R² up to 0.090). - Moderators (service): No significant differences among pollinator groups (LRT p = 0.312), between fruits and vegetables (LRT p = 0.344), across experimental scales (LRT p = 0.296), cropping environments (LRT p = 0.335), or climates (LRT p = 0.950). - Pollination deficit: Overall effect not significant (mean = 2%; 95% CI: −2% to 5%; p = 0.390). Weak positive effect on organoleptic traits, particularly size (TM p = 0.054). Negative effect on nutritional traits, significant for macronutrients (TM p = 0.045) and marginal for micronutrients (TM p = 0.085). Effects more evident in fruit crops (TM p = 0.056); no improvement with pollinator group moderator (LRT p = 0.212). No variation by scale (LRT p = 0.334), environment (LRT p = 0.550), or climate (LRT p = 0.215). - Crop groups: Pollination did not increase quality in stimulant crops, nuts, and spices. - Geographic and temporal coverage: 48 countries; studies from 1968–2023 with increasing publications over time; both field and greenhouse settings.

The meta-analysis demonstrates that animal pollination substantially enhances organoleptic and marketability-related attributes of fruits and vegetables—such as size, shape, firmness, shelf life, and commercial grade—while having smaller and more variable effects on nutritional content. These results address the research question by quantifying pollination’s contribution to multi-dimensional food quality beyond yield, showing that approximately one-fourth of overall quality is attributable to animal pollination across diverse crops and contexts. Mechanistically, improvements in organoleptic traits are consistent with fertilization-driven phytohormonal pathways: auxin-mediated promotion of gibberellins after successful ovule fertilization stimulates seed formation and fruit growth, increasing size and preventing malformations, and enhancing firmness and storability. Conversely, the weaker or negative responses in nutrient concentrations can be explained by dilution effects due to rapid cell expansion and higher water content as fruit size increases, compounded by variability in synthesis and instability of certain compounds and by environmental and agronomic influences. The absence of a significant overall pollination deficit, alongside a weak signal for size, suggests that current pollinator activity—wild and/or managed—often suffices to maximize quality, though small deficits may reflect landscape-level service declines. Managed pollinators may mitigate deficits in greenhouse and monoculture systems, while reliance on them underscores vulnerability where wild pollinators are reduced. Implications are substantial for agriculture and the food industry: quality standards and consumer preferences are strongly tied to appearance and perishability. Pollination-related improvements can increase market access and value, whereas suboptimal quality elevates waste along supply chains, with economic and environmental costs. Integrating quality benefits with yield in valuation can reveal large economic contributions of pollinators (e.g., strawberries in the EU). Protecting pollinator services is therefore vital for food security and market sustainability.

This global synthesis shows that animal pollination significantly enhances the organoleptic and marketable quality of many food crops, with smaller and context-dependent effects on nutritional content. Overall, current pollination levels typically suffice to achieve near-optimal quality, though weak signals of deficits highlight the need for continued conservation and management of pollinators. The study contributes a comprehensive, multi-level meta-analytic assessment of pollination impacts on food quality across 48 crops and 48 countries, quantifying trait-specific effects and testing multiple moderators. Future research should: (1) expand evidence on understudied high-value crops (stimulants, spices); (2) evaluate cultivar-level variability; (3) disentangle roles of vertebrate pollinators; (4) examine context-dependency and interactions with environmental and agronomic factors using manipulative designs; and (5) integrate consumer perceptions and market standards into valuations. Policy and management should enhance floral and nesting resources and reduce pressures on wild and managed pollinators across landscapes to safeguard both yield and quality outcomes.

- Evidence gaps for stimulants and spices limited inference for these economically important groups. - Cultivar effects, a known source of variability in quality and pollination responses, could not be tested due to insufficient replication. - Strong imbalance between insect and vertebrate pollinator studies prevented taxon-specific conclusions for vertebrates. - Context-dependency and interactions with other biotic/abiotic factors remain poorly resolved. - Collinearity and incomplete combinations among moderators necessitated separate models without interactions. - Nutritional outcomes showed greater uncertainty, potentially influenced by dilution effects and environmental/agronomic variability.