Philippa Kaur

Distinctive temperature responses along a substrate gradient
Within the temperature range of ~14 to ~34 °C in our incubations, the observed AORs at the ambient substrate level (AORambient, see Methods) varied over 3 orders of magnitude, from 0.5 to ~4000 nM d−1, across a wide spectrum of ambient ammonium levels ranging from 14 nM to 96 μM (Fig. 1). Three different types of temperature response of AORambient patterns at estuarine, shelf, and sea basin stations were observed: (I) a positive response with a Topt of ≥34 °C (Fig. 1a, b); (II) a negative response, which has never been reported before, with a Topt of ≤14 °C (Fig. 1d–f); and (III) a dome-shaped response with a Topt of 20–29 °C (Fig. 1c, g–i).
The Type I pattern was observed at two of the three estuarine stations (JLR1 and JLR2, Fig. 1a, b) where ammonium concentrations were high (≥24 μM), and the AOR increased linearly as the temperature increased from 14 to 34 °C. In these cases, the Topt was equal to or higher than the maximum experimental temperature of 34 °C (Fig. 1a, b). The Type II pattern was observed at the shelf stations (N1, M1, and M2), where NH4+ concentrations ranged from 45 to 550 nM (Fig. 1d–f). In contrast to the Type I pattern, the Topt of the Type II pattern was equal to or lower than the minimum experimental temperature of 14 °C, showing a continuously decreasing AOR as temperature increased. The Type III pattern was observed at station JLR3 (outer estuary), N2 (shelf), N3 and J1 (basin), for which the Topt of the AOR varied from 20 to 29 °C, with rates decreasing toward both higher and lower temperatures (Fig. 1c, g–i). The NH4+ concentrations of the Type III stations ranged from 14 to 5000 nM. Nevertheless, the highest Topt values were observed at coastal sites with the highest ambient ammonium concentrations (Fig. 1).
Substrate regulates AOR and its thermal optimum temperature
For those stations with low ammonium concentrations, the AOR at in situ temperature increased when the substrate was enriched (AORenriched, additions of 2000 nM 15NH4+) (Fig. 1f, i). Meanwhile, the Topt of the AOR shifted significantly toward higher values (t test, p  Q10Vmax (Supplementary Fig. 2). This criterion was fulfilled in both J1 and JLR4 cases (Fig. 2c–f). We see the positive shift of Topt due to the ammonium addition (up to ~100 nM at J1 station and up to ~10 μM at JLR4) can be closely predicted (Fig. 3c, d; Supplementary Fig. 2). Overall, the DAMM model successfully predicts the entire thermal response curve, including rates and Topt, except when the manipulated temperatures exceed Topt-sat (Fig. 3; Supplementary Fig. 2). AOR drops significantly when temperature is greater than the Topt-sat, so heat-impaired biological enzyme activity27,28 might result in deviations from the relationship between Vmax (Km) values and temperature from the Arrhenius law.
Fig. 3: Validation plot for the rate predictions and observations.

a, b Scatter plot of the predicted rates via the Dual Arrhenius and Michaelis–Menten kinetics model (DAMM model) and the measured rates under different substrate concentrations and temperatures (below the optimum temperature in substrate-saturated conditions, Topt-sat). Linear regressions between the model predictions and the measurements are presented (two-sided t test was used to generate the p value (95% confidence) to measure the strength of correlation coefficient. p values are uncorrected). c, d The rate patterns (dots) against temperature under different substrate concentrations. Curves stand for the predicted rates derived from the DAMM model and the symbols represent the measured rates. The shades denote the uncertainty of model prediction. The dashed black vertical lines represent the Topt-sat. The measured rates in (a, c) are presented as mean values, instead of standard deviation the given bars indicate the variation range of two independent experiments. The measured rates in (b, d) and the predicted rates in (a–d) are expressed as the mean values ± SD (n = 10 in (a, c); n = 48 in (b, d); independent experiments).

Full size image

The substrate-dependent thermal optimum is attributable to the effect of temperature on biochemical kinetics and the structural stability of the enzymes. Increasing temperature promotes catalytic rate, thus, Vmax increases due to increasing kinetic energy of reactants and rates of collision, as well as higher structural flexibility of enzymes27,29. However, higher structural flexibility (lower stability) also results in active sites with a reduced ability for ligand recognition and binding, therefore, lower kinetic efficiency. Accordingly, one important physiological response of an organism to rising temperature will be a reduction in substrate affinity (Supplementary Table 2), and thus higher substrate demands (i.e. higher Km value)25,29,30,31. In other words, higher substrate levels help to compensate for enzyme structural stability losses and so promote growth rates at higher temperature. Note that some other microbes may respond differently to temperature, with Q10Km ≤ Q10Vmax for instance. This may lead to predictable yet unidirectional rate increases in response to warming (without substrate-regulated Topt) until the Topt-sat is reached, regardless of substrate changes.
Similarly, nutrient-dependent Topt has been reported for phytoplankton growth in pure cultures previously. For instance, Thomas et al.32 indicated that the Topt for growth of a marine diatom was a saturating function of major nutrient (nitrate and phosphate) concentration, and that the Topt could decrease by 3–6 °C at low concentrations relative to that at saturated nutrient levels. In addition to studies of pure cultures, field studies have also suggested that organisms may tolerate higher temperature stresses when nutrients are more abundant. For example, kelp (Laminaria saccarina) with high nitrogen reserves have more capacity for thermal adaptation33, while corals with symbionts limited by phosphate are more susceptible to heat-induced bleaching34. Although these examples are functionally and taxonomically distant from AOA and ammonia-oxidizing bacteria (AOB), strong similarities in substrate/nutrient regulation characteristics may imply a similar mechanism of enzymatic thermal responses between chemoautotrophs and photoautotrophs.
Nevertheless, the higher thermal optimum of AO in the estuarine system (e.g., JRL1, JLR2, and JLR3) than in the offshore environment (e.g., N3 and J1) can be explained by a substrate-regulated Topt. Note that field AOR represents explicitly the collective activity of the AO community composed of AOA and AOB, which may have distinctive thermal tolerances and affinities for substrate. Therefore, community structure very likely plays a role in modulating the thermal response patterns of community AOR in the field environments, in addition to substrate concentration.
Rate proportion and community thermal optimum
To further examine to what extent the community structure (proportions of AOA and AOB) might shape the thermal response patterns of community AOR observed in the field, we added allylthiourea (ATU) to inhibit the activity of AOB for rate discrimination (see Methods; Supplementary Discussions). Results showed that the inhibitory efficiency of AOR was gradually reduced with increasing offshore distance (Fig. 4). That is, from the estuary (JLR4, JLR1, JLR2, and JLR3) to the shelf (N1 and N2) and the sea basin (N3), the relative contribution of AOB to the community AOR dropped from as high as ~100% in the upper estuary down to ~70% in the shelf transition zone, and near 0% in the basin. Meanwhile, the AOA/AOB gene copies data (see Supplementary Methods) from estuary to sea basin (Fig. 4) also clearly show that the abundance of AOA relative to AOB increased exponentially with increasing offshore distance. A similar offshore pattern of community distribution was also observed in other regions, such as from the Pearl River estuary to the South China Sea35, and from the freshwater region of the Chesapeake Bay to the coastal and open ocean water column36. This pattern suggests that AOB strongly prefer substrate-replete niches, and vice versa for AOA20,37, agreeing well with our M–M experimental data that the substrate saturation condition for AOB-dominated water at JLR4 was several orders of magnitude higher than that for AOA-dominated water from J1 (Supplementary Table 2). The Km values of AOB in JLR4 varied from 7 to 55 μM in accordance with varying temperatures from 10 to 37 °C, while the Km values of AOA in J1 varied from 13 to 44 nM over a similar temperature range (Supplementary Table 2). Results were supportive of previous pure culture and field studies which showed the minimum ammonium demand for AOB is >1 μM and Km values range from 28 to 4000 μM38,39,40,41, while minimum ammonium demand and Km value for AOA are  Q10Vmax, which also results in a reduction in α during warming for both AOA and AOB. Yet, the relative reduction in AOA specific affinity as temperature increases is more significant (Fig. 5a), suggesting AOAs are more competitive in low temperature environments relative to AOBs, and so may not be favored in a warming ocean. On the other hand, the specific affinity of AOBs is insensitive to temperature change, suggesting their adaptation to nearshore environments with greater temperature fluctuation. The seaward gradient in temperature fluctuations and ammonium concentrations determine the nitrifier community, thus, thermal response pattern of community AOR observed in the field.
Fig. 5: Thermal response projections in near- and offshore regions.

a The thermal responses of specific affinity at the J1 and JLR4 stations. Data are expressed as the mean values ± SD (n = 10 in J1 station; n = 48 in JLR4 station; independent experiments). b Normalized warming-driven variations in ammonia oxidation rates. Rate changed (%) is relative to the ammonia oxidation rate (AOR) at in situ temperature. The mean increase (nearshore hollow dots) is denoted by the dashed line and the mean reduction (offshore, solid dots) is denoted by the solid black line. The shaded area represents the 4 °C increase in temperature mentioned in the IPCC study. Note that the stations where the surface salinities are lower than 32 are classified as nearshore station, and the others are classified as offshore stations. The data in (b) are presented as mean values, instead of standard deviation the given bars indicate the variation range of two independent experiments.

Full size image

To predict the trends of AOR in different geographic spaces in warming ocean, we compile the available marine AOR data to examine the AOR changes empirically. If we assume the biogeographic distribution of AO community remains unchanged and consider solely the warming effect on AOR relative to the onsite temperature, we found the thermal responses of AOR in nearshore and offshore are quite different (Fig. 5b). More specifically, the higher Topt of these AO communities in nearshore regimes allows ocean warming to promote coastal AOR when the temperature change increment is More

1.
Flemming, H.-C. et al. Biofilms: an emergent form of bacterial life. Nat. Rev. Microbiol. 14, 563. https://doi.org/10.1038/nrmicro.2016.94 (2016).
CAS  Article  PubMed  Google Scholar 
2.
Boddey, J. A., Flegg, C. P., Day, C. J., Beacham, I. R. & Peak, I. R. Temperature-regulated microcolony formation by Burkholderia pseudomallei requires pilA and enhances association with cultured human cells. Infect. Immunity 74, 5374–5381. https://doi.org/10.1128/iai.00569-06 (2006).
CAS  Article  Google Scholar 

3.
Lister, J. L. & Horswill, A. R. Staphylococcus aureus biofilms: recent developments in biofilm dispersal. Front. Cell. Infect. Microbiol. https://doi.org/10.3389/fcimb.2014.00178 (2014).
Article  PubMed  PubMed Central  Google Scholar 

4.
Whitchurch, C. B., Tolker-Nielsen, T., Ragas, P. C. & Mattick, J. S. Extracellular DNA required for bacterial biofilm formation. Science 295, 1487–1487. https://doi.org/10.1126/science.295.5559.1487 (2002).
CAS  Article  PubMed  Google Scholar 

5.
Foster, T. J., Geoghegan, J. A., Ganesh, V. K. & Höök, M. Adhesion, invasion and evasion: the many functions of the surface proteins of Staphylococcus aureus. Nat. Rev. Microbiol. 12, 49. https://doi.org/10.1038/nrmicro3161 (2013).
CAS  Article  Google Scholar 

6.
Mack, D. et al. The intercellular adhesin involved in biofilm accumulation of Staphylococcus epidermidis is a linear beta-1,6-linked glucosaminoglycan: purification and structural analysis. J. Bacteriol. 178, 175–183 (1996).
CAS  Article  Google Scholar 

7.
Limoli, D. H., Jones, C. J. & Wozniak, D. J. Bacterial extracellular polysaccharides in biofilm formation and function. Microbiol. Spectrum. https://doi.org/10.1128/microbiolspec.MB-0011-2014 (2015).
Article  Google Scholar 

8.
Gallaher, T. K., Wu, S., Webster, P. & Aguilera, R. Identification of biofilm proteins in non-typeable Haemophilus Influenzae. BMC Microbiol. 6, 65. https://doi.org/10.1186/1471-2180-6-65 (2006).
CAS  Article  PubMed  PubMed Central  Google Scholar 

9.
Hu, W. et al. DNA builds and strengthens the extracellular matrix in Myxococcus xanthus biofilms by interacting with exopolysaccharides. PLoS ONE 7, e51905. https://doi.org/10.1371/journal.pone.0051905 (2012).
ADS  CAS  Article  PubMed  PubMed Central  Google Scholar 

10.
Boles, B. R. & Horswill, A. R. Staphylococcal biofilm disassembly. Trends Microbiol. 19, 449–455. https://doi.org/10.1016/j.tim.2011.06.004 (2011).
CAS  Article  PubMed  PubMed Central  Google Scholar 

11.
Reen, F. J. et al. Bile signalling promotes chronic respiratory infections and antibiotic tolerance. Sci. Rep. 6, 29768 (2016).
ADS  CAS  Article  Google Scholar 

12.
Duanis-Assaf, D., Steinberg, D., Chai, Y. & Shemesh, M. The LuxS based quorum sensing governs lactose induced biofilm formation by Bacillus subtilis. Front. Microbiol. https://doi.org/10.3389/fmicb.2015.01517 (2016).
Article  PubMed  PubMed Central  Google Scholar 

13.
Le, K. Y. & Otto, M. Quorum-sensing regulation in staphylococci-an overview. Front. Microbiol. 6, 1174–1174. https://doi.org/10.3389/fmicb.2015.01174 (2015).
Article  PubMed  PubMed Central  Google Scholar 

14.
Qi, L. et al. Relationship between antibiotic resistance, biofilm formation, and biofilm-specific resistance in Acinetobacter baumannii. Front. Microbiol. 7, 483–483. https://doi.org/10.3389/fmicb.2016.00483 (2016).
Article  PubMed  PubMed Central  Google Scholar 

15.
O’Callaghan, A. & van Sinderen, D. Bifidobacteria and their role as members of the human gut microbiota. Front. Microbiol. 7, 925–925. https://doi.org/10.3389/fmicb.2016.00925 (2016).
Article  PubMed  PubMed Central  Google Scholar 

16.
Hill, C. et al. The International Scientific Association for Probiotics and Prebiotics consensus statement on the scope and appropriate use of the term probiotic. Nat. Rev. Gastroenterol. Hepatol. 11, 506. https://doi.org/10.1038/nrgastro.2014.66 (2014).
Article  PubMed  Google Scholar 

17.
Sánchez, B., Ruiz, L., Gueimonde, M., Ruas-Madiedo, P. & Margolles, A. Adaptation of bifidobacteria to the gastrointestinal tract and functional consequences. Pharmacol. Res. 69, 127–136. https://doi.org/10.1016/j.phrs.2012.11.004 (2013).
Article  PubMed  Google Scholar 

18.
Holm, R., Müllertz, A. & Mu, H. Bile salts and their importance for drug absorption. Int. J. Pharm. 453, 44–55. https://doi.org/10.1016/j.ijpharm.2013.04.003 (2013).
CAS  Article  PubMed  Google Scholar 

19.
Islam, K. B. M. S. et al. Bile acid is a host factor that regulates the composition of the cecal microbiota in rats. Gastroenterology 141, 1773–1781. https://doi.org/10.1053/j.gastro.2011.07.046 (2011).
CAS  Article  PubMed  Google Scholar 

20.
Begley, M., Gahan, C. G. & Hill, C. The interaction between bacteria and bile. FEMS Microbiol. Rev. 29, 625–651. https://doi.org/10.1016/j.femsre.2004.09.003 (2005).
CAS  Article  PubMed  Google Scholar 

21.
Ruiz, L., Margolles, A. & Sanchez, B. Bile resistance mechanisms in Lactobacillus and Bifidobacterium. Front. Microbiol. 4, 396. https://doi.org/10.3389/fmicb.2013.00396 (2013).
Article  PubMed  PubMed Central  Google Scholar 

