Ecology

Rook, G. A. W. Review series on helminths, immune modulation and the hygiene hypothesis: The broader implications of the hygiene hypothesis. Immunology 126, 3–11 (2009).CAS 
Article 

Google Scholar 
Von Hertzen, L., Hanski, I. & Haahtela, T. Natural immunity. Biodiversity loss and inflammatory diseases are two global megatrends that might be related. EMBO Rep. 12, 1089–1093 (2011).Article 

Google Scholar 
Von Hertzen, L. & Haahtela, T. Disconnection of man and the soil: Reason for the asthma and atopy epidemic?. J. Allergy Clin. Immunol. 117, 334–344 (2006).Article 

Google Scholar 
Hanski, I. et al. Environmental biodiversity, human microbiota, and allergy are interrelated. Proc. Natl. Acad. Sci. U. S. A. 109, 8334–8339 (2012).ADS 
CAS 
Article 

Google Scholar 
Haahtela, T. et al. Immunological resilience and biodiversity for prevention of allergic diseases and asthma. Allergy Eur. J. Allergy Clin. Immunol. https://doi.org/10.1111/all.14895 (2021).Article 

Google Scholar 
Rook, G. A. W. et al. Mycobacteria and other environmental organisms as immunomodulators for immunoregulatory disorders. Springer Semin. Immunopathol. 25, 237–255 (2004).CAS 
Article 

Google Scholar 
Fyhrquist, N. et al. Acinetobacter species in the skin microbiota protect against allergic sensitization and inflammation. J. Allergy Clin. Immunol. 134, 1301-1309.e11 (2014).CAS 
Article 

Google Scholar 
Ottman, N. et al. Soil exposure modifies the gut microbiota and supports immune tolerance in a mouse model. J. Allergy Clin. Immunol. 143, 1198-1206.e12 (2019).CAS 
Article 

Google Scholar 
Nurminen, N. et al. Nature-derived microbiota exposure as a novel immunomodulatory approach. Fut. Microbiol. 13, 737–744 (2018). CAS 
Article 

Google Scholar 
Shaffer, M. & Lozupone, C. Prevalence and source of fecal and oral bacteria on infant, child, and adult hands. mSystems 3, 1–12 (2018).Article 

Google Scholar 
Grönroos, M. et al. Short-term direct contact with soil and plant materials leads to an immediate increase in diversity of skin microbiota. MicrobiologyOpen https://doi.org/10.1002/mbo3.645 (2019).Article 

Google Scholar 
Roslund, M. I. et al. Biodiversity intervention enhances immune regulation and health-associated commensal microbiota among daycare children. Sci. Adv. 6, 7–105 (2020).Article 

Google Scholar 
Roslund, M. I. et al. Long-term biodiversity intervention shapes health-associated commensal microbiota among urban day-care children. Environ. Int. 157, 7008 (2021).Article 

Google Scholar 
Lax, S. et al. Longitudinal analysis of microbial interaction between humans and the indoor environment. Science (80-.). 345, 1048–1052 (2014).ADS 
CAS 
Article 

Google Scholar 
Flies, E. J., Clarke, L. J., Brook, B. W. & Jones, P. Urbanisation reduces the abundance and diversity of airborne microbes-but what does that mean for our health? A systematic review. Sci. Total Environ. 738, 140337 (2020).ADS 
CAS 
Article 

Google Scholar 
Ege, M. J. et al. Exposure to environmental microorganisms and childhood asthma.. Science 364, 701–709 (2011).CAS 

Google Scholar 
Li, H. et al. Spatial and seasonal variation of the airborne microbiome in a rapidly developing city of China. Sci. Total Environ. 665, 61–68 (2019).ADS 
CAS 
Article 

Google Scholar 
Chase, J. et al. Geography and location are the primary drivers of office microbiome composition. mSystems 1, 1–18 (2016).
Google Scholar 
Danko, D. et al. A global metagenomic map of urban microbiomes and antimicrobial resistance. Cell 184, 3376-3393.e17 (2021).CAS 
Article 

Google Scholar 
Hui, N. et al. Soil microbial communities are shaped by vegetation type and park age in cities under cold climate. Environ. Microbiol. 19, 1281–1295 (2017).Article 

Google Scholar 
Mhuireach, G. et al. Urban greenness influences airborne bacterial community composition. Sci. Total Environ. 571, 680–687 (2016).ADS 
CAS 
Article 

Google Scholar 
Franzetti, A., Gandolfi, I., Gaspari, E., Ambrosini, R. & Bestetti, G. Seasonal variability of bacteria in fine and coarse urban air particulate matter. Appl. Microbiol. Biotechnol. 90, 745–753 (2011).CAS 
Article 

Google Scholar 
Mhuireach, G., Wilson, H. & Johnson, B. R. Urban aerobiomes are influenced by season, vegetation, and individual site characteristics. EcoHealth 18, 331–344 (2021).Article 

Google Scholar 
Mahnert, A., Moissl-Eichinger, C. & Berg, G. Microbiome interplay: Plants alter microbial abundance and diversity within the built environment. Front. Microbiol. 6, 1–11 (2015).Article 

Google Scholar 
Ruokolainen, L. et al. Green areas around homes reduce atopic sensitization in children. Allergy Eur. J. Allergy Clin. Immunol. 70, 195–202 (2015).CAS 
Article 

Google Scholar 
Kirjavainen, P. V. et al. Farm-like indoor microbiota in non-farm homes protects children from asthma development. Nat. Med. 25, 1089–1095 (2019).CAS 
Article 

Google Scholar 
Nurminen, N. et al. Land cover of early-life environment modulates the risk of type 1 diabetes. Diabetes Care 44, 1506–1514 (2021).Article 

Google Scholar 
Parajuli, A. et al. Yard vegetation is associated with gut microbiota composition. Sci. Total Environ. 713, 136707 (2020).ADS 
CAS 
Article 

Google Scholar 
Köberl, M., Dita, M., Martinuz, A., Staver, C. & Berg, G. Members of Gammaproteobacteria as indicator species of healthy banana plants on Fusarium wilt-infested fields in Central America. Sci. Rep. 7, 1–9 (2017).Article 

Google Scholar 
Delanghe, L. et al. The role of lactobacilli in inhibiting skin pathogens. Biochem. Soc. Trans. 5, 617–627. https://doi.org/10.1042/bst20200329 (2021).CAS 
Article 

Google Scholar 
George, F. et al. Occurrence and dynamism of lactic acid bacteria in distinct ecological niches: A multifaceted functional health perspective. Front. Microbiol. 9, 1–15 (2018).CAS 
Article 

Google Scholar 
Yu, A. O., Leveau, J. H. J. & Marco, M. L. Abundance, diversity and plant-specific adaptations of plant-associated lactic acid bacteria. Environ. Microbiol. Rep. 12, 16–29 (2020).CAS 
Article 

Google Scholar 
Parajuli, A. et al. Urbanization reduces transfer of diverse environmental microbiota indoors. Front. Microbiol. 9, 1405 (2018).Article 

Google Scholar 
Parajuli, A. et al. The abundance of health-associated bacteria is altered in PAH polluted soils—Implications for health in urban areas?. PLoS One 7, 1–18. https://doi.org/10.1371/journal.pone.0187852 (2017).CAS 
Article 

Google Scholar 
Vari, H. K. et al. Associations between land cover categories, gaseous PAH levels in ambient air and endocrine signaling predicted from gut bacterial metagenome of the elderly. Chemosphere 265, 1559 (2021).Article 

Google Scholar 
Orsini, F., Kahane, R., Nono-Womdim, R. & Gianquinto, G. Urban agriculture in the developing world: A review. Agron. Sustain. Dev. 33, 695–720 (2013).Article 

Google Scholar 
Hui, N. et al. Diverse environmental microbiota as a tool to augment biodiversity in urban landscaping materials. Front. Microbiol. 10, 1–10 (2019).CAS 
Article 

Google Scholar 
Puhakka, R. et al. Greening of daycare yards with biodiverse materials affords well-being, play and environmental relationships. Int. J. Environ. Res. Public Health 16, 2948 (2019).Article 

Google Scholar 
Burmeister, A. R. & Marriott, I. The interleukin-10 family of cytokines and their role in the CNS. Front. Cell. Neurosci. 12, 1–13 (2018).Article 

Google Scholar 
Opal, S. M. & DePalo, V. A. Anti-inflammatory cytokines. Chest 117, 1162–1172 (2000).CAS 
Article 

Google Scholar 
Kuwabara, T., Ishikawa, F., Kondo, M. & Kakiuchi, T. The role of IL-17 and related cytokines in inflammatory autoimmune diseases. Mediators Inflamm. 2017, 4598 (2017).Article 

Google Scholar 
Li, M. O., Wan, Y. Y., Sanjabi, S., Robertson, A. K. L. & Flavell, R. A. Transforming growth factor-β regulation of immune responses. Annu. Rev. Immunol. 24, 99–146 (2006).CAS 
Article 

Google Scholar 
Prudhomme, G. J. & Piccirillo, C. A. The inhibitory effects of transforming growth factor-beta-1 (TGF-β1) in autoimmune diseases. J. Autoimmun. 14, 23–42 (2000).CAS 
Article 

Google Scholar 
Esebanmen, G. E. & Langridge, W. H. R. The role of TGF-beta signaling in dendritic cell tolerance. Immunol. Res. 65, 987–994 (2017).CAS 
Article 

Google Scholar 
Honkanen, J. et al. IL-17 immunity in human type 1 diabetes. J. Immunol. 185, 1959–1967 (2010).CAS 
Article 

Google Scholar 
Torpy, F. et al. Testing the single-pass VOC removal efficiency of an active green wall using methyl ethyl ketone (MEK). Air Qual. Atmos. Heal. 11, 163–170 (2018).CAS 
Article 

Google Scholar 
Roslund, M. I. et al. Endocrine disruption and commensal bacteria alteration associated with gaseous and soil PAH contamination among daycare children. Environ. Int. 130, 104894 (2019).CAS 
Article 

Google Scholar 
Schloss, P. D., Gevers, D. & Westcott, S. L. Reducing the effects of PCR amplification and sequencing Artifacts on 16s rRNA-based studies. PLoS One 6, 1789 (2011).
Google Scholar 
Kozich, J., Westcott, S., Baxter, N., Highlander, S. & Schloss, P. Development of a dual-index sequencing strategy and curation pipeline for analyzing amplicon sequence data on the MiSeq Illumina sequencing platform. Appl. Environ. Microbiol. 79, 5112–5120 (2013).ADS 
CAS 
Article 

Google Scholar 
Soininen, L., Grönroos, M., Roslund, M. I. & Sinkkonen, A. Long-term storage affects resource availability and occurrence of bacterial taxa linked to pollutant degradation and human health in landscaping materials. Urban For. Urban Green. 60, 1789 (2021).Article 

Google Scholar 
Quast, C. et al. The SILVA ribosomal RNA gene database project: Improved data processing and web-based tools. Nucleic Acids Res. 41, 590–596 (2013).Article 

Google Scholar 
Huse, S. M., Welch, D. M., Morrison, H. G. & Sogin, M. L. Ironing out the wrinkles in the rare biosphere through improved OTU clustering. Environ. Microbiol. 12, 1889–1898 (2010).CAS 
Article 

