Protocol for assessing wildlife exposure and risk from pesticides in tropical food chains with limited biomonitoring data Authors: Shaorong Chen, Zijian Li Published: 2026 ABSTRACT SummaryPesticides threaten wildlife through terrestrial food chains, but data on wildlife exposure are scarce. Here, we present a protocol to assess pesticide health risks in wildlife using an integrated physiologically-based kinetic approach. We describe steps for data compilation, regional contamination scoring, and three-tiered exposure estimation using tissue, plant, and soil. We then detail procedures for ecological risk calculation, spatial mapping, and uncertainty analysis. This protocol enables traceable, quantitative, spatially resolved risk assessments for herbivores, carnivores, and omnivores.For complete details on the use and execution of this protocol, please refer to Chen and Li.1Subject areas: Bioinformatics, Health Sciences, Environmental sciences Before you begin Background Pesticides are widely used in agricultural, forestry, public health, and urban settings, but they can enter the environment through spray drift, runoff, leaching, volatilization, atmospheric deposition, and transboundary or long-distance transport.2,3,4 Consequently, pesticide residues may reach wildlife habitats and accumulate in environmental media such as soil, water, vegetation, and prey organisms.5,6,7,8 Through food-web transfer, herbivores, carnivores, and omnivores may be exposed via diet, drinking water, inhalation, dermal contact, or ingestion of contaminated plants and prey.9,10,11 Wildlife are essential for maintaining ecosystem structure, function, biodiversity, and stability.12,13 Therefore, assessing pesticide-related health risks in wildlife is important for protecting ecological health.13 However, biomonitoring data for free-ranging wildlife are often limited and may only include pesticide concentrations measured in whole bodies or internal organs such as liver, kidney, muscle, or fat.14,15 These data gaps make conventional risk assessment difficult. To address this limitation, this protocol develops a physiologically based kinetic (PBK) modeling framework that integrates limited internal biomonitoring data with modeled external exposure estimates.16 This internal–external exposure approach supports wildlife health risk assessment under data-limited conditions and provides a transparent method for evaluating pesticide-related risks across wildlife species and trophic groups. Innovation The innovation of this protocol is the development of an integrated modeling approach to track pesticide transfer through the terrestrial trophic food chain, from grass and other environmental media to herbivores, carnivores, and omnivores. The model can estimate pesticide distribution at each trophic level and use contamination in lower-level organisms to infer dietary exposure in higher-level wildlife. In addition, the framework can apply limited biomonitoring data, such as pesticide concentrations in tissues or organs, to back-calculate external exposure from internal body burdens. This internal–external modeling approach supports wildlife risk assessment under data-limited conditions. For each equation used in the protocol, the source and attribution type (i.e., whether directly adapted from the literature, modified from literature, or newly developed in this protocol) are summarized in Table S1. Model assumptions This protocol relies on a set of core assumptions that underpin the exposure and risk calculations: Steady-state condition Pesticide concentrations in wildlife tissues and environmental media are assumed to be at steady state (i.e., input equals elimination over the exposure period). This assumption enables the use of biotransfer factors (BTFs) and the back-calculation of daily intake from tissue residues. It is appropriate for chronic, continuous exposure scenarios but may overestimate or underestimate doses under episodic or pulsed exposure. Linear dose-response relationship The calculation of hazard quotients (HQ) and ecological carcinogenic risks (ECR) assumes linearity between the external dose (ADD) and the toxicological endpoint, consistent with the USEPA’s low-dose linear extrapolation for carcinogens and threshold-based approach for non-carcinogens. Species extrapolation When species-specific BTFs or physiological parameters are unavailable, values from closely related species or allometric scaling are used. This introduces uncertainty that is quantified in Step 22 but may still bias estimates if the target species differs substantially in metabolic capacity or tissue composition. Tissue concentration represents dietary exposure. For Tier 1, the measured liver concentration is assumed to reflect the dietary intake of the pesticide, with minimal contribution from non-dietary routes (inhalation, dermal absorption, drinking water). This assumption is reasonable for lipophilic pesticides with high oral bioavailability but may underestimate exposure for highly volatile or water-soluble compounds. Plant uptake is transpiration-driven. The soil-plant transfer model (Step 10) assumes that pesticide movement from soil to leaves is dominated by transpiration-driven mass flow, with diffusion and sorption described by the model parameters. This assumption is most valid for moderately hydrophobic compounds (log K∼ow∼ between 1 and 4) and under non-water-stressed conditions. Global mean concentration is representative. The contamination score (S) is calculated relative to a global mean concentration for each pesticide, weighted by the number of sampling sites. This global mean is assumed to represent a reasonable central tendency, although it is influenced by the uneven geographic distribution of data. Key resources table REAGENT or RESOURCESOURCEIDENTIFIERChemicals, peptides, and recombinant proteinsPesticide standards (for analytical confirmation if applicable)Sigma-Aldrich or equivalentVarious (CAS-specific)Deposited dataA dataset of pesticide residues in soil, wild plants and wild mammals from global nature reserves and undisturbed ecosystems, based on published research (2000–2024)Chen and Li (2026)1https://doi.org/10.1016/j.isci.2026.115870Software and algorithmsMicrosoft Excel 2021Microsofthttps://excel.cloud.microsoft/R (Version: 4.2.2)R Core Teamhttps://www.r-project.org/R Studio (Version: 2025.09.2+418)R Studio Teamhttp://www.rstudio.com/; RRID: SCR_000432ArcGIS (Version: 10.8)Environmental Systems Research Institutehttps://www.arcgis.com/index.html; RRID SCR_011081Origin 2024OriginLab Corporationhttps://www.originlab.com/;RRID: SCR_014212OtherReference manager (Mendeley)VariousMendeley: RRID: SCR_002750 Step-by-step method details Collecting data Timing: 2–3 months, depending on the data volume This section describes a systematic workflow (Figure 1) to retrieve and compile published data on pesticide concentrations in natural soils, wild plants, and wild mammals (herbivores, carnivores, omnivores). Follow the steps below to build a dataset suitable for the subsequent tiered risk assessment. 1.Define search strategy and data sources.a.Use Web of Science Core Collection as the primary source. (Secondary sources such as PubMed or Scopus may be added for broader coverage.)b.Time period: The search was limited to publications dated from January 1, 2000 to December 31, 2024.c.Construct keyword combinations following a three-group pattern: (environmental medium) + (pesticide terms) + (contamination indicators). Within each group, use “OR” to combine synonyms; connect groups with “AND”. The recommended keyword sets are:i.Wild mammals: “wild herbivores” OR “wild omnivores” OR “wild mammals” OR “wild animals” combined with pesticide-related terms (“pesticide”, “insecticide”, “herbicide”, “fungicide”) and contamination indicators (“residuals”, “pollution”, “concentration”).ii.Soils: “ecological soil” OR “highland soil” OR “mountain soil” OR “forest soil” OR “grassland soil” OR “hilly soil” OR “plain soil” combined with the same pesticide and contamination terms as above.iii.Wild plants: “wild plants” OR “vegetation” OR “shrubs” OR “wildflowers” OR “weeds” OR “trees” combined with the same pesticide and contamination terms. Define search strategy and data sources.a.Use Web of Science Core Collection as the primary source. (Secondary sources such as PubMed or Scopus may be added for broader coverage.)b.Time period: The search was limited to publications dated from January 1, 2000 to December 31, 2024.c.Construct keyword combinations following a three-group pattern: (environmental medium) + (pesticide terms) + (contamination indicators). Within each group, use “OR” to combine synonyms; connect groups with “AND”. The recommended keyword sets are:i.Wild mammals: “wild herbivores” OR “wild omnivores” OR “wild mammals” OR “wild animals” combined with pesticide-related terms (“pesticide”, “insecticide”, “herbicide”, “fungicide”) and contamination indicators (“residuals”, “pollution”, “concentration”).ii.Soils: “ecological soil” OR “highland soil” OR “mountain soil” OR “forest soil” OR “grassland soil” OR “hilly soil” OR “plain soil” combined with the same pesticide and contamination terms as above.iii.Wild plants: “wild plants” OR “vegetation” OR “shrubs” OR “wildflowers” OR “weeds” OR “trees” combined with the same pesticide and contamination terms. a.Use Web of Science Core Collection as the primary source. (Secondary sources such as PubMed or Scopus may be added for broader coverage.) Use Web of Science Core Collection as the primary source. (Secondary sources such as PubMed or Scopus may be added for broader coverage.) b.Time period: The search was limited to publications dated from January 1, 2000 to December 31, 2024. Time period: The search was limited to publications dated from January 1, 2000 to December 31, 2024. c.Construct keyword combinations following a three-group pattern: (environmental medium) + (pesticide terms) + (contamination indicators). Within each group, use “OR” to combine synonyms; connect groups with “AND”. The recommended keyword sets are:i.Wild mammals: “wild herbivores” OR “wild omnivores” OR “wild mammals” OR “wild animals” combined with pesticide-related terms (“pesticide”, “insecticide”, “herbicide”, “fungicide”) and contamination indicators (“residuals”, “pollution”, “concentration”).ii.Soils: “ecological soil” OR “highland soil” OR “mountain soil” OR “forest soil” OR “grassland soil” OR “hilly soil” OR “plain soil” combined with the same pesticide and contamination terms as above.iii.Wild plants: “wild plants” OR “vegetation” OR “shrubs” OR “wildflowers” OR “weeds” OR “trees” combined with the same pesticide