Philippa Kaur

1.
Perry, A. L., Low, P. J., Ellis, J. R. & Reynolds, J. D. Climate change and distribution shifts in marine fishes. Science 308, 1912–1915 (2005).
CAS  Google Scholar 
2.
Albouy, C. et al. Projected climate change and the changing biogeography of coastal Mediterranean fishes. J. Biogeogr. 40, 534–547 (2013).
Google Scholar 

3.
Hoegh-Guldberg, O. et al. Impacts of 1.5 °C global warming on natural and human systems. in Special Report on Global Warming of 1.5 °C (eds Masson-Delmotte, V. et al.) 212–251 (WMO, 2018).

4.
Hughes, T. P. et al. Spatial and temporal patterns of mass bleaching of corals in the Anthropocene. Science 359, 80–83 (2018).
CAS  Google Scholar 

5.
Smale, D. A. et al. Marine heatwaves threaten global biodiversity and the provision of ecosystem services. Nat. Clim. Change 9, 306–312 (2019).
Google Scholar 

6.
EUROSTAT EU Fishery Economic Report 2010. European Union Mediterranean and Black Sea Fishing Fleet (FIRMS, 2010).

7.
Abulafia, D. The Great Sea: a Human History of the Mediterranean (Oxford Univ. Press, 2011).

8.
Giakoumi, S. et al. Ecoregion-based conservation planning in the Mediterranean: dealing with large-scale heterogeneity. PLoS ONE 8, e76449 (2013).
CAS  Google Scholar 

9.
Katsanevakis, S. et al. Invading the Mediterranean Sea: biodiversity patterns shaped by human activities. Front. Mar. Sci. 1, 32 (2014).
Google Scholar 

10.
Pinardi, N., Arneri, E., Crise, A., Ravaioli, M. & Zavatarelli, M. in The Sea Vol. 14 (eds Robinson, A. & Brink, K.) 1243–1330 (Harvard Univ. Press, 2006).

11.
Lejeusne, C., Chevaldonné, P., Pergent-Martini, C., Boudouresque, C. F. & Perez, T. Climate change effects on a miniature ocean: the highly diverse, highly impacted Mediterranean Sea. Trends Ecol. Evol. 25, 250–260 (2010).
Google Scholar 

12.
Cramer, W. et al. Climate change and interconnected risks to sustainable development in the Mediterranean. Nat. Clim. Change 8, 972–980 (2018).
Google Scholar 

13.
Bintanja, R. & van der Linden, E. C. The changing seasonal climate in the Arctic. Sci. Rep. 3, 1556 (2013).
CAS  Google Scholar 

14.
Kharin, V. V., Zwiers, F. W., Zhang, X. & Wehner, M. Changes in temperature and precipitation extremes in the CMIP5 ensemble. Clim. Change 119, 345–357 (2013).
Google Scholar 

15.
Thibeault, J. M. & Seth, A. Changing climate extremes in the Northeast United States: observations and projections from CMIP5. Clim. Change 127, 273–287 (2014).
Google Scholar 

16.
Both, C. et al. Avian population consequences of climate change are most severe for long-distance migrants in seasonal habitats. Proc. R. Soc. Lond. B Biol. Sci. 2010, 1259–1266 (1685).
Google Scholar 

17.
Chesson, P. et al. Resource pulses, species interactions, and diversity maintenance in arid and semi-arid environments. Oecologia 141, 236–253 (2004).
Google Scholar 

18.
Tonkin, J. D., Bogan, M. T., Bonada, N., Rios-Touma, B. & Lytle, D. A. Seasonality and predictability shape temporal species diversity. Ecology 98, 1201–1216 (2017).
Google Scholar 

19.
Hiddink, J. G. & Ter Hofstede, R. Climate induced increases in species richness of marine fishes. Glob. Change Biol. 14, 453–460 (2008).
Google Scholar 

20.
Rutterford, L. A. et al. Future fish distributions constrained by depth in warming seas. Nat. Clim. Change 5, 569–573 (2015).
Google Scholar 

21.
Colloca, F., Scarcella, G. & Libralato, S. Recent trends and impacts of fisheries exploitation on Mediterranean stocks and ecosystems. Front. Mar. Sci. 4, 244 (2017).
Google Scholar 

22.
Barange, M. et al. Impacts of Climate Change on Fisheries and Aquaculture: Synthesis of Current Knowledge, Adaptation and Mitigation Options. FAO Fisheries and Aquaculture Technical Paper No. 627 (FAO, 2018).