22.
Price, C. E., Reid, S. J., Driessen, A. J. & Abratt, V. R. The Bifidobacterium longum NCIMB 702259T ctr gene codes for a novel cholate transporter. Appl. Environ. Microbiol. 72, 923–926. https://doi.org/10.1128/aem.72.1.923-926.2006 (2006).
CAS  Article  PubMed  PubMed Central  Google Scholar 

23.
Gueimonde, M., Garrigues, C., van Sinderen, D., de los Reyes-Gavilan, C. G. & Margolles, A. Bile-inducible efflux transporter from Bifidobacterium longum NCC2705, conferring bile resistance. Appl. Environ. Microbiol. 75, 3153–3160. https://doi.org/10.1128/aem.00172-09 (2009).
CAS  Article  PubMed  PubMed Central  Google Scholar 

24.
Ruiz, L., Zomer, A., O’Connell-Motherway, M., van Sinderen, D. & Margolles, A. Discovering novel bile protection systems in Bifidobacterium breve UCC2003 through functional genomics. Appl. Environ. Microbiol. 78, 1123–1131. https://doi.org/10.1128/aem.06060-11 (2012).
CAS  Article  PubMed  PubMed Central  Google Scholar 

25.
Ruiz, L., Sánchez, B., Ruas-Madiedo, P., De Los Reyes-Gavilán, C. G. & Margolles, A. Cell envelope changes in Bifidobacterium animalis ssp. lactis as a response to bile. FEMS Microbiol. Lett. 274, 316–322. https://doi.org/10.1111/j.1574-6968.2007.00854.x (2007).
CAS  Article  PubMed  Google Scholar 

26.
Gómez Zavaglia, A., Kociubinski, G., Pérez, P., Disalvo, E. & De Antoni, G. Effect of bile on the lipid composition and surface properties of bifidobacteria. J. Appl. Microbiol. 93, 794–799. https://doi.org/10.1046/j.1365-2672.2002.01747.x (2002).
Article  PubMed  Google Scholar 

27.
An, H. et al. Integrated transcriptomic and proteomic analysis of the bile stress response in a centenarian-originated probiotic Bifidobacterium longum BBMN68. Mol. Cell. Proteom. 13, 2558–2572. https://doi.org/10.1074/mcp.M114.039156 (2014).
CAS  Article  Google Scholar 

28.
Sanchez, B., de los Reyes-Gavilan, C. G. & Margolles, A. The F1F0-ATPase of Bifidobacterium animalis is involved in bile tolerance. Environ. Microbiol. 8, 1825–1833. https://doi.org/10.1111/j.1462-2920.2006.01067.x (2006).
CAS  Article  PubMed  Google Scholar 

29.
Sanchez, B., Noriega, L., Ruas-Madiedo, P., de los Reyes-Gavilan, C. G. & Margolles, A. Acquired resistance to bile increases fructose-6-phosphate phosphoketolase activity in Bifidobacterium. FEMS Microbiol. Lett. 235, 35–41. https://doi.org/10.1016/j.femsle.2004.04.009 (2004).
CAS  Article  PubMed  Google Scholar 

30.
Sanchez, B. et al. Proteomic analysis of global changes in protein expression during bile salt exposure of Bifidobacterium longum NCIMB 8809. J. Bacteriol. 187, 5799–5808. https://doi.org/10.1128/jb.187.16.5799-5808.2005 (2005).
CAS  Article  PubMed  PubMed Central  Google Scholar 

31.
Noriega, L., Gueimonde, M., Sanchez, B., Margolles, A. & de los Reyes-Gavilan, C. G. Effect of the adaptation to high bile salts concentrations on glycosidic activity, survival at low PH and cross-resistance to bile salts in Bifidobacterium. Int. J. Food Microbiol. 94, 79–86. https://doi.org/10.1016/j.ijfoodmicro.2004.01.003 (2004).
CAS  Article  PubMed  Google Scholar 

32.
Tanaka, H., Hashiba, H., Kok, J. & Mierau, I. Bile salt hydrolase of Bifidobacterium longum-biochemical and genetic characterization. Appl. Environ. Microbiol. 66, 2502–2512. https://doi.org/10.1128/aem.66.6.2502-2512.2000 (2000).
CAS  Article  PubMed  PubMed Central  Google Scholar 

33.
Noriega, L., Cuevas, I., Margolles, A. & de los Reyes-Gavilán, C. G. Deconjugation and bile salts hydrolase activity by Bifidobacterium strains with acquired resistance to bile. Int. Dairy J. 16, 850–855. https://doi.org/10.1016/j.idairyj.2005.09.008 (2006).
CAS  Article  Google Scholar 

34.
Ambalam, P., Kondepudi, K. K., Nilsson, I., Wadstrom, T. & Ljungh, A. Bile enhances cell surface hydrophobicity and biofilm formation of bifidobacteria. Appl. Biochem. Biotechnol. 172, 1970–1981. https://doi.org/10.1007/s12010-013-0596-1 (2014).
CAS  Article  PubMed  Google Scholar 

35.
Pumbwe, L. et al. Bile salts enhance bacterial co-aggregation, bacterial-intestinal epithelial cell adhesion, biofilm formation and antimicrobial resistance of Bacteroides fragilis. Microb. Pathog. 43, 78–87. https://doi.org/10.1016/j.micpath.2007.04.002 (2007).
CAS  Article  PubMed  Google Scholar 

36.
Lebeer, S., Verhoeven, T. L., Perea Velez, M., Vanderleyden, J. & De Keersmaecker, S. C. Impact of environmental and genetic factors on biofilm formation by the probiotic strain Lactobacillus rhamnosus GG. Appl. Environ. Microbiol. 73, 6768–6775. https://doi.org/10.1128/aem.01393-07 (2007).
CAS  Article  PubMed  PubMed Central  Google Scholar 

37.
Macfarlane, S. & Macfarlane, G. T. Composition and metabolic activities of bacterial biofilms colonizing food residues in the human gut. Appl. Environ. Microbiol. 72, 6204–6211. https://doi.org/10.1128/aem.00754-06 (2006).
CAS  Article  PubMed  PubMed Central  Google Scholar 

38.
Macfarlane, M. J. H. G. T. M. S. Bacterial growth and metabolism on surfaces in the large intestine. Microb. Ecol. Health Dis. 12, 64–72. https://doi.org/10.1080/089106000750060314 (2000).
Article  Google Scholar 

39.
Pereira, C. S., Thompson, J. A. & Xavier, K. B. AI-2-mediated signalling in bacteria. FEMS Microbiol. Rev. 37, 156–181. https://doi.org/10.1111/j.1574-6976.2012.00345.x (2013).
CAS  Article  PubMed  Google Scholar 

40.
Hammer, B. K. & Bassler, B. L. Quorum sensing controls biofilm formation in Vibrio cholerae. Mol. Microbiol. 50, 101–104. https://doi.org/10.1046/j.1365-2958.2003.03688.x (2003).
CAS  Article  PubMed  Google Scholar 

41.
Solano, C., Echeverz, M. & Lasa, I. Biofilm dispersion and quorum sensing. Curr. Opin. Microbiol. 18, 96–104. https://doi.org/10.1016/j.mib.2014.02.008 (2014).
CAS  Article  PubMed  Google Scholar 

42.
Sun, Z., He, X., Brancaccio, V. F., Yuan, J. & Riedel, C. U. Bifidobacteria exhibit LuxS-dependent autoinducer 2 activity and biofilm formation. PLoS ONE 9, e88260. https://doi.org/10.1371/journal.pone.0088260 (2014).
ADS  CAS  Article  PubMed  PubMed Central  Google Scholar 

43.
Christiaen, S. E. et al. Autoinducer-2 plays a crucial role in gut colonization and probiotic functionality of Bifidobacterium breve UCC2003. PLoS ONE 9, e98111. https://doi.org/10.1371/journal.pone.0098111 (2014).
ADS  CAS  Article  PubMed  PubMed Central  Google Scholar 

44.
Yuan, J. et al. A proteome reference map and proteomic analysis of Bifidobacterium longum NCC2705. Mol. Cell. Proteom. 5, 1105–1118. https://doi.org/10.1074/mcp.M500410-MCP200 (2006).
CAS  Article  Google Scholar 

45.
D’Urzo, N. et al. Acidic pH strongly enhances in vitro biofilm formation by a subset of hypervirulent ST-17 Streptococcus agalactiae strains. Appl. Environ. Microbiol. 80, 2176–2185. https://doi.org/10.1128/aem.03627-13 (2014).
Article  PubMed  PubMed Central  Google Scholar 

46.
O’Neill, E. et al. Association between methicillin susceptibility and biofilm regulation in Staphylococcus aureus isolates from device-related infections. J. Clin. Microbiol. 45, 1379–1388. https://doi.org/10.1128/jcm.02280-06 (2007).
CAS  Article  PubMed  PubMed Central  Google Scholar 

47.
Hung, D. T., Zhu, J., Sturtevant, D. & Mekalanos, J. J. Bile acids stimulate biofilm formation in Vibrio cholerae. Mol. Microbiol. 59, 193–201. https://doi.org/10.1111/j.1365-2958.2005.04846.x (2006).
CAS  Article  PubMed  Google Scholar 

48.
Maze, A., O’Connell-Motherway, M., Fitzgerald, G. F., Deutscher, J. & van Sinderen, D. Identification and characterization of a fructose phosphotransferase system in Bifidobacterium breve UCC2003. Appl. Environ. Microbiol. 73, 545–553. https://doi.org/10.1128/aem.01496-06 (2007).
CAS  Article  PubMed  Google Scholar 

49.
Lanigan, N., Bottacini, F., Casey, P. G., O’Connell Motherway, M. & van Sinderen, D. Genome-wide search for genes required for bifidobacterial growth under iron-limitation. Front. Microbiol. 8, 964. https://doi.org/10.3389/fmicb.2017.00964 (2017).
Article  PubMed  PubMed Central  Google Scholar 

50.
Ruiz, L., Motherway, M. O., Lanigan, N. & van Sinderen, D. Transposon mutagenesis in Bifidobacterium breve: construction and characterization of a Tn5 transposon mutant library for Bifidobacterium breve UCC2003. PLoS ONE 8, e64699. https://doi.org/10.1371/journal.pone.0064699 (2013).
ADS  CAS  Article  PubMed  PubMed Central  Google Scholar 

51.
Fanning, S. et al. Bifidobacterial surface-exopolysaccharide facilitates commensal-host interaction through immune modulation and pathogen protection. Proc. Natl. Acad. Sci. USA. 109, 2108–2113. https://doi.org/10.1073/pnas.1115621109 (2012).
ADS  Article  PubMed  Google Scholar 

52.
Alonso-Casajus, N. et al. Glycogen phosphorylase, the product of the glgP Gene, catalyzes glycogen breakdown by removing glucose units from the nonreducing ends in Escherichia coli. J. Bacteriol. 188, 5266–5272. https://doi.org/10.1128/jb.01566-05 (2006).
CAS  Article  PubMed  PubMed Central  Google Scholar 

53.
Nocek, B. P., Gillner, D. M., Fan, Y., Holz, R. C. & Joachimiak, A. Structural basis for catalysis by the mono- and dimetalated forms of the dapE-encoded N-succinyl-L, L-diaminopimelic acid desuccinylase. J. Mol. Biol. 397, 617–626. https://doi.org/10.1016/j.jmb.2010.01.062 (2010).
CAS  Article  PubMed  PubMed Central  Google Scholar 

54.
Ethapa, T. et al. Multiple factors modulate biofilm formation by the anaerobic pathogen Clostridium difficile. J. Bacteriol. 195, 545–555. https://doi.org/10.1128/jb.01980-12 (2013).
Article  PubMed  PubMed Central  Google Scholar 

55.
Donlan, R. M. Biofilms: microbial life on surfaces. Emerg. Infect. Dis. 8, 881–890. https://doi.org/10.3201/eid0809.020063 (2002).
Article  PubMed  PubMed Central  Google Scholar 

56.
Bottacini, F., Ventura, M., van Sinderen, D. & O’Connell Motherway, M. Diversity, ecology and intestinal function of bifidobacteria. Microbial Cell Fact. https://doi.org/10.1186/1475-2859-13-s1-s4 (2014).
Article  Google Scholar 

57.
Legrand-Defretin, V., Juste, C., Henry, R. & Corring, T. Ion-pair high-performance liquid chromatography of bile salt conjugates: Application to pig bile. Lipids 26, 578–583. https://doi.org/10.1007/bf02536421 (1991).
CAS  Article  PubMed  Google Scholar 

58.
Sanchez, B. et al. Adaptation and response of Bifidobacterium animalis subsp. lactis to bile: a proteomic and physiological approach. Appl. Environ. Microbiol. 73, 6757–6767. https://doi.org/10.1128/aem.00637-07 (2007).
CAS  Article  PubMed  PubMed Central  Google Scholar 

59.
Ruas-Madiedo, P., Hernandez-Barranco, A., Margolles, A. & de los Reyes-Gavilan, C. G. A bile salt-resistant derivative of Bifidobacterium animalis has an altered fermentation pattern when grown on glucose and maltose. Appl. Environ. Microbiol. 71, 6564–6570. https://doi.org/10.1128/aem.71.11.6564-6570.2005 (2005).
CAS  Article  PubMed  PubMed Central  Google Scholar 

60.
Ruiz, L. et al. The cell-envelope proteome of Bifidobacterium longum in an in vitro bile environment. Microbiology 155, 957–967. https://doi.org/10.1099/mic.0.024273-0 (2009).
CAS  Article  PubMed  Google Scholar 

61.
Wang, G. et al. Functional role of oppA encoding an oligopeptide-binding protein from Lactobacillus salivarius Ren in bile tolerance. J. Ind. Microbiol. Biotechnol. 42, 1167–1174. https://doi.org/10.1007/s10295-015-1634-5 (2015).
CAS  Article  PubMed  Google Scholar 

62.
Lebeer, S. et al. Impact of luxS and suppressor mutations on the gastrointestinal transit of Lactobacillus rhamnosus GG. Appl. Environ. Microbiol. 74, 4711–4718. https://doi.org/10.1128/aem.00133-08 (2008).
CAS  Article  PubMed  PubMed Central  Google Scholar 

63.
Wilson, C. M., Aggio, R. B., O’Toole, P. W., Villas-Boas, S. & Tannock, G. W. Transcriptional and metabolomic consequences of LuxS inactivation reveal a metabolic rather than quorum-sensing role for LuxS in Lactobacillus reuteri 100–23. J. Bacteriol. 194, 1743–1746. https://doi.org/10.1128/jb.06318-11 (2012).
Article  PubMed  PubMed Central  Google Scholar 

64.
Rezzonico, F. & Duffy, B. Lack of genomic evidence of AI-2 receptors suggests a non-quorum sensing role for luxS in most bacteria. BMC Microbiol. 8, 154. https://doi.org/10.1186/1471-2180-8-154 (2008).
CAS  Article  PubMed  PubMed Central  Google Scholar 

65.
Giddens, S. R. et al. Mutational activation of niche-specific genes provides insight into regulatory networks and bacterial function in a complex environment. Proc. Natl. Acad. Sci. USA 104, 18247. https://doi.org/10.1073/pnas.0706739104 (2007).
ADS  Article  PubMed  Google Scholar 

66.
Thompson, A. P. et al. Glycolysis and pyrimidine biosynthesis are required for replication of adherent–invasive Escherichia coli in macrophages. Microbiology 162, 954–965. https://doi.org/10.1099/mic.0.000289 (2016).
CAS  Article  PubMed  Google Scholar 

67.
Sambrook, J. & Russell, D. Molecular Cloning: A Laboratory Manual 2001 Cold Spring Harbor (Cold Spring Harbor Laboratory Press, New York, 2001).
Google Scholar 