Google Scholar 
Edgar, R. C., Haas, B. J., Clemente, J. C., Quince, C. & Knight, R. UCHIME improves sensitivity and speed of chimera detection. Bioinformatics 27, 2194–2200 (2011).CAS 
Article 

Google Scholar 
Wang, Q., Garrity, G. M., Tiedje, J. M., Cole, J. R. & Al, W. E. T. Naive Bayesian classifier for rapid assignment of rRNA sequences into the new bacterial taxonomy. Appl. Environ. Microbiol. 73, 5261–5267 (2007).ADS 
CAS 
Article 

Google Scholar 
R Core Team. R: A Language and Environment for Statistical Computing. (2020).Oksanen, J. et al. vegan: Community Ecology Package. (2019).Huang, F. L. Alternatives to multilevel modeling for the analysis of clustered data. J. Exp. Educ. 84, 175–196 (2016).Article 

Google Scholar 
Moen, E. L., Fricano-Kugler, C. J., Luikart, B. W. & O’Malley, A. J. Analyzing clustered data: Why and how to account for multiple observations nested within a study participant?. PLoS ONE 11, 1–17 (2016).
Google Scholar 
Twisk, J. et al. Different ways to estimate treatment effects in randomised controlled trials. Contemp. Clin. Trials Commun. 10, 80–85 (2018).Chapat, L., Chemin, K., Dubois, B., Bourdet-Sicard, R. & Kaiserlian, D. Lactobacillus casei reduces CD8+ T cell-mediated skin inflammation. Eur. J. Immunol. 34, 2520–2528 (2004).CAS 
Article 

Google Scholar 
Kaur, K. & Rath, G. Formulation and evaluation of UV protective synbiotic skin care topical formulation. J. Cosmet. Laser Ther. 21, 332–342 (2019).Article 

Google Scholar 
Rong, J. et al. Skin resistance to UVB-induced oxidative stress and hyperpigmentation by the topical use of Lactobacillus helveticus NS8-fermented milk supernatant. J. Appl. Microbiol. 123, 511–523 (2017).CAS 
Article 

Google Scholar 
Yuan, J. et al. Microbial volatile compounds alter the soil microbial community. Environ. Sci. Pollut. Res. 24, 22485–22493 (2017).CAS 
Article 

Google Scholar 
Abis, L. et al. Reduced microbial diversity induces larger volatile organic compound emissions from soils. Sci. Rep. 10, 1–15 (2020).Article 

Google Scholar 
Duffy, E. & Morrin, A. Endogenous and microbial volatile organic compounds in cutaneous health and disease. TrAC Trends Anal. Chem. 111, 163–172 (2019).CAS 
Article 

Google Scholar 
Lemfack, M. C. et al. Novel volatiles of skin-borne bacteria inhibit the growth of Gram-positive bacteria and affect quorum-sensing controlled phenotypes of Gram-negative bacteria. Syst. Appl. Microbiol. 39, 503–515 (2016).CAS 
Article 

Google Scholar 
Ahmed, M. & Gaffen, S. L. IL-17 in obesity and adipogenesis. Cytokine Growth Factor Rev. 21, 449–453 (2010).CAS 
Article 

Google Scholar  More

Pollution gradients, soil series, and glasshouse conditions for the empirical studySoils used for this experiment were collected from a wood preservation site (6 ha). In this site, the use of creosote, and various Cu-based salts has resulted in soil Cu-contamination over the whole site and large patches of PAH in smaller areas59. Former studies have shown the ecotoxic impact of this contamination on vegetation biomass, cover, richness and diversity60 and the reduction of soil enzymatic activity61. In February 2016, 65 kg of soil were collected from two areas of the site, one known for its contamination in Cu, and the other previously identified for its contamination in PAH59. An additional control soil was collected from the grassland next to the site. The control corresponds to an alluvial sandy soil (Fluviosol – Eutric Gleysols, World Reference Base for soil resources) and the contaminated soils were developed on this Fluviosol. Soils were transferred to a glasshouse nearby and spread out thinly on a tarpaulin for 15 days to ensure complete air drying. Ten samples of each soil were analysed for their PAH, Cu, C, N, and P concentrations. N concentration was higher in the control soils, while polluted soils showed higher P-availability (Supplementary Table 1). Regarding PAH, the 16 regulatory PAH were quantified. The range of soil properties and contamination values (889 ± 10 mg Cu.kgsoil−1 and 657 ± 331 mg PAH.kgsoil−1 for the first contaminated soil, 4276 ± 209 mg Cu.kg−1 and 3142 ± 419 mg PAH.kgsoil−1 for the second contaminated soil) showed Cu and PAH contamination in both cases, with higher contamination of the second soil. In this study, these soils are referred to as Cu-PAH soil and HIGH-Cu-PAH soil, respectively.To create the soil series, both soils were mixed by combining one third and two-thirds of air-dried contaminated soils with the control soil (March 2016) giving seven soil treatments: Control, 1/3 Cu-PAH, 2/3 Cu-PAH, Cu-PAH, 1/3 HIGH-Cu-PAH, 2/3 HIGH-Cu-PAH, and HIGH-Cu-PAH. Each of these seven soils was divided into 25 pots (10 × 10 × 15 cm) containing 800 mg of soil, giving a total of 175 pots. In order to inoculate all potted soils with similar micro-organism populations (especially Rhizobium populations), 1 g of control soil was added to the pots with undiluted polluted soils and vice versa. All pots were watered and weighed to determine their water holding capacity and left for 2 weeks to enable micro-organisms populations to react.To ensure that the environment was as homogeneous as possible during the whole experiment, a whitened glasshouse was used to favour diffuse and homogeneous solar radiation, and to limit differences in temperature and Vapour Pressure Deficit (VPD). In the case of a temperature increase above 25 °C, the glasshouse was also cooled by automatic ventilation and misting was used to avoid an increase of VPD above 1 kPa. An air temperature and humidity probe (U23 Prov V2 ®Hobo) was used to monitor VPD variations (kPa) during the whole course of the experiment.Plant cultivation and monitoring of plant developmentThe dwarf bean (Phaseolus vulgaris, cv. Oxinel, ® Vilmorin) was chosen as a model plant species because of its known plasticity of biomass allocation, both for wild and selected genotypes29. This plasticity has been detected in response to soil resources29 and also light regimes62. In addition, it is a species commonly used as bio-assays in ecotoxicology due to its sensitivity to soil pollution (see ref. 26 using soils from the same site as this study). Seeds of similar weight [0.22; 0.30 g] were selected to avoid large differences in seed reserves. After soaking for 4 h in tap water, three seeds were sown in each of the pots on March 21. Germination took 11–16.2 days depending on the soil treatment and this time increased with soil contamination. As a large majority of the seeds germinated, 1 seedling per pot was selected randomly and kept for the experiment. For each soil, we planned to harvest five plants at five different development stages (stage 1: end of cotyledon opening, stage 2: first trifoliate leaf, 3: second trifoliate leaf, 4: 3–4 trifoliate leaves, 5: 5–6 trifoliate leaves) giving 25 plants for each soil. Pots were watered every 2 or 3 days and weighed to maintain the water holding capacity (WHC) of soil at 60%63. Plants were harvested for analysis when they reached the desired development stage. In most phytotoxic soils, plants did not reach the fourth or fifth stage by the end of the experiment, thus they were harvested and classified into their real development stage at harvest (see for instance Table 1 shows that most plants of the High-Cu-PAH soil did not grow and were classified in development stage 1).Biomass partitioning, root and shoot (specific) areasOn the day of harvest, plant parts were separated (stem, leaves, and roots). Roots were washed gently with water and nodule numbers were counted. All organs were scanned and analysed to determine their area (software Winfolia for leaves and stems, WinRhizo for roots, Regents Instruments, Quebec, Canada). Then all plant samples were dried and weighed. The whole process determined the dry biomass of plant parts, their area, as well as Specific Leaf Area (SLA, cm2.g−1) and Specific Root Area (SRA, cm2.g−1). Analysis of SLA and SRA is important because: (i) they may also be involved in plant response along resource gradients to maintain a functional equilibrium. For instance, SLA can increase strength in the shade to maintain light capture area13, and SRA can increase to maintain water uptake during water stress64; and (ii) they may be impacted by soil pollution. A decrease in SRA is part of the root syndrome in phytotoxic soils because of decreasing root elongation and root thickening43.Indicators of resource acquisitionTo estimate light capture and potential acquisition of photo-chemical energy, we assessed chlorophyll a, b and another carotenoid synthesis by determining their leaf concentrations (See Supplementary Information for more details regarding corresponding methods).Water uptake and transpiration: To limit water evaporation, the soil in each pot was covered with a small plastic sheet (10 × 10 cm). At each watering, the mass of water added to maintain the pot at 60% of SWHC was recorded as the amount of water taken up and transpired since the last watering. The last 10 days before harvest were considered for analysis of plant transpiration. Independently of soil treatments, the amount of water transpired could be impacted by the leaf area (and the number of stomata), and by the variation of VPD occurring in the glasshouse despite cooling and water misting. Therefore, the weight of water transpired per leaf area, per day and per kPa of VPD (({{{{{{rm{mg}}}}}}}_{{{{{{{rm{H}}}}}}}_{2}0}).cm−2.day−1. kPaVPD−1) was calculated.Nitrogen acquisition and Symbiotic Nitrogen Fixation (SNF): To estimate N acquisition by plants, their leaf N concentration and an indicator of their SNF were determined. After drying and grinding (Retsch PM4 planetary grinder, Retsch, Haan, Germany), leaf N concentration was measured by an elemental analyser (NA 1500 NCS, Carlo Erba, Milan, Italy) for a subset of 112 samples encompassing all soil treatments and a wide range of plant size. Regarding SNF, most plants in this experiment did not initiate nodulation (because of toxicity and their small size). It was planned to use 15N soil labelling and the isotopic dilution method to estimate the efficiency of the SNF65, but it was not applicable in our experiment because of the low nodulation and the small amount of N derived from the atmosphere. Instead, at harvest, roots were cleaned gently and the number of nodules was determined.Statistics and reproducibilityAll pots were placed randomly in the glasshouse at the beginning of the experiment. They were moved randomly every 15 days to avoid any spatial dependency between sample units. All statistical analyses were performed with R software (R Core Team, 2016). Regarding the biomass of plant parts, stems and leaves were considered together in a single shoot compartment when analysing the results. Bean stems are also photosynthetically active, and separate analyses for leaves gave consistent results.All measured plant traits (SRA, SLA, water transpiration, nodule number, leaf N and chlorophyll concentration) can vary with plant ontogeny and plant size. Thus, variation of these traits (dependent variables) was analysed considering both plant size and soil treatments (explanatory variables). Shoot biomass was used as a surrogate for size for aboveground traits (SLA, water transpiration, leaf N, and chlorophyll concentration). Root biomass was used for belowground traits (root nodule number). Water transpiration was analysed by ANCOVA (shoot biomass as a regressor, soil treatment as a covariate). Leaf nitrogen and chlorophyll concentrations were first analysed by segmented linear modelling (segmented package) because of a radical change in the relationship between shoot biomass and these leaf traits at some size threshold. Then soil treatment effect on these traits was analysed on the residuals of the segmented relationships by ANOVA. Note that similar responses were observed for the different kinds of pigments, and only the results for chlorophylla+b concentrations are reported in this study. Similarly, we used a segmented linear model (segmented package) for the relationship between root nodule number and root biomass, and the soil treatment effect was analysed on this first model residuals. As to SLA and SRA, plants were grouped according to their shoot and root biomass tertile respectively. Then for each tertile, ANOVA was used to test the difference between SLA and SRA with soil treatment. When performing ANCOVAs, in the case of significant effects of soil treatment and interaction with plant size, post-hoc pairwise comparisons were used to test the difference of intercept or slopes between soil treatments (emmeans package). When performing ANOVAs, Post-hoc Tukey HSD pairwise comparisons were performed in case of significant effect of soil treatment. All variables were log-transformed when necessary to respect the condition of application of linear modelling.When investigating allometric relationships between root and shoot biomass, or root and shoot areas, the interest is related to the analysis of how root biomass (or area) scales against shoot biomass (or area), rather than predicting the value of one variable from another. Standard Major Axis (SMA) regression (smatr package) on log-transformed variables was used accordingly to study this allometric scaling and its changes with soil treatments66. When changes in α scaling exponent with soil treatment are significant, estimation of differences in β (proportionality coefficient) between treatments is not enabled by SMA regression66. In that case, after estimating α with SMA regression, we estimated β value for each soil treatment using non-linear least square modelling (see Supplementary Table 2) because changes in β values have a biological meaning in our context (delay in early root development).Meta-analysis of literature dealing with plant biomass partitioning and modelling of changes in root: shoot ratioCollection of published studies and case studies: we used the ISI Web of Science database to locate published studies on the effect of soil pollution on plant biomass partitioning. We entered a general query made using the combination of two phrases, one regarding biomass partitioning, and the other regarding soil pollution. We used several equivalent phrases regarding both terms, leading to the following query: (“biomass partitioning” OR “biomass partition” OR “biomass allocation” OR “root: shoot”) AND (“pollution” OR “contamination” OR “heavy metals” OR “PAH” OR “phytoremediation” OR “phytomanagement”). Some additional studies were picked out from the reference list found in the studies collected from our query. From the first selection of 53 potential studies (from their title and summary), the final collection made after careful reading of the entire studies comprised only 15 references (Table 2). From these studies, we identified 25 case studies suitable for the meta-analysis that was conducted in a large variety of geographical locations and climates. Studies and case studies were excluded from our database when MR: MS could not be calculated, when they dealt with air pollution (not our subject), when no phytotoxic effects were shown (no decrease in plant growth), when no statistical analyses or tests had been done for the reported results regarding MR: MS and root and shoot parts. When plant growth was reported both in hydroponic and for growth in soil substrates, we assumed that results from soil substrates where more suitable for analysing the biomass partitioning response. When other treatments were used (for instance mycorrhizae inoculation), we averaged the response to these treatments at each level of soil pollution. Finally, we considered one case study as being the unique combination of one team of researchers, one studied plant species, and one contaminant at stake. In one study, we made an exception and considered two case studies for two populations of the same plant species being exposed to the same contaminant, the two populations being reported as being metallicolous and non-metallicolous and which showed contrasting responses.The statistics and information recorded: we aimed to answer three questions: Is there any general pattern (increase or decrease) of the MR: MS ratio reported in the literature? Do changes in MR: MS depend on some explicative factors (for instance the contaminant type)? and Do MR: MS variations depend on pollution effect on plant size? This last question is important in our study which aimed to distinguish changes due to simple allometric effects rather than plant response. For each case study, the main indicator of biomass partitioning available was the MR: MS ratio, either provided directly, or calculated from root and shoot biomasses. Total plant dry biomasses were also recorded. Then, we calculated two statistics to enable the comparison of studies that were not originally designed to be compared. Firstly, we calculated the effect size metric (referred to as relative response in this study) to estimate the effect of pollution on the MR: MS ratio as follows:$${{{Relative}}},{M}_{R}:{M}_{S}={{{{{rm{log }}}}}}big({M}_{R}:{M}_{S{{_}}{{{polluted}}}}/{M}_{R}:{M}_{S{{_}}{{{control}}}}big)$$
(3)
Values close to 0 are associated with a negligible effect of the treatment, while negative and positive values indicate negative and positive effects of the treatment, respectively. The relative response is a reliable approach to quantify the effects of treatment compared to control and is regularly used in plant science (e.g. ref. 67). Secondly, we calculated a phytotoxic effect by normalisation of the effect of pollution on plant growth as follows:$$Phytotoxicity,(biomass,loss)=1-big({biomass}_{polluted}/{biomass}_{control}big)$$
(4)
Values close to 0 are associated with a negligible effect on plant growth, while values close to 1 indicate a strong decrease in plant growth. Additionally, we report relevant information to analyse its potential influence on MR: MS results. We reported the plant species involved, its functional group (monocotyledonous grass, dicotyledonous forb, and woody species), its life cycle (annual, perennial), the experiment duration (and if several measurements were made at different times) and the kind of contaminant at stake (see Table 2).The meta-analysis: regarding the general pattern of MR: MS changes, they were classified on the basis of the statistical results reported in the different studies as follows: (i) stable: no change of the MR: MS value was reported; (ii) variable: increase or decrease of the MR: MS ratio was reported for a pollution treatment compared to the control, and other treatments with a higher level of pollution showed no or opposite effects; (iii) increase: increase of the MR: MS ratio was reported for pollution treatment, and other treatments with higher levels of pollution also showed an increase compared to the control; (iv): decrease: an opposite situation to increase described above. Additionally, we tested the effect of the contaminant type, plant functional type, and plant life cycle on the relative MR: MS by ANOVA. Finally, we tested the dependence of the relative MR: MS with biomass loss for case studies showing an increase or decrease in this ratio by using linear modelling. This was done to compare results among case studies (one averaged value per case study). For results within a case study (when several soil pollution levels were available per case study), relative MR: MS and biomass loss compared to the control was calculated for each pollution level. Rate of relative MR: MS changes (Δ MR: MS /Δ biomass loss) was calculated and compared to 0.Modelling of changes in root: shoot ratio with exposure to pollution stress: we modelled the changes in the root: shoot ratio compared to a control situation without exposure to excess contaminants. We considered the change of allometric relationships (Eq. 1.) between root and shoot parts by the three potential drivers related to soil effect and plant response, and we followed the following steps. First, we calculated the growth of shoot parts as follows:$$S={gr}, .,left(1-{{gr}}_{{{{{{{mathrm{decrease}}}}}}}}right),.,d$$
(5)
gr represents plant growth rate (it can concern shoot biomass or area) per day; d is the duration of the growing period (in days); grdecrease is the phytotoxic effect on plant growth (interval [0,0.8] is considered here); S is the number of shoot parts produced after the corresponding duration d.Second, we calculated corresponding root parts as follows$$R={{upbeta }},.,left(1-{{{upbeta }}}_{{{{decrease}}}}right),.,{S}^{{{upalpha }}.(1+{{upalpha }}_{{{{increase}}}})}$$
(6)
With β and α the parameters of the allometric relationship of a given plant species in a control soil; βdecrease (the interval [0;0.5] is considered) is the effect of pollution stress on the early root development; αincrease (the interval [0;0.5] is considered) corresponds to plant response with increasing biomass partitioning in favour of roots; and R is the number of root parts produced.Finally, changes in root: shoot ratios were calculated by dividing the root: shoot ratio obtained on polluted soils by the root: shoot ratio obtained in a control situation (grdecrease; βdecrease; and αincrease set to 0).Reporting summaryFurther information on research design is available in the Nature Research Reporting Summary linked to this article. More