23.
IPCC Climate Change 2014: Synthesis Report (eds Core Writing Team, Pachauri, R. K. & Meyer, L. A.) (IPCC, 2014).

24.
Rice, J. C. & Garcia, S. M. Fisheries, food security, climate change, and biodiversity: characteristics of the sector and perspectives on emerging issues. ICES J. Mar. Sci. 68, 1343–1353 (2011).
Google Scholar 

25.
Free, C. M. et al. Impacts of historical warming on marine fisheries production. Science 363, 979–983 (2019).
CAS  Google Scholar 

26.
Ben Rais Lasram, F. et al. The Mediterranean Sea as a ‘cul‐de‐sac’ for endemic fishes facing climate change. Glob. Change Biol. 16, 3233–3245 (2010).
Google Scholar 

27.
Yeager, L. A., Deith, M. C., McPherson, J. M., Williams, I. D. & Baum, J. K. Scale dependence of environmental controls on the functional diversity of coral reef fish communities. Glob. Ecol. Biogeogr. 26, 1177–1189 (2017).
Google Scholar 

28.
Zenetos, A. et al. Annotated list of marine alien species in the Mediterranean with records of the worst invasive species. Mediterr. Mar. Sci. 6, 63–118 (2005).
Google Scholar 

29.
The State of Mediterranean and Black Sea Fisheries: 2018 (FAO & General Fisheries Commission for the Mediterranean, 2018).

30.
Scientific Technical and Economic Committee for Fisheries (STECF) The 2015 Annual Economic Report on the EU Fishing Fleet (STECF-15-07) (Publications Office of the European Union, 2015).

31.
Albouy, C. et al. FishMed: traits, phylogeny, current and projected species distribution of Mediterranean fishes, and environmental data. Ecology 96, 2312–2313 (2015).
Google Scholar 

32.
Beuvier, J. et al. Modeling the Mediterranean Sea interannual variability during 1961–2000: focus on the Eastern Mediterranean Transient. J. Geophys. Res. Oceans 115, C08017 (2010).
Google Scholar 

33.
Arnell, N. W. Climate change and global water resources: SRES emissions and socio-economic scenarios. Glob. Environ. Change 14, 31–52 (2004).
Google Scholar 

34.
Adloff, F. et al. Mediterranean Sea response to climate change in an ensemble of twenty first century scenarios. Clim. Dynam. 45, 2775–2802 (2015).
Google Scholar 

35.
Whitehead, P. J. P., Bauchot, M., Hureau, J., Nielsen, J. & Tortonese, E. Fishes of the North-Eastern Atlantic and the Mediterranean Vol. 1 (UNESCO, 1984).

36.
Gonçalves, A. R., Von Zuben, F. J. & Banerjee, A. Multi-label structure learning with Ising model selection. In Proc. Twenty-Fourth International Joint Conference on Artificial Intelligence 3525–3531 (AAAI Press, 2015).