and contamination terms. Construct keyword combinations following a three-group pattern: (environmental medium) + (pesticide terms) + (contamination indicators). Within each group, use “OR” to combine synonyms; connect groups with “AND”. The recommended keyword sets are:i.Wild mammals: “wild herbivores” OR “wild omnivores” OR “wild mammals” OR “wild animals” combined with pesticide-related terms (“pesticide”, “insecticide”, “herbicide”, “fungicide”) and contamination indicators (“residuals”, “pollution”, “concentration”).ii.Soils: “ecological soil” OR “highland soil” OR “mountain soil” OR “forest soil” OR “grassland soil” OR “hilly soil” OR “plain soil” combined with the same pesticide and contamination terms as above.iii.Wild plants: “wild plants” OR “vegetation” OR “shrubs” OR “wildflowers” OR “weeds” OR “trees” combined with the same pesticide and contamination terms. i.Wild mammals: “wild herbivores” OR “wild omnivores” OR “wild mammals” OR “wild animals” combined with pesticide-related terms (“pesticide”, “insecticide”, “herbicide”, “fungicide”) and contamination indicators (“residuals”, “pollution”, “concentration”). Wild mammals: “wild herbivores” OR “wild omnivores” OR “wild mammals” OR “wild animals” combined with pesticide-related terms (“pesticide”, “insecticide”, “herbicide”, “fungicide”) and contamination indicators (“residuals”, “pollution”, “concentration”). ii.Soils: “ecological soil” OR “highland soil” OR “mountain soil” OR “forest soil” OR “grassland soil” OR “hilly soil” OR “plain soil” combined with the same pesticide and contamination terms as above. Soils: “ecological soil” OR “highland soil” OR “mountain soil” OR “forest soil” OR “grassland soil” OR “hilly soil” OR “plain soil” combined with the same pesticide and contamination terms as above. iii.Wild plants: “wild plants” OR “vegetation” OR “shrubs” OR “wildflowers” OR “weeds” OR “trees” combined with the same pesticide and contamination terms. Wild plants: “wild plants” OR “vegetation” OR “shrubs” OR “wildflowers” OR “weeds” OR “trees” combined with the same pesticide and contamination terms. Note: Inclusion and exclusion criteria. Only include studies that report empirically measured pesticide concentrations. Exclude review articles, meta-analyses, modeling studies, and secondary analyses. Keep only samples from natural, non-degraded, non-urbanized environments (e.g., nature reserves, mountains, forests, grasslands). Discard any study from agricultural fields, orchards, or areas with direct pesticide application. 2.Export search results and retrieve full texts.a.Export citations in RIS or CSV format.b.Download full texts or supplementary data files when available.c.Use a reference manager (e.g., Mendeley, EndNote, or Zotero) to organize the retrieved records. Export search results and retrieve full texts.a.Export citations in RIS or CSV format.b.Download full texts or supplementary data files when available.c.Use a reference manager (e.g., Mendeley, EndNote, or Zotero) to organize the retrieved records. a.Export citations in RIS or CSV format. Export citations in RIS or CSV format. b.Download full texts or supplementary data files when available. Download full texts or supplementary data files when available. c.Use a reference manager (e.g., Mendeley, EndNote, or Zotero) to organize the retrieved records. Use a reference manager (e.g., Mendeley, EndNote, or Zotero) to organize the retrieved records. 3.Select studies based on eligibility criteria.a.Apply the following inclusion criteria to select eligible studies:i.Verify that sampling occurred in natural, undisturbed settings (e.g., nature reserves, forests, grasslands) — not on agricultural land or areas with direct pesticide use.ii.For wild mammals, confirm that liver tissue concentrations are reported; if multiple organs are given, prioritize liver data. Assign each mammal to a dietary guild: herbivore, carnivore, or omnivore.iii.Ensure that the study provides numeric concentration values and basic sampling metadata.iv.Confirm that the study was published in a peer-reviewed journal, with sampling years from January 1, 2000 to December 31, 2024.v.For soils, confirm that pesticide concentrations in natural soils are reported, along with numerical data and sampling metadata (location, coordinates, year).vi.For wild plants, confirm that pesticide concentrations in wild plants (herbs, shrubs, wildflowers, weeds, trees) are reported, along with numerical data and metadata.b.Apply the following exclusion criteria to all matrices: Exclude review articles, meta-analyses, pure modeling or predictive studies, and national monitoring summaries without original concentration data. Select studies based on eligibility criteria.a.Apply the following inclusion criteria to select eligible studies:i.Verify that sampling occurred in natural, undisturbed settings (e.g., nature reserves, forests, grasslands) — not on agricultural land or areas with direct pesticide use.ii.For wild mammals, confirm that liver tissue concentrations are reported; if multiple organs are given, prioritize liver data. Assign each mammal to a dietary guild: herbivore, carnivore, or omnivore.iii.Ensure that the study provides numeric concentration values and basic sampling metadata.iv.Confirm that the study was published in a peer-reviewed journal, with sampling years from January 1, 2000 to December 31, 2024.v.For soils, confirm that pesticide concentrations in natural soils are reported, along with numerical data and sampling metadata (location, coordinates, year).vi.For wild plants, confirm that pesticide concentrations in wild plants (herbs, shrubs, wildflowers, weeds, trees) are reported, along with numerical data and metadata.b.Apply the following exclusion criteria to all matrices: Exclude review articles, meta-analyses, pure modeling or predictive studies, and national monitoring summaries without original concentration data. a.Apply the following inclusion criteria to select eligible studies:i.Verify that sampling occurred in natural, undisturbed settings (e.g., nature reserves, forests, grasslands) — not on agricultural land or areas with direct pesticide use.ii.For wild mammals, confirm that liver tissue concentrations are reported; if multiple organs are given, prioritize liver data. Assign each mammal to a dietary guild: herbivore, carnivore, or omnivore.iii.Ensure that the study provides numeric concentration values and basic sampling metadata.iv.Confirm that the study was published in a peer-reviewed journal, with sampling years from January 1, 2000 to December 31, 2024.v.For soils, confirm that pesticide concentrations in natural soils are reported, along with numerical data and sampling metadata (location, coordinates, year).vi.For wild plants, confirm that pesticide concentrations in wild plants (herbs, shrubs, wildflowers, weeds, trees) are reported, along with numerical data and metadata. Apply the following inclusion criteria to select eligible studies:i.Verify that sampling occurred in natural, undisturbed settings (e.g., nature reserves, forests, grasslands) — not on agricultural land or areas with direct pesticide use.ii.For wild mammals, confirm that liver tissue concentrations are reported; if multiple organs are given, prioritize liver data. Assign each mammal to a dietary guild: herbivore, carnivore, or omnivore.iii.Ensure that the study provides numeric concentration values and basic sampling metadata.iv.Confirm that the study was published in a peer-reviewed journal, with sampling years from January 1, 2000 to December 31, 2024.v.For soils, confirm that pesticide concentrations in natural soils are reported, along with numerical data and sampling metadata (location, coordinates, year).vi.For wild plants, confirm that pesticide concentrations in wild plants (herbs, shrubs, wildflowers, weeds, trees) are reported, along with numerical data and metadata. i.Verify that sampling occurred in natural, undisturbed settings (e.g., nature reserves, forests, grasslands) — not on agricultural land or areas with direct pesticide use. Verify that sampling occurred in natural, undisturbed settings (e.g., nature reserves, forests, grasslands) — not on agricultural land or areas with direct pesticide use. ii.For wild mammals, confirm that liver tissue concentrations are reported; if multiple organs are given, prioritize liver data. Assign each mammal to a dietary guild: herbivore, carnivore, or omnivore. For wild mammals, confirm that liver tissue concentrations are reported; if multiple organs are given, prioritize liver data. Assign each mammal to a dietary guild: herbivore, carnivore, or omnivore. iii.Ensure that the study provides numeric concentration values and basic sampling metadata. Ensure that the study provides numeric concentration values and basic sampling metadata. iv.Confirm that the study was published in a peer-reviewed journal, with sampling years from January 1, 2000 to December 31, 2024. Confirm that the study was published in a peer-reviewed journal, with sampling years from January 1, 2000 to December 31, 2024. v.For soils, confirm that pesticide concentrations in natural soils are reported, along with numerical data and sampling metadata (location, coordinates, year). For soils, confirm that pesticide concentrations in natural soils are reported, along with numerical data and sampling metadata (location, coordinates, year). vi.For wild plants, confirm that pesticide concentrations in wild plants (herbs, shrubs, wildflowers, weeds, trees) are reported, along with numerical data and metadata. For wild plants, confirm that pesticide concentrations in wild plants (herbs, shrubs, wildflowers, weeds, trees) are reported, along with numerical data and metadata. b.Apply the following exclusion criteria to all matrices: Exclude review articles, meta-analyses, pure modeling or predictive studies, and national monitoring summaries without original concentration data. Apply the following exclusion criteria to all matrices: Exclude review articles, meta-analyses, pure modeling or predictive studies, and national monitoring summaries without original concentration data. Note: Diet classification of wild mammals follows standard ecological definitions: herbivores (plant-based diet), carnivores (meat-based diet), and omnivores (mixed diet). 