We deployed state-of-the-art instrumentation to Finnish wetland, Siikaneva (61°49’59.4“N 24°11’32.5“E, 162 m a.s.l.) where is located a class II ecosystem ICOS (European Integrated Carbon Observation System) station40 and to SMEAR II station (Station for Measuring Ecosystem-Atmosphere Relations)41, in Hyytiälä (61°50’47.1“N 24°17’43.2“E, 181 m a.s.l.) and investigated all the relevant components that are known to influence the new particle formation. The observations were performed on 10th May–15th June 2016. We monitored direct VOC and CH4 emissions from wetland and the concentrations of oxidation products of VOCs, SO2, and O3. We monitored concentrations and chemical composition of atmospheric clusters, aerosols, and air ions from the smallest sizes (0.5 nm) up to 40 nm approaching sizes which can be activated to CCN. As a reference, we utilized SMEAR II station in Hyytiälä, located 5 km east of these measurements. The SMEAR II station is monitoring over 1200 variables, including also the ones measured in the Siikaneva wetland.The Hyytiälä site is a relatively homogeneous Scots pine stand surrounded by evergreen coniferous forests41, while the Siikaneva site is located in a pristine boreal fen. Peat started to accumulate in Siikaneva after the latest ice age about 9000 years ago and peat depth at the measurement site is approximately 4 meters42,43. Siikaneva fen is characterized by relatively flat topography with a number of vegetation communities and some surface patterning featuring drier hummocks and wetter lawns.The measurement site consisted of a small hut containing all the instrumentation, which was equipped with sampling inlets at heights of approximately 1.5 m and 3 m. The CI-APi-TOF and APi-TOF, NAIS, PSM, O3 measurements were conducted with the inlet at 1.5 m, while all the meteorological, CH4, CO2, and VOC data were obtained at 3 m.Data sets from the SMEAR II station at Hyytiälä can be obtained from the AVAA smartSMEAR website (https://avaa.tdata.fi/web/smart)44. A detailed description of the SMEAR II station at Hyytiälä can be found elsewhere41,45. Siikaneva station is part of ICOS (European Integrated Carbon Observation System) network that includes two classes of Ecosystem stations, referred to as Class 1 (complete) and Class 2 (basic) stations. They differ in costs of construction, operation, and maintenance due to the reduced number of variables measured at the Class 2 stations. Siikaneva station is classified as the class 2 ecology site.Air temperature and relative humidity (RH) were measured with Rotronic HC2 sensor (Rotronic AG, Switzerland) at 2-meter height in Siikaneva. The air temperature was measured at 2 min and RH one minute time resolution. Photosynthetically active radiation (PAR) was measured once in a minute by a Li-Cor Li-190SZ quantum sensor (LI-COR, Inc., USA). Wind speed and direction were measured with Metek USA-1/Gill HS 50 anemometer at 3 meters height. The averaging period for all auxiliary measurements was 30 minutes.VOC concentrations were measured with a proton transfer time-of-flight mass spectrometer (PTR-TOF, Ionicon) which consists of a proton transfer reaction ion source (PTR) and a TOF-MS46. The PTR instrument is described in detail in literature47,48 and only short description is given here. The PTR consists of a H3O + ion source (hollow cathode discharge in water vapor) and a drift tube where protonated water is mixed with the sample and protons are transferred to the VOC species according to Eq. 1:$${{{{{{rm{H}}}}}}}_{3}{{{{{{rm{O}}}}}}}^{+}+{{{{{rm{VOC}}}}}}to {{{{{{rm{VOCH}}}}}}}^{+}+{{{{{{rm{H}}}}}}}_{2}{{{{{rm{O}}}}}}$$
(1)
This charging mechanism works for VOCs with higher proton affinity than that of water, most atmospheric VOC fulfill this requirement47.The ionized VOCH+ are then passed to the TOF and the mass is determined with an accuracy of 20ppt and resolving power of 3000Th/Th. The VOC is identified using the accurate mass and the prior made calibration. The concentrations of VOCs can be computed from the calibration as the ratio of sample to reagent ion using equation Eq. 2:$$[{{{{{rm{VOC}}}}}}]=[{{{{{{rm{VOC}}}}}}}^{+}]/([{{{{{{rm{H}}}}}}}_{3}{{{{{{rm{O}}}}}}}^{+}]cdot {{{{{rm{kt}}}}}})$$
(2)
where [H3O +] is the concentration of H3O + in the absence of reacting neutrals, k is the reaction coefficient of the proton transfer reaction and t is the average time the ions spend in the reaction region47. Product kt is obtained from calibration.Terpene and isoprene emissions are depended on temperature and light49. Accordingly, an increase in both concentrations is observed when approaching summer, indicating an increase in biogenic emissions (Supplementary note 6 and 7. Fig. S10-S12).The chemical composition of air ions was measured with atmospheric pressure interface (APi) time of flight mass spectrometer50 (APi-TOF, Tofwerk AG). The sample was driven to the instrument through 10 mm electropolished stainless steel tube with a flow rate of 6lpm. The sample was further introduced to APi through a critical orifice with a sample flow of 0.8 l min−1, ions are transported into the TOF to determine their mass to charge ratio(m/Q). The ion beam is focused by two guiding quadrupoles and an ion lens assembly, in three separate differentially pumped chambers, leading into the TOF. The instrument has resolving power of >3000 Th/Th and mass accuracy More