37.
Galar, M., Fernandez, A., Barrenechea, E., Bustince, H. & Herrera, F. A review on ensembles for the class imbalance problem: bagging-, boosting-, and hybrid-based approaches. IEEE Trans. Syst. Man. Cybern. C Appl. Rev. 42, 463–484 (2012).
Google Scholar 

38.
Gravel, D. et al. Bringing Elton and Grinnell together: a quantitative framework to represent the biogeography of ecological interaction networks. Ecography 42, 401–415 (2019).
Google Scholar 

39.
Dayton, P. K. Competition, disturbance, and community organization: the provision and subsequent utilization of space in a rocky intertidal community. Ecol. Monogr. 41, 351–389 (1971).
Google Scholar 

40.
Dormann, C. F. et al. Biotic interactions in species distribution modelling: 10 questions to guide interpretation and avoid false conclusions. Glob. Ecol. Biogeogr. 27, 1004–1016 (2018).
Google Scholar 

41.
Dormann, C. F., Fründ, J. & Schaefer, H. M. Identifying causes of patterns in ecological networks: opportunities and limitations. Annu. Rev. Ecol. Evol. Syst. 48, 559–584 (2017).
Google Scholar 

42.
Golding, N., Nunn, M. A. & Purse, B. V. Identifying biotic interactions which drive the spatial distribution of a mosquito community. Parasites Vectors 8, 367 (2015).
Google Scholar 

43.
Harris, D. J. Generating realistic assemblages with a joint species distribution model. Methods Ecol. Evol. 6, 465–473 (2015).
Google Scholar 

44.
Thorson, J. T. et al. Joint dynamic species distribution models: a tool for community ordination and spatio-temporal monitoring. Glob. Ecol. Biog. 25, 1144–1158 (2016).
Google Scholar 

45.
Azaele, S., Muneepeerakul, R., Rinaldo, A. & Rodriguez-Iturbe, I. Inferring plant ecosystem organization from species occurrences. J. Theor. Biol. 262, 323–329 (2010).
CAS  Google Scholar 

46.
Harris, D. J. Inferring species interactions from co‐occurrence data with Markov networks. Ecology 97, 3308–3314 (2016).
Google Scholar 

47.
Cheng, J., Levina, E., Wang, P. & Zhu, J. A sparse Ising model with covariates. Biometrics 70, 943–953 (2014).
Google Scholar 

48.
Lindberg, O. Markov Random Fields in Cancer Mutation Dependencies. MSc Thesis, Univ. Turku (2016).

49.
Clark, N. J., Wells, K. & Lindberg, O. Unravelling changing interspecific interactions across environmental gradients using Markov random fields. Ecology 99, 1277–1283 (2018).
Google Scholar 

50.
Lee, J. D. & Hastie, T. J. Learning the structure of mixed graphical models. J. Comput. Graph. Stat. 24, 230–253 (2015).
Google Scholar 

51.
Wood, S. N. Thin plate regression splines. J. R. Stat. Soc. B Stat. Methodol. 65, 95–114 (2003).
Google Scholar 

52.
Kammann, E. & Wand, M. P. Geoadditive models. J. R. Stat. Soc. C Appl. Stat. 52, 1–18 (2003).
Google Scholar 

53.
McInerny, G. J. & Purves, D. W. Fine‐scale environmental variation in species distribution modelling: regression dilution, latent variables and neighbourly advice. Methods Ecol. Evol. 2, 248–257 (2011).
Google Scholar 

54.
Givan, O., Parravicini, V., Kulbicki, M. & Belmaker, J. Trait structure reveals the processes underlying fish establishment in the Mediterranean. Glob. Ecol. Biogeogr. 26, 142–153 (2017).
Google Scholar 

55.
Mouillot, D. et al. Functional over-redundancy and high functional vulnerability in global fish faunas on tropical reefs. Proc. Natl Acad. Sci. USA 111, 13757–13762 (2014).
CAS  Google Scholar 

56.
Tsirogiannis, C. & Sandel, B. PhyloMeasures: a package for computing phylogenetic biodiversity measures and their statistical moments. Ecography 39, 709–714 (2015).
Google Scholar 

57.
Webb, C. O., Ackerly, D. D. & Kembel, S. W. Phylocom: software for the analysis of phylogenetic community structure and trait evolution. Bioinformatics 24, 2098–2100 (2008).
CAS  Google Scholar 

58.
Newman, M. E. Modularity and community structure in networks. Proc. Natl Acad. Sci. USA 103, 8577–8582 (2006).
CAS  Google Scholar 