68.
O’Riordan, K. & Fitzgerald, G. F. Molecular characterisation of a 575-kb cryptic plasmid from Bifidobacterium breve NCFB 2258 and determination of mode of replication. FEMS Microbiol. Lett. 174, 285–294. https://doi.org/10.1111/j.1574-6968.1999.tb13581.x (1999).
CAS  Article  PubMed  Google Scholar 

69.
Alessandri, G. et al. Ability of bifidobacteria to metabolize chitin-glucan and its impact on the gut microbiota. Sci. Rep. 9, 5755–5755. https://doi.org/10.1038/s41598-019-42257-z (2019).
ADS  CAS  Article  PubMed  PubMed Central  Google Scholar 

70.
Duranti, S. et al. Bifidobacterium bifidum and the infant gut microbiota: an intriguing case of microbe-host co-evolution. Environ. Microbiol. 21, 3683–3695. https://doi.org/10.1111/1462-2920.14705 (2019).
CAS  Article  PubMed  Google Scholar 

71.
Fredheim, E. G. et al. Biofilm formation by Staphylococcus haemolyticus. J Clin Microbiol 47, 1172–1180. https://doi.org/10.1128/jcm.01891-08 (2009).
CAS  Article  PubMed  PubMed Central  Google Scholar  More

Forest sites and soil sampling
The sites used in this study are part of a permanent network of existing mature forest plots across Europe established in 2011–2012 (see Baeten et al.54 for detailed descriptions). We included four sites ranging over a large climatic gradient: North Karelia (Finland), Białowieża (Poland), Râşca (Romania), and Colline Metallifere (Italy), which correspond to typical boreal forests, hemiboreal mixed broadleaved-coniferous, montane mixed beech, and Mediterranean thermophilous, respectively (Supplementary Table 1, Supplementary Fig. 1). At each site, we selected 30 m × 30 m forest plots dominated by either one tree species (monospecific stands) or by three co-dominating tree species, hereafter referred to as mixed stands, resulting in a total of 34 species combinations (species were considered co-dominant if they composed >15% of the stand; see Supplementary Data file 1 for plot and tree species information). Each site differed in total species numbers, species identity, and species combinations (Supplementary Table 1). There were two replicates per tree species for the monospecific plots of each site, except for Picea abies and Quercus robur, which were only replicated once and Betula pendula which had no mono-specific plot in Białowieża. There was a minimum of three mixed species plot replicates per site that were composed of any of the target species present at the site (Supplementary Table 1), i.e., the replicate mixed plots at each site did not necessarily have the same tree species combinations. There were 64 plots in total. The sampling design with the total plot number, their distribution over four forest ecosystems, and including a wide range of tree species is well suited to address the generality of our hypothesis that microbial responses to DRW cycles are modified by tree species mixing but poorly suited to identify site-specific patterns with plot numbers too limiting within specific sites for robust testing.
Within each plot, we selected five tree triplets, a triplet being a triangle of three tree individuals within a maximum distance of 8 m from each other and no obstructing tree individuals within the triangle. Each triplet was composed of either the same species in the monospecific stands (monospecific triplet) or the three tree species present in the mixed stands (mixed triplet). At the estimated tree individual size weighted (based on individual diameter at breast height) center within the triangle, we collected five soil cores from the topsoil (10 cm deep, 5.3 cm diameter) after the litter layer had been removed. The five soil cores were spaced at roughly 35 cm from each other circling the center point (approximate sampled area 50 cm × 50 cm). A depth of 10 cm was selected because it is the standard topsoil sampling depth in soil ecology, has the highest soil microbial activity, and is under the most influence from the plant community19. All soil cores from each sampling location (i.e., tree triplet) within a plot were then sieved together through a 2 mm sieve and air-dried immediately after sampling for transportation and experiment preparation.
Experimental design
The soils collected from the 64 forest plots at the four sites were split into six replicate microcosms, yielding a total of 384 microcosms that were housed at the Montpellier European Ecotron CNRS in Montpellier, France. Each microcosm contained 95 g dry weight of soil in a glass vial (soil volume 51–72 ml; air volume 259–279 ml), initially incubated at 80% of water holding capacity (WHC) using deionized water, 25 °C, no light, and 40% relative air humidity (the vials were covered with Parafilm® to allow gas exchange but to prevent soil desiccation) for 3 weeks to reactivate the microbial community (Supplementary Fig. 3). After this acclimation period, half of the microcosms (192, i.e., n = 3 per plot) was assigned to a drying-rewetting (DRW) treatment and the other half (192, i.e., n = 3 per plot) to a control treatment. Maximum microbial mineralization activity appears to be reached between 60% and 80% WHC55. We chose 80% to ensure soils were entirely and homogeneously humid; very sandy soils with a low WHC, such as those from the Polish site, were not completely wetted at the typically chosen 60% WHC. Each treatment replicate was housed in a 2 m3 individual growth chamber (n = 6). Within each chamber, the microcosms were randomly distributed on a single shelf and re-randomized weekly. The DRW treatment was defined as two DRW cycles while the soils in the control treatment were maintained at 80% WHC throughout the experiment (Supplementary Fig. 3). Water content was adjusted gravimetrically 2–3 times a week.
Due to the large latitudinal distribution and varying soil and climate conditions of the sites (Supplementary Table 1), the soil microbial communities do not necessarily have the same degree of drought history and adaptation56. We therefore applied a site-specific drought treatment representative of each of the four study sites, i.e., site-specific drought intensity and duration. We used the permanent wilting point as a water stress threshold indicator since there is not a known microbial equivalent. The permanent wilting point was measured using a pressure plate extractor (1500F2, Soilmoisture Equipment Corp., Santa Barbara, USA) at pF 4.2 (15.5 bar) for the plots with the fastest and slowest drying soils of each site. The soil drying speed, i.e., the number of days it took for the soil to dry from 80% WHC down to constant weight, was measured gravimetrically for each plot using a subsample of soil that was subsequently excluded from the experiment. We then averaged the permanent wilting point values per site and designated this average as the drought intensity: Colline Metallifere 11% H2O g−1 dry soil, Râşca 30%, Białowieża 12%, and North Karelia 12%. The beginning of the drought was considered the moment the soil water content arrived at this threshold. The drought duration was calculated using the forest drought history data from Grossiord et al.56 as the average annual number of days the relative extractable water (REW) dropped below 0.2 (unitless) over the 1997–2010 period. REW is the ratio of available soil water to maximum extractable water (i.e., WHC), ranging between the field capacity (REW 1.0) and the permanent wilting point (REW 0.0)56. Plants are in non-water limited conditions when REW is >0.4 and water limited when REW is 2 mm diameter) and fine roots (≤2 mm in diameter). Fine roots were further separated into tree and understory roots. Tree fine roots were further divided into dead (which are hollow, brittle, and dark-colored) and live fine roots, which were then sorted by species (based on distinct color, architecture, morphology, and mycorrhizal types) and subsequently further divided based on their functions into absorptive and transport roots66. Ectomycorrhizal root tips were counted on absorptive tree roots using a binocular. All absorptive tree fine-root samples were scanned with a flat-bed scanner (resolution of 800 dpi) and scans were then analyzed using WinRhizo (Regent Instruments, Quebec, Canada, 2009) to quantify root length, surface area, volume, and diameter. Coarse root samples were also scanned to obtain coarse root volume, which was used together with the stone mass to calculate fine-earth volume (cm−3) of each soil sample. Root samples were dried (minimum 72 h, 40 °C) and weighed. Carbon and nitrogen concentrations of milled absorptive fine-root samples were measured for samples pooled at the plot level using dry combustion (Elementar Vario El Cube). Absorptive root analysis results are provided in Supplementary Table 2. We chose absorptive root traits rather than the commonly used leaf traits to characterize functional trait characteristics of tree communities, because the majority of soil microorganisms are intimately associated to the rhizosphere and thus root traits67. CWM is a measure of the relative species abundance weighted trait values. FDis is a measure of the abundance weighted mean distance between the “trait space’ of individual species. Both indices were calculated with the same standard root chemical and morphological traits (Supplementary Table 2) using the R function ‘dbFD’ in the FD package (version 1.0-1268). Due to difficulties in differentiating between Quercus species in root samples from some of the Italian plots, we were unable to determine mean absorptive root trait values at the plot level. We therefore used mean root trait values at the site level calculated from the mono-specific stands. Although the root trait values were not at the plot level, we were still able to determine the CWM and FDis indices at plot level since the root traits values were reported to tree species relative abundance in each plot. The relative abundance of each tree species was calculated using the basal areas of the tree individuals used in the five plot tree triplets (three tree individuals per tree triplet). Within each plot, the basal areas of a tree species (including five or fifteen tree individuals depending on whether the plot was mixed or mono-specific, respectively) were summed and then reported to the total basal area of the 15 tree individual, giving the relative basal area of each tree species within each plot. In order to synthesize this data, we incorporated them in a principal component analysis (PCA) and extracted the first axis scores (explaining 52.8% of the variance; Supplementary Fig. 2). Although the evidence supporting a universal root economics spectrum (RES) for woody species is inconsistent69,70,71, we consider our PCA1 axis as an acquisitive to conservative trait gradient with lower scores represented acquisitive root traits (high N content, specific root length, and ectomycorrhizal colonization intensity) and higher scores represented conservative traits (large diameter and high tissue density). The FDis was calculated following Laliberté & Legendre64 based on all traits at the plot level. The mono-specific stands had a FDis value of zero, which limits FDis variability for half of the plots. Accordingly, there was one single FDis value per plot that was used in our statistical analyses.
Since soil microbial resistance and recovery are tied to soil parameters and resource availability17,18, we also included major topsoil parameters (0–10 cm) known to affect microbial activity and/or community composition (Supplementary Table 2) measured previously during the FunDivEurope project54 at the plot level. Similar to the CWM absorptive root traits, we incorporated the topsoil variables into a PCA using the function ‘prcomp’ from the factoextra package (version 1.0.662) and extracted the first axis scores (explaining 52.5% of the variance; Supplementary Fig. 2) for a synthetic soil parameter measure for each individual plot. High PC1 scores are associated with higher pH, carbon content, and clay content and lower bulk density, the inverse is correlated with low PC1 scores.
We used generalized mixed-effects linear models (two-sided) using the lme4 package (version 1.1–2172) to assess the effects of the DRW treatment and the influence of the tree species number on microbial C and N-related parameters. The root FDis, root CWM PC1, and soil PC1 variables were included with the treatment × tree species interaction as explanatory variables. For the response variables (instantaneous CO2 and N2O fluxes measured five times over the experiment and cumulative fluxes, DOC, and TDN leaching, qCO2, and resistance and recovery indices), extreme values were removed (±3 times the IQR of all values for each variable). The soil collection site and plot as well as the growth chambers used for the incubation were included as random variables with plot nested within site. We did not include any climatic variables from the different sites, because they were highly correlated to site, which was already a random effect in the model. The model structure was as follows: response variable ~ Root FDis + Root CWM PC1 + Soil PCA axis + Treatment * Tree species number * Flux measurement time + (1|Chamber) + (1 | Site/Plot). The “Flux measurement time” variable, which identifies the times the five flux measurements were taken (i.e., beginning, drought 1, rewetting 1, drought 2, rewetting 2; Supplementary Fig. 3), was used only in the models that looked at the temporal dynamics of CO2 and N2O fluxes. For the analysis of resistance and recovery indices, we did not keep the “Treatment” variable in the model since these indices were calculated using both the DRW and control treatment results (see above). Additionally, for the resistance and recovery indices, instead of a “Flux measurement time” variable, a “Cycle” variable was included to distinguish the microbial activity resistance and recovery of the first and second cycles; the “Cycle” result indicates the change between the first and second cycle. Model residuals were plotted to test for normality, and data was transformed (log2 or BoxCox) when normality was not met. We also verified for data homogeneity and model probability (Q–Q plots). In order to identify the most parsimonious model we used the R software (version 3.5.3) and the “dredge” function in the MuMIn package (version 1.43.673) which uses the lowest Akaike information criteria (AIC) to rank all possible models with all possible combinations of the explanatory variables in the full model.
The data presented here is tied to specific spatial and temporal ecological conditions (e.g., forest drought history, tree species presence, microbial community composition, and soil property heterogeneity) which are susceptible to change. This makes exact study replication challenging and underlines the importance of including a wide range of conditions (e.g., multiple forest types, tree species, tree species combinations, climatic conditions, and soil types) as done here in order to explore general, potentially reproducible, trends oppose to site-specific trends.
Reporting summary
Further information on research design is available in the Nature Research Reporting Summary linked to this article. More

Selection criteria
Salinity is the measure of the concentration of dissolved (soluble) salts in water from all sources, and it can be measured by a range of parameters (including dissolved solids fractions, total dissolved solids, chloride, electrical conductivity, salinity) and units (including ppm, mg L−1, µS cm−1, dS m−1). A primary data collection focus here was given to EC measurements, since this is the most widely reported salinity parameter, and a main aim of this database is to provide comparable data across various scales. However, total dissolved solids (TDS) is also a common salinity parameter, particularly for groundwater quality measurements. The relationship of TDS and EC is correlated and can be determined using a conversion factor19. Regional conversion factors have been shown to produce better correlations than global factors, since the relationship between EC and TDS depends on a range of factors that may vary spatially, e.g. with climate, temperature, dissolved ion concentrations and ionic strength20. Thus, for optimizing data inclusion, a dataset containing TDS measurements was included, but only if a regional conversion factor could be found in the literature (see Methods and Technical Validation for further description on conversion and correlation analyses).
Multiple selection criteria were applied for each monitoring location and water type sampled. Surface waters were divided into the following categories: (i) river; and (ii) lake/reservoir. A sampling location was included if there were at least 30 measurements within the selected time period (1980–2019). For groundwater, we included all measurements at each location, if reported sampling depth information was available. The reason for this less stringent sampling frequency criterion for each groundwater location was due to the general limitation of high frequency groundwater monitoring compared to surface water monitoring21,22. Additionally, low temporal resolution groundwater data could provide valuable input for first order salinity assessments, model calibration and/or hypothesis testing23. An important variable for interpreting groundwater EC is however sample depth, since this has large implications on, for example, withdrawal depths for different sectoral water use, as well as for estimation of the freshwater/saltwater lens24. This thus motivates the depth availability criterion over sampling frequency for groundwaters. In addition to these criteria, all samples also had to have date and coordinate (latitude, longitude) information for qualifying inclusion in the database (see Fig. 2 for a schematic flowchart of the data selection and processing steps).
Fig. 2

Data selection and harmonisation flowchart. The figure illustrates the processing and harmonizing steps of each dataset (divided into surface and groundwater parts) after initial data collection.