Gienapp, P., Reed, T. E. & Visser, M. E. Why climate change will invariably alter selection pressures on phenology. Proc. R. Soc. B 281, 20141611 (2014).Article 
PubMed 
PubMed Central 

Google Scholar 
Parmesan, C. Influences of species, latitudes and methodologies on estimates of phenological response to global warming. Glob. Change Biol. 13, 1860–1872 (2007).ADS 
Article 

Google Scholar 
Root, T. L. et al. Fingerprints of global warming on wild animals and plants. Nature 421, 57–60 (2003).ADS 
CAS 
Article 
PubMed 

Google Scholar 
Thackeray, S. J. et al. Phenological sensitivity to climate across taxa and trophic levels. Nature 535, 241–245 (2016).ADS 
CAS 
Article 
PubMed 

Google Scholar 
Thackeray, S. J. et al. Trophic level asynchrony in rates of phenological change for marine, freshwater and terrestrial environments. Glob. Change Biol. 16, 3304–3313 (2010).ADS 
Article 

Google Scholar 
Blondel, J., Dias, P. C., Perret, P., Maistre, M. & Lambrechts, M. M. Selection-based biodiversity at a small spatial scale in a low-dispersing insular bird. Science 285, 1399–1402 (1999).CAS 
Article 
PubMed 

Google Scholar 
Serreze, M. C. & Barry, R. G. Processes and impacts of Arctic amplification: a research synthesis. Glob. Planet. Change 77, 85–96 (2011).ADS 
Article 

Google Scholar 
Inouye, D. W., Barr, B., Armitage, K. B. & Inouye, B. D. Climate change is affecting altitudinal migrants and hibernating species. Proc. Natl Acad. Sci. 97, 1630–1633 (2000).ADS 
CAS 
Article 
PubMed 
PubMed Central 

Google Scholar 
Walther, G.-R. et al. Ecological responses to recent climate change. Nature 416, 389–395 (2002).ADS 
CAS 
Article 
PubMed 

Google Scholar 
Reneerkens, J. et al. Effects of food abundance and early clutch predation on reproductive timing in a high Arctic shorebird exposed to advancements in arthropod abundance. Ecol. Evol. 6, 7375–7386 (2016).Article 
PubMed 
PubMed Central 

Google Scholar 
Bonamour, S., Chevin, L.-M., Charmantier, A. & Teplitsky, C. Phenotypic plasticity in response to climate change: the importance of cue variation. Philos. Trans. R. Soc. B Biol. Sci. 374, 20180178 (2019).Article 

Google Scholar 
Saalfeld, S. T. & Lanctot, R. B. Multispecies comparisons of adaptability to climate change: a role for life-history characteristics? Ecol. Evol. 7, 10492–10502 (2017).Article 
PubMed 
PubMed Central 

Google Scholar 
Visser, M. E. et al. Variable responses to large-scale climate change in European Parus populations. Proc. R. Soc. Lond. B Biol. Sci. 270, 367–372 (2003).Article 

Google Scholar 
Wolkovich, E. M. et al. Warming experiments underpredict plant phenological responses to climate change. Nature 485, 494–497 (2012).ADS 
CAS 
Article 
PubMed 

Google Scholar 
Asch, M. V., Tienderen, P. H. V., Holleman, L. J. M. & Visser, M. E. Predicting adaptation of phenology in response to climate change, an insect herbivore example. Glob. Change Biol. 13, 1596–1604 (2007).ADS 
Article 

Google Scholar 
Silverin, B. et al. Ambient temperature effects on photo induced gonadal cycles and hormonal secretion patterns in Great Tits from three different breeding latitudes. Horm. Behav. 54, 60–68 (2008).CAS 
Article 
PubMed 

Google Scholar 
Kharouba, H. M. & Wolkovich, E. M. Disconnects between ecological theory and data in phenological mismatch research. Nat. Clim. Change 10, 406–415 (2020).ADS 
Article 

Google Scholar 
Visser, M. E. & Gienapp, P. Evolutionary and demographic consequences of phenological mismatches. Nat. Ecol. Evol. 3, 879–885 (2019).Article 
PubMed 
PubMed Central 

Google Scholar 
Silverin, B., Massa, R. & Stokkan, K. A. Photoperiodic adaptation to breeding at different latitudes in great tits. Gen. Comp. Endocrinol. 90, 14–22 (1993).CAS 
Article 
PubMed 

Google Scholar 
Phillimore, A. B. et al. Passerines may be sufficiently plastic to track temperature‐mediated shifts in optimum lay date. Glob. Chang Biol. 22, 3259–3272 (2016).ADS 
Article 
PubMed 

Google Scholar 
Bailey, L. D. & van de Pol, M. climwin: an R toolbox for climate window analysis. PLoS ONE 11, e0167980 (2016).Article 
CAS 
PubMed 
PubMed Central 

Google Scholar 
van de Pol, M. et al. Identifying the best climatic predictors in ecology and evolution. Methods Ecol. Evol. 7, 1246–1257 (2016).Article 

Google Scholar 
Van de Pol, M. & Bailey, L. D. Quantifying the climatic sensitivity of individuals, populations, and species. Eff. Clim. Change Birds 44; pp. 44–59 (2019).Culina, A. et al. Connecting the data landscape of long‐term ecological studies: the SPI‐Birds data hub. J. Anim. Ecol. https://doi.org/10.1111/1365-2656.13388 (2020).Verhagen, I., Tomotani, B. M., Gienapp, P. & Visser, M. E. Temperature has a causal and plastic effect on timing of breeding in a small songbird. J. Exp. Biol. 223, jeb218784 (2020).Article 
PubMed 

Google Scholar 
Buse, A., Dury, S. J., Woodburn, R. J. W., Perrins, C. M. & Good, J. E. G. Effects of elevated temperature on multi-species interactions: the case of Pedunculate Oak, Winter Moth and Tits. Funct. Ecol. 13, 74–82 (1999).Article 

Google Scholar 
Van Noordwijk, A. J., McCleery, R. H. & Perrins, C. M. Selection for the timing of great tit breeding in relation to caterpillar growth and temperature. J. Anim. Ecol. 64, 451–458 (1995).Article 

Google Scholar 
Visser, M. E., Holleman, L. J. M. & Gienapp, P. Shifts in caterpillar biomass phenology due to climate change and its impact on the breeding biology of an insectivorous bird. Oecologia 147, 164–172 (2006).ADS 
Article 
PubMed 

Google Scholar 
Foden, W. B. et al. Identifying the world’s most climate change vulnerable species: a systematic trait-based assessment of all birds, amphibians and corals. PLOS ONE 8, e65427 (2013).ADS 
CAS 
Article 
PubMed 
PubMed Central 

Google Scholar 
Pacifici, M. et al. Assessing species vulnerability to climate change. Nat. Clim. Change 5, 215–224 (2015).ADS 
Article 

Google Scholar 
Williams, S. E., Shoo, L. P., Isaac, J. L., Hoffmann, A. A. & Langham, G. Towards an integrated framework for assessing the vulnerability of species to climate change. PLOS Biol. 6, e325 (2008).Article 
CAS 
PubMed Central 

Google Scholar 
Dhondt, A. A., Eyckerman, R., Moermans, R. & Hublé, J. Habitat and laying date of Great and Blue Tit Parus major and P. caeruleus. Ibis 126, 388–397 (1984).Article 

Google Scholar 
Bourgault, P., Thomas, D., Perret, P. & Blondel, J. Spring vegetation phenology is a robust predictor of breeding date across broad landscapes: a multi-site approach using the Corsican blue tit (Cyanistes caeruleus). Oecologia 162, 885–892 (2010).ADS 
Article 
PubMed 

Google Scholar 
Blondel, J., Dias, P. C., Maistre, M. & Perret, P. Habitat Heterogeneity and Life-History Variation of Mediterranean Blue Tits (Parus caeruleus). Auk 110, 511–520 (1993).Article 

Google Scholar 
Blondel, J., Dervieux, A., Maistre, M. & Perret, P. Feeding ecology and life history variation of the blue tit in Mediterranean deciduous and sclerophyllous habitats. Oecologia 88, 9–14 (1991).ADS 
Article 
PubMed 

Google Scholar 
Vatka, E., Orell, M. & Rytkönen, S. Warming climate advances breeding and improves synchrony of food demand and food availability in a boreal passerine. Glob. Change Biol. 17, 3002–3009 (2011).ADS 
Article 

Google Scholar 
Massa, B., Cusimano, C. A., Margagliotta, B. & Galici, R. Reproductive characteristics and differential response to seasonal temperatures of blue and great tits (Cyanistes caeruleus & Parus major) in three neighbouring mediterranean habitats. Rev. Ecol. 66, 157–172 (2011).
Google Scholar 
Both, C., van Asch, M., Bijlsma, R. G., van den Burg, A. B. & Visser, M. E. Climate change and unequal phenological changes across four trophic levels: constraints or adaptations? J. Anim. Ecol. 78, 73–83 (2009).Article 
PubMed 

Google Scholar 
Nolet, B. A. et al. Faltering lemming cycles reduce productivity and population size of a migratory Arctic goose species. J. Anim. Ecol. 82, 804–813 (2013).Article 
PubMed 
PubMed Central 

Google Scholar 
Petanidou, T. et al. Variable flowering phenology and pollinator use in a community suggest future phenological mismatch. Acta Oecologica 59, 104–111 (2014).ADS 
Article 

Google Scholar 
McLean, N., van der Jeugd, H. P. & van de. Pol, M. High intra-specific variation in avian body condition responses to climate limits generalisation across species. PLOS ONE 13, e0192401 (2018).Article 
CAS 
PubMed 
PubMed Central 

Google Scholar 
van de Pol, M. & Wright, J. A simple method for distinguishing within-versus between-subject effects using mixed models. Anim. Behav. 77, 753–758 (2009).Article 

Google Scholar 
Radchuk, V. et al. Adaptive responses of animals to climate change are most likely insufficient. Nat. Commun. 10, 3109 (2019).ADS 
Article 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Burgess, M. D. et al. Tritrophic phenological match–mismatch in space and time. Nat. Ecol. Evol. 2, 970–975 (2018).Article 
PubMed 