59.
Csárdi, G. & Nepusz, T. The igraph software package for complex network research. InterJournal Complex Syst. 1965, 1–9 (2006).
Google Scholar 

60.
Pellissier, L. et al. Comparing species interaction networks along environmental gradients. Biol. Rev. 93, 785–800 (2017).
Google Scholar 

61.
Miller, P. J., Lubke, G. H., McArtor, D. B. & Bergeman, C. S. Finding structure in data using multivariate tree boosting. Psychol. Methods 21, 583–602 (2016).
Google Scholar 

62.
R Development Core Team R: A Language and Environment for Statistical Computing (R Foundation for Statistical Computing, 2017).

63.
Clark, N. J., Wells, K. & Lindberg, O. MRFcov: Markov random fields with additional covariates. R package version 1.0 https://github.com/nicholasjclark/MRFcov (2018).

64.
Wood, S. N. Fast stable restricted maximum likelihood and marginal likelihood estimation of semiparametric generalized linear models. J. R. Stat. Soc. B Stat. Methodol. 73, 3–36 (2011).
Google Scholar 

65.
Wickham, H. & Francois, R. dplyr: A grammar of data manipulation. R package version 0.7.2 https://CRAN.R-project.org/package=dplyr (2017).

66.
Wickham, H. ggplot2: Elegant Graphics for Data Analysis (Springer, 2016).

67.
Paradis, E., Claude, J. & Strimmer, K. APE: analyses of phylogenetics and evolution in R language. Bioinformatics 20, 289–290 (2004).
CAS  Google Scholar  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

Study site
This study was conducted in Shenmu County of Shaanxi Province in China (110° 05′–110° 30′ E, 39° 27′–39° 15′ N), which is located in the continental arid and Semiarid areas. The annual average temperature is 8.9 °C. The annual frost-free period is 130 days, the mean annual precipitation of the area is about 396 mm, and potential evaporation is 1,790 mm.
The research plot is in the heartland of Shendong coal mining subsidence area in Shenmu County, typical steppe landscape, soil impoverishment, and fragile ecological environment. The basic physical soil properties in the test site were measured (Table 1). According to the SL237-1999 engineering classification standard of the Geotechnical Test Regulations, the soil in the test area was named as low liquid limit silt (ML). Major plant species under natural conditions in the study area include S. psammophila, Caragana microphylla Lam., H. rhamnoides L., Artemisia ordosica Krasch., Agriophyllum squarrosum (Linn. ) Moq., and Lespedeza bicolor Turcz.
Table 1 Basic physical properties of soil in the test area.
Full size table

Root sampling
The straight roots of 4 years old of S. psammophila and H. rhamnoides L. were used as materials, and applied instantaneous axial small injured force (corresponding to 30% of the average ultimate force of the radial level, less than the elastic ultimate force, and the deformation is recoverable elastic deformation) and instantaneous axial large injured force (corresponding to 70% of the average ultimate force of the radial level, greater than the elastic ultimate force, and the deformation is irreversible plastic deformation) without leaving the plant body, to understand the survival rate, the change in root diameter and tensile strength of straight roots withstand transient tensile injured after self-repair.
As the layers of soil could interfere with the anti-tension force and anti-tension strength of roots, we excavated the roots without leaving the plant body and selected the roots which are distributed in the same soil layer. Straight roots with uniform diameter ranged from 1 to 4.5 mm, roots segments of 100 mm were selected from the root systems. To sufficiently attribute the tensile ability of roots, the selected roots were divided into seven diameter classes with 0.5 mm interval. To ensure the parallelism of the test, each diameter class selected eight test roots and eight control roots.
In the test, the soil around test roots was removed and the position of test roots was kept unchanged so that the exposed length of the roots reached the test requirements, the test roots were shaded, and sprayed water to maintain moisture. And each test root length was greater than 100 mm (Fig. 1), three points of A, B and C was selected along the root, and the diameters were measured by the cross method using an electronic Vernier caliper with an accuracy of 0.01 mm. B was the midpoint of the test root, A, C were the ends of the 30 mm from the midpoint. The diameters of control roots were measured in the same way.
Figure 1

The schematic diagram of test root.