Full size image

Data collection and sources
Data was collected from both surface water and groundwater monitoring locations using a combination of data sources, including: (i) global datasets, (ii) regional datasets, and (iii) individual river basins and groundwater aquifers datasets. The regional data includes datasets spanning multiple river basins and/or groundwater aquifers, both within the same region, but also cross-regionally. Most of these data are provided by governmental organizations or cross-regional data portal platforms under environmental protection agencies or National water quality monitoring programs. The local/individual basins datasets consist of monitoring data for individual basins and were usually found through governmental agencies, river basin management commissions, research organizations, as well as provided by individual researchers. Each data source is listed and described shortly below (the data source abbreviations were defined by us, for easy reference to the database terminology). A full list of the corresponding data (including their spatial and temporal resolution) for each of these sources (including their URL), divided by water type, is given in online-only Table 1.
For the here presented database, we focused on combining and harmonizing EC datasets from already available, open data sources. The reason for this is that EC is often included in broader environmental monitoring websites and/or water quality datasets, which are not identifiable as salinity datasets, but rather in general water quality terms. We thus wanted to extract the salinity data component, and facilitate the reuse of harmonized EC data for salinity-specific applications. Most of the dataset included in our database have original licenses that permit unrestricted reuse. Where this was not the case, or if information was lacking, we requested and were granted permission from the data owners to release the data under the CC-BY license.
Although we acknowledge the potential of valuable datasets in the scientific literature, this was not a data focus type, since this requires a different data search and extraction approach. We only incorporated pre-extracted datasets from literature reviews and synthesis when shared from individual researchers (reached through communication within our research community, e.g. during workshops and conferences and within own networks and communication channels). The following subsections provide an overview of the global, regional and local salinity datasets included in our developed database.
Global salinity dataset
The Global River Chemistry Dataset (GLORICH) includes multiple water quality parameters for river locations around the world, assembled by researchers from Hamburg University25,26. This data is publicly available and was downloaded as a zip file from PANGEA. The dataset includes 1.27 million samples of major compounds, nutrients, carbon species and physical properties. We extracted Specific Conductivity data (another terminology for EC) from the “hydrochemistry” csv file and paired it with station information (“Sampling_locations” file), for all stations that fulfilled our selection criteria.
Regional salinity datasets:
(1)
Data for Europe was collected from the European Environment Agency’s water quality database; Waterbase. Waterbase contains multiple water quality parameters for rivers, lakes and groundwater bodies throughout Europe. We extracted relevant EC and station information data using the raw disaggregated water quality data file: “Waterbase_v2018_1_T_WISE4_DisaggregatedData” and the parameter code for EC (“EEA_3142-01-6”, specified as Specific Conductance). The water types were identified and distinguished from the column parameterWaterBodyCategory, where “RW” is river, “LW” is lake and “GW” is groundwater location. Site information was extracted from the file: “Waterbase_v2018_1_WISE4_MonitoringSite_DerivedData”. The groundwater EC data was matched with depth information, using the parameterSampleDepth parameter.

(2)
The Water Quality Portal (WQP) for surface and groundwaters across the United States contains a range of water quality data for surface and groundwaters across the US. The data portal is established by the United States Geological Survey (USGS), the Environmental Protection Agency (EPA), and the National Water Quality Monitoring Council (NWQMC). The data originated from state, federal, tribal, and local agencies. Data was downloaded in bulk, for Specific conductance, for all available sites included under the search criteria (i) streams, (ii) lake, reservoir, impoundment and (iii) subsurface. Station information was additionally downloaded and paired with the salinity data.

(3)
Groundwater data for the US was also gathered from the Dissolved-Solids Dataset (Qi & Harris 2017)27, by downloading the “Dissolved solids” csv file and combining it with depth information from the “AquiferDepthSources” excel file. This data is published by the ScienceBase Catalog, provided by the USGS and contains EC (and other geochemical) data that was collected with the purpose of assessing brackish groundwaters across the United States. The original dataset contains a compilation of water-quality samples from 33 sources for almost 384,000 groundwater wells across the continental U.S., Alaska, Hawaii, Puerto Rico, the U.S. Virgin Islands, Guam, and American Samoa, dating back to the early 18th century.

(4)
Groundwater data from Colorado was collected from the Department of Agriculture and Agricultural Chemicals & Groundwater Protection section (Co Gov). Data was downloaded directly from the site using a search query of statewide inorganic quality monitoring data, and selecting the parameter Specific Conductance (Lab), for all available years. Site coordinate (latitude, longitude) information was not available online, but when requested via email, it was submitted to us, by their groundwater monitoring specialists (Karl Mauch, personal email communication). In addition, data on well sampling depth estimations were also provided via email, and the perforated interval measure (the interval between top and bottom of perforated section where the pump is installed) was recommended and used as depth information.

(5)
Groundwater data from California was downloaded from the GeoTracker Groundwater Ambient Monitoring and Assessment Program (GAMA), provided by the California state open data portal. The dataset includes multiple groundwater quality data from the GAMA Domestic Well (DW) and Priority Basin (PB) programs, covering locations throughout the state. The column “well_depth” was the only depth information available, and was included (and converted from feet to meters) as the Depth parameter.

(6)
Groundwater monitoring data from the Ohio Environmental Protection Agency (Ohio EPA) was downloaded from their ambient groundwater monitoring program. Monitoring of groundwater wells was established in the late 1960s and today covers more than 300 wells. Also here, the “well_depth” parameter was the only depth information available, and was included (and converted from feet to meters) as the Depth parameter.

(7)
The groundwater database from the Texas Water Development Board (TWDB) was also utilized to download water quality data. EC data was downloaded in bulk by groundwater aquifer (in total nine datasets). Well depths were converted from feet to meters and where multiple measurements for the same day and well was reported, daily averages were calculated. A total of 404 wells fulfilled the selection criteria and were included in the main groundwater database.

(8)
Data for South Africa was collected from the Department of Water and Sanitation (DWS), Republic of South Africa28. Both surface- and groundwaters are monitored, as a part of their National Chemical Monitoring Program. Monitoring stations and their data can be viewed and downloaded through the Water quality data exploration tool. However, due to the large amount of data for surface waters, we requested and recieved raw water quality data from the Resource Quality Information Services national monitoring programs for specific rivers and dams, through E-mail.

(9)
Surface water monitoring data for a large part of Australia is provided by the Australian Government, Bureau of Meteorology (AU Gov). Data can be queried at the Water Data Online portal, and search criteria can be specified. Conducted search criteria of all stations with EC data resulted in 1,333 stations. However, since data can only be downloaded as one by one station, we sent an email through the help desk system requesting a bulk download of all available data. The data was then provided as daily means recorded at midnight and as csv files (one file per station), with a metadata summary file included (with station information). From this, all files were combined and stations that fulfilled the selection criteria were then included in the main database. The separation between river and lake/reservoir locations were determined from the datafile “long_name” column, which always included the water type as well as the actual name of the monitoring location.

(10)
Surface water data for Australia was also synthesized from the Queensland Government Open Data Portal (QLD AU Gov). Data from QLD AU Gov was collected from the ambient estuary water quality monitoring program, which includes tidal rivers, streams and inshore waters of Central Queensland, monitored from 1993–2013. Data is available for 12 different drainage basins, reported as Specific Conductance at 25 °C. Data was downloaded as individual csv-files for each drainage basin (containing multiple sampling locations), and then combined and extracted according to the selection criteria.

(11)
Groundwater data for Australia was gathered from the Australian Government Bioregional Assessment Program (BAP). The data is provided through a collaboration between the Department of the Environment and Energy, the Bureau of Meteorology, CSIRO and Geoscience Australia. The dataset contains EC measurements of groundwater bores in the Namoi sub-region. The data is collected from groundwater bores that fell within the data management acquisition area as provided by the Bioregional Assessment to the Namoi NSW Office of Water. All data were downloaded in one csv-file.

(12)
Another groundwater dataset from Australia was collected, using the groundwater data portal from WaterConnect, which provides data from the Department for Environment and Water, for South Australia. Data was here queried by region, and then one file containing EC data for all sampled wells and one file containing site information were downloaded, for each region (in total 12 regions). The “Latest_Depth (m)” was used for depth information and all stations with both depth and EC measurements for a given data were included.

(13)
Additional groundwater data from Australia was downloaded using the Australian Groundwater Explorer tool (AU GwEX). Data was here search for by parameters Water level and Salinity and downloaded by region (in total 8 regions) and combined. Water levels and EC data was linked to the NGIS bore data to get the location and attributes of the measurement wells.

(14)
Data for New Zealand was gathered from New Zealand’s Hydro Web Portal for Hydrometric and Water Quality data (NIWA). This platform provides river water quality data under the National Institute of Water and Atmospheric Research. Data was queried by searching for all available data under the parameter conductivity and time-series, in their map interphase (resulting in 77 locations of timeseries data). Each dataset was then added for bulk export, using the export tab and a download link, via the map-interface platform.

(15)
Surface water quality data from the Government of Canada (Ca Gov) was downloaded from the National Long-term Water Quality Monitoring Data portal. The data include both rivers and lakes monitored for a set of physio-chemical variables, including specific conductance. Data was downloaded as csv-files.

(16)
River data was also synthesized from the Government of Ontario for multiple rivers, monitored between 2000–2016. The data is collected by the Provincial (Stream) Water Quality Monitoring Network (PWQMN), who measures water quality in rivers and streams across Ontario. Data was downloaded as individual excel files for each year, and then combined with site information.

(17)
Groundwater data from Argentina was downloaded from the repository of open public data of the Argentinian Republic (Dat.ar). The data is provided by the Federal Groundwater Information System SIFAS-SISAG and contains groundwater well measurements from April 2015. The data was downloaded as a main csv-file and translated from Spanish.

(18)
Groundwater data was also collected from Cambodia, using the online well database of Cambodia (WellMap). WellMap is an initiative of the Ministry of Rural Development of Cambodia, supported by the Water and Sanitation Program of the World Bank (WSP). The database is provided as a Microsoft Access Database and consists of water quality data collected from rural wells throughout the Country. Data was queried and extracted using the RODBC R package, that allows R interfacing to database systems. UTM coordinates were re-projected and converted to latitude and longitude, as decimal degrees, using the functions “proj4string” and “spTransform” in R.

(19)
Data from Mexico Government (MX Gov), was downloaded and translated (from Spanish) from one main csv-file, containing both water quality and site information data. The data included both surface water locations (original classification was rivers, streams, dams, which were reclassified to the here used terminology) and groundwater locations, monitored since 2012.

(20)
Groundwater data from Bangladesh was provided by M.M. Rahman (TH Cologne, University of Applied Sciences, Institute for Technology and Resources Management in the Tropics and Subtropics). The data was collected and shared by M.M. Rahman, and include electrical conductivity and depth data synthesized from both literature and governmental sources (see specifications and references in online-only Table 1).

(21)
Groundwater EC and level data from the Swedish geological Survey (SGU) was downloaded, on a county basis, for all 21 counties in Sweden, from environmental monitoring data. EC data was extracted from environmental monitoring files, with one file per county (queried using county specific codes and a URL link to each dataset) and combined with well water level data (downloaded in the same way as the salinity data) using matching coordinates. All stations with water level information were translated to English and were included in the main groundwater database.

Salinity datasets from individual river basins and groundwater aquifers:
(1)
Data for river locations within the Danoube river basin was collected from the Danube River Basin Water Quality Database. This database is provided by the International Commission for Protection of the Danube River (ICPDR) Information System Danubis (ICPDR). The database provides geochemical data for the major rivers in the Danube River Basin and waters are sampled at a minimum frequency of 12 times per year. The data was accessed through creating an account, and then performing a data search, for all available years and stations for the conductivity parameter, and exporting the resulting data as a csv file.

(2)
Data for the lower Murray Darling river basin was accessed through the Water Connect data portal (Waterconnect). All stations within the river basin that fulfilled the data selection criteria (six stations) were included and downloaded, one by one (using a combination of the historical EC daily readings and the Site summary files).

(3)
Groundwater TDS data for the Nile Delta aquifer (van Engelen et al.)29 was provided by Joeri van Engelen. These data include three datasets consisting of TDS measurements, synthesized from literature, collected with the selection criteria of including measurement data from less than 250 m depth. Two of these datasets had unspecific dates, and samples were thus assumed to be from the 1st of each reported month (see further specification of the data in van Engelen et al.29). The TDS data was then converted to EC, using a regional specific conversion factor, from literature sources (see section Conversions of TDS to EC for specifics on how this was done).

Data processing and harmonization
The overall objective with this database is to facilitate data reuse and research efforts within different fields of salinity research. For this purpose, the harmonization of data was a main part of the database construction. The flowchart (Fig. 2) illustrates the data selection criteria, data processing and harmonization of each sampling location and its associated dataset before it was added to the main database. All processing was done in R, version 3.6.0, using mainly the data.table and dplyr R packages. First, harmonization and fixing of data with regards to missing values and other uninterpretable field values and/or symbols preventing the appropriate reading of data files (i.e., special symbols like “***” or erroneous changes in field separators, e.g. from “,” to “;”) were done, e.g. by setting it to the standard missing data value (i.e., NA values) and by fixing or excluding rows which could not be read properly. Additionally, assumed erroneous data values for reported salinity values and depth (such as negative values, 999 and 9999, as well as depth values of zero) were removed.
Since information on sampling water type and parameter nomenclature and reported units differs between regions and organizations, we re-classified water types into the three mentioned categories (river, lake/reservoir, groundwater). Where needed, we also re-named and converted other parameters and their associated units, according to the database variables listed in Table 1.
Table 1 Variable names and descriptions, including reported units, of the salinity database.
Full size table

Different spatial and temporal conversions were also made (see Fig. 2). For instance, where multiple measurements per day were available, these were averaged into daily values, using the data.table package, and grouping by Station_ID and Date (see Table 1 for parameter definitions). Depth conversions were also common and included conversions from feet or centimeter to meters. Regarding spatial harmonization, each sample coordinates were converted to decimal degrees and re-projected to WGS 1984, if needed, using the “SpatialPoints”, “proj4string“ and the “spTransform” function of the rgdal R-package. If country information was missing, this was assigned from coordinates of each station using the package map.where, or extracted from country codes (if available) using the function “countrycode”. Continent information was then assigned from country names, also using the “countrycode” function, by matching country name with continent.
For assisting studies that might be interested specifically in coastal regions and applications, we also quantified if a sampling location was coastal or not. This analysis was done in ArcMap, using the “Near Table” analysis tool. The distance from all sampling locations to the coastline was computed, (using vector data from Natural Earth: https://www.naturalearthdata.com/downloads/10m-physical-vectors/). All locations within 10 km from the coastline were classified as being coastal. The identification of coastal stations was then included in each database summary file, under the column “Coastal_location” (see Table 1).
Conversions of TDS to EC
We considered the inclusion of additional groundwater data, where TDS measurements could be converted to EC. The relationship between EC and other measured salinity parameters (e.g. TDS) is depending on a range of conditions, such as temperature, climate and concentrations of ionic and undissociated species18. This relationship is commonly estimated according to Eq. (1).

$$EC=frac{TDS}{f}$$
(1)

where EC is in µS cm−1, TDS in mg L−1 and f is a conversion factor19,30. Commonly, predefined conversion factors without proper site-specific validation are used, but such estimation may be highly uncertain, due to the conditions mentioned above20. Instead, it has been shown that the use of region-specific conversion factors may be more representative, since these have been developed from measured relationships between EC and TDS under more local-reginal conditions19,20.
Due to reported improved predictability of EC-TDS relationships when using region-specific conversion factors (f), we included additional groundwater TDS measurements only for regions with available reported region-specific f values. This resulted in the inclusion of three additional groundwater datasets to the final database; one from Idaho31, one from California32 and one from Egypt29. Together these datasets added 3,477 sampling locations and a total of 9,654 measurements to the groundwater database. Both the original TDS data, as well as the converted EC values are included in the database.
For the two TDS groundwater datasets from the United States, TDS was converted to EC using the region-specific conversion factor f of 0.65. This conversion factor has been developed for the continental United States, by the US Geological Survey and is widely used cross-regionally within the US20,33. For the TDS groundwater data from Egypt (from the Nile delta)29, we converted TDS to EC using the region-specific conversion factor f of 0.64. This factor value has been derived from local measurement data in the Nile delta itself34.
For validation of our approach of predicting EC from TDS, we used regional-conversion factor f values on other groundwater datasets that had both TDS and EC measurements reported. These datasets, including data from both the US and from Australia, showed strong correlations between predicted and measured EC (Fig. 3; R2 of 0.91–0.99), supporting the approach of using TDS and region-specific conversion factors to estimate EC (see Technical validation section).
Fig. 3

Validation of converted TDS to EC for groundwaters. Time-series plot and scatter correlations of measured vs. predicted electrical conductivity (EC), using regional conversion factors. Panel (a) shows an example time-series from the groundwater station with the highest number of measurements (estimated from the “max” function in R) in Australia (data source: Water connect, n = 538) and panel (b) shows its corresponding scatter correlation (R2 = 0.99). Panel (c) shows the correlation between measured and converted EC for the full dataset of all groundwater stations from Water connect (n= 37,819, R2 = 0.98). Panel (d) and (e) shows correlations between measured and predicted EC data, for groundwaters in Texas (data source: TWDB, n = 59,985, R2 = 0.91) respectively California (data source: GAMA, n = 4,706, R2 = 0.98). All scatterplots were done in R, using the “ggscatter” function from the ggpubr package and estimating correlation coefficients using the “pearson” function.