Google Scholar 
Martin, R. O., Cunningham, S. J. & Hockey, P. A. R. Elevated temperatures drive fine-scale patterns of habitat use in a savanna bird community. Ostrich 86, 127–135 (2015).Article 

Google Scholar 
Latimer, C. E. & Zuckerberg, B. Forest fragmentation alters winter microclimates and microrefugia in human-modified landscapes. Ecography 40, 158–170 (2017).Article 

Google Scholar 
Frey, S. J. K. et al. Spatial models reveal the microclimatic buffering capacity of old-growth forests. Sci. Adv. 2, e1501392 (2016).ADS 
Article 
PubMed 
PubMed Central 

Google Scholar 
Zellweger, F. et al. Forest microclimate dynamics drive plant responses to warming. Science 368, 772–775 (2020).ADS 
CAS 
Article 
PubMed 

Google Scholar 
Hansen, M. C. et al. High-resolution global maps of 21st-century forest cover change. Science 342, 850–853 (2013).ADS 
CAS 
Article 
PubMed 

Google Scholar 
Tang, H. et al. Characterizing global forest canopy cover distribution using spaceborne lidar. Remote Sens. Environ. 231, 111262 (2019).ADS 
Article 

Google Scholar 
Visser, M. E. & Both, C. Shifts in phenology due to global climate change: the need for a yardstick. Proc. Biol. Sci. 272, 2561–2569 (2005).PubMed 
PubMed Central 

Google Scholar 
Samplonius, J. M. et al. Strengthening the evidence base for temperature-mediated phenological asynchrony and its impacts. Nat. Ecol. Evol. 5, 155–164 (2021).Article 
PubMed 

Google Scholar 
Langmore, N. E., Bailey, L. D., Heinsohn, R. G., Russell, A. F. & Kilner, R. M. Egg size investment in superb fairy-wrens: helper effects are modulated by climate. Proc. Biol. Sci. 283, 20161875 (2016).PubMed 
PubMed Central 

Google Scholar 
del Hoyo, J. et al. Handbook of the Birds of the World and BirdLife International Illustrated Checklist of the Birds of the World (Lynx Edicions/Birdlife International, 2016).Both, C. et al. Large–scale geographical variation confirms that climate change causes birds to lay earlier. Proc. R. Soc. Lond. B Biol. Sci. 271, 1657–1662 (2004).Article 

Google Scholar 
Haylock, M. R. et al. A European daily high-resolution gridded data set of surface temperature and precipitation for 1950–2006. J. Geophys. Res. Atmospheres 113, D20119 (2008).ADS 
Article 

Google Scholar 
Klok, E. J. & Klein Tank, A. M. G. Updated and extended European dataset of daily climate observations. Int. J. Climatol. 29, 1182–1191 (2009).Article 

Google Scholar 
Simmonds, E. G., Cole, E. F. & Sheldon, B. C. Cue identification in phenology: a case study of the predictive performance of current statistical tools. J. Anim. Ecol. 88, 1428–1440 (2019).Article 
PubMed 
PubMed Central 

Google Scholar 
Samplonius, J. M. et al. Phenological sensitivity to climate change is higher in resident than in migrant bird populations among European cavity breeders. Glob. Change Biol. 24, 3780–3790 (2018).ADS 
Article 

Google Scholar 
Slagsvold, T. Annual and geographical variation in the time of breeding of the great tit Parus major and the Pied Flycatcher Ficedula hypoleuca in relation to environmental phenology and spring temperature. Ornis Scand. Scand. J. Ornithol. 7, 127–145 (1976).Article 

Google Scholar 
Haest, B., Hüppop, O. & Bairlein, F. Challenging a 15‐year‐old claim: the North Atlantic Oscillation index as a predictor of spring migration phenology of birds. Glob. Change Biol. 24, 1523–1537 (2018).ADS 
Article 

Google Scholar 
Rosseel, Y. lavaan: an R package for structural equation modeling. J. Stat. Softw. 48, 1–36 (2012).Article 

Google Scholar 
Metzger, M. J., Bunce, R. G. H., Jongman, R. H. G., Mücher, C. A. & Watkins, J. W. A climatic stratification of the environment of Europe. Glob. Ecol. Biogeogr. 14, 549–563 (2005).Article 

Google Scholar 
Rousset, F. & Ferdy, J.-B. Testing environmental and genetic effects in the presence of spatial autocorrelation. Ecography 37, 781–790 (2014).Article 

Google Scholar 
R. Core Team. R: A Language and Environment for Statistical Computing. (R Foundation for Statistical Computing, 2018).Bailey, L. D. et al. Bird populations most exposed to climate change are less sensitive to climatic variation, Zenodo, https://doi.org/10.5281/zenodo.5747634 (2022).Bailey, L. D. et al. Bird populations most exposed to climate change are less sensitive to climatic variation, LiamDBailey/baileyetal2021, https://doi.org/10.5281/zenodo.6027546 (2022). More

Fosberg, F. R. & Chapman, V. J. Mangrove Vegetation. Taxon 26, 113 (1977).Article 

Google Scholar 
Vo, Q. T., Kuenzer, C., Vo, Q. M., Moder, F. & Oppelt, N. Review of valuation methods for mangrove ecosystem services. Ecol. Indic. 23, 431–446 (2012).Article 

Google Scholar 
Costanza, R. et al. The value of the world’s ecosystem services and natural capital. Nature 387, 253–260 (1997).ADS 
CAS 
Article 

Google Scholar 
Duke, N. C. et al. A world without mangroves?. Science. 317, 41b–42b (2007).Article 

Google Scholar 
Barbier, E. B. et al. The value of estuarine and coastal ecosystem services. Ecol. Monogr. 81, 169–193 (2011).Article 

Google Scholar 
Saderne, V. et al. Total alkalinity production in a mangrove ecosystem reveals an overlooked Blue Carbon component. Limnol. Oceanogr. Lett. https://doi.org/10.1002/lol2.10170 (2020).Allen, J. R. L. Morphodynamics of Holocene salt marshes: a review sketch from the Atlantic and Southern North Sea coasts of Europe. Quat. Sci. Rev. 19, 1155–1231 (2000).ADS 
Article 

Google Scholar 
Fagherazzi, S. et al. Tidal networks 1. Automatic network extraction and preliminary scaling features from digital terrain maps. Water Resour. Res. 35, 3891–3904 (1999).ADS 
Article 

Google Scholar 
D’Alpaos, A., Lanzoni, S., Mudd, S. M. & Fagherazzi, S. Modeling the influence of hydroperiod and vegetation on the cross-sectional formation of tidal channels. Estuar. Coast. Shelf Sci. 69, 311–324 (2006).ADS 
Article 

Google Scholar 
D’Alpaos, A. & Marani, M. Reading the signatures of biologic-geomorphic feedbacks in salt-marsh landscapes. Adv. Water Resour. 93, 265–275 (2016).ADS 
Article 

Google Scholar 
Schwarz, C. et al. Self-organization of a biogeomorphic landscape controlled by plant life-history traits. Nat. Geosci. 11, 672–677 (2018).ADS 
CAS 
Article 

Google Scholar 
Mariotti, G. & Canestrelli, A. Long-term morphodynamics of muddy backbarrier basins: fill in or empty out? Water Resour. Res. 53, 7029–7054 (2017).ADS 
Article 

Google Scholar 
Stark, J., Van Oyen, T., Meire, P. & Temmerman, S. Observations of tidal and storm surge attenuation in a large tidal marsh. Limnol. Oceanogr. 60, 1371–1381 (2015).ADS 
Article 

Google Scholar 
Montgomery, J., Bryan, K., Horstman, E. & Mullarney, J. Attenuation of tides and surges by mangroves: contrasting case studies from New Zealand. Water 10, 1119 (2018).Article 

Google Scholar 
Temmerman, S. et al. Vegetation causes channel erosion in a tidal landscape. Geology 35, 631–634 (2007).ADS 
Article 

Google Scholar 
van Maanen, B., Coco, G. & Bryan, K. R. On the ecogeomorphological feedbacks that control tidal channel network evolution in a sandy mangrove setting. Proc. R. Soc. A Math. Phys. Eng. Sci. 471, 20150115 (2015).
Google Scholar 
Bij de Vaate, I., Brückner, M. Z. M., Kleinhans, M. G. & Schwarz, C. On the Impact of Salt Marsh Pioneer Species-Assemblages on the Emergence of Intertidal Channel Networks. Water Resour. Res. 56, (2020).Bouma, T. J. et al. Density-dependent linkage of scale-dependent feedbacks: a flume study on the intertidal macrophyte Spartina anglica. Oikos 118, 260–268 (2009).Article 

Google Scholar 
Schwarz, C. et al. Impacts of salt marsh plants on tidal channel initiation and inheritance. J. Geophys. Res. Earth Surf. 119, 385–400 (2014).ADS 
Article 

Google Scholar 
Mcowen, C. J. et al. A global map of saltmarshes. Biodivers. Data J. 5, (2017).Spalding, M. World Atlas of Mangroves. World Atlas of Mangroves https://doi.org/10.4324/9781849776608 (2010).Fromard, F., Vega, C. & Proisy, C. Half a century of dynamic coastal change affecting mangrove shorelines of French Guiana. A case study based on remote sensing data analyses and field surveys. in. Mar. Geol. 208, 265–280 (2004).ADS 
Article 

Google Scholar 
Proisy, C. et al. Mud bank colonization by opportunistic mangroves: a case study from French Guiana using lidar data. Cont. Shelf Res. 29, 632–641 (2009).ADS 
Article 

Google Scholar 
Balke, T. et al. Windows of opportunity: thresholds to mangrove seedling establishment on tidal flats. Mar. Ecol. Prog. Ser. 440, 1–9 (2011).ADS 
Article 

Google Scholar 
Tomlinson, P. B. The botany of mangroves. Bot. Mangroves https://doi.org/10.2307/2996392 (1986).Article 

Google Scholar 
Duke, N. C., Ball, M. C. & Ellison, J. C. Factors influencing biodiversity and distributional gradients in mangroves. Glob. Ecol. Biogeogr. Lett. 7, 27–47 (1998).Article 

Google Scholar 
Swales, A., Bentley, S. J. & Lovelock, C. E. Mangrove-forest evolution in a sediment-rich estuarine system: Opportunists or agents of geomorphic change? Earth Surf. Process. Landf. 40, 1672–1687 (2015).ADS 
Article 

Google Scholar 
Nardin, W. et al. Dynamics of a fringe mangrove forest detected by Landsat images in the Mekong River Delta, Vietnam. Earth Surf. Process. Landf. 41, 2024–2037 (2016).ADS 
Article 

Google Scholar 
Proffitt, C. E., Travis, S. E. & Edwards, K. R. Genotype and elevation influence Spartina alterniflora colonization and growth in a created salt marsh. Ecol. Appl. 13, 180–192 (2003).Article 

Google Scholar 
van Wesenbeeck, B. K. et al. Potential for sudden shifts in transient systems: distinguishing between local and landscape-scale processes. Ecosystems 11, 1133–1141 (2008).Article 

Google Scholar 
Ranwell, D. S. Spartina salt marshes in southern England 3. Rates of establishment, succession and nutrient supply at Bridgewater Bay, Somerset. J. Ecol. 52, 95–105 (1964).Article 