Full size image

Root treatments
At the beginning of the plant growing season, in 2019 May. Test roots were applied two instantaneous axial injured forces without leaving the plant body, then covered soil growth. The process of excavation and covering soil was also carried out on the control roots in the same way (Fig. 1). By August, after a 3-month growth period, the test roots self-repaired for 3 months and excavated the test roots and control roots again, observed the survival rate, measured the root diameter and the tensile strength of test root (Fig. 2).
Figure 2

Excavation phase of test roots.

Full size image

Root tensile tests were conducted by a homemade portable instrument (Fig. 3). The instrument is composed with a platform (Part A), a root clap (Part B) fixed on the platform, a HG 100 digital display type push–pull meter (Part C), a moveable root clap (Part D) which is connected with Part C, a crank handle (Part E) which is used to move Part C and a Vernier caliper (Part F) which is connected with Part C to control the loading rate of the load. The test root is clapped by the two root pads. To make sure the test root not slip, we put a rubber pad inside each of the root clap. When the crank handle is turned, Part C is moved away from Part B and the force acted on the test root is recorded. The accuracy of HG 100 digital display type push–pull meter is 0.05 N. After selecting the test root, carefully excavated the soil under the test root and placed the instrument (length 50 cm, width 13 cm, height 20 cm). Fixed the points a and c of the test root at the jaws of the clamp so that the test root was in tension. The axial direction was pulled, and the length of the instrument was placed in the same direction as the root growth direction. Using 50 mm/min loading rate applied force injury by reading the Vernier caliper moving rate, stopped after the degree of injury urging force of the design, marked the ends of the test root segment and backfilled them, and marked on the ground to be dug again. The treatment method of the parallel control test roots was the same.
Figure 3

HG 100 digital display type push–pull meter and self-made portable test instrument.

Full size image

Determination of injury force
The recoverable elastic deformation occurs in the root system before the elastic limit point. After the elastic limit point, the root system undergoes irreversible plastic deformation. Previous researches indicate that the elastic limit of the 0–8 mm straight root of S. psammophila is about 40% of the ultimate tensile strength, and that of H. rhamnoides L. is about 60% of the ultimate tensile strength30,31. Tests have shown that when the injury force reaches 80% of the average ultimate tensile strength, more test roots break when applying the injury forces. To observe impact of varied injury force on the self-repair of roots, we selected two levels of injury force in this study, the small injury force was 30% of the ultimate tensile force (less than the elastic limit point), and the large injury force was 70% of the ultimate tensile force (greater than the elastic limit point).
Due to the uneven root diameter along the axial direction, it was impossible to determine the fracture point at which the test root may be damaged before the test. To guarantee data quality, the measured number was not recorded when the test root was fractured in the experiment. The ultimate force was calculated from the regression equation according to the average root diameter of each test root test segment, the average root diameter of each test root was the mean of the root diameters of the three points A, B, and C. According to the root diameter of each test root, the ultimate tensile force was calculated by the regression equation, and the corresponding small injury force and large injury force were determined (Table 2).
Table 2 Ultimate anti-fracture force and its regression equation with root diameter of two plants.
Full size table

Data analysis
The data were analyzed using SPSS 15.0 for Windows. The test roots and the parallel control roots were excavated after self-repaired for 3 months, observed the root shape, color and elasticity. If the root turned black, dry and begins to fall off, the root was dead. For the roots that survived, the root diameter and ultimate tensile strength were measured again, and the tensile strength was calculated using Eq. 1.

$$ {text{P}} = 4{text{F}}/left( {uppi {text{D}}^{2} } right) $$
(1)

where P is the tensile strength (MPa), F is the tensile force (N), D is the root diameter (mm). More