Full size image More

1.
Stearns, S. C. The evolution of life histories. (Oxford, 1992).
2.
Ellison, P. T. Endocrinology, energetics, and human life history: a synthetic model. Horm. Behav. 91, 97–106 (2017).
PubMed  Google Scholar 

3.
Stoehr, A. M. & Kokko, H. Sexual dimorphism in immunocompetence: what does life-history theory predict?. Behav. Ecol. 17, 751–756 (2006).
Google Scholar 

4.
Subramanian, S. V., Özaltin, E. & Finlay, J. E. Height of nations: a sioeconomic analysis of cohort differences and patterns among women in 54 low- to middle-income countries. PLoS ONE 6, e18962 (2011).
ADS  CAS  PubMed  PubMed Central  Google Scholar 

5.
Stulp, G. & Barrett, L. Evolutionary perspectives on human height variation. Biol. Rev. 91, 206–234 (2016).
PubMed  Google Scholar 

6.
Georgiev, A. V., Kuzawa, C. W. & McDade, T. W. Early developmental exposures shape trade-offs between acquired and innate immunity in humans. Evol. Med. Public Health 2016, 256–269 (2016).
PubMed  PubMed Central  Google Scholar 

7.
Krams, I. et al. Reproduction compromises adaptive immunity in a cyprinid fish. Ecol. Res. 32, 559–566 (2017).
CAS  Google Scholar 

8.
Baumard, N. Psychological origins of the industrial revolution. Behav. Brain Sci. 1, 1–47 (2018).
Google Scholar 

9.
Krams, I. et al. A head start for life history development? Family income mediates associations between height and immune response in men. Am. J. Phys. Anthropol. 168, 421–427 (2019).
PubMed  Google Scholar 

10.
Luoto, S. An updated theoretical framework for human sexual selection: from ecology, genetics, and life history to extended phenotypes. Adapt. Hum. Behav. Physiol. 5, 48–102 (2019).
Google Scholar 

11.
Noble, K. G. et al. Family income, parental education and brain structure in children and adolescents. Nature Neurosci. 18, 773–778 (2015).
CAS  PubMed  Google Scholar 

12.
Zeki, A. et al. Sustained economic hardship and cognitive function: The coronary artery risk development in young adults study. Am. J. Prev. Med. 52, 1–9 (2017).
Google Scholar 

13.
Stotz, K. Why developmental niche construction is not selective niche construction: and why it matters. Interface Focus 7, 20160157 (2017).
PubMed  PubMed Central  Google Scholar 

14.
Said-Mohamed, R., Pettifor, J. M. & Norris, S. A. Life history theory hypotheses on child growth: potential implications for short and long-term child growth, development and health. Am. J. Phys. Anthropol. 165, 4–19 (2018).
PubMed  Google Scholar 

15.
Worthman, C. M. & Trang, K. Dynamics of body time, social time and life history at adolescence. Nature 554(7693), 451–457 (2018).
ADS  CAS  PubMed  Google Scholar 

16.
Cabeza de Baca, T., Wahl, R. A., Barnett, M. A., Figueredo, A. J. & Ellis, B. J. Adversity, adaptive calibration, and health: the case of disadvantaged families. Adapt. Hum. Behav. Physiol. 2(2), 93–115 (2016).
Google Scholar 

17.
Bateson, P. & Gluckman, P. Plasticity, robustness, development and evolution (Cambridge University Press, Cambridge, 2011).
Google Scholar 

18.
Morisaki, N. et al. Ecological analysis of secular trends in low birth weight births and adult height in Japan. J. Epidemiol. Comm. Health 71, 1014–1018 (2017).
Google Scholar 

19.
Puts, D. A. Beauty and the beast: mechanisms of sexual selection in humans. Evol. Hum. Behav. 31, 157–175 (2010).
Google Scholar 

20.
Tarka, M., Guenther, A., Niemelä, P. T., Nakagawa, S. & Noble, D. W. Sex differences in life history, behavior, and physiology along a slow-fast continuum: a meta-analysis. Behav. Ecol. Sociobiol. 72(8), 132 (2018).
PubMed  PubMed Central  Google Scholar 

21.
Khramtsova, E. A., Davis, L. K. & Stranger, B. E. The role of sex in the genomics of human complex traits. Nature Rev. Genet. 20, 173–190 (2019).
CAS  PubMed  Google Scholar 

22.
Luoto, S., Krams, I. & Rantala, M. J. A life history approach to the female sexual orientation spectrum: evolution, development, causal mechanisms, and health. Arch. Sex. Behav. 48(5), 1273–1308. https://doi.org/10.1007/s10508-018-1261-0 (2019).
Article  PubMed  Google Scholar 

23.
Wells, J. C. K. Sexual dimorphism in body composition across human populations: associations with climate and proxies for short- and long-term energy supply. Am. J. Hum. Biol. 24, 411–419 (2012).
PubMed  Google Scholar 

24.
García-Martínez, D., Torres-Tamayo, N., Torres-Sanchez, I., García-Río, F. & Bastir, M. Morphological and functional implications of sexual dimorphism in the human skeletal thorax. Am. J. Phys. Anthropol. 161(3), 467–477. https://doi.org/10.1002/ajpa.23051 (2016).
Article  PubMed  Google Scholar 

25.
Fischer, B. & Mitteroecker, P. Allometry and sexual dimorphism in the human pelvis. Anatomic. Record 300(4), 698–705 (2017).
Google Scholar 

26.
Lassek, W. D. & Gaulin, S. J. Costs and benefits of fat-free muscle mass in men: relationship to mating success, dietary requirements, and native immunity. Evol. Hum. Behav. 30(5), 322–328 (2009).
Google Scholar 

27.
Massy-Westropp, N. M., Gill, T. K., Taylor, A. W., Bohannon, R. W. & Hill, C. L. Hand grip strength: age and gender stratified normative data in a population-based study. BMC Res. Notes 4(1), 127 (2011).
PubMed  PubMed Central  Google Scholar 

28.
Samal, A., Subramani, V. & Marx, D. B. An analysis of sexual dimorphism in the human face. J Vis. Comm Image Represent 18, 453–463 (2007).
Google Scholar 

29.
McDade, T. W. Life history theory and the immune system: Steps toward a human ecological immunology. Am. J. Phys. Anthropol. 122, 100–125 (2003).
Google Scholar 

30.
Rigby, N. & Kulathinal, R. J. Genetic architecture of sexual dimorphism in humans. J. Cell. Physiol. 230(10), 2304–2310 (2015).
CAS  PubMed  Google Scholar 

31.
Stringer, S., Polderman, T. & Posthuma, D. Majority of human traits do not show evidence for sex-specific genetic and environmental effects. Sci. Rep. 7(1), 8688 (2017).
ADS  PubMed  PubMed Central  Google Scholar 

32.
Grasgruber, P., Sebera, M., Hrazdíra, E., Cacek, J. & Kalina, T. Major correlates of male height: a study of 105 countries. Econom. Hum. Biol. 21, 172–195 (2016).
CAS  Google Scholar 

33.
Perkins, J. M., Subramanian, S. V., Davey Smith, G. & Özaltin, E. Adult height, nutrition, and population health. Nutr. Rev. 74, 149–165 (2016).
PubMed  PubMed Central  Google Scholar 

34.
Hämäläinen, A., Immonen, E., Tarka, M. & Schuett, W. Evolution of sex-specific pace-of-life syndromes: causes and consequences. Behav. Ecol. Sociobiol. 72(3), 50 (2018).
Google Scholar 

35.
Immonen, E., Hämäläinen, A., Schuett, W. & Tarka, M. Evolution of sex-specific pace-of-life syndromes: genetic architecture and physiological mechanisms. Behav. Ecol. Sociobiol. 72(3), 60. https://doi.org/10.1007/s00265-018-2462-1 (2018).
Article  PubMed  PubMed Central  Google Scholar 

36.
Phalane, K. G., Tribe, C., Steel, H. C., Cholo, M. C. & Coetzee, V. Facial appearance reveals immunity in African men. Sci. Rep. 7(1), 7443 (2017).
ADS  PubMed  PubMed Central  Google Scholar 

37.
Luoto, S., Rantala, M. J. & Krams, I. England first, America second: the ecological predictors of life history and innovation [Commentary]. Behav. Brain Sci. 42, 1. https://doi.org/10.1017/S0140525X19000165 (2019).
Article  Google Scholar 

38.
Lourenço, A. M., Levy, A. M., Caetano, L. C., Carraro Abrahão, A. A. & Prado, J. C. Influence sexual dimorphism on the persistence of blood parasites in infected Calomys callosus. Res. Vet. Sci. 85, 515–521. https://doi.org/10.1016/j.rvsc.2008.01.008 (2008).
Article  PubMed  Google Scholar 

39.
Klein, S. L. & Roberts, C. W. (Eds.) Sex Hormones and Immunity to Infection. (Springer Verlag, 2010).

40.
Klein, S. L. Sex influences immune responses to viruses, and efficacy of prophylaxis and treatments for viral diseases. BioEssays 34, 1050–1059 (2012).
CAS  PubMed  PubMed Central  Google Scholar 

41.
Giefing-Kröll, C., Berger, P., Lepperdinger, G. & Grubeck-Loebenstein, B. How sex and age affect immune responses, susceptibility to infections, and response to vaccination. Aging Cell 14(3), 309–321 (2015).
PubMed  PubMed Central  Google Scholar 

42.
Xirocostas, Z. A., Everingham, S. E. & Moles, A. T. The sex with the reduced sex chromosome dies earlier: a comparison across the tree of life. Bio. Lett. 16, 20190867. https://doi.org/10.1098/rsbl.2019.0867 (2020).
Article  Google Scholar 

43.
Krams, I. A. et al. Body height affects the strength of immune response in young men, but not young women. Sci. Rep. 4, 6223 (2014).
ADS  CAS  PubMed  PubMed Central  Google Scholar 

44.
Skrinda, I. et al. Body height, immunity, facial and vocal attractiveness in young men. Naturwissenschaften 101, 1017 (2014).
ADS  CAS  PubMed  Google Scholar 

45.
Rantala, M. J. et al. Adiposity, compared with masculinity, serves as a more valid cue to immunocompetence in human mate choice. Proc. R. Soc. B. 280, 20122495 (2013).
PubMed  Google Scholar 

46.
Pļaviņa, L. & Kārkliņa, H. Sieviešu galveno antropometrisko parametru izvērtējums dažādos postnatālās ontoģenēzes periodos. Rīgas Stradiņa universitāte 2014. gada zinātniskā konference: Tēzes, Rīgā, 2014. gada 10. un 11. aprīlī. Rīga: RSU. 31. lpp. (2014).

47.
Rantala, M., J., Coetzee, V., Moore, F. R., Skrinda, I., Kecko, S., Krama, T., Kivleniece, I. & Krams, I. Facial attractiveness is related to women’s cortisol and body fat, but not with immune responsiveness. Biol. Lett. 9, 20130255 (2013).

48.
Pawłowski, B., Nowak, J., Borkowska, B., Augustyniak, D. & Drulis-Kawa, Z. Body height and immune efficacy: testing body stature as a signal of biological quality. Proc. R. Soc. B. 284, 20171372 (2017).
PubMed  Google Scholar 

49.
Petry, L. J., Weems, L. B. & Livingstone, J. N. Relationship of stress, distress, and the immunological response to a recombinant hepatitis-B vaccine. J Family Pract. 32, 481–486 (1991).
CAS  Google Scholar 

50.
Jabaaij, L. et al. Influence of perceived psychological stress and distress on antibody response to low dose rDNA hepatitis B vaccine. J. Psychosomat. Res. 37(4), 361–369 (1993).
CAS  Google Scholar 

51.
Jabaaij, L. et al. Modulation of immune response to rDNA hepatitis B vaccination by psychological stress. J. Psychosomat. Res. 41, 129–137 (1996).
CAS  Google Scholar 

52.
Ellis, B. J. & Del Giudice, M. Developmental adaptation to stress: an evolutionary perspective. Ann. Rev. Psychol. 70(1), 111–139 (2019).
Google Scholar 

53.
LaBeaud, A. D., Malhotra, I., King, M. J., King, C. L. & King, C. H. Do antenatal parasite infections devalue childhood vaccination?. PLoS Negl. Trop. Diseases 3(5), e442 (2009).
Google Scholar 

54.
Cooper, P. J. et al. Human infection with Ascaris lumbricoides is associated with suppression of the interleukin-2 response to recombinant cholera toxin B subunit following vaccination with the live oral cholera vaccine CVD 103-HgR. Infect. Immun. 69, 1574–1580 (2001).
CAS  PubMed  PubMed Central  Google Scholar 

55.
Elias, D., Britton, S., Aseffa, A., Engers, H. & Akuffo, H. Poor immunogenicity of BCG in helminth infected population is associated with increased in vitro TGF-beta production. Vaccine 26, 3897–3902 (2008).
CAS  PubMed  Google Scholar 

56.
Djuardi, Y., Wammes, L. J., Supali, T., Sartono, E. & Yazdanbakhsh, M. Immunological footprint: the development of a child’s immune system in environments rich in microorganisms and parasites. Parasitology 138(12), 1508–1518 (2011).
PubMed  Google Scholar 

57.
Blackwell, A. D., Snodgrass, J. J., Madimenos, F. C. & Sugiyama, L. S. Life history, immune function, and intestinal helminths: trade-offs among immunoglobulin E, C-reactive protein, and growth in an Amazonian population. Am. J. Hum. Biol. 22(6), 836–848 (2010).
PubMed  PubMed Central  Google Scholar 

58.
Cao, J. et al. Early-life exposure to widespread environmental toxicants and health risk: a focus on the immune and respiratory systems. Ann. Glob. Health 82(1), 119–131 (2016).
PubMed  Google Scholar 

59.
Lander, R. L. et al. Factors influencing growth and intestinal parasitic infections in preschoolers attending philanthropic daycare centers in Salvador, Northeast Region of Brazil. Cadernos Saúde Pública, Rio de Janeiro 28(11), 2177–2188 (2012).
Google Scholar 

60.
Anuar, T. S., Salleh, F. M. & Moktar, N. Soil-transmitted helminth infections and associated risk factors in three Orang Asli tribes in Peninsular Malaysia. Sci. Rep. 4, 4101 (2014).
ADS  CAS  PubMed  PubMed Central  Google Scholar 

61.
Hotez, P. J. & Gurwith, M. Europe’s neglected infections of poverty. Int. J. Infect. Diseas. 15, e611–e619 (2011).
Google Scholar 

62.
McDade, T. W. et al. Genome-wide analysis of DNA methylation in relation to socioeconomic status during development and early adulthood. Am. J. Phys. Anthropol. 1, 1–9 (2019).
Google Scholar 

63.
Needham, B. L. et al. Life course socioeconomic status and DNA methylation in genes related to stress reactivity and inflammation: the multi-ethnic study of atherosclerosis. Epigenetics 10(10), 958–969 (2015).
PubMed  PubMed Central  Google Scholar 