Google Scholar 
van Wesenbeeck, B. K., van de Koppel, J., Herman, P. M. J. & Bouma, T. J. Does scale dependent feedback explain spatial complexity in salt marsh ecosystems? Oikos 117, 152–159 (2008).Article 

Google Scholar 
Taylor, C. M. & Hastings, A. Finding optimal control strategies for invasive species: a density-structured model for Spartina alterniflora. J. Appl. Ecol. 41, 1049–1057 (2004).Article 

Google Scholar 
Vandenbruwaene, W. et al. Flow interaction with dynamic vegetation patches: Implications for biogeomorphic evolution of a tidal landscape. J. Geophys. Res. Earth Surf. 116, 1–13 (2011).Article 

Google Scholar 
Mobberley, D. G. Taxonomy and distribution of the genus Spartina. (Iowa State University, 1953).Gourgue, O. et al. A Convolution Method to Assess Subgrid-Scale Interactions Between Flow and Patchy Vegetation in Biogeomorphic Models. J. Adv. Model. Earth Syst. 127, 1–25 (2021).Zong, L. & Nepf, H. Spatial distribution of deposition within a patch of vegetation. Water Resour. Res. 47, (2011).Suyadi, Gao, J., Lundquist, C. J. & Schwendenmann, L. Characterizing landscape patterns in changing mangrove ecosystems at high latitudes using spatial metrics. Estuar. Coast. Shelf Sci. 215, 1–10 (2018).ADS 
Article 

Google Scholar 
Best, S. N. et al. Do salt marshes survive sea level rise? Modelling wave action, morphodynamics and vegetation dynamics. Environ. Model. Softw. 109, 152–166 (2018).Article 

Google Scholar 
Chen, Y., Li, Y., Cai, T., Thompson, C. & Li, Y. A comparison of biohydrodynamic interaction within mangrove and saltmarsh boundaries. Earth Surf. Process. Landf. 41, 1967–1979 (2016).ADS 
Article 

Google Scholar 
Xie, D. et al. Mangrove diversity loss under sea-level rise triggered by bio-morphodynamic feedbacks and anthropogenic pressures. Environ. Res. Lett. 15, 114033 (2020).ADS 
Article 

Google Scholar 
Steel, T. J. & Pye, K. The development of salt marsh tidal creek networks: evidence from the UK. In Proceedings of the Canadian Coastal Conference 1, 267–280 (1997).Fagherazzi, S. & Sun, T. A stochastic model for the formation of channel networks in tidal marshes. Geophys. Res. Lett. 31, L21503 (2004).ADS 
Article 

Google Scholar 
D’Alpaos, A., Lanzoni, S., Marani, M., Fagherazzi, S. & Rinaldo, A. Tidal network ontogeny: channel initiation and early development. J. Geophys. Res. 110, F02001 (2005).ADS 

Google Scholar 
Marani, M. et al. On the drainage density of tidal networks. Water Resour. Res. 39, 1040 (2003).ADS 

Google Scholar 
Liu, Z. et al. Efficient tidal channel networks alleviate the drought-induced die-off of salt marshes: Implications for coastal restoration and management. Sci. Total Environ. 749, 141493 (2020).ADS 
CAS 
PubMed 
Article 
PubMed Central 

Google Scholar 
Kearney, W. S. et al. Salt marsh vegetation promotes efficient tidal channel networks. Nat. Commun. 7, 12287 (2016).ADS 
CAS 
PubMed 
PubMed Central 
Article 

Google Scholar 
Hood, W. G. Applying tidal landform scaling to habitat restoration planning, design, and monitoring. Estuar. Coast. Shelf Sci. 244, 106060 (2020).Article 

Google Scholar 
Horstman, E., Dohmen-Janssen, C., Geomorphology, T. B.- & 2015, undefined. Tidal-scale flow routing and sedimentation in mangrove forests: Combining field data and numerical modelling. ElsevierCoco, G. et al. Morphodynamics of tidal networks: Advances and challenges. Mar. Geol. 346, 1–16 (2013).ADS 
Article 

Google Scholar 
Geng, L., Gong, Z., Zhou, Z., Lanzoni, S. & D’Alpaos, A. Assessing the relative contributions of the flood tide and the ebb tide to tidal channel network dynamics. Earth Surf. Process. Landf. 45, 237–250 (2020).ADS 
Article 

Google Scholar 
Andutta, F. P., Wang, X. H., Li, L. & Williams, D. Hydrodynamics and Sediment Transport in a Macro-tidal Estuary: Darwin Harbour, Australia. in 111–129 (Springer, Dordrecht, 2014). https://doi.org/10.1007/978-94-007-7019-5_7Elmqvist, T. & Cox, P. A. The Evolution of Vivipary in Flowering Plants. Oikos 77, 3 (1996).Article 

Google Scholar 
Zhang, X., Leonardi, N., Donatelli, C. & Fagherazzi, S. Fate of cohesive sediments in a marsh-dominated estuary. Adv. Water Resour. 125, 32–40 (2019).ADS 
Article 

Google Scholar 
Nardin, W. & Edmonds, D. A. Optimum vegetation height and density for inorganic sedimentation in deltaic marshes. Nat. Geosci. 7, 722–726 (2014).ADS 
CAS 
Article 

Google Scholar 
Swales, A., Bentley, S. J., Lovelock, C. & Bell, R. G. Sediment Processes and Mangrove-Habitat Expansion on a Rapidly-Prograding Muddy Coast, New Zealand. In Coastal Sediments ’07 1441–1454 (American Society of Civil Engineers, 2007). https://doi.org/10.1061/40926(239)111Wang, F., Lu, X., Sanders, C. J. & Tang, J. Tidal wetland resilience to sea level rise increases their carbon sequestration capacity in United States. Nat. Commun. 10, 1–11 (2019).ADS 
Article 
CAS 

Google Scholar 
Kristensen, E., Bouillon, S., Dittmar, T. & Marchand, C. Organic carbon dynamics in mangrove ecosystems: A review. Aquat. Bot. 89, 201–219 (2008).CAS 
Article 

Google Scholar 
Fagherazzi, S. et al. Fluxes of water, sediments, and biogeochemical compounds in salt marshes. Ecol. Process 2, 1–16 (2013).Article 

Google Scholar 
Kirchner, J. W. Statistical inevitability of Horton’s laws and the apparent randomness of stream channel networks. Geology 21, 591–594 (1993).ADS 
Article 

Google Scholar 
Vandenbruwaene, W., Meire, P. & Temmerman, S. Formation and evolution of a tidal channel network within a constructed tidal marsh. Geomorphology (2012).Marani, M. et al. Patterns in tidal environments: salt-marsh channel networks and vegetation. in Geoscience and Remote Sensing Symposium. IEEE 5 3269–3271 (2003).Horstman, E. M., Karin R. B., and Julia C. M. “Drag variations, tidal asymmetry and tidal range changes in a mangrove creek system.” Earth Surf. Process. Landf. (2021).R. Core, Team. R: A language and environment for statistical computing. (2013).Lillesand, T. M. & Kiefer, R. W. Remote Sensing and Image Interpretation. John Willey & Sons. Inc, USA. (1994).Vandenbruwaene, W., Bouma, T. J., Meire, P. & Temmerman, S. Bio-geomorphic effects on tidal channel evolution: impact of vegetation establishment and tidal prism change. Earth Surf. Process. Landforms 38, 122–132 (2013).ADS 
Article 

Google Scholar 
Stefanon, L., Carniello, L., D’Alpaos, A. & Lanzoni, S. Experimental analysis of tidal network growth and development. Cont. Shelf Res. 30, 950–962 (2010).ADS 
Article 

Google Scholar 
Braat, L., Leuven, J. R. F. W., Lokhorst, I. R. & Kleinhans, M. G. Effects of estuarine mudflat formation on tidal prism and large-scale morphology in experiments. Earth Surf. Process. Landf. 44, 417–432 (2019).ADS 
Article 

Google Scholar 
Kleinhans, M. G. et al. Turning the tide: Comparison of tidal flow by periodic sea level fluctuation and by periodic bed tilting in scaled landscape experiments of estuaries. Earth Surf. Dyn. 5, 731–756 (2017).ADS 
Article 

Google Scholar 
Paola, C., Straub, K., Mohrig, D. & Reinhardt, L. The ‘unreasonable effectiveness’ of stratigraphic and geomorphic experiments. Earth-Sci. Rev. 97, 1–43 (2009).ADS 
Article 

Google Scholar 
Kleinhans, M. G., Leuven, J. R. F. W., Braat, L. & Baar, A. Scour holes and ripples occur below the hydraulic smooth to rough transition of movable beds. Sedimentology 64, 1381–1401 (2017).Article 

Google Scholar 
Lokhorst, I. R., Lange, S. I., Buiten, G., Selaković, S. & Kleinhans, M. G. Species selection and assessment of eco‐engineering effects of seedlings for biogeomorphological landscape experiments. Earth Surf. Process. Landf. 44, 2922–2935 (2019).ADS 
Article 

Google Scholar 
Widdows, J. et al. Inter-comparison between five devices for determining erodability of intertidal sediments. Cont. Shelf Res. 27, 1174–1189 (2007).ADS 
Article 

Google Scholar 
Verney, R., Brun-Cottan, J. C., Lafite, R., Deloffre, J. & Taylor, J. A. Tidally-induced shear stress variability above intertidal mudflats in the macrotidal seine estuary. Estuaries and Coasts 29, 653–664 (2006).Article 

Google Scholar 
Wu, W., Perera, C., Smith, J. & Sanchez, A. Critical shear stress for erosion of sand and mud mixtures. J. Hydraul. Res. 56, 96–110 (2018).Article 

Google Scholar 
Wolters, M., Garbutt, A., Bekker, R. M., Bakker, J. P. & Carey, P. D. Restoration of salt-marsh vegetation in relation to site suitability, species pool and dispersal traits. J. Appl. Ecol. 45, 904–912 (2007).Article 

Google Scholar  More

Here at the Germplasm Bank of Wild Species of China at the Kunming Institute of Botany, we want to preserve the seeds of as many wild plants as possible from across China’s vast land area. I work on developing the best techniques to freeze plant seeds and tissues at ultracold temperatures, to maintain their viability for years. The idea is that if we plant these seeds again in hundreds of years, a plant will grow.The picture shows me taking a sample of embryos from the seeds of a magnolia tree out of a liquid-nitrogen cryopreservation tank to test whether they’ll regrow when thawed. I dress in protective equipment from head to toe to protect me from the nitrogen, which has a temperature of −196 °C.When I came to the institute in 2009 as a PhD student, it had just purchased its first liquid-nitrogen cryopreservation system, but no one knew how to operate it. I was the one to work it out.Over the years, human activities and climate change have had a negative impact on plant biodiversity. The ultimate goal of the plant seed bank is to collect and preserve all wild plant species in China that are endangered, rare or valuable. We want to save these species before they go extinct. We’ve collected seeds from nearly 11,000 plant species, but that’s only one-third of what grows in China.Many wild plants have genes that help them to survive in harsh environments and make them disease- or drought-resistant. In the future, we might need these genetic materials to breed new crops that can better adapt to the changing climate.I am constantly amazed by how diverse and beautiful seeds are. Some of them are brightly coloured and others are star-shaped. I feel proud when I see the unfrozen seeds germinate in test tubes and gradually grow into large plants. We have three plants in the seed-bank lobby that we cultivated from preserved tissues, and they are all now taller than me. More