4.Save the relevant information in a spreadsheet.a.Concentration data fields: Pesticide name, Average concentration and its unit.b.Record metadata fields including sampling site name and coordinates, sampling year, and the reported limit of detection (LOD) or limit of quantification (LOQ).c.Standardize all concentration units as follows: Convert animal liver concentrations to mg·kg−1 lipid weight (if not provided, perform conversion using species-specific lipid content data). Convert soil concentrations to mg·kg−1 dry weight. Convert wild plant concentrations to mg·kg−1 wet weight.d.For studies without a reported mean, substitute the median or the midpoint of the reported range. Replace non-detects with the corresponding LOD or LOQ to prevent downward bias.e.Assign a unique CAS number to each pesticide via the PubChem database (or CAS Common Chemistry) to ensure chemical accuracy. Save the relevant information in a spreadsheet.a.Concentration data fields: Pesticide name, Average concentration and its unit.b.Record metadata fields including sampling site name and coordinates, sampling year, and the reported limit of detection (LOD) or limit of quantification (LOQ).c.Standardize all concentration units as follows: Convert animal liver concentrations to mg·kg−1 lipid weight (if not provided, perform conversion using species-specific lipid content data). Convert soil concentrations to mg·kg−1 dry weight. Convert wild plant concentrations to mg·kg−1 wet weight.d.For studies without a reported mean, substitute the median or the midpoint of the reported range. Replace non-detects with the corresponding LOD or LOQ to prevent downward bias.e.Assign a unique CAS number to each pesticide via the PubChem database (or CAS Common Chemistry) to ensure chemical accuracy. a.Concentration data fields: Pesticide name, Average concentration and its unit. Concentration data fields: Pesticide name, Average concentration and its unit. b.Record metadata fields including sampling site name and coordinates, sampling year, and the reported limit of detection (LOD) or limit of quantification (LOQ). Record metadata fields including sampling site name and coordinates, sampling year, and the reported limit of detection (LOD) or limit of quantification (LOQ). c.Standardize all concentration units as follows: Convert animal liver concentrations to mg·kg−1 lipid weight (if not provided, perform conversion using species-specific lipid content data). Convert soil concentrations to mg·kg−1 dry weight. Convert wild plant concentrations to mg·kg−1 wet weight. Standardize all concentration units as follows: Convert animal liver concentrations to mg·kg−1 lipid weight (if not provided, perform conversion using species-specific lipid content data). Convert soil concentrations to mg·kg−1 dry weight. Convert wild plant concentrations to mg·kg−1 wet weight. d.For studies without a reported mean, substitute the median or the midpoint of the reported range. Replace non-detects with the corresponding LOD or LOQ to prevent downward bias. For studies without a reported mean, substitute the median or the midpoint of the reported range. Replace non-detects with the corresponding LOD or LOQ to prevent downward bias. e.Assign a unique CAS number to each pesticide via the PubChem database (or CAS Common Chemistry) to ensure chemical accuracy. Assign a unique CAS number to each pesticide via the PubChem database (or CAS Common Chemistry) to ensure chemical accuracy. Note: Recommended column headings for the Excel spreadsheet include: Reference Number, Country Name, Mainland/Region Name, Year of Sampling, Sample Size (or Sampling Area), Pesticide Name, CAS Number, Average Concentration, LOD, LOQ, Unit, Basis (lipid/dry/wet weight). 5.Compile the dataset and save in Excel format.a.After completing screening, data extraction, and quality control, consolidate all eligible studies into a single Excel file (.xlsx). Compile the dataset and save in Excel format.a.After completing screening, data extraction, and quality control, consolidate all eligible studies into a single Excel file (.xlsx). a.After completing screening, data extraction, and quality control, consolidate all eligible studies into a single Excel file (.xlsx). After completing screening, data extraction, and quality control, consolidate all eligible studies into a single Excel file (.xlsx). Note: In a representative application of this protocol (Chen & Li, 2026), the final dataset comprised 100 studies from 30 countries, distributed as: 15 herbivore studies, 13 carnivore studies, 10 omnivore studies, 47 soil studies, and 15 wild plant studies. Your own dataset composition will depend on the available literature. Pause point: After compiling the dataset, save the Excel file and archive all source references. The compiled dataset can be stored and used later for contamination score calculations. Resuming the protocol requires only opening the Excel file and proceeding to Step 6. Detailed instructions for the method, adapted from Chen and Li1 with permission (© 2026, Elsevier Inc.) Calculating the regional contamination scores Timing: 10–15 h In this step, a computational routine is applied to the preprocessed dataset to derive regional pollution scores, enabling quantitative comparisons of contamination levels across different spatial units (e.g., countries).17,18 6.Calculate the contamination score (S): Calculate the contamination score (S): CRITICAL: S is a unit-free comparative index, not a toxicological limit. If S > 0, the region’s contamination exceeds the worldwide mean; if S ≤ 0, it does not. Note: M denotes number of sampling sites within the country,N denotes number of pesticides detected at a given sampling site m, Cn,m denotes the adjusted concentration of pesticide n at site m: If the pesticide is detected, use the reported concentration Cn,m. If not detected, substitute with the LOD (or LOQ if LOD unavailable). logCn¯ denotes the global mean Log10 concentration for pesticide n, global central tendency calculated in step 7 of this protocol. Compute S separately for each matrix: soil (Ssoil), wild plants (Splant), and each animal guild (Sanimal) (herbivores, carnivores, omnivores). By calculating separate scores for soil, plants, and each animal group, one can follow the upward movement of pesticides from non-living media through vegetation and into wildlife. 7.Calculate the global central tendency. Calculate the global central tendency. The overall average pesticide concentration is computed across all sampling sites worldwide, weighted by the number of sites: Note: Q is the total number of global sampling points across all countries. Cn denotes the concentration of pesticide n in a specific species within a given country/region. All concentrations are log-transformed to reduce skewness. Non-detects are temporarily replaced with LOD (or LOQ) for the global mean calculation. The global mean is calculated once for each pesticide and then used in Step 6 for all countries. These scores serve only for relative ranking across countries; they are not used as inputs in the exposure calculations of Tiers 1–3. Instead, the S scores are used later in the Mapping section to generate comparative pollution maps (Step 19a). The contamination score S is a comparative indicator, not an absolute risk measure. It should be interpreted alongside the exposure and risk estimates from later steps. Maps of contamination scores (see Step 18–20) can help identify regions where pesticide levels are notably higher than the global baseline, but do not directly indicate ecological harm. A stratified strategy for estimating wildlife exposure levels Timing: 10–15 days A three-tier approach is applied. Adopt a hierarchical evidence-based approach wherein empirical tissue measurements take precedence over dietary reconstructions, and both supersede soil-to-plant modeling. Anticipate expanding uncertainty bounds as one moves from Tier 1 through Tier 3. 8.Calculate Tier 1 average daily dose from measured wildlife tissue concentrations.Note: Use this tier when empirical pesticide concentrations measured in wildlife liver are available for a given species and region. The goal is to back-calculate the average daily dose (ADD) by combining a PBK approach with species-specific biotransfer factors (BTFs).a.Retrieve the BTF for the target speciesi.BTFs are species- and tissue-specific. For this study, BTFs for liver are taken from the supplemental information or from reported in PBK models for wildlife.ii.If a BTF is not directly available for the exact species, use allometric scaling or the closest related species, and document the uncertainty.Note:BTF is defined as the steady-state tissue concentration (mg·kg−1) divided by the corresponding daily intake per kg of diet. Its unit is essentially (mg·kg−1 tissue) per (mg·kg−1 intake).b.Obtain the body weight (BW) of the species.i.Use species-average body weight in kg. Values are provided in the supplemental information can be sourced from standard wildlife biology references.c.Calculate the average daily dose (ADD)ADDspecies=Cspecies,tissueBTFspecies,tissue×BWspeciesADDherbivores,1=Cherbivores,tissueBTFherbivores,tissue×BWherbivoresADDcarnivores,1=Ccarnivores,tissueBTFcarnivores,tissue×BWcarnivoresADDomnivores,1=Comnivores,tissueBTFomnivores,tissue×BWomnivoresNote:Cspecies,tissue(mg kg−1) denotes measured pesticide concentration in the liver (steady-state, lipid-weight basis); BTFspecies,tissuedenotes biotransfer factor for liver of mammals; BWspecies (kg) denotes body weight of species. Apply the general ADD equation separately to each feeding group as shown above.d.Document all input and output values.i.Record the source of the concentration data (references, country, species), the BTF value and its source, the body weight used, and the resulting ADD value. Calculate Tier 1 average daily dose from measured wildlife tissue concentrations.Note: Use this tier when empirical pesticide concentrations measured in wildlife liver are available for a given species and region. The goal is to back-calculate the average daily dose (ADD) by combining a PBK approach with species-specific biotransfer factors (BTFs).a.Retrieve the BTF for the target speciesi.BTFs are species- and tissue-specific. For this study, BTFs for liver are taken from the supplemental information or from reported in PBK models for wildlife.ii.If a BTF is not directly available for the exact species, use allometric scaling or the closest related species, and document the uncertainty.Note:BTF is defined as the steady-state tissue concentration (mg·kg−1) divided by the corresponding daily intake per kg of diet. Its unit is essentially (mg·kg−1 tissue) per (mg·kg−1 intake).b.Obtain the body weight (BW) of the species.i.Use species-average body weight in kg. Values are provided in the supplemental information can be sourced from standard wildlife biology references.c.Calculate the average daily dose (ADD)ADDspecies=Cspecies,tissueBTFspecies,tissue×BWspeciesADDherbivores,1=Cherbivores,tissueBTFherbivores,tissue×BWherbivoresADDcarnivores,1=Ccarnivores,tissueBTFcarnivores,tissue×BWcarnivoresADDomnivores,1=Comnivores,tissueBTFomnivores,tissue×BWomnivoresNote:Cspecies,tissue(mg kg−1) denotes measured pesticide concentration in the liver (steady-state, lipid-weight basis); BTFspecies,tissuedenotes biotransfer factor for liver of mammals; BWspecies (kg) denotes body weight of species. Apply the general ADD equation separately to each feeding group as shown above.d.Document all input and output values.i.Record the source of the concentration data (references, country, species), the BTF value and its source, the body weight used, and the resulting ADD value. Note: Use this tier when empirical pesticide concentrations measured in wildlife liver are available for a given species and region. The goal is to back-calculate the average daily dose (ADD) by combining a PBK approach with species-specific biotransfer factors (BTFs). a.Retrieve the BTF for the target speciesi.BTFs are species- and tissue-specific. For this study, BTFs for liver are taken from the supplemental information or from reported in PBK models for wildlife.ii.If a BTF is not directly available for the exact species, use allometric scaling or the closest related species, and document the uncertainty.Note:BTF is defined as the steady-state tissue concentration (mg·kg−1) divided by the corresponding daily intake per kg of diet. Its unit is essentially (mg·kg−1 tissue) per (mg·kg−1 intake). Retrieve the BTF for the target speciesi.BTFs are species- and tissue-specific. For this study, BTFs for liver are taken from the supplemental information or from reported in PBK models for wildlife.ii.If a BTF is not directly available for the exact species, use allometric scaling or the closest related species, and document the uncertainty.Note:BTF is defined as the steady-state tissue concentration (mg·kg−1) divided by the corresponding daily intake per kg of diet. Its unit is essentially (mg·kg−1 tissue) per (mg·kg−1 intake). i.BTFs are species- and tissue-specific. For this study, BTFs for liver are taken from the supplemental information or from reported in PBK models for wildlife. BTFs are species- and tissue-specific. For this study, BTFs for liver are taken from the supplemental information or from reported in PBK models for wildlife. ii.If a BTF is not directly available for the exact species, use allometric scaling or the closest related species, and document the uncertainty.Note:BTF is defined as the steady-state tissue concentration (mg·kg−1) divided by the corresponding daily intake per kg of diet. Its unit is essentially (mg·kg−1 tissue) per (mg·kg−1 intake). If a BTF is not directly available for the exact species, use allometric scaling or the closest related species, and document the uncertainty.Note:BTF is defined as the steady-state tissue concentration (mg·kg−1) divided by the corresponding daily intake per kg of diet. Its unit is essentially (mg·kg−1 tissue) per (mg·kg−1 intake). Note:BTF is defined as the steady-state tissue concentration (mg·kg−1) divided by the corresponding daily intake per kg of diet. Its unit is essentially (mg·kg−1 tissue) per (mg·kg−1 intake). b.Obtain the body weight (BW) of the species.i.Use species-average body weight in kg. Values are provided in the supplemental information can be sourced from standard wildlife biology references. Obtain the body weight (BW) of the species.i.Use species-average body weight in kg. Values are provided in the supplemental information can be sourced from standard wildlife biology references. i.Use species-average body weight in kg. Values are provided in the supplemental information can be sourced from standard wildlife biology references. Use species-average body weight in kg. Values are provided in the supplemental information can be sourced from standard wildlife biology references. c.Calculate the average daily dose (ADD)ADDspecies=Cspecies,tissueBTFspecies,tissue×BWspeciesADDherbivores,1=Cherbivores,tissueBTFherbivores,tissue×BWherbivoresADDcarnivores,1=Ccarnivores,tissueBTFcarnivores,tissue×BWcarnivoresADDomnivores,1=Comnivores,tissueBTFomnivores,tissue×BWomnivoresNote:Cspecies,tissue(mg kg−1) denotes measured pesticide concentration in the liver (steady-state, lipid-weight basis); BTFspecies,tissuedenotes biotransfer factor for liver of mammals; BWspecies (kg) denotes body weight of species. Apply the general ADD equation separately to each feeding group as shown above. Calculate the average daily dose (ADD)ADDspecies=Cspecies,tissueBTFspecies,tissue×BWspeciesADDherbivores,1=Cherbivores,tissueBTFherbivores,tissue×BWherbivoresADDcarnivores,1=Ccarnivores,tissueBTFcarnivores,tissue×BWcarnivoresADDomnivores,1=Comnivores,tissueBTFomnivores,tissue×BWomnivoresNote:Cspecies,tissue(mg kg−1) denotes measured pesticide concentration in the liver (steady-state, lipid-weight basis); BTFspecies,tissuedenotes biotransfer factor for liver of mammals; BWspecies (kg) denotes body weight of species. Apply the general ADD equation separately to each feeding group as shown above. Note:Cspecies,tissue(mg kg−1) denotes measured pesticide concentration in the liver (steady-state, lipid-weight basis); BTFspecies,tissuedenotes biotransfer factor for liver of mammals; BWspecies (kg) denotes body weight of species. Apply the general ADD equation separately to each feeding group as shown above. d.Document all input and output values.i.Record the source of the concentration data (references, country, species), the BTF value and its source, the body weight used, and the resulting ADD value. Document all input and output values.i.Record the source of the concentration data (references, country, species), the BTF value and its source, the body weight used, and the resulting ADD value. i.Record the source of the concentration data (references, country, species), the BTF value and its source, the body weight used, and the resulting ADD value. Record the source of the concentration data (references, country, species), the BTF value and its source, the body weight used, and the resulting ADD value. 9.Calculate Tier 2 average daily dose from measured plant concentrations.Note: Use this tier when liver concentration data are lacking for wildlife in a country/region, but measured pesticide concentrations in wild plants (leaves, stems, grasses, roots, seeds – excluding pollen and nectar) are available. Exposure doses are then estimated via food chain transfer from plants to herbivores, and then to carnivores/omnivores.a.Calculate ADD for herbivores directly from plant concentrations.ADDherbivores,2=Cwildplant·IRherbivoresBWherbivoresNote:Cwildplant (mg kg−1) denotes measured pesticide concentration in wild plants (wet weight); IRherbivores(kg day−1) denotes daily intake rate of the herbivore; BWherbivores(kg) denotes body weight of the herbivore.b.Predict herbivore tissue concentrations (muscle, fat).Cherbivores,tissue=BTFherbivores,tissue×IRherbivores×CwildplantNote:BTFherbivores,tissueis the biotransfer factor for that tissue (obtained from supplementary data). Carnivores and omnivores generally consume specific tissues of herbivores, most commonly muscle and fat. Accordingly, pesticide concentrations in both muscle and fat are required to calculate a weighted average concentration.c.Calculate the weighted average concentration in herbivores.C¯herbivores=∑i=1n(Cherbivores,tissue×wtissue)∑i=1nwtissueNote:C¯herbivores (mg kg−1) denotes weighted average of concentrations in herbivores. wtissue denotes weighting factor for a certain tissue within the body; n denotes number of data points. The herbivore’s weighted concentration then allows calculation of the carnivore’s exposure dose (ADDcarnivores,2).d.Calculate ADD for carnivores that consume herbivores.ADDcarnivores,2=C¯herbivores×IRcarnivoresBWcarnivorese.Calculate the weighted average concentration in carnivores.C¯carnivores=∑i=1n(Ccarnivores,tissue×wtissue)∑i=1nwtissueNote:C¯carnivores (mg kg−1) denotes weighted average of concentrations in carnivores. wtissue denotes tissue weighting factor; Ccarnivores,tissue denotes concentration of tissue (muscle and fat).f.Calculate ADD for omnivores.ADDomnivores,2=(fp·Cplant+fh·C¯herbivores+fc·C¯carnivores)·IRomnivoresBWomnivoresNote:fp, fh, fc denotes dietary proportions of plant-based food, herbivore prey, and carnivore prey, respectively (each between 0 and 1; sum ≤ 1). Obtain these from ecological literature or the supplementary tables. IRomnivores(kg day−1) denotes daily food intake of omnivores. BWomnivores(kg) denotes body weight of omnivorous animals. Calculate Tier 2 average daily dose from measured plant concentrations.Note: Use this tier when liver concentration data are lacking for wildlife in a country/region, but measured pesticide concentrations in wild plants (leaves, stems, grasses, roots, seeds – excluding pollen and nectar) are available. Exposure doses are then estimated via food chain transfer from plants to herbivores, and then to carnivores/omnivores.a.Calculate ADD for herbivores directly from plant concentrations.ADDherbivores,2=Cwildplant·IRherbivoresBWherbivoresNote:Cwildplant (mg kg−1) denotes measured pesticide concentration in wild plants (wet weight); IRherbivores(kg day−1) denotes daily intake rate of the herbivore; BWherbivores(kg) denotes body weight of the herbivore.b.Predict herbivore tissue concentrations (muscle, fat).Cherbivores,tissue=BTFherbivores,tissue×IRherbivores×CwildplantNote:BTFherbivores,tissueis the biotransfer factor for that tissue (obtained from supplementary data). Carnivores and omnivores generally consume specific tissues of herbivores, most commonly muscle and fat. Accordingly, pesticide concentrations in both muscle and fat are required to calculate a weighted average concentration.c.Calculate the weighted average concentration in herbivores.C¯herbivores=∑i=1n(Cherbivores,tissue×wtissue)∑i=1nwtissueNote:C¯herbivores (mg kg−1) denotes weighted average of concentrations in herbivores. wtissue denotes weighting factor for a certain tissue within the body; n denotes number of data points. The herbivore’s weighted concentration then allows calculation of the carnivore’s exposure dose (ADDcarnivores,2).d.Calculate ADD for carnivores that consume herbivores.ADDcarnivores,2=C¯herbivores×IRcarnivoresBWcarnivorese.Calculate the weighted average concentration in carnivores.C¯carnivores=∑i=1n(Ccarnivores,tissue×wtissue)∑i=1nwtissueNote:C¯carnivores (mg kg−1) denotes weighted average of concentrations in carnivores. wtissue denotes tissue weighting factor; Ccarnivores,tissue denotes concentration of tissue (muscle and fat).f.Calculate ADD for