64.
Kubzansky, L., Seeman, T. E. & Glymour, M. M. Biological pathways linking social conditions and health: plausible mechanisms and emerging puzzles. In Social epidemiology (eds Berkman, L. F. et al.) 512–561 (Oxford University Press, Oxford, 2014).
Google Scholar 

65.
Gaulin, S. J. & Boster, J. S. Human marriage systems and sexual dimorphism in stature. Am. J. Phys. Anthropol. 89(4), 467–475 (1992).
CAS  PubMed  Google Scholar 

66.
Polo, P., Fernandez, A., Muñoz-Reyes, J. A., Dufey, M. & Buunk, A. P. Intrasexual competition and height in adolescents and adults. Evol. Psychol. 16(1), 1474704917749172 (2018).
CAS  PubMed  Google Scholar 

67.
Cornwallis, C. K. & Uller, T. Towards an evolutionary ecology of sexual traits. Trends Ecol. Evol. 25(3), 145–152 (2010).
PubMed  Google Scholar 

68.
Jewell, S. L., Luecken, L. J., Gress-Smith, J., Crnic, K. A. & Gonzales, N. A. Economic stress and cortisol among postpartum low-income Mexican American women: buffering influence of family support. Behav. Med. 41(3), 138–144 (2015).
PubMed  PubMed Central  Google Scholar 

69.
Serwinski, B., Salavecz, G., Kirschbaum, C. & Steptoe, A. Associations between hair cortisol concentration, income, income dynamics and status incongruity in healthy middle-aged women. Psychoneuroendocrinol. 67, 182–188 (2016).
CAS  Google Scholar 

70.
Ursache, A., Merz, E. C., Melvin, S., Meyer, J. & Noble, K. G. Socioeconomic status, hair cortisol and internalizing symptoms in parents and children. Psychoneuroendocrinol. 78, 142–150 (2017).
CAS  Google Scholar 

71.
Pepper, G. V. & Nettle, D. The behavioural constellation of deprivation: causes and consequences. Behav. Brain Sci. 40, e314 (2017).
PubMed  Google Scholar 

72.
Burns, V. E., Carroll, D., Ring, C., Harrison, L. K. & Drayson, M. Stress, coping, and hepatitis B antibody status. Psychosom. Med. 64(2), 287–293 (2002).
PubMed  Google Scholar 

73.
Glaser, R. & Kiecolt-Glaser, J. K. Stress-induced immune dysfunction: Implications for health. Nat. Rev. Immunol. 5(3), 243–251 (2005).
CAS  PubMed  Google Scholar 

74.
O’Connor, T. G. et al. Prenatal maternal anxiety predicts reduced adaptive immunity in infants. Brain Behav. Immun. 32, 21–28 (2013).
CAS  PubMed  PubMed Central  Google Scholar 

75.
Hayward, S. E. et al. A systematic review of the impact of psychosocial factors on immunity: implications for enhancing BCG response against tuberculosis. Soc. Sci. Med. Popul. Health. 10, 100522 (2020).
Google Scholar 

76.
Cohen, B. E., Edmondson, D. & Kronish, I. M. State of the art review: depression, stress, anxiety, and cardiovascular disease. Am. J. Hypertens. 28(11), 1295–1302 (2015).
CAS  PubMed  PubMed Central  Google Scholar 

77.
Golbidi, S., Frisbee, J. C. & Laher, I. Chronic stress impacts the cardiovascular system: animal models and clinical outcomes. Am. J. Physiol. Heart Circ. Physiol. 308(12), 1476–1498 (2015).
Google Scholar 

78.
Cozma, S. et al. Salivary cortisol and α-amylase: subclinical indicators of stress as cardiometabolic risk. Braz. J. Med. Biol. Res. 50(2), e5577 (2017).
CAS  PubMed  PubMed Central  Google Scholar 

79.
Steptoe, A. & Kiwimäki, M. Stress and cardiovascular disease. Nature Rev. Cardiol. 9, 360–370 (2012).
CAS  Google Scholar 

80.
Kivimäki, M. & Kawachi, I. Work stress as a risk factor for cardiovascular disease. Curr. Cardiol. Rep. 17(9), 630 (2015).
PubMed  Google Scholar 

81.
Sephton, S. E. et al. Diurnal cortisol rhythm as a predictor of lung cancer survival. Brain Behav. Immun. 30(Suppl), S163-170 (2012).
PubMed  Google Scholar 

82.
Spiegel, D. Minding the body: psychotherapy and cancer survival. Br. J. Health Psychol. 19(3), 465–485 (2014).
PubMed  Google Scholar 

83.
Garland, E. L., Beck, A. C., Lipschitz, D. L. & Nakamura, Y. Dispositional mindfulness predicts attenuated waking salivary cortisol levels in cancer survivors: a latent growth curve analysis. J. Cancer Survivorship 9, 215 (2015).
Google Scholar 

84.
McEwen, B. S. & Gianaros, P. J. Central role of the brain in stress and adaptation: links to socioeconomic status, health, and disease. Ann. N. Y. Acad. Sci. 1186, 190–222 (2010).
ADS  PubMed  PubMed Central  Google Scholar 

85.
Del Giudice, M., Ellis, B. J. & Shirtcliff, E. A. The adaptive calibration model of stress responsivity. Neurosci. Biobehav. Rev. 35(7), 1562–1592 (2011).
PubMed  Google Scholar 

86.
Rantala, M. J., Luoto, S., Krams, I. & Karlsson, H. Depression subtyping based on evolutionary psychiatry: proximate mechanisms and ultimate functions. Brain Behav. Immun. 69, 603–617 (2018).
PubMed  Google Scholar 

87.
Rantala, M. J., Luoto, S., Krama, T. & Krams, I. Eating disorders: an evolutionary psychoneuroimmunological approach. Front. Psychol. 10, 2200. https://doi.org/10.3389/fpsyg.2019.02200 (2019).
Article  PubMed  PubMed Central  Google Scholar 

88.
Greff, M. J. et al. Hair cortisol analysis: an update on methodological considerations and clinical applications. Clin. Biochem. 63, 1–9 (2018).
PubMed  Google Scholar 

89.
Khoury, J. E., Enlow, M. B., Plamondon, A. & Lyons-Ruth, K. The Association between adversity and hair cortisol levels in humans: a meta-analysis. Psychoneuroendocrinol. 103, 104–117 (2019).
CAS  Google Scholar 

90.
Doyle, H. H. & Murphy, A. Z. Sex differences in innate immunity and its impact on opioid pharmacology. J. Neurosci. Res. 95(1–2), 487–489 (2017).
CAS  PubMed  PubMed Central  Google Scholar 

91.
Fish, E. N. The X-files in immunity: sex-based differences predispose immune responses. Nature Rev. Immunol. 8, 737–744 (2008).
CAS  Google Scholar 

92.
Hao, S. et al. Modulation of 17b-estradiol on the number and cytotoxicity of NK cells in vivo related to MCM and activating receptors. Int. Immunopharmacol. 7, 1765–1775 (2007).
CAS  PubMed  Google Scholar 

93.
Ashcroft, G. S., Greenwell-Wild, T., Horan, M. A., Wahl, S. M. & Ferguson, M. W. Topical estrogen accelerates cutaneous wound healing in aged humans associated with an altered inflammatory response. Am. J. Pathol. 155, 1137–1146 (1999).
CAS  PubMed  PubMed Central  Google Scholar 

94.
Krams, I. et al. Heterophil/lymphocyte ratios predict the magnitude of humoral immune response to a novel antigen in great tits (Parus major). Comp. Biochem. Physiol. A: Mol. Integrat. Physiol. 161, 422–428 (2012).
CAS  Google Scholar 

95.
Stoll, M. L. Interactions of the innate and adaptive arms of the immune system in the pathogenesis of spondyloarthritis. Clin. Exp. Rheumatol. 29, 322–330 (2011).
CAS  PubMed  PubMed Central  Google Scholar 

96.
Klasing, K. C. & Leshchinsky, T. V. Functions, costs, and bene- fits of the immune system during development and growth. Ostrich 69, 2817–2832 (1999).
Google Scholar 

97.
McDade, T. W., Georgiev, A. V. & Kuzawa, C. V. Trade-offs between acquired and innate immune defenses in humans. Evol. Med. Publ. Health 2016(1), 1–16 (2016).
Google Scholar 

98.
Elia, M. Organ and tissue contribution to metabolic rate. in Energy Metabolism: Tissue Determinants and Cellular Corollaries (eds. McKinney, J. M. & Tucker, H. N.) 61–80. (Raven, 1992).

99.
Muehlenbein, M. P., Hirschtick, J. L. & Bonner, J. Z. Toward quantifying the usage costs of human immunity: altered metabolic rates and hormone levels during acute immune activation in men. Am. J. Human Biol. 22, 546–556 (2010).
Google Scholar 

100.
Taylor, S. E., Lehman, B. J., Kiefe, C. I. & Seeman, T. E. Relationship of early life stress and psychological functioning to adult C-reactive protein in the coronary artery risk development in young adults study. Biol. Psych. 60, 819–824 (2006).
CAS  Google Scholar 

101.
Danese, A. et al. Adverse childhood experiences and adult risk factors for age-related disease: depression, inflammation, and clustering of metabolic risk markers. Arch. Pediatr. Adolesc. Med. 163, 1135–1143 (2009).
PubMed  PubMed Central  Google Scholar 

102.
Miller, G. E. et al. Low early-life social class leaves a biological residue manifested by decreased glucocorticoid and increased proinflammatory signaling. Proc. Natl. Acad. Sci. USA 106, 14716–14721 (2009).
ADS  CAS  PubMed  Google Scholar 

103.
Miller, G. E. & Chen, E. Harsh family climate in early life presages the emergence of a proinflammatory phenotype in adolescence. Psychol. Sci. 21, 848–856 (2010).
PubMed  PubMed Central  Google Scholar 

104.
Archer, J. The reality and evolutionary significance of human psychological sex differences. Biol. Rev. 94(4), 1381–1415 (2019).
PubMed  Google Scholar 

105.
Hartling, C. et al. Interaction of HPA axis genetics and early life stress shapes emotion recognition in healthy adults. Psychoneuroendocrinol. 99, 28–37 (2019).
Google Scholar 

106.
Frankenhuis, W. E., Nettle, D. & Dall, S. R. A case for environmental statistics of early-life effects. Phil. Trans. R. Soc. B 374(1770), 20180110 (2019).
PubMed  Google Scholar 

107.
Foo, Y. Z. et al. Immune function during early adolescence positively predicts adult facial sexual dimorphism in both men and women. Evol. Hum. Behav. 1, 1. https://doi.org/10.1016/j.evolhumbehav.2020.02.002 (2020).
Article  Google Scholar 

108.
Cohen, S., Miller, G. E. & Rabin, B. S. Psychological stress and antibody response to immunization: a critical review of the human literature. Psychosom. Med. 63, 7–18 (2001).
CAS  PubMed  Google Scholar 

109.
Rantala, M. J. et al. Evidence for the stress-linked immunocompetence handicap hypothesis in humans. Nature Comm. 3, 694 (2012).
ADS  Google Scholar 

110.
Tyrrell, J. et al. Height, body mass index, and socioeconomic status: mendelian randomisation study in UK Biobank. Br. Med. J. 352, 582 (2016).
Google Scholar 

111.
Lavrinoviča, I., Lavriņenko, O., & Teivāns-Treinovskis, J. Population income differentiation and its influence on the crime. in Proceedings of the XIII International Scientific Conference Sustainable Business under Changing Economic Conditions (Dotkus, W., Holger, B., Žilys, J., Rozīte, M., Rumpīte, D., & Vīksne, I., Eds.) Rīga, Latvia: School of Business Administration Turība, pp. 242–251. Retrieved from: https://aurora.turiba.lv/bti/Editor/Manuscript/Proceeding/ (2012).

112.
R Core Team. R. A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna. https://www.R-project.org/ (2018).

113.
Wood, S. N. Fast stable restricted maximum likelihood and marginal likelihood estimation of semiparametric generalized linear models. J. R. Stat. Soc. B 73, 3–36 (2011).
MathSciNet  MATH  Google Scholar  More

1.
Bevins, S. N., Pedersen, K., Lutman, M. W., Gidlewiski, T. & Deliberto, T. J. Consequences associated with the recent range expansion of nonnative feral swine. Bioscience 64, 291–299 (2014).
Google Scholar 
2.
Corn, J. L. & Jordan, T. R. Development of the National feral swine map, 1982–2016. Wildl. Soc. B. 41, 758–763 (2017).
Google Scholar 

3.
McClure, M. L. et al. Modeling and mapping the probability of occurrence of invasive wild pigs across the Contiguous United States. PLoS ONE 10, e0133771 (2015).
PubMed  PubMed Central  Google Scholar 

4.
Snow, N. P., Jarzyna, M. A. & VerCauteren, K. C. Interpreting and predicting the spread of invasive wild pigs. J. Appl. Ecol. 54, 2022–2032 (2017).
Google Scholar 

5.
Anderson, A., Slootmaker, C., Harper, E., Holderieath, J. & Shwiff, S. A. Economic estimates of feral swine damage and control in 11 US states. Crop Prot. 89, 89–94 (2016).
Google Scholar 

6.
Holderieath, J. J. et al. Valuing the absence of feral swine in the United States: a partial equilibrium approach. Crop Prot. 112, 63–66 (2018).
Google Scholar 

7.
Pimentel, D., Zuniga, R. & Morrison, D. Update on the environmental and economic costs associated with alien-invasive species in the United States. Ecol. Econ. 52, 273–288 (2005).
Google Scholar 

8.
Barrios-Garcia, N. M. & Balairi, S. A. Impact of wild boar (Sus scrofa) in its introduced and native range: a review. Biol. Invasions 14, 2283–2300 (2012).
Google Scholar 

9.
Seward, N. K., VerCauteren, K. C., Witmer, G. W. & Engeman, R. M. Feral swine impacts on agriculture and the environment. Sheep Goat Res. J. 19, 34–40 (2004).
Google Scholar 

10.
Centner, T. J. & Shuman, R. M. Governmental provisions to manage and eradicate feral swine in areas of the United States. Ambio 44, 121–130 (2015).
PubMed  Google Scholar 

11.
United States Department of Agriculture. Final Environmental Impact Statement Feral swine damage management: a national approach. https://www.aphis.usda.gov/aphis/resources/pests-diseases/feral-swine/feral-swine-eis (2015).

12.
Engeman, R. M. et al. Locating and eliminating feral swine from a large area of fragmented mixed forest and agriculture habitats in north-central USA. Environ. Sci. Pollut. R. 26, 1654–1660 (2019).
Google Scholar 

13.
Graves, H. B. Behavior and ecology and wild and feral swine (Sus scrofa). J. Anim. Sci. 58, 482–492 (1984).
Google Scholar 

14.
Comer, C. E. & Mayer, J. J. Wild pig reproductive biology. In: Mayer, J.J & Brisbin, I.L. (eds) Wild pigs: biology, damage, control techniques and management. Savannah River National Laboratory SRNL-RP-2009-00869, Aiken, South Carolina, USA, 51–75 (2009).