Biogeographic pattern of LOSSThe original forest plot data aggregated at 0.25 degree show large spatial variations (Fig. 1a) across the continents, with the greatest LOSS in Asia & Australia (mean ± 1 SE; 6.5 ± 0.5 Mg ha−1 y−1) > South America (4.9 ± 0.2 Mg ha−1 y−1) and Africa (4.6 ± 0.2 Mg ha−1 y−1) > North America (2.3 ± 0.1 Mg ha−1 y−1 in boreal and 2 ± 0.1 Mg ha−1 y−1 in temperate)36 (Fig. 1b; Supplementary Fig. 5a). This pattern was robust to bootstrapping (1000 iterations) to randomly select 90% of plots for estimating the probability distribution of the mean continental values (Supplementary Fig. 5b). The upscaled gridded LOSS maps generated by our random forest algorithm (see Methods) over the spatial domain of our full datasets shows hotspots of high LOSS in Southern Asia & Australia ( > 6 Mg ha−1 y−1), Northwestern South America (Amazon) ( > 5 Mg ha−1 y−1), and the western coast of North America ( >3 Mg ha−1 y−1)10,36,37,38 (Supplementary Fig. 6a). These patterns were robust to two bootstrapping approaches – based on the sampled biomes of each point feature and also randomly sampling 90% data with replacement (see Methods) (Fig. 2a; Supplementary Fig. 6b). The uncertainty (coefficient of variance – CV; %mean) was generally low ( 10%), despite the larger sample size (n  > 500 at 0.25 degree) (Fig. 2b; Supplementary Fig. 6c), likely because of potential effects of forest recovery or regrowth following past disturbance16 as well as the small plot size (i.e., 0.067 ha in each plot)39.Fig. 1: Map of sample locations and biomass loss to mortality (LOSS) data.a Sampling sites. A total of 2676 samples were collected and aggregated into 814 grids at 0.25 degree that were used for geospatial modeling. b The median and interquartile range of LOSS across continents—North America, South America, Africa, and Asia & Australia.Full size imageFig. 2: Map of biomass loss to mortality (LOSS) and its uncertainty across continents.a, b Ensemble mean of LOSS a and its uncertainty (coefficient of variation, b across continents at 0.25 degree derived from forest plot data using the bootstrapped (10 iterations) approach by randomly sampling 90% plots with replacement. c, d Ensemble mean of LOSS c and its uncertainty (coefficient of variation, d across continents at 0.5 degree derived from six dynamic vegetation models (DGVMs, ORCHIDEE, CABLE-POP, JULES, LPJ-GUESS, LPJmL, and SEIB-DGVM). Coefficient of variation was quantified as the standard deviation divided by the mean predicted value as a measure of prediction accuracy. e The difference of LOSS between ensemble mean of DGVMs and ensemble mean of LOSS derived from forest plots data across continents at 0.5 degree, quantified as difference between c and a, whereby LOSS in Fig. 2a is resampled at 0.5 degree.Full size imageDrivers of LOSSMean annual temperature (MAT), aridity index (the ratio of precipitation to potential evapotranspiration), and precipitation seasonality were identified as the dominant predictors of LOSS across continents (Supplementary Fig. 7a), with positive relationships with LOSS (Fig. 3a)10,36. In contrast to local-scale studies40,41, wood density, forest stand density, and soil conditions were poor predictors of LOSS when all data were used. These relationships were largely driven by the spatial pattern of LOSS and climate gradients, whereby LOSS and MAT, aridity index, and precipitation seasonality were high in tropical forests (Supplementary Fig. 8). This motivated us to examine the drivers of LOSS in tropical vs non-tropical biomes (Supplementary Fig. 7b, c; Fig. 3b–d). With a smaller gradient in climate within wet tropical forests, soil properties such as nutrient content and cation exchange capacity (CEC) were significant predictors of LOSS (Supplementary Fig. 7b; Fig. 3b)42. In wet tropical forests, the relationships between soil nutrient content and CEC and LOSS were positive (Fig. 3b) and thus appeared to support the pattern of higher mortality in more productive tropical forests growing over nutrient rich soils42,43. In non-tropical regions, basal area or a competition index based on the degree of crowding within stocked areas44 (see Methods) were the dominant predictors of LOSS, especially in extra-tropical North America (Supplementary Fig. 7c; Fig. 3c, d). This result highlights the role of stand competition in driving the spatial patterns of LOSS44,45. This pattern also supports the existence of a spatial tradeoff between faster growth and higher mortality because of resource limitations or younger death, whereby competition plays the fundamental role13,45. In contrast to other studies15,46, forest age (available in boreal and temperate forests in North America) was not a good predictor of LOSS (Supplementary Fig. 9), likely because of our focus on mature and old-growth forests (i.e., age > 80 and 100 years in boreal and temperate forests, respectively).Fig. 3: Standardized response coefficients (mean ± 95% CIs) between LOSS and dominant environmental drivers.The scales analyzed were at continents a, tropical regions b vs non-tropical regions c, d. The response coefficients were quantified by linear mixed model which account for each plot as a random effect. Panels c and d used basal area and stand density index (SDI) as competition index, respectively. SDI was defined as the degree of crowding within stocked areas and quantified as a function of tree density and the quadratic mean diameter in centimeters. Basal area is strongly correlated with total biomass and higher LOSS in higher basal area may be merely because of its correlations. Thus, we used another competition metrics – SDI to further confirm the role of competition in LOSS. The error bars denote the 95% confidence interval. *P  130%) in western boreal forests in North America (Fig. 2d).Conventional emergent constraintWe first used the conventional emergent constraint approach27 to constrain the projected (2015–2099) NPP and HR across continents. This approach was conducted by building the statistic (linear) relationship between the historical LOSS averaged at forest-plot scale (derived from original plot data of LOSS) or continental scale (derived from the map of LOSS) and projected NPP and HR summed across continents (see Methods and Supplementary Fig. 4 for details). We found that the emergent constraint approach worked well in North America, where the relationship between historical LOSS and projected NPP and HR was significant (the scenario of using original plot data of LOSS: R2 = 0.68 and P = 0.04 for grid-level NPP; R2 = 0.97 and P = 0.0001 for grid-level HR; the scenario of using map of LOSS at continent scale: R2 = 0.7 and P = 0.04 for grid-level NPP; R2 = 0.95 and P = 0.0008 for grid-level HR) (Supplementary Fig. 11a; Supplementary Fig. 12a). This emergent constraint approach was less effective, however, for other continents, where tropical forests are predominant (all P  > 0.05; Supplementary Fig. 11b, c, d; Supplementary Fig. 12b, c, d). These results suggest a weak linear relationships when observations are lumped or averaged at broad continental scales for tropical continents, thus highlighting the importance of spatial scale and non-linear relationships in emergent constraint25. We interpret the result that this LOSS emergent constraint works better in North America than in the tropical forests, by a better representation of forest plot distribution and couplings of LOSS and NPP and HR across space in North America.Machine learning constraintTo overcome this limitation, we trained a machine learning algorithm34 to reproduce the emerging relationship between historical LOSS and projected NPP and HR at grid level in each DGVM by incorporating all grid values without or with climate predictors, expressed as NPPpro or HRpro = f(LOSShis) or f(LOSShis, MATpro, MAPpro), respectively, where pro refers to projected variables, his refers to historical variables, and MAT and MAP is mean annual temperature precipitation, respectively (see Methods). The results show consistently positive non-linear relationships between LOSShis and NPPpro or HRpro across DGVMs (Supplementary Fig. 3). Our machine learning algorithms can surrogate well the results of process-based models between the historical LOSS and the projected NPP and HR (R  > 0.65 and R  > 0.9 in both scenarios without climate effects and with climate effects, respectively; see Methods) (Supplementary Fig. 13). After including the observed LOSShis (derived from LOSS) in the machine learning algorithm, we were able to generate spatially explicit constrained estimates34 of projected NPP and HR, and then compare them with the scenario without the constraint (Supplementary Fig. 14; Supplementary Fig. 15). These patterns essentially show a lower NPPpro or HRpro in locations of overestimated LOSShis in DGVMs, consistent with the positive relationship between LOSShis and NPPpro or HRpro (Supplementary Fig. 3).Our results show that most DGVMs overestimate tree mortality, particularly in tropical regions (Fig. 2c, e). Thus, if modeled mortality is over-estimated, we expect that NPP is over-estimated as well. Ultimately, we used a bootstrap approach to generate 100 maps of mean value of LOSS with its distribution following the values of the average and 2 times of standard deviation of LOSS maps as a conservative constraint (see Methods). Then the 100 maps of mean value of LOSS were used to constrain the projected NPP or HR as ensemble means in our ML constraint and the uncertainty of the constraint was assessed. Our bootstrapping constraint approach by LOSS reduces this common bias of models and decreases projected NPP down to 7.9, 2.3, 2 Pg C y−1 in South America, Africa and Asia & Australia, compared to original NPP values of 9, 2.4, 2.3 Pg C y−1 (Fig. 4a). The reason for this is that NPP or growth is strongly positively correlated with LOSS across space in both inventory data and DGVMs (Supplementary Figs. 2 and 3; Supplementary Fig. 16). The constant mortality parameter used in most models may be too large if modelers have tuned this parameter to obtain reasonable biomass stocks, thus compensating for an overestimate of NPP in absence of modeled competition between individuals and nutrients (e.g. phosphorus) limitations in tropical forests13. Likewise, HRpro showed similar patterns with NPPpro because of coupling of HR and NPP and LOSS at broad spatial and long term scales20,21, despite the likely decoupling of the instantaneous rate of HR and NPP and LOSS at local and short-term scales22,23. Thus, we also constrained a decrease in projected grid-level HR with values of 6.5, 1.9, 1.7 Pg C y−1 in South America, Africa and Asia & Australia compared to 7, 1.9, 1.8 Pg C y−1 in the original model ensemble (Fig. 4b). Taken together, our results constrain a weaker future tropical forest carbon sink from observation-based LOSS estimates down to 1.4, 0.4, 0.3 Pg C y−1 in South America, Africa and Asia & Australia as compared to 2, 0.5, 0.5 Pg C y−1 in the original models. The projected sink is reduced in particular over the Amazon basin, while North America showed an enhanced future carbon sink (1.1 and 0.8 Pg C y−1 after and before constraint, respectively). The constraint by the machine learning approach significantly reduced the model spread in grid-level NPPpro and HRpro generally in tropical regions and particularly in South America (Fig. 4; Table 1). This was in contrast to the case of constraint at the whole North America scale (Fig. 4; Table 1), presumably because of spatial trade-off or compensation from regions of mortality overestimation (i.e., eastern North America—temperate zones) vs underestimation (i.e., boreal zones). To this end, we further divided the whole North America into temperate and boreal forests and found the significant effects of the ML constraint (Supplementary Fig. 17). These results highlight the importance of spatial scale in the ML constraint approach. We thus recommend accounting for the role of spatial trade-off in our ML constraint approach or using our ML constraint approach at broad spatial scales whereby the effect of spatial trade-off is minimal. We also caution that the bootstrapping (100 times) approach used in our ML constraint increases the sample size and could have increased the significant difference with and without LOSS constraint. Overall, the uncertainty of the ML constraint was low in the bootstrapping approach (Supplementary Fig. 18).Fig. 4: Projected grid-level NPP and grid heterotrophic respiration (HR) across continents.a, b Projected (2015–2099) grid-level NPP a and grid-level HR b across continents quantified by six dynamic vegetation models—DGVMs (ORCHIDEE, CABLE-POP, JULES, LPJ-GUESS, LPJmL, and SEIB-DGVM). The y axes are the minimum, mean, and maximum values in six DGVMs. ‘DGVMs’ refers to the scenario before constraint and ‘DGVMs + Observation’ refers to the scenario after constraint without climate predictors. The constraint was achieved by using the observational maps (n = 100; through a bootstrapping approach; see Methods for details) of LOSS derived from forest plots data to feed into the trained ML (random forest) model. Reported are ensemble means of constraint. The constraint effect was significant when North America were divided into temperate and boreal forests (see results of Supplementary Fig. 17). *P  More