omnivores.ADDomnivores,2=(fp·Cplant+fh·C¯herbivores+fc·C¯carnivores)·IRomnivoresBWomnivoresNote:fp, fh, fc denotes dietary proportions of plant-based food, herbivore prey, and carnivore prey, respectively (each between 0 and 1; sum ≤ 1). Obtain these from ecological literature or the supplementary tables. IRomnivores(kg day−1) denotes daily food intake of omnivores. BWomnivores(kg) denotes body weight of omnivorous animals. Note: Use this tier when liver concentration data are lacking for wildlife in a country/region, but measured pesticide concentrations in wild plants (leaves, stems, grasses, roots, seeds – excluding pollen and nectar) are available. Exposure doses are then estimated via food chain transfer from plants to herbivores, and then to carnivores/omnivores. a.Calculate ADD for herbivores directly from plant concentrations.ADDherbivores,2=Cwildplant·IRherbivoresBWherbivoresNote:Cwildplant (mg kg−1) denotes measured pesticide concentration in wild plants (wet weight); IRherbivores(kg day−1) denotes daily intake rate of the herbivore; BWherbivores(kg) denotes body weight of the herbivore. Calculate ADD for herbivores directly from plant concentrations.ADDherbivores,2=Cwildplant·IRherbivoresBWherbivoresNote:Cwildplant (mg kg−1) denotes measured pesticide concentration in wild plants (wet weight); IRherbivores(kg day−1) denotes daily intake rate of the herbivore; BWherbivores(kg) denotes body weight of the herbivore. Note:Cwildplant (mg kg−1) denotes measured pesticide concentration in wild plants (wet weight); IRherbivores(kg day−1) denotes daily intake rate of the herbivore; BWherbivores(kg) denotes body weight of the herbivore. b.Predict herbivore tissue concentrations (muscle, fat).Cherbivores,tissue=BTFherbivores,tissue×IRherbivores×CwildplantNote:BTFherbivores,tissueis the biotransfer factor for that tissue (obtained from supplementary data). Carnivores and omnivores generally consume specific tissues of herbivores, most commonly muscle and fat. Accordingly, pesticide concentrations in both muscle and fat are required to calculate a weighted average concentration. Predict herbivore tissue concentrations (muscle, fat).Cherbivores,tissue=BTFherbivores,tissue×IRherbivores×CwildplantNote:BTFherbivores,tissueis the biotransfer factor for that tissue (obtained from supplementary data). Carnivores and omnivores generally consume specific tissues of herbivores, most commonly muscle and fat. Accordingly, pesticide concentrations in both muscle and fat are required to calculate a weighted average concentration. Note:BTFherbivores,tissueis the biotransfer factor for that tissue (obtained from supplementary data). Carnivores and omnivores generally consume specific tissues of herbivores, most commonly muscle and fat. Accordingly, pesticide concentrations in both muscle and fat are required to calculate a weighted average concentration. c.Calculate the weighted average concentration in herbivores.C¯herbivores=∑i=1n(Cherbivores,tissue×wtissue)∑i=1nwtissueNote:C¯herbivores (mg kg−1) denotes weighted average of concentrations in herbivores. wtissue denotes weighting factor for a certain tissue within the body; n denotes number of data points. The herbivore’s weighted concentration then allows calculation of the carnivore’s exposure dose (ADDcarnivores,2). Calculate the weighted average concentration in herbivores.C¯herbivores=∑i=1n(Cherbivores,tissue×wtissue)∑i=1nwtissueNote:C¯herbivores (mg kg−1) denotes weighted average of concentrations in herbivores. wtissue denotes weighting factor for a certain tissue within the body; n denotes number of data points. The herbivore’s weighted concentration then allows calculation of the carnivore’s exposure dose (ADDcarnivores,2). Note:C¯herbivores (mg kg−1) denotes weighted average of concentrations in herbivores. wtissue denotes weighting factor for a certain tissue within the body; n denotes number of data points. The herbivore’s weighted concentration then allows calculation of the carnivore’s exposure dose (ADDcarnivores,2). d.Calculate ADD for carnivores that consume herbivores.ADDcarnivores,2=C¯herbivores×IRcarnivoresBWcarnivores Calculate ADD for carnivores that consume herbivores.ADDcarnivores,2=C¯herbivores×IRcarnivoresBWcarnivores e.Calculate the weighted average concentration in carnivores.C¯carnivores=∑i=1n(Ccarnivores,tissue×wtissue)∑i=1nwtissueNote:C¯carnivores (mg kg−1) denotes weighted average of concentrations in carnivores. wtissue denotes tissue weighting factor; Ccarnivores,tissue denotes concentration of tissue (muscle and fat). Calculate the weighted average concentration in carnivores.C¯carnivores=∑i=1n(Ccarnivores,tissue×wtissue)∑i=1nwtissueNote:C¯carnivores (mg kg−1) denotes weighted average of concentrations in carnivores. wtissue denotes tissue weighting factor; Ccarnivores,tissue denotes concentration of tissue (muscle and fat). Note:C¯carnivores (mg kg−1) denotes weighted average of concentrations in carnivores. wtissue denotes tissue weighting factor; Ccarnivores,tissue denotes concentration of tissue (muscle and fat). f.Calculate ADD for omnivores.ADDomnivores,2=(fp·Cplant+fh·C¯herbivores+fc·C¯carnivores)·IRomnivoresBWomnivoresNote:fp, fh, fc denotes dietary proportions of plant-based food, herbivore prey, and carnivore prey, respectively (each between 0 and 1; sum ≤ 1). Obtain these from ecological literature or the supplementary tables. IRomnivores(kg day−1) denotes daily food intake of omnivores. BWomnivores(kg) denotes body weight of omnivorous animals. Calculate ADD for omnivores.ADDomnivores,2=(fp·Cplant+fh·C¯herbivores+fc·C¯carnivores)·IRomnivoresBWomnivoresNote:fp, fh, fc denotes dietary proportions of plant-based food, herbivore prey, and carnivore prey, respectively (each between 0 and 1; sum ≤ 1). Obtain these from ecological literature or the supplementary tables. IRomnivores(kg day−1) denotes daily food intake of omnivores. BWomnivores(kg) denotes body weight of omnivorous animals. Note:fp, fh, fc denotes dietary proportions of plant-based food, herbivore prey, and carnivore prey, respectively (each between 0 and 1; sum ≤ 1). Obtain these from ecological literature or the supplementary tables. IRomnivores(kg day−1) denotes daily food intake of omnivores. BWomnivores(kg) denotes body weight of omnivorous animals. 10.Calculate Tier 3 average daily dose from soil concentrations via plant bioaccumulation modeling.Note: Reserve for data-impoverished contexts where only edaphic measurements are at hand. The workflow necessitates an upstream modeling step—phyto-accumulation simulation—prior to executing the Tier 2 trophic cascade.a.Calculate the plant bioaccumulation factor (BAF) using a mechanistic soil-plant transfer model.19i.Calculate the leaf-specific transpiration flux QH(TAir, RHAir) (L·d−1).QH(TAir,RHAir)=ρAircpLAIρWaterΔHVapγ(sΔT+VPDRS+RAir)(86400sd)ii.Calculate the slope of saturated vapor pressure versus temperature (s) (Pa·K−1):s=2.50×106×e17.2(TAir−273.15)TAir−35.85(TAir−35.85)2iii.Calculate the vapor pressure deficit (VPD) (Pa).VPD=610.78(1−RHAir100)e17.2(TAir−273.15)TAir−35.85iv.Calculate the leaf-air temperature difference (ΔT):ΔT=Tleaf−TAir=31.145−0.1TAirv.Calculate the soil-water partition coefficient (kd) (mg·kg−1mg·L−1).kd=Koc×focvi.Calculate the leaf-air distribution coefficient (KLA) (mgkg−1mgL−1).KLA=WL+LL(Kow)ccρocKAWvii.Compute the BAFplantH(TAir,RHAir) (dimensionless) using the values from i–vi:BAFplantH(TAir,RHAir)=QH(TAir,RHAir)×1kdALH·g·1000KLA+λLH·MLHNote: The BAF quantifies the equilibrium ratio of foliar to edaphic concentrations. The numerator captures transpiration-mediated uptake; the denominator aggregates elimination pathways (stomatal diffusion, photolytic degradation, metabolic transformation, and growth dilution). Detailed definitions of all parameters are: ALH (m2) denotes leaf surface area. g denotes vapor conduction; g denotes vapor conductance; KLA (mgkg−1mgL−1) denotes leaf-air distribution coefficient; λLH (d−1) denotes total loss rate of pesticides within leaves, encompassing photodegradation, plant metabolism, and dilution due to plant growth; MLH (kg) denotes mass of a single leaf. This model assumes steady-state conditions, with pesticide movement from soil to leaves driven mainly by transpiration and limited by diffusion; VPD (Pa) denotes vapour pressure deficit. (K) denotes temperature difference between the leaf and the air, assumed to be due to the continuous transport of water; kd (mg·kg−1mg·L−1) denotes soil-water partition coefficient, governing the distribution of analyte between solid and aqueous soil phases. WL (L⋅kg−1), LL (kg⋅kg−1) denote foliar water and lipid contents, respectively; cc (dimensionless) denote empirical correction factor; KOW (dimensionless), KAW (dimensionless) denote octanol-water partition coefficient and the air-water partition coefficient of pesticides. If specific meteorological or physiological parameters are unavailable, use the default values listed in Table S2. This should be noted in the uncertainty analysis.b.Calculate the simulated plant concentration.CPlantmodeldry=Csoil×BAFplantH(TAir,RHAir)CPlantmodelwet=CPlantmodeldry×(1−moisturecontent)Note: The soil concentration(Csoil) (mg·kg−1dry weight): measured pesticide concentration in soil. BAFplantH(TAir,RHAir) denotes the plant bioaccumulation factor (dimensionless) calculated in Step 10a. The modeled botanical concentration initially emerges on a dry-mass basis; convert to fresh-weight units for Tier 2 dietary calculations using species-appropriate moisture content. Because herbivores and omnivores consume fresh plant material, convert the dry-weight concentration to wet-weight concentration using species-specific moisture content. Use a moisture content of 0.8 for herbaceous plants and 0.5 for woody plants as defaults; adjust if species-specific data are available.c.Use the simulated plant concentration as input for Tier 2.Substitute the Cplant_model value obtained in the previous step (corrected for moisture content to fresh weight) into Step 9 (Tier 2), then complete all steps in Tier 2 to obtain the ADD results for herbivorous, carnivorous and omnivorous animals.Substitute CPlantmodelwet for CPlant in Step 9. Then follow all sub-steps of Tier 2 to obtain ADD estimates for herbivores, carnivores, and omnivores.ADDherbivores,3=CPlantmodelwet·IRherbivoresBWherbivoresADDcarnivores,3=C¯herbivores,PM×IRcarnivoresBWcarnivoresADDomnivores,3=(fp·CPlantmodelwet+fh·C¯herbivores,PM+fc·C¯carnivores,PM)×IRomnivoresBWomnivoresWhere C¯herbivores,PM and C¯carnivores,PM are the weighted average concentrations derived from the simulated plant concentrations (as calculated in Steps 9c and 9e).Note: Model parameter availability – The Trapp