15.
West, B. C., Cooper, A. L. & Armstrong, J. B. Managing wild pigs: a technical guide. Human-Wildl. Integr. Monogr. 1, 1–55 (2009).
Google Scholar 

16.
Kay, S. L. et al. Quantifying drivers of wild pig movement across multiple spatial and temporal scales. Mov. Ecol. 5, 14 (2017).
PubMed  PubMed Central  Google Scholar 

17.
Geisser, H. & Reyer, H. Efficacy of hunting, feeding, and fencing to reduce crop damage by wild boars. J. Wildl. Manag. 68, 939–946 (2004).
Google Scholar 

18.
Wang, S. W., Curtis, P. D. & Lassoie, J. P. Farmer perceptions of crop damage by wildlife in Jigme Singye Wangchuck National Park, Bhutan. Wildlife Soc. B. 34, 389–395 (2006).
CAS  Google Scholar 

19.
Morelle, K. & Lejeune, P. Seasonal variations of wild boar Sus scrofa distribution in agricultural landscapes: a species distribution modelling approach. Eur. J. Wildlife Res. 61, 45–56 (2015).
Google Scholar 

20.
Virgos, E. Factors affecting wild boar (Sus scrofa) occurrence in highly fragmented Mediterranean landscapes. Can. J. Zool. 80, 430–435 (2002).
Google Scholar 

21.
Michel, N. L., LaForge, M. P., Van Beest, F. M. & Brook, R. K. Spatiotemporal trends in Canadian domestic wild boar production and habitat predict wild pig distribution. Landscape Urban Plan. 165, 30–38 (2017).
Google Scholar 

22.
Keuling, O., Stier, N. & Roth, M. How does hunting influence activity and spatial usage in wild boar Sus scrofa L.?. Eur. J. Wildl. Res. 54, 729–737 (2008).
Google Scholar 

23.
Campbell, T. A. & Long, D. B. Activity patterns of wild boars (Sus scrofa) in southern Texas. Southwest. Nat. 55, 564–600 (2010).
Google Scholar 

24.
Fischer, J. W. et al. Effects of simulated removal activities on movements and space use of feral swine. Eur. J. Wildl. Res. 62, 285–292 (2016).
Google Scholar 

25.
Hernandez, F. A. et al. Invasive ecology of wild pigs (Sus scrofa) in Florida, USA: the role of humans in the expansion and colonization of an invasive wild ungulate. Biol. Invasions 20, 1865–1880 (2018).
Google Scholar 

26.
McCann, B. E. et al. Molecular population structure for feral swine in the United States. J. Wildl. Manag. 82, 821–832 (2018).
Google Scholar 

27.
Pepin, K. M., Davis, A. J., Cunningham, F. L., VerCauteren, K. C. & Ekery, D. C. Potential effects of incorporating fertility control into typical culling regimes in wild pig populations. PLoS ONE 12, e0183441 (2017).
PubMed  PubMed Central  Google Scholar 

28.
Wilcox, J. T., Aschehoug, E. T., Scott, C. A. & Van Vuren, D. H. A test of the judas technique as a method for eradicating feral pigs. Trans. West. Sect. Wildl. Soc. 40, 120–126 (2004).
Google Scholar 

29.
McCann, B. E. & Garcelon, D. K. Eradication of feral pigs from Pinnacles National Monument. J. Wildl. Manag. 72, 1287–1295 (2008).
Google Scholar 

30.
Parkes, J. P. et al. Rapid eradication of feral pigs (Sus scrofa) from Santa Cruz Island, California. Biol. Conserv. 143, 634–641 (2010).
Google Scholar 

31.
Williams, B. L., Holtfreter, R. W., Ditchkoff, S. S. & Grand, J. B. Efficiency of time-lapse intervals and simple baits for camera surveys of wild pigs. J. Wildl. Manag. 75, 655–659 (2011).
Google Scholar 

32.
Engeman, R. M., Massei, G., Sage, M. & Gentle, M. N. Monitoring wild pig populations: a review of methods. Environ. Sci. Pollut. R. 20, 8077–8091 (2013).
CAS  Google Scholar 

33.
Davis, A. J. et al. Quantifying site-level usage and certainty of absence for an invasive species though occupancy analysis of camera-trap data. Biol. Invasions 20, 877–890 (2018).
Google Scholar 

34.
Peine, J. D. & Farmer, J. A. Wild hog management program at Great Smoky Mountain National Park. Proceedings of the 14 thVertebrate Pest Management Conference 14, 221–227 (1990).

35.
Saunders, G., Kay, B. & Nicol, H. Factors affecting bait uptake and trapping success for feral pigs (Sus scrofa) in Kosciusko National Park. Wildl. Res. 20, 653–665 (1993).
Google Scholar 

36.
Phillips, L. M., Smith, M. D. & Johnson, D. K. Effects of opportunistic shooting on trap visitation by wild pigs. Proceedings of the 15th Wildlife Damage Management Conference 15, 37–38 (2013).

37.
Bowman, B., Belant, J. L., Beyer, D. E. & Martel, D. Characterizing nontarget species use at bait sites for white-tailed deer. Hum.-Wildl. Interact. 9, 110–118 (2015).
Google Scholar 

38.
Schley, L., Dufrene, M., Krier, A. & Frantz, A. C. Patterns of crop damage by wild boar (Sus scrofa) in Luxembourg over a 10-year period. Eur. J. Wildl. Res. 54, 589–599 (2008).
Google Scholar 

39.
Engeman, R. M., Terry, J., Stephens, L. R. & Gruver, K. S. Prevalence and amount of feral swine damage to three row crops at planting. Crop Prot. 112, 252–256 (2018).
Google Scholar 

40.
United States Department of Agriculture. Field Crops Usual Planting and Harvesting Dates (October 2010). https://usda.library.cornell.edu/concern/publications/vm40xr56k?locale=en (2010).

41.
R: A language and environment for statistical computing, R Core Team, R Foundation for Statistical Computing, Vienna, Austria, 2017, https://www.R-project.org.

42.
Zhang, Y. Likelihood-based and Bayesian methods for Tweedie compound Poisson linear mixed models. Stat. Comput. 23, 743–757 (2013).
MathSciNet  CAS  MATH  Google Scholar 

43.
Bates, D., Mächler, M., Bolker, B. M. & Walker, S. C. Fitting linear mixed-effects models using lme4. J. Stat. Softw. 67, 1–48 (2015).
Google Scholar 

44.
MuMIn. R package version 1.43.17 (2020)

45.
Burnham, K. P. & Anderson, D. R. Model Selection and Multimodel Inference: a Practical Information-Theoretic Approach (Springer, New York, 2002).
Google Scholar 

46.
Lukacs, P. M., Burnham, K. P. & Anderson, D. R. Model selection bias and Freedman’s paradox. Ann. I. Stat. Math. 62, 117–125 (2010).
MathSciNet  MATH  Google Scholar 

47.
Nakagawa, S. & Freckleton, R. P. Model averaging, missing data and multiple imputation: a case study for behavioural ecology. Behav. Ecol. Sociobiol. 65, 103–116 (2010).
Google Scholar 

48.
Grueber, C. E., Nakagawa, S., Laws, R. J. & Jamieson, I. G. Multimodel inference in ecology and evolution: challenges and solutions. Evol. Biol. 24, 699–711 (2011).
CAS  Google Scholar 

49.
Bodenchuk, M. Method-specific costs of feral swine removal in a large metapopulation: the Texas experience. Proceedings of the 26 thVertebrate Pest Conference 26, 269–271 (2014).

50.
Davis, A. J., Leland, B., Bodenchuk, M., VerCauteren, K. C. & Pepin, K. Costs and effectiveness of damage management of an overabundant species. Wildlife Res. 45, 696–705 (2018).
Google Scholar 

51.
Massei, G., Genov, P. V., Staines, B. W. & Gorman, M. L. Mortality of wild boar, Sus scrofa, in a Mediterranean area in relation to sex and age. J. Zool. 242, 394–400 (1997).
Google Scholar 

52.
Castillo-Contreras, R., Carvalho, J., Serrano, E., Mentaberre, G. & Fernandez-Aguilar, X. Urban wild boars prefer fragmented areas with food resources near natural corridors. Sci. Total Environ. 615, 282–288 (2018).
ADS  CAS  PubMed  Google Scholar 

53.
Gonzalez-Crespo, C., Serrano, E., Cahill, S., Castillo-Contreras, R. & Cabaneros, L. Stochastic assessment of management strategies for a Mediterranean peri-urban wild boar population. PLoS ONE 13, e0202289 (2018).
PubMed  PubMed Central  Google Scholar 

54.
Van Vuren, D. Diurnal activity and habitat use by feral pigs on Santa Cruz Island, California. Calif. Fish Game 70, 140–144 (1984).
Google Scholar 

55.
Baber, D. W. & Coblentz, B. E. Density, home range, habitat use, and reproduction in feral pigs on Santa Catalina Island. J. Mammal. 67, 512–525 (1984).
Google Scholar 

56.
Dexter, N. The influence of pasture distribution and temperature on habitat selection by feral pigs in a semi-arid environment. Wildlife Res. 25, 547–559 (1998).
Google Scholar 

57.
Choquenot, D. & Ruscoe, W. S. Landscape complementation and food limitation of large herbivores: habitat-related constraints on the foraging efficiency of wild pigs. J. Anim. Ecol. 72, 14–26 (2003).
Google Scholar 

58.
Dardaillon, M. Seasonal variations in habitat selection and spatial distribution of wild boar (Sus scrofa) in the Camargue, southern France. Behav. Process. 13, 251–268 (1986).
CAS  Google Scholar 

59.
Waithman, J. Guide to Hunting Wild Pigs in California (California Department of Fish and Game, Sacramento, 2001).
Google Scholar 

60.
Snow, N. P. et al. Bait preference of free-ranging feral swine for delivery of a novel toxicant. PLoS ONE 11, e0146712 (2016).
PubMed  PubMed Central  Google Scholar 

61.
Brivio, F. et al. An analysis of intrinsic and extrinsic factors affecting the activity of a nocturnal species: the wild boar. Mamm. Biol. 84, 73–81 (2017).
Google Scholar 

62.
Choquenot, D., Hone, J. & Saunders, G. Using aspects of predator-prey theory to evaluate helicopter shooting for feral pig control. Wildl. Res. 26, 251–261 (1999).
Google Scholar 

63.
Pepin, K. M., Snow, N. P. & VerCauteren, K. C. Optimal bait density for delivery of acute toxicants to vertebrate pests. J. Pest Sci. 93, 723–735 (2020).
Google Scholar 

64.
Lancia, R. A., Bishir, J. W., Conner, M. C. & Rosenberry, C. S. Use of catch-effort to estimate population size. Wildl. Soc. B. 24, 731–737 (1996).
Google Scholar 

65.
McIlroy, J. C. & Gifford, E. J. The ‘Judas’ pig technique for controlling feral pigs. Wildl. Res. 24, 483–491 (1997).
Google Scholar 

66.
Ditchkoff, S. S., Jolley, D. B., Sparklin, B. D., Hanson, L. B. & Mitchell, M. S. Reproduction in a population of wild pigs (Sus scrofa) subjected to lethal control. J. Wildl. Manag. 76, 1235–1240 (2012).
Google Scholar 

67.
Brook, R. K. & Van Beest, F. M. Feral wild boar distribution and perceptions of risk on central Canadian Prairies. Wildl. Soc. B. 38, 486–494 (2014).
Google Scholar 

68.
Stolle, K., Van Beest, F. M., Wal, E. D. & Brook, R. K. Diurnal and nocturnal activity patterns of invasive wild boar (Sus scrofa) in Saskatchewan, Canada. Can. Field Nat. 129, 76–79 (2015).
Google Scholar  More

1.
WWF. Living Planet Report 2016. Risk and Resilience in a New Era. (Gland, Switzerland: WWW International, 2016).
2.
Johnson, C. N. et al. Biodiversity losses and conservation responses in the Anthropocene. Science 356, 270–275 (2017).
ADS  CAS  PubMed  Google Scholar 

3.
Dirzo, R. et al. Defaunation in the Anthropocene. Science 345, 401–406 (2014).
ADS  CAS  PubMed  Google Scholar 

4.
IPBES. Summary for policymakers of the global assessment report on biodiversity and ecosystem services of the Intergovernmental Science-Policy Platform on Biodiversity and Ecosystem Services. (IPBES secretariat, 2019).

5.
Sax, D. F. & Gaines, S. D. Species diversity: from global decreases to local increases. Trends Ecol. Evol. 18, 561–566 (2003).
Google Scholar 

6.
Dornelas, M. et al. Assemblage time series reveal biodiversity change but not systematic loss. Science 344, 296–299 (2014).
ADS  CAS  PubMed  Google Scholar 

7.
Conrad, K. F., Warren, M. S., Fox, R., Parsons, M. S. & Woiwod, I. P. Rapid declines of common, widespread British moths provide evidence of an insect biodiversity crisis. Biol. Conserv. 132, 279–291 (2006).
Google Scholar 

8.
Goulson, D., Lye, G. C. & Darvill, B. Decline and conservation of bumble bees. Annu. Rev. Entomol. 53, 191–208 (2008).
CAS  PubMed  Google Scholar 

9.
Thomas, J. A. Butterfly communities under threat. Science 353, 216–218 (2016).
ADS  CAS  PubMed  Google Scholar 

10.
Hallmann, C. A. et al. More than 75 percent decline over 27 years in total flying insect biomass in protected areas. PLoS ONE 12, e0185809 (2017).
PubMed  PubMed Central  Google Scholar 

11.
Chamberlain, D. E. & Fuller, R. J. Local extinctions and changes in species richness of lowland farmland birds in England and Wales in relation to recent changes in agricultural land-use. Agric. Ecosyst. Environ. 78, 1–17 (2000).
Google Scholar 

12.
Inger, R. et al. Common European birds are declining rapidly while less abundant species’ numbers are rising. Ecol. Lett. 18, 28–36 (2015).
PubMed  Google Scholar 

13.
Rosenberg, K. V. et al. Decline of the North American avifauna. Science 366, 120–124 (2019).
ADS  CAS  PubMed  Google Scholar 

14.
Haase, P. et al. Moderate warming over the past 25 years has already reorganized stream invertebrate communities. Sci. Total Environ. 658, 1531–1538 (2019).
ADS  CAS  PubMed  Google Scholar 

15.
Baranov, V., Jourdan, J., Pilotto, F., Wagner, R. & Haase, P. Complex and nonlinear climate-driven changes in freshwater insect communities over 42 years. Conserv. Biol. https://doi.org/10.1111/cobi.13477 (2020).