Mann, J. Behavioral sampling methods for cetaceans: A review and critique. Mar. Mammal Sci. 15, 102–122 (1999).Article 

Google Scholar 
Pompanon, F. et al. Who is eating what: Diet assessment using next generation sequencing. Mol. Ecol. https://doi.org/10.1111/j.1365-294X.2011.05403.x (2012).Article 

Google Scholar 
Deagle, B. E. et al. Counting with DNA in metabarcoding studies: How should we convert sequence reads to dietary data?. Mol. Ecol. 28, 391–406 (2019).Article 

Google Scholar 
Berry, T. E. et al. DNA metabarcoding for diet analysis and biodiversity: A case study using the endangered Australian sea lion (Neophoca cinerea). Ecol. Evol. 7, 5435–5453 (2017).Article 

Google Scholar 
Brassea-Pérez, E., Schramm, Y., Heckel, G., Chong-Robles, J. & Lago-Lestón, A. Metabarcoding analysis of the Pacific harbor seal diet in Mexico. Mar. Biol. 166, 1–14 (2019).Article 

Google Scholar 
Ford, M. J. et al. Estimation of a killer whale (Orcinus orca) population’s diet using sequencing analysis of DNA from feces. PLoS ONE 11, e0144956 (2016).Article 

Google Scholar 
Thomas, A. C., Deagle, B. E., Eveson, J. P., Harsch, C. H. & Trites, A. W. Quantitative DNA metabarcoding: Improved estimates of species proportional biomass using correction factors derived from control material. Mol. Ecol. Resour. 16, 714–726 (2016).CAS 
Article 

Google Scholar 
Deagle, B. E., Chiaradia, A., Mcinnes, J. & Jarman, S. N. Pyrosequencing faecal DNA to determine diet of little penguins: is what goes in what comes out? https://doi.org/10.1007/s10592-010-0096-6.Ando, H. et al. Methodological trends and perspectives of animal dietary studies by noninvasive fecal DNA metabarcoding. Environ. DNA 2, 391–406 (2020).Article 

Google Scholar 
Günther, B., Fromentin, J., Metral, L. & Arnaud-haond, S. Metabarcoding confirms the opportunistic foraging behaviour of Atlantic bluefin tuna and reveals the importance of gelatinous prey. PeerJ 9, e11757. https://doi.org/10.7717/peerj.11757 (2021).Article 

Google Scholar 
Simon, M., Hanson, M. B., Murrey, L., Tougaard, J. & Ugarte, F. From captivity to the wild and back: An attempt to release keiko the killer whale. Mar. Mammal Sci. 25, 693–705 (2009).Article 

Google Scholar 
Moore, M. et al. Rehabilitation and release of marine mammals in the United States: Risks and benefits. Mar. Mammal Sci. 23, 731–750 (2007).Article 

Google Scholar 
Leray, M. et al. A new versatile primer set targeting a short fragment of the mitochondrial COI region for metabarcoding metazoan diversity: Application for characterizing coral reef fish gut contents. Front. Zool. 10, 1–14 (2013).Article 

Google Scholar 
Geller, J., Meyer, C. & Parker, M. Redesign of PCR primers for mitochondrial cytochrome c oxidase subunit I for marine invertebrates and application in all-taxa biotic surveys. Mol. Ecol. Resour. 13(5), 851–861. https://doi.org/10.1111/1755-0998.12138 (2013).CAS 
Article 

Google Scholar 
Blaxter, M. L. et al. A molecular evolutionary framework for the phylum Nematoda. Nature https://doi.org/10.1038/32160 (1998).Article 

Google Scholar 
Sinniger, F. et al. Worldwide analysis of sedimentary DNA reveals major gaps in taxonomic knowledge of deep-sea benthos. Front. Mar. Sci. 3, 1–14 (2016).Article 

Google Scholar 
Brandt, M. I. et al. Bioinformatic pipelines combining denoising and clustering tools allow for more comprehensive prokaryotic and eukaryotic metabarcoding. Mol. Ecol. Resour. 21(6), 1904–1921 (2021).Article 

Google Scholar 
Martin, M. Cutadapt removes adapter sequences from high-throughput sequencing reads. EMBnet.journal https://doi.org/10.14806/ej.17.1.200 (2011).Article 

Google Scholar 
Callahan, B. J. et al. DADA2: High-resolution sample inference from Illumina amplicon data. Nat. Methods 13, 581–583 (2016).CAS 
Article 

Google Scholar 
Antich, A., Palacin, C., Wangensteen, O. S. & Turon, X. To denoise or to cluster, that is not the question: Optimizing pipelines for COI metabarcoding and metaphylogeography. BMC Bioinform. 22, 1–25 (2021).Article 

Google Scholar 
Mahé, F., Rognes, T., Quince, C., de Vargas, C. & Dunthorn, M. Swarmv2: Highly-scalable and high-resolution amplicon clustering. PeerJ 2015, 1–12 (2015).
Google Scholar 
Quast, C. et al. The SILVA ribosomal RNA gene database project: Improved data processing and web-based tools. Nucleic Acids Res. https://doi.org/10.1093/nar/gks1219 (2013).Article 

Google Scholar 
Machida, R. J., Leray, M., Ho, S.-L. & Knowlton, N. Metazoan mitochondrial gene sequence reference datasets for taxonomic assignment of environmental samples. Sci. Data 4, 170027 (2017).CAS 
Article 

Google Scholar 
Wang, Q., Garrity, G. M., Tiedje, J. M. & Cole, J. R. Naıve Bayesian classifier for rapid assignment of rRNA sequences.pdf. Appl. Environ. Microbiol. 73, 5261–5267 (2007).ADS 
CAS 
Article 

Google Scholar 
Davis, N. M., Di Proctor, M., Holmes, S. P., Relman, D. A. & Callahan, B. J. Simple statistical identification and removal of contaminant sequences in marker-gene and metagenomics data. Microbiome https://doi.org/10.1186/s40168-018-0605-2 (2018).Article 

Google Scholar 
Wangensteen, O. S., Palacín, C., Guardiola, M. & Turon, X. DNA metabarcoding of littoral hardbottom communities: High diversity and database gaps revealed by two molecular markers. PeerJ 2018, 1–30 (2018).
Google Scholar 
Schnell, I. B., Bohmann, K. & Gilbert, M. T. P. Tag jumps illuminated – reducing sequence-to-sample misidentifications in metabarcoding studies. Mol. Ecol. Resour. 15, 1289–1303 (2015).CAS 
Article 

Google Scholar 
Song, X. et al. A new deep-sea hydroid (Cnidaria:Hydrozoa ) from the Bering Sea Basin reveals high genetic relevance to Arctic and adjacent shallow-water species. Polar Biol. 39, 461–471 (2016).Article 

Google Scholar 
Frøslev, T. G. et al. Algorithm for post-clustering curation of DNA amplicon data yields reliable biodiversity estimates. Nat. Commun. 8, 1–11 (2017).Article 

Google Scholar 
Vacquié-Garcia, J., Lydersen, C., Ims, R. A. & Kovacs, K. M. Habitats and movement patterns of white whales Delphinapterus leucas in Svalbard, Norway in a changing climate. Mov. Ecol. 6, 1–12 (2018).Article 

Google Scholar 
Kastelein, R. A., Nieuwstraten, S. H. & Verstegen, M. W. A. Passage time of carmine red dye through the digestion tract . In The Biology of the Harbour Porpoise 235–245 (1997).Lesage, V., Lair, S., Turgeon, S. & Beland, P. Diet of St. Lawrence Estuary Beluga (Delphinapterus leucas) in a changing ecosystem. Can. Field-Nat. 134, 21–35 (2020).Article 

Google Scholar 
Bluhm, B. A. & Gradinger, R. Regional variability in food availability for arctic marine mammals. Ecol. Appl. 18, S77–S96 (2008).Article 

Google Scholar 
Quakenbush, L. T. et al. Diet of beluga whales, Delphinapterus leucas, in Alaska from stomach contents, March-November. Mar. Fish. Rev. 77, 70–84 (2015).Article 

Google Scholar 
Choy, E. S. et al. Variation in the diet of beluga whales in response to changes in prey availability: Insights on changes in the Beaufort Sea ecosystem. Mar. Ecol. Prog. Ser. 647, 195–210 (2020).ADS 
CAS 
Article 

Google Scholar 
Mychek-Londer, J. G., Chaganti, S. R. & Heath, D. D. Metabarcoding of native and invasive species in stomach contents of Great Lakes fishes. PLoS ONE 15, 1–22 (2020).Article 

Google Scholar 
Nedreaas, K. Food and feeding habits of young saithe, Pollachius virens (L.), on the coast of Western Norway. Fisk. Skr. Ser. Havundersokelser 18, 263–301 (1987).
Google Scholar 
Højgaard, D. P. Food and parasitic nematodes of saithe, Pollachius virens (L.), from the Faroe Islands. Sarsia 84, 473–478 (1999).Article 

Google Scholar 
Ekbaum, E. Notes on the occurrence of Acanthocephala in Pacific fishes: I. Echinorhynchus gadi (Zoega) Müller in salmon and E. lageniformis sp. nov. and Corynosoma strumosum (Rudolphi) in two species of flounder. Parasitology 30, 267–274 (1938).Article 

Google Scholar 
Baptista-Fernandes, T. et al. Human gastric hyperinfection by Anisakis simplex: A severe and unusual presentation and a brief review. Int. J. Infect. Dis. 64, 38–41 (2017).Article 

Google Scholar 
Hubert, B., Bacou, J. & Belveze, H. Epidemiology of human anisakiasis: Incidence and sources in France. Am. J. Trop. Med. Hyg. 40, 301–303 (1989).CAS 
Article 

Google Scholar 
Hays, R., Measures, L. N. & Huot, J. Capelin (Mallotus villosus) and herring (Clupea harengus) as paratenic hosts of Anisakis simplex, a parasite of beluga (Delphinapterus leucas) in the St. Lawrence estuary. Can. J. Zool. 78, 1411–1417 (1998).Article 

Google Scholar 
Yanong, R. P. E. Nematode (Roundworm) Infections in Fish Vol. 1, 1–9 (2002).Jauniaux, T. et al. Post-mortem findings and causes of death of harbour porpoises (Phocoena phocoena) stranded from 1990 to 2000 along the coastlines of Belgium and Northern France. J. Compar. Pathol. 126, 243–253 (2002).CAS 
Article 

Google Scholar  More