model requires multiple environmental and plant physiological parameters. If certain parameters are unavailable, default values provided in Table S2 may be used, and this should be reflected in the uncertainty analysis.Note: Spatial heterogeneity – Soil sampling points may be unevenly distributed. It is recommended to apply Tier 3 within smaller spatial units (e.g., watersheds, provinces/states) to reduce aggregation bias.Note: Tier 3 is the most indirect approach among the three-tiered strategy. Sources of error include measurement errors in soil concentrations, uncertainties in model parameters, and accumulated errors from the subsequent Step 9. A sensitivity analysis (e.g., Monte Carlo simulation) and uncertainty propagation are strongly recommended. Calculate Tier 3 average daily dose from soil concentrations via plant bioaccumulation modeling.Note: Reserve for data-impoverished contexts where only edaphic measurements are at hand. The workflow necessitates an upstream modeling step—phyto-accumulation simulation—prior to executing the Tier 2 trophic cascade.a.Calculate the plant bioaccumulation factor (BAF) using a mechanistic soil-plant transfer model.19i.Calculate the leaf-specific transpiration flux QH(TAir, RHAir) (L·d−1).QH(TAir,RHAir)=ρAircpLAIρWaterΔHVapγ(sΔT+VPDRS+RAir)(86400sd)ii.Calculate the slope of saturated vapor pressure versus temperature (s) (Pa·K−1):s=2.50×106×e17.2(TAir−273.15)TAir−35.85(TAir−35.85)2iii.Calculate the vapor pressure deficit (VPD) (Pa).VPD=610.78(1−RHAir100)e17.2(TAir−273.15)TAir−35.85iv.Calculate the leaf-air temperature difference (ΔT):ΔT=Tleaf−TAir=31.145−0.1TAirv.Calculate the soil-water partition coefficient (kd) (mg·kg−1mg·L−1).kd=Koc×focvi.Calculate the leaf-air distribution coefficient (KLA) (mgkg−1mgL−1).KLA=WL+LL(Kow)ccρocKAWvii.Compute the BAFplantH(TAir,RHAir) (dimensionless) using the values from i–vi:BAFplantH(TAir,RHAir)=QH(TAir,RHAir)×1kdALH·g·1000KLA+λLH·MLHNote: The BAF quantifies the equilibrium ratio of foliar to edaphic concentrations. The numerator captures transpiration-mediated uptake; the denominator aggregates elimination pathways (stomatal diffusion, photolytic degradation, metabolic transformation, and growth dilution). Detailed definitions of all parameters are: ALH (m2) denotes leaf surface area. g denotes vapor conduction; g denotes vapor conductance; KLA (mgkg−1mgL−1) denotes leaf-air distribution coefficient; λLH (d−1) denotes total loss rate of pesticides within leaves, encompassing photodegradation, plant metabolism, and dilution due to plant growth; MLH (kg) denotes mass of a single leaf. This model assumes steady-state conditions, with pesticide movement from soil to leaves driven mainly by transpiration and limited by diffusion; VPD (Pa) denotes vapour pressure deficit. (K) denotes temperature difference between the leaf and the air, assumed to be due to the continuous transport of water; kd (mg·kg−1mg·L−1) denotes soil-water partition coefficient, governing the distribution of analyte between solid and aqueous soil phases. WL (L⋅kg−1), LL (kg⋅kg−1) denote foliar water and lipid contents, respectively; cc (dimensionless) denote empirical correction factor; KOW (dimensionless), KAW (dimensionless) denote octanol-water partition coefficient and the air-water partition coefficient of pesticides. If specific meteorological or physiological parameters are unavailable, use the default values listed in Table S2. This should be noted in the uncertainty analysis.b.Calculate the simulated plant concentration.CPlantmodeldry=Csoil×BAFplantH(TAir,RHAir)CPlantmodelwet=CPlantmodeldry×(1−moisturecontent)Note: The soil concentration(Csoil) (mg·kg−1dry weight): measured pesticide concentration in soil. BAFplantH(TAir,RHAir) denotes the plant bioaccumulation factor (dimensionless) calculated in Step 10a. The modeled botanical concentration initially emerges on a dry-mass basis; convert to fresh-weight units for Tier 2 dietary calculations using species-appropriate moisture content. Because herbivores and omnivores consume fresh plant material, convert the dry-weight concentration to wet-weight concentration using species-specific moisture content. Use a moisture content of 0.8 for herbaceous plants and 0.5 for woody plants as defaults; adjust if species-specific data are available.c.Use the simulated plant concentration as input for Tier 2.Substitute the Cplant_model value obtained in the previous step (corrected for moisture content to fresh weight) into Step 9 (Tier 2), then complete all steps in Tier 2 to obtain the ADD results for herbivorous, carnivorous and omnivorous animals.Substitute CPlantmodelwet for CPlant in Step 9. Then follow all sub-steps of Tier 2 to obtain ADD estimates for herbivores, carnivores, and omnivores.ADDherbivores,3=CPlantmodelwet·IRherbivoresBWherbivoresADDcarnivores,3=C¯herbivores,PM×IRcarnivoresBWcarnivoresADDomnivores,3=(fp·CPlantmodelwet+fh·C¯herbivores,PM+fc·C¯carnivores,PM)×IRomnivoresBWomnivoresWhere C¯herbivores,PM and C¯carnivores,PM are the weighted average concentrations derived from the simulated plant concentrations (as calculated in Steps 9c and 9e).Note: Model parameter availability – The Trapp model requires multiple environmental and plant physiological parameters. If certain parameters are unavailable, default values provided in Table S2 may be used, and this should be reflected in the uncertainty analysis.Note: Spatial heterogeneity – Soil sampling points may be unevenly distributed. It is recommended to apply Tier 3 within smaller spatial units (e.g., watersheds, provinces/states) to reduce aggregation bias.Note: Tier 3 is the most indirect approach among the three-tiered strategy. Sources of error include measurement errors in soil concentrations, uncertainties in model parameters, and accumulated errors from the subsequent Step 9. A sensitivity analysis (e.g., Monte Carlo simulation) and uncertainty propagation are strongly recommended. Note: Reserve for data-impoverished contexts where only edaphic measurements are at hand. The workflow necessitates an upstream modeling step—phyto-accumulation simulation—prior to executing the Tier 2 trophic cascade. a.Calculate the plant bioaccumulation factor (BAF) using a mechanistic soil-plant transfer model.19i.Calculate the leaf-specific transpiration flux QH(TAir, RHAir) (L·d−1).QH(TAir,RHAir)=ρAircpLAIρWaterΔHVapγ(sΔT+VPDRS+RAir)(86400sd)ii.Calculate the slope of saturated vapor pressure versus temperature (s) (Pa·K−1):s=2.50×106×e17.2(TAir−273.15)TAir−35.85(TAir−35.85)2iii.Calculate the vapor pressure deficit (VPD) (Pa).VPD=610.78(1−RHAir100)e17.2(TAir−273.15)TAir−35.85iv.Calculate the leaf-air temperature difference (ΔT):ΔT=Tleaf−TAir=31.145−0.1TAirv.Calculate the soil-water partition coefficient (kd) (mg·kg−1mg·L−1).kd=Koc×focvi.Calculate the leaf-air distribution coefficient (KLA) (mgkg−1mgL−1).KLA=WL+LL(Kow)ccρocKAWvii.Compute the BAFplantH(TAir,RHAir) (dimensionless) using the values from i–vi:BAFplantH(TAir,RHAir)=QH(TAir,RHAir)×1kdALH·g·1000KLA+λLH·MLHNote: The BAF quantifies the equilibrium ratio of foliar to edaphic concentrations. The numerator captures transpiration-mediated uptake; the denominator aggregates elimination pathways (stomatal diffusion, photolytic degradation, metabolic transformation, and growth dilution). Detailed definitions of all parameters are: ALH (m2) denotes leaf surface area. g denotes vapor conduction; g denotes vapor conductance; KLA (mgkg−1mgL−1) denotes leaf-air distribution coefficient; λLH (d−1) denotes total loss rate of pesticides within leaves, encompassing photodegradation, plant metabolism, and dilution due to plant growth; MLH (kg) denotes mass of a single leaf. This model assumes steady-state conditions, with pesticide movement from soil to leaves driven mainly by transpiration and limited by diffusion; VPD (Pa) denotes vapour pressure deficit. (K) denotes temperature difference between the leaf and the air, assumed to be due to the continuous transport of water; kd (mg·kg−1mg·L−1) denotes soil-water partition coefficient, governing the distribution of analyte between solid and aqueous soil phases. WL (L⋅kg−1), LL (kg⋅kg−1) denote foliar water and lipid contents, respectively; cc (dimensionless) denote empirical correction factor; KOW (dimensionless), KAW (dimensionless) denote octanol-water partition coefficient and the air-water partition coefficient of pesticides. If specific meteorological or physiological parameters are unavailable, use the default values listed in Table S2. This should be noted in the uncertainty analysis. Calculate the plant bioaccumulation factor (BAF) using a mechanistic soil-plant transfer model.19i.Calculate the leaf-specific transpiration flux QH(TAir, RHAir) (L·d−1).QH(TAir,RHAir)=ρAircpLAIρWaterΔHVapγ(sΔT+VPDRS+RAir)(86400sd)ii.Calculate the slope of saturated vapor pressure versus temperature (s) (Pa·K−1):s=2.50×106×e17.2(TAir−273.15)TAir−35.85(TAir−35.85)2iii.Calculate the vapor pressure deficit (VPD) (Pa).VPD=610.78(1−RHAir100)e17.2(TAir−273.15)TAir−35.85iv.Calculate the leaf-air temperature difference (ΔT):ΔT=Tleaf−TAir=31.145−0.1TAirv.Calculate the soil-water partition coefficient (kd) (mg·kg−1mg·L−1).kd=Koc×focvi.Calculate the leaf-air distribution coefficient (KLA) (mgkg−1mgL−1).KLA=WL+LL(Kow)ccρocKAWvii.Compute the BAFplantH(TAir,RHAir) (dimensionless) using the values from i–vi:BAFplantH(TAir,RHAir)=QH(TAir,RHAir)×1kdALH·g·1000KLA+λLH·MLHNote: The BAF quantifies the equilibrium ratio of foliar to edaphic concentrations. The numerator captures transpiration-mediated uptake; the denominator aggregates elimination pathways (stomatal diffusion, photolytic degradation, metabolic transformation, and growth dilution). Detailed definitions of all parameters are: ALH (m2) denotes leaf surface area. g denotes vapor conduction; g denotes vapor conductance; KLA (mgkg−1mgL−1) denotes leaf-air distribution coefficient; λLH (d−1) denotes total loss rate of pesticides within leaves, encompassing photodegradation, plant metabolism, and dilution due to plant growth; MLH (kg) denotes mass of a single leaf. This model assumes steady-state conditions, with pesticide movement from soil to leaves driven mainly by transpiration and limited by diffusion; VPD (Pa) denotes vapour pressure deficit. (K) denotes temperature difference between the leaf and the air, assumed to be due to the continuous transport of water; kd (mg·kg−1mg·L−1) denotes soil-water partition coefficient, governing the distribution of analyte between solid and aqueous soil phases. WL (L⋅kg−1), LL (kg⋅kg−1) denote foliar water and lipid contents, respectively; cc (dimensionless) denote empirical correction factor; KOW (dimensionless), KAW (dimensionless) denote octanol-water partition coefficient and the air-water partition coefficient of pesticides. If specific meteorological or physiological parameters are unavailable, use the default values listed in Table S2. This should be noted in the uncertainty analysis. i.Calculate the leaf-specific transpiration flux QH(TAir, RHAir) (L·d−1).QH(TAir,RHAir)=ρAircpLAIρWaterΔHVapγ(sΔT+VPDRS+RAir)(86400sd) Calculate the leaf-specific transpiration flux QH(TAir, RHAir) (L·d−1).QH(TAir,RHAir)=ρAircpLAIρWaterΔHVapγ(sΔT+VPDRS+RAir)(86400sd) ii.Calculate the slope of saturated vapor pressure versus temperature (s) (Pa·K−1):s=2.50×106×e17.2(TAir−273.15)TAir−35.85(TAir−35.85)2 Calculate the slope of saturated vapor pressure versus temperature (s) (Pa·K−1):s=2.50×106×e17.2(TAir−273.15)TAir−35.85(TAir−35.85)2 iii.Calculate the vapor pressure deficit (VPD) (Pa).VPD=610.78(1−RHAir100)e17.2(TAir−273.15)TAir−35.85 Calculate the vapor pressure deficit (VPD) (Pa).VPD=610.78(1−RHAir100)e17.2(TAir−273.15)TAir−35.85 iv.Calculate the leaf-air temperature difference (ΔT):ΔT=Tleaf−TAir=31.145−0.1TAir Calculate the leaf-air temperature difference (ΔT):ΔT=Tleaf−TAir=31.145−0.1TAir v.Calculate the soil-water partition coefficient (kd) (mg·kg−1mg·L−1).kd=Koc×foc Calculate the soil-water partition coefficient (kd) (mg·kg−1mg·L−1).kd=Koc×foc vi.Calculate the leaf-air distribution coefficient (KLA) (mgkg−1mgL−1).KLA=WL+LL(Kow)ccρocKAW Calculate the leaf-air distribution coefficient (KLA) (mgkg−1mgL−1).KLA=WL+LL(Kow)ccρocKAW vii.Compute the BAFplantH(TAir,RHAir) (dimensionless) using the values from i–vi:BAFplantH(TAir,RHAir)=QH(TAir,RHAir)×1kdALH·g·1000KLA+λLH·MLHNote: The BAF quantifies the equilibrium ratio of foliar to edaphic concentrations. The numerator captures transpiration-mediated uptake; the denominator aggregates elimination pathways (stomatal diffusion, photolytic degradation, metabolic transformation, and growth dilution). Detailed definitions of all parameters are: ALH (m2) denotes leaf surface area. g denotes vapor conduction; g denotes vapor conductance; KLA (mgkg−1mgL−1) denotes leaf-air distribution coefficient; λLH (d−1) denotes total loss rate of pesticides within leaves, encompassing photodegradation, plant metabolism, and dilution due to plant growth; MLH (kg) denotes mass of a single leaf. This model assumes steady-state conditions, with pesticide movement from soil to leaves driven mainly by transpiration and limited by diffusion; VPD (Pa) denotes vapour pressure deficit. (K) denotes temperature difference between the leaf and the air, assumed to be due to the continuous transport of water; kd (mg·kg−1mg·L−1) denotes soil-water partition coefficient, governing the distribution of analyte between solid and aqueous soil phases. WL (L⋅kg−1), LL (kg⋅kg−1) denote foliar water and lipid contents, respectively; cc (dimensionless) denote empirical correction factor; KOW (dimensionless), KAW (dimensionless) denote octanol-water partition coefficient and the air-water partition coefficient of pesticides. If specific meteorological or physiological parameters are unavailable, use the default values listed in Table S2. This should be noted in the uncertainty analysis. Compute the BAFplantH(TAir,RHAir) (dimensionless) using the values from i–vi:BAFplantH(TAir,RHAir)=QH(TAir,RHAir)×1kdALH·g·1000KLA+λLH·MLHNote: The BAF quantifies the equilibrium ratio of foliar to edaphic concentrations. The numerator captures transpiration-mediated uptake; the denominator aggregates elimination pathways (stomatal diffusion, photolytic degradation, metabolic transformation, and growth dilution). Detailed definitions of all parameters are: ALH (m2) denotes leaf surface area. g denotes vapor conduction; g denotes vapor conductance; KLA (mgkg−1mgL−1) denotes leaf-air distribution coefficient; λLH (d−1) denotes total loss rate of pesticides within leaves, encompassing photodegradation, plant metabolism, and dilution due to plant growth; MLH (kg) denotes mass of a single leaf. This model assumes steady-state conditions, with pesticide movement from soil to leaves driven mainly by transpiration and limited by diffusion; VPD (Pa) denotes vapour pressure deficit. (K) denotes temperature difference between the leaf and the air, assumed to be due to the continuous transport of water; kd (mg·kg−1mg·L−1) denotes soil-water partition coefficient, governing the distribution of analyte between solid and aqueous soil phases. WL (L⋅kg−1), LL (kg⋅kg−1) denote foliar water and lipid contents, respectively; cc (dimensionless) denote empirical correction factor; KOW (dimensionless), KAW (dimensionless) denote octanol-water partition coefficient and the air-water partition coefficient of pesticides. If specific meteorological or physiological parameters are unavailable, use the default values listed in Table S2. This should be noted in the uncertainty analysis. Note: The BAF quantifies the equilibrium ratio of foliar to edaphic concentrations. The numerator captures transpiration-mediated uptake; the denominator aggregates elimination pathways (stomatal diffusion, photolytic degradation, metabolic transformation, and growth dilution). Detailed definitions of all parameters are: ALH (m2) denotes leaf surface area. g denotes vapor conduction; g denotes vapor conductance; KLA (mgkg−1mgL−1) denotes leaf-air distribution coefficient; λLH (d−1) denotes total loss rate of pesticides within leaves, encompassing photodegradation, plant metabolism, and dilution due to plant growth; MLH (kg) denotes mass of a single leaf. This model assumes steady-state conditions, with pesticide movement from soil to leaves driven mainly by transpiration and limited by diffusion; VPD (Pa) denotes vapour pressure deficit. (K) denotes temperature difference between the leaf and the air, assumed to be due to the continuous transport of water; kd (mg·kg−1mg·L−1) denotes soil-water partition coefficient, governing the distribution of analyte between solid and aqueous soil phases. WL (L⋅kg−1), LL (kg⋅kg−1) denote foliar water and lipid contents, respectively; cc (dimensionless) denote empirical correction factor; KOW (dimensionless), KAW (dimensionless) denote octanol-water partition coefficient and the air-water partition coefficient of pesticides. If specific meteorological or physiological parameters are unavailable, use the default values listed in Table S2. This should be noted in the uncertainty analysis. b.Calculate the simulated plant concentration.CPlantmodeldry=Csoil×BAFplantH(TAir,RHAir)CPlantmodelwet=CPlantmodeldry×(1−moisturecontent)Note: The soil concentration(Csoil) (mg·kg−1dry weight): measured pesticide concentration in soil. BAFplantH(TAir,RHAir) denotes the plant bioaccumulation factor (dimensionless) calculated in Step 10a. The modeled botanical concentration initially emerges on a dry-mass basis; convert to fresh-weight units for Tier 2 dietary calculations using species-appropriate moisture content. Because herbivores and omnivores consume fresh plant material, convert the dry-weight concentration to wet-weight concentration using species-specific moisture content. Use a moisture content of 0.8 for herbaceous plants and 0.5 for woody plants as defaults; adjust if species-specific data are available. Calculate the simulated plant concentration.CPlantmodeldry=Csoil×BAFplantH(TAir,RHAir)CPlantmodelwet=CPlantmodeldry×(1−moisturecontent)Note: The soil concentration(Csoil) (mg·kg−1dry weight): measured pesticide concentration in soil. BAFplantH(TAir,RHAir) denotes the plant bioaccumulation factor (dimensionless) calculated in Step 10a. The modeled botanical concentration initially emerges on a dry-mass basis; convert to fresh-weight units for Tier 2 dietary calculations using species-appropriate moisture content. Because herbivores and omnivores consume fresh plant material, convert the dry-weight concentration to wet-weight concentration using species-specific moisture content. Use a moisture content of 0.8 for herbaceous plants and 0.5 for woody plants as defaults; adjust if species-specific data are available. Note: The soil concentration(Csoil) (mg·kg−1dry weight): measured pesticide concentration in soil. BAFplantH(TAir,RHAir) denotes the plant bioaccumulation factor (dimensionless) calculated in Step 10a. The modeled botanical concentration initially emerges on a dry-mass basis; convert to fresh-weight units for Tier 2 dietary calculations using species-appropriate moisture content. Because herbivores and omnivores consume fresh plant material, convert the dry-weight concentration to wet-weight concentration using species-specific moisture content. Use a moisture content of 0.8 for herbaceous plants and 0.5 for woody plants as defaults; adjust if species-specific data are available. c.Use the simulated plant concentration as input for Tier 2.Substitute the Cplant_model value obtained in the previous step (corrected for moisture content to fresh weight) into Step 9 (Tier 2), then complete all steps in Tier 2 to obtain the ADD results for herbivorous, carnivorous and omnivorous animals.Substitute CPlantmodelwet for CPlant in Step 9. Then follow all sub-steps of Tier 2 to obtain ADD estimates for herbivores, carnivores, and omnivores.ADDherbivores,3=CPlantmodelwet·IRherbivoresBWherbivoresADDcarnivores,3=C¯herbivores,PM×IRcarnivoresBWcarnivoresADDomnivores,3=(fp·CPlantmodelwet+fh·C¯herbivores,PM+fc·C¯carnivores,PM)×IRomnivoresBWomnivoresWhere C¯herbivores,PM and C¯carnivores,PM are the weighted average concentrations derived from the simulated plant concentrations (as calculated in Steps 9c and 9e).Note: Model parameter availability – The Trapp model requires multiple environmental and plant physiological parameters. If certain parameters are unavailable, default values provided in Table S2 may be used, and this should be reflected in the uncertainty analysis.Note: Spatial heterogeneity – Soil sampling points may be unevenly distributed. It is recommended to apply Tier 3 within smaller spatial units (e.g., watersheds, provinces/states) to reduce aggregation bias.Note: Tier 3 is the most indirect approach among the three-tiered strategy. Sources of error include measurement errors in soil concentrations, uncertainties in model parameters, and accumulated errors from the subsequent Step 9. A sensitivity analysis (e.g., Monte Carlo simulation) and uncertainty propagation are strongly recommended. Use the simulated plant concentration as input for Tier 2. Substitute the Cplant_model value obtained in the previous step (corrected for moisture content to fresh weight) into Step 9 (Tier 2), then complete all steps in Tier 2 to obtain the ADD results for herbivorous, carnivorous