16.
Martinho, F. et al. Does the flatfish community of the Mondego estuary (Portugal) reflect environmental changes? J. Appl. Ichthyol. 26, 843–852 (2010).
Google Scholar 

17.
Knapp, S., Kühn, I., Stolle, J. & Klotz, S. Changes in the functional composition of a Central European urban flora over three centuries. Perspect. Plant Ecol. Evol. Syst. 12, 235–244 (2010).
Google Scholar 

18.
Förster, A., Becker, T., Gerlach, A., Meesenburg, H. & Leuschner, C. Long-term change in understorey plant communities of conventionally managed temperate deciduous forests: effects of nitrogen deposition and forest management. J. Veg. Sci. 28, 747–761 (2017).
Google Scholar 

19.
Steinbauer, M. J. et al. Accelerated increase in plant species richness on mountain summits is linked to warming. Nature 556, 231–234 (2018).
ADS  CAS  PubMed  Google Scholar 

20.
Vellend, M. et al. Global meta-analysis reveals no net change in local-scale plant biodiversity over time. Proc. Natl Acad. Sci. USA 110, 19456–19459 (2013).
ADS  CAS  PubMed  Google Scholar 

21.
Primack, R. B. et al. Biodiversity gains? The debate on changes in local- vs global-scale species richness. Biol. Conserv. 219, A1–A3 (2018).
Google Scholar 

22.
Blowes, S. A. et al. The geography of biodiversity change in marine and terrestrial assemblages. Science 366, 339–345 (2019).
ADS  CAS  PubMed  Google Scholar 

23.
Bowler, D. E. et al. Cross-realm assessment of climate change impacts on species’ abundance trends. Nat. Ecol. Evol. 1, s41559–016 (2017). 0067–016.
Google Scholar 

24.
Gibson-Reinemer, D. K., Sheldon, K. S. & Rahel, F. J. Climate change creates rapid species turnover in montane communities. Ecol. Evol. 5, 2340–2347 (2015).
PubMed  PubMed Central  Google Scholar 

25.
Domisch, S. et al. Modelling distribution in European stream macroinvertebrates under future climates. Glob. Change Biol. 19, 752–762 (2013).
ADS  Google Scholar 

26.
Mirtl, M. et al. Genesis, goals and achievements of Long-Term Ecological Research at the global scale: a critical review of ILTER and future directions. Sci. Total Environ. 626, 1439–1462 (2018).
ADS  CAS  PubMed  Google Scholar 

27.
Parmesan, C. et al. Poleward shifts in geographical ranges of butterfly species associated with regional warming. Nature 399, 579–583 (1999).
ADS  CAS  Google Scholar 

28.
Walther, G.-R. et al. Ecological responses to recent climate change. Nature 416, 389–395 (2002).
ADS  CAS  PubMed  PubMed Central  Google Scholar 

29.
Parmesan, C. & Yohe, G. A globally coherent fingerprint of climate change impacts across natural systems. Nature 421, 37–42 (2003).
ADS  CAS  Article  Google Scholar 

30.
Pöyry, J., Luoto, M., Heikkinen, R. K., Kuussaari, M. & Saarinen, K. Species traits explain recent range shifts of Finnish butterflies. Glob. Change Biol. 15, 732–743 (2009).
ADS  Google Scholar 

31.
Seibold, S. et al. Arthropod decline in grasslands and forests is associated with landscape-level drivers. Nature 574, 671–674 (2019).
ADS  CAS  PubMed  Google Scholar 

32.
Sánchez-Bayo, F. & Wyckhuys, K. A. G. Worldwide decline of the entomofauna: a review of its drivers. Biol. Conserv. 232, 8–27 (2019).
Google Scholar 

33.
Simmons, B. I. et al. Worldwide insect declines: an important message, but interpret with caution. Ecol. Evol. 9, 3678–3680 (2019).
PubMed  PubMed Central  Google Scholar 

34.
Valtonen, A. et al. Long‐term species loss and homogenization of moth communities in Central Europe. J. Anim. Ecol. 86, 730–738 (2017).
PubMed  Google Scholar 

35.
Thomas, C. D. Local diversity stays about the same, regional diversity increases, and global diversity declines. Proc. Natl Acad. Sci. USA 110, 19187–19188 (2013).
ADS  CAS  PubMed  Google Scholar 

36.
Larsen, S., Chase, J. M., Durance, I. & Ormerod, S. J. Lifting the veil: richness measurements fail to detect systematic biodiversity change over three decades. Ecology 99, 1316–1326 (2018).
PubMed  Google Scholar 

37.
Olden, J. D. & Poff, N. L. Toward a mechanistic understanding and prediction of biotic homogenization. Am. Nat. 162, 442–460 (2003).
PubMed  Google Scholar 

38.
Winter, M. et al. Plant extinctions and introductions lead to phylogenetic and taxonomic homogenization of the European flora. Proc. Natl Acad. Sci. USA 106, 21721–21725 (2009).
ADS  CAS  PubMed  Google Scholar 

39.
Hillebrand, H. et al. Biodiversity change is uncoupled from species richness trends: Consequences for conservation and monitoring. J. Appl. Ecol. 55, 169–184 (2018).
Google Scholar 

40.
Antão, L. H., Pöyry, J., Leinonen, R. & Roslin, T. Contrasting latitudinal patterns in diversity and stability in a high-latitude species-rich moth community. Glob. Ecol. Biogeogr. 29, 896–907 (2020).
Google Scholar 

41.
Hastings, A. et al. Transient phenomena in ecology. Science 361, eaat6412 (2018).
PubMed  Google Scholar 

42.
Gaüzère, P., Iversen, L. L., Barnagaud, J.-Y., Svenning, J.-C. & Blonder, B. Empirical predictability of community responses to climate change. Front. Ecol. Evol. 6, 186 (2018).

43.
Essl, F. et al. Socioeconomic legacy yields an invasion debt. Proc. Natl Acad. Sci. USA 108, 203–207 (2011).
ADS  CAS  PubMed  Google Scholar 

44.
Monteith, D. T. et al. Biological responses to the chemical recovery of acidified fresh waters in the UK. Environ. Pollut. 137, 83–101 (2005).
CAS  PubMed  Google Scholar 

45.
Rose, R. et al. Evidence for increases in vegetation species richness across UK Environmental Change Network sites linked to changes in air pollution and weather patterns. Ecol. Indic. 68, 52–62 (2016).
CAS  Google Scholar 

46.
Kuemmerle, T. et al. Hotspots of land use change in Europe. Environ. Res. Lett. 11, 064020 (2016).
ADS  Google Scholar 

47.
Didham, R. K., Tylianakis, J. M., Hutchison, M. A., Ewers, R. M. & Gemmell, N. J. Are invasive species the drivers of ecological change? Trends Ecol. Evol. 20, 470–474 (2005).
PubMed  Google Scholar 

48.
Martínez‐Abraín, A., Jiménez, J. & Oro, D. Pax Romana: ‘refuge abandonment’ and spread of fearless behavior in a reconciling world. Anim. Conserv. 22, 3–13 (2019).
Google Scholar 

49.
Kröel-Dulay, G. et al. Increased sensitivity to climate change in disturbed ecosystems. Nat. Commun. 6, 1–7 (2015).
Google Scholar 

50.
Simmons, B. I. et al. Worldwide insect declines: an important message, but interpret with caution. Ecol. Evol. 9, 3678–3680 (2019).
PubMed  PubMed Central  Google Scholar 

51.
Cardinale, B. J., Gonzalez, A., Allington, G. R. H. & Loreau, M. Is local biodiversity declining or not? A summary of the debate over analysis of species richness time trends. Biol. Conserv. 219, 175–183 (2018).
Google Scholar 

52.
Gonzalez, A. et al. Estimating local biodiversity change: a critique of papers claiming no net loss of local diversity. Ecology 97, 1949–1960 (2016).
PubMed  Google Scholar 

53.
Habel, J. C. et al. Butterfly community shifts over two centuries. Conserv. Biol. 30, 754–762 (2016).
PubMed  Google Scholar 

54.
Soga, M. & Gaston, K. J. Shifting baseline syndrome: causes, consequences, and implications. Front. Ecol. Environ. 16, 222–230 (2018).
Google Scholar 

55.
Silliman, B. R. et al. Are the ghosts of nature’s past haunting ecology today? Curr. Biol. 28, R532–R537 (2018).
CAS  PubMed  Google Scholar 

56.
Battarbee, R. W. et al. Recovery of UK lakes from acidification: An assessment using combined palaeoecological and contemporary diatom assemblage data. Ecol. Indic. 37, 365–380 (2014).
CAS  Google Scholar 

57.
Barnosky, A. D. et al. Merging paleobiology with conservation biology to guide the future of terrestrial ecosystems. Science 355, eaah4787 (2017).

58.
Albrecht, J. et al. Logging and forest edges reduce redundancy in plant-frugivore networks in an old-growth European forest. J. Ecol. 101, 990–999 (2013).
Google Scholar 

59.
Kareiva, P., Marvier, M. & Silliman, B. Effective Conservation Science: Data Not Dogma. (Oxford University Press, 2017).

60.
Haase, P. et al. The next generation of site-based long-term ecological monitoring: Linking essential biodiversity variables and ecosystem integrity. Sci. Total Environ. 613–614, 1376–1384 (2018).
ADS  PubMed  Google Scholar 

61.
Heffernan, J. B. et al. Macrosystems ecology: understanding ecological patterns and processes at continental scales. Front. Ecol. Environ. 12, 5–14 (2014).
Google Scholar 

62.
Harvey, J. A. et al. International scientists formulate a roadmap for insect conservation and recovery. Nat. Ecol. Evol. 4, 174–176 (2020).
PubMed  Google Scholar 

63.
Hallett, L. et al. codyn: Community Dynamics Metrics. R package version 2.0.0. (2018).

64.
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, (2008).

65.
Kendall, M. G. Rank correlation methods. (1948).

66.
Mann, H. B. Nonparametric Tests Against Trend. Econometrica 13, 245–259 (1945).
MathSciNet  MATH  Google Scholar 

67.
Venerables, W. N. & Ripley, B. D. Modern applied statistics with S. (new york: Springer, 2002).

68.
Hamed, K. H. & Rao, A. R. A modified Mann-Kendall trend test for autocorrelated data. J. Hydrol. 204, 182–196 (1998).
ADS  Google Scholar 

69.
Daufresne, M., Lengfellner, K. & Sommer, U. Global warming benefits the small in aquatic ecosystems. Proc. Natl Acad. Sci. USA 106, 12788–12793 (2009).
ADS  CAS  PubMed  Google Scholar 

70.
Viechtbauer, W. Conducting meta-analyses in R with the metafor package. J. Stat. Softw. 36, 1–48 (2010).
Google Scholar 

71.
Everaert, G., Deschutter, Y., De Troch, M., Janssen, C. R. & De Schamphelaere, K. Multimodel inference to quantify the relative importance of abiotic factors in the population dynamics of marine zooplankton. J. Mar. Syst. 181, 91–98 (2018).
Google Scholar 

72.
Anderson, D. R. & Burnham, K. P. Avoiding pitfalls when using information-theoretic methods. J. Wildl. Manag. 66, 912–918 (2002).
Google Scholar 

73.
Calcagno, V. glmulti: Model selection and multimodel inference made easy. R package version 1.0.7. (2013).

74.
EEA. Biogeographical regions and Marine regions and subregions under the Marine Strategy Framework Directive. https://www.eea.europa.eu/data-and-maps (2016). More

Data collection takes place in the Seine estuary sector extending from Ouistreham (Coordinates in projection world geodesic system 1984 or WGS84, 49°17′N 0°16′W) to Antifer (49°40′20″N 0°11′21″E) and from the Pont de Normandie (49°26′09″N 0°16′28″E) to roughly 20 meter-depth offshore to the west (Fig. 1). This 20 meter-depth limit delimitates the area considered as part of the nursery grounds23. The survey follows a fixed stratified sampling design. The stratification is based on bathymetry and distance to the mouth estuary. In total, 47 hauls are distributed across 12 sectors. Haul positions are randomly drawn in each sector. Due to rocky outcrop and the presence of many shipwrecks in the area, hauls’ locations are later assessed based on recommendations from professional fishers operating in the area and adjusted where needed. Morin and Schlaich23 provided a standardized sampling protocol for nursery zones from 1995 to 2017. In 2018, the protocol was updated in order to obtain a standardized sampling protocol on a national scale and to comply with the French Marine Strategy Framework Directive (MSFD) survey plan19. Differences in the two protocols for this particular survey are highlighted where needed. Sampling occurred once a year from 1995 to 2002, then from 2008 to 2010 and from 2017 to 2019. The two first periods strictly follow the first protocol. Only the last years are susceptible to changes due to protocol updates.
Fig. 1

Geographical extent and sectors of the NOURSEINE survey displaying the position of hauls performed across all years. Sectors are originally established from the distance to the estuary and the bathymetry.

Full size image

Sampling is carried out with a 20 mm mesh size beam trawl of 2 or 3 m wide depending on the sectors, with a 0.50 m vertical opening. The beam trawl is equipped with ground chains. Each haul lasts 15 minutes and is done against the tide at speed between 2.5 to 2.8 knots. From 2018 onward, a length of 7 minutes for the 2 m beam trawl was applied, in line with the updated national protocol. Shooting and hauling coordinates, times and depths of each haul are systematically noted. Using two different fishing gear may cause differences in the catchability of individuals, leading to differences in population characteristics estimates. An intercalibration exercise was implemented and results are presented in Riou’s work24. Data on flounder and sole captures were used to draw the comparison. Briefly, they showed that there were no differences in the mean density nor in the size structure for these two species. Therefore, the density values are considered comparable no matter the gear used in this protocol.
The period of reference for sampling is at the end of summer or beginning of autumn. Sampling dates scope from August 25 to September 30 over the time series. The juvenile stages here regroup individuals of age 1 and age 0. The latter corresponds to individuals who settled in the estuary on the year of the survey. Fish from age 0 group had their first period of growth over the summer. Sampling in late summer or early autumn ensures good catchability by the 20 mm mesh size beam trawl providing an accurate image of the fish distribution and abundance. Each survey day, 12 to 15 trawl stations are performed. In total, 40 to 47 stations are sampled each year. In 1996, 63 stations were surveyed as replicates were done. Hauls of a given station locate themselves relatively close to each other throughout the surveys.
After each haul, the content of the trawl is emptied on deck, and a total or partial sorting is carried out depending on the volume and homogeneity of the capture. All taxa, both fish and benthos, are sorted, identified, counted and weighted. Fishes of commercial value and all others flatfish are measured. Otoliths are collected on the main commercial fish species (sole, plaice, flounder, dab, pouting, large whiting and European bass) for later age group determination in the laboratory. In 1999, the sampling was incomplete and only commercial fish and invertebrates (King scallop and lobster) were sampled. The year was kept in the dataset to ensure continuity for these taxa.
Sorting the capture can be separated into three different steps (Fig. 2):
1.
Total capture weighting: when the hauls are emptied on the deck, the whole capture is distributed in several baskets/box in order to weight it.

2.
Fish and large taxa sorting: All fish and large taxa of invertebrates easily identified (edible crab, common spider crab, large cephalopods) are sorted, identified, numbered, measured (for fish) and weighted (total weight per taxa). Depending on the size of the capture, subsampling might be necessary. Operations are performed on the subsample in such a case. If visual identification is too difficult (for instance due to a large mud proportion), the capture is washed using a 5 mm sieve. The weight ratio between the total capture and the subsample form a “division” variable that allows the calculation of density. Another subsampling may be needed if a taxon has a high abundance. In that case, for practical reasons, only a subsample of the individuals are numbered, measured and weighted.

3.
Benthic fauna sorting: What is left from the second step is weighted before the sorting operation. All taxa constituting benthic fauna are sorted, identified, numbered and weighted (total weight per taxa). Some taxa may be measured (whelk, scallop). As for step 2, a subsample might be necessary before sorting according to the quantity of benthic fauna.

Fig. 2

Sketch of the capture sorting process used during the NOURSEINE surveys (adapted from19).

Full size image

All observations are manually recorded on fieldwork paper books before being checked and registered in the NOURSEINE database.
The NOURSEINE database consists of all information on fish and benthic taxa collected in a given haul, together with haul and survey information. Throughout the survey period, some changes on the level of identification are observed: while all fish taxa were normally considered and processed, sampling was reduced to commercial taxa solely during the 1999 survey. Changes in human operators may lead to mis-identifications and irregular records of a same taxa through the dataset. To provide a readily exploitable dataset, taxa clustering was applied to keep a homogenous record in the time series. Changes were mostly applied to benthic taxa (Table 1). In hauls where several taxa were clustered, abundance and weight were summed to calculate taxa density accordingly. Out of the 161 taxa initially recorded in the database, 138 remained after clustering.
Table 1 Outcome of the clustering process applied to homogenize the dataset.
Full size table

As only raw abundance is available first-hand, taxa densities are calculated based on trawled surface but also takes into account if the haul has been partially sorted or not. The formula to calculate the density of individuals per surface unit is:

$$Density=frac{(Raw,abundanceast Divisionast Coefficient)}{Trawled,surface}$$

where Division is a factor used to elevate the abundance if the whole haul was not sorted. The same formula with abundance replaced by the capture’s weight gives the captured weight per surface unit.
The database is reworked and corrected in an R script before being provided here. The coordinates of each haul are given at the beginning and the end of the fishing operation in degrees, minutes and seconds. They are converted in decimal degrees. It is in this R script that the taxa density is calculated, along with the mean weight of the capture, and that taxa clustering happens.
Efforts have been made to detect and correct any typos that potentially slipped through the first correction when data are entered in the database. More