and omnivorous animals. Substitute CPlantmodelwet for CPlant in Step 9. Then follow all sub-steps of Tier 2 to obtain ADD estimates for herbivores, carnivores, and omnivores.ADDherbivores,3=CPlantmodelwet·IRherbivoresBWherbivoresADDcarnivores,3=C¯herbivores,PM×IRcarnivoresBWcarnivoresADDomnivores,3=(fp·CPlantmodelwet+fh·C¯herbivores,PM+fc·C¯carnivores,PM)×IRomnivoresBWomnivoresWhere C¯herbivores,PM and C¯carnivores,PM are the weighted average concentrations derived from the simulated plant concentrations (as calculated in Steps 9c and 9e).Note: Model parameter availability – The Trapp model requires multiple environmental and plant physiological parameters. If certain parameters are unavailable, default values provided in Table S2 may be used, and this should be reflected in the uncertainty analysis.Note: Spatial heterogeneity – Soil sampling points may be unevenly distributed. It is recommended to apply Tier 3 within smaller spatial units (e.g., watersheds, provinces/states) to reduce aggregation bias.Note: Tier 3 is the most indirect approach among the three-tiered strategy. Sources of error include measurement errors in soil concentrations, uncertainties in model parameters, and accumulated errors from the subsequent Step 9. A sensitivity analysis (e.g., Monte Carlo simulation) and uncertainty propagation are strongly recommended. Note: Model parameter availability – The Trapp model requires multiple environmental and plant physiological parameters. If certain parameters are unavailable, default values provided in Table S2 may be used, and this should be reflected in the uncertainty analysis. Note: Spatial heterogeneity – Soil sampling points may be unevenly distributed. It is recommended to apply Tier 3 within smaller spatial units (e.g., watersheds, provinces/states) to reduce aggregation bias. Note: Tier 3 is the most indirect approach among the three-tiered strategy. Sources of error include measurement errors in soil concentrations, uncertainties in model parameters, and accumulated errors from the subsequent Step 9. A sensitivity analysis (e.g., Monte Carlo simulation) and uncertainty propagation are strongly recommended. Exposure and ecological risk assessment Timing: 3–5 days Based on the ADD (mg·kg−1·day−1) calculated previously, the non-carcinogenic risk (hazard quotient, HQ) and carcinogenic risk (ecological carcinogenic risk, ECR) to wildlife were assessed. All risk values are presented in the common logarithmic (Log10) format and are ultimately visualised on maps to illustrate their spatial distribution. 11.Obtain toxicological benchmarks for each pesticide.For the target pesticide, two toxicity parameters are required, both obtained from the USEPA, IRIS database or other authoritative sources (see supplemental information, Table S3):a.Oral reference dose (RfD, mg·kg−1·day−1): chronic oral threshold for non-cancer effects, representing the daily intake unlikely to produce appreciable adverse outcomes over a lifetime.b.Oral slope factor (OSF, per mg·kg−1·day−1): carcinogenic potency factor, providing an upper-bound estimate of lifetime excess cancer probability per unit chronic dose.Note: If multiple toxicity endpoint data are available for a given pesticide, the most sensitive value (i.e. the lowest RfD or highest OSF) should be selected. If an RfD or OSF is not available for a given pesticide, the corresponding risk cannot be calculated and this omission should be noted in the results. Obtain toxicological benchmarks for each pesticide. For the target pesticide, two toxicity parameters are required, both obtained from the USEPA, IRIS database or other authoritative sources (see supplemental information, Table S3):a.Oral reference dose (RfD, mg·kg−1·day−1): chronic oral threshold for non-cancer effects, representing the daily intake unlikely to produce appreciable adverse outcomes over a lifetime.b.Oral slope factor (OSF, per mg·kg−1·day−1): carcinogenic potency factor, providing an upper-bound estimate of lifetime excess cancer probability per unit chronic dose.Note: If multiple toxicity endpoint data are available for a given pesticide, the most sensitive value (i.e. the lowest RfD or highest OSF) should be selected. If an RfD or OSF is not available for a given pesticide, the corresponding risk cannot be calculated and this omission should be noted in the results. a.Oral reference dose (RfD, mg·kg−1·day−1): chronic oral threshold for non-cancer effects, representing the daily intake unlikely to produce appreciable adverse outcomes over a lifetime. Oral reference dose (RfD, mg·kg−1·day−1): chronic oral threshold for non-cancer effects, representing the daily intake unlikely to produce appreciable adverse outcomes over a lifetime. b.Oral slope factor (OSF, per mg·kg−1·day−1): carcinogenic potency factor, providing an upper-bound estimate of lifetime excess cancer probability per unit chronic dose.Note: If multiple toxicity endpoint data are available for a given pesticide, the most sensitive value (i.e. the lowest RfD or highest OSF) should be selected. If an RfD or OSF is not available for a given pesticide, the corresponding risk cannot be calculated and this omission should be noted in the results. Oral slope factor (OSF, per mg·kg−1·day−1): carcinogenic potency factor, providing an upper-bound estimate of lifetime excess cancer probability per unit chronic dose.Note: If multiple toxicity endpoint data are available for a given pesticide, the most sensitive value (i.e. the lowest RfD or highest OSF) should be selected. If an RfD or OSF is not available for a given pesticide, the corresponding risk cannot be calculated and this omission should be noted in the results. Note: If multiple toxicity endpoint data are available for a given pesticide, the most sensitive value (i.e. the lowest RfD or highest OSF) should be selected. If an RfD or OSF is not available for a given pesticide, the corresponding risk cannot be calculated and this omission should be noted in the results. 12.Calculate Hazard Quotient (HQ) for non-carcinogenic risk. Calculate Hazard Quotient (HQ) for non-carcinogenic risk. Note: ADD is the average daily dose (mg·kg−1·day−1) estimated from Tier 1–3.RfD is the reference dose (mg·kg−1·day−1). As a rule of thumb, an HQ at or below 1 means no appreciable non-cancer hazard; an HQ above 1 signals a possible concern, and the larger the HQ, the higher the risk. For pesticides without an RfD, skip the HQ calculation. When multiple pesticides or exposure pathways are present, consider calculating a Hazard Index (sum of HQs) as a conservative screening metric. This protocol reports individual HQs, but the cumulative approach can be added. 13.Calculate Ecological Carcinogenic Risk (ECR). Calculate Ecological Carcinogenic Risk (ECR). Note: ECR estimates the incremental lifetime cancer probability attributable to chronic chemical exposure. Conventionally, probabilities below 10−6 fall within the negligible range, whereas those surpassing 10−4 trigger unacceptable risk designations requiring management response. Intermediate values demand context-specific deliberation. For pesticides without an OSF, ECR cannot be calculated. Report as “insufficient data”. 14.Log-transform HQ and ECR. Log-transform HQ and ECR. For ease of visualisation and comparison, all HQ and ECR values have been converted to a logarithmic scale with a base of 10: Note: If HQ or ECR equals 0 (rare in practice, as ADD is usually >0), set it to a small placeholder (e.g., 1×10−10) before log transformation, or leave as missing data. The warning threshold for non-carcinogenic risk is defined as Log10HQ > 0 (HQ > 1), and for carcinogenic risk it is Log10ECR > −4 (ECR > 10−4). 15.Compile results by species, country, and exposure pathway.a.Record the species name (Latin name or standard English common name recommended) and its dietary category (herbivore, carnivore, or omnivore).b.Record the pesticide name and its CAS number.c.Record the ADD value (mg·kg−1·day−1), along with the tier level (Tier 1/2/3) from which it was derived and the exposure medium (soil, plant, or animal tissue) on which it is based.d.Record the corresponding toxicity parameters: RfD and OSF, including their sources (e.g., USEPA IRIS database or relevant literature).e.Calculate and record the HQ and its common Log10 HQ. If the RfD is missing, do not calculate HQ.f.Calculate and record the ECR and its common Log10 ECR. If the OSF is missing, do not calculate ECR.g.Assign risk interpretation: classify each species-chemical combination into one of five categories—(i) non-cancer hazard only, (ii) carcinogenic concern only, (iii) co-occurring risks, (iv) no detectable risk, or (v) data insufficient—based on threshold exceedance. Document all data deficiencies preventing risk calculation. Compile results by species, country, and exposure pathway.a.Record the species name (Latin name or standard English common name recommended) and its dietary category (herbivore, carnivore, or omnivore).b.Record the pesticide name and its CAS number.c.Record the ADD value (mg·kg−1·day−1), along with the tier level (Tier 1/2/3) from which it was derived and the exposure medium (soil, plant, or animal tissue) on which it is based.d.Record the corresponding toxicity parameters: RfD and OSF, including their sources (e.g., USEPA IRIS database or relevant literature).e.Calculate and record the HQ and its common Log10 HQ. If the RfD is missing, do not calculate HQ.f.Calculate and record the ECR and its common Log10 ECR. If the OSF is missing, do not calculate ECR.g.Assign risk interpretation: classify each species-chemical combination into one of five categories—(i) non-cancer hazard only, (ii) carcinogenic concern only, (iii) co-occurring risks, (iv) no detectable risk, or (v) data insufficient—based on threshold exceedance. Document all data deficiencies preventing risk calculation. a.Record the species name (Latin name or standard English common name recommended) and its dietary category (herbivore, carnivore, or omnivore). Record the species name (Latin name or standard English common name recommended) and its dietary category (herbivore, carnivore, or omnivore). b.Record the pesticide name and its CAS number. Record the pesticide name and its CAS number. c.Record the ADD value (mg·kg−1·day−1), along with the tier level (Tier 1/2/3) from which it was derived and the exposure medium (soil, plant, or animal tissue) on which it is based. Record the ADD value (mg·kg−1·day−1), along with the tier level (Tier 1/2/3) from which it was derived and the exposure medium (soil, plant, or animal tissue) on which it is based. d.Record the corresponding toxicity parameters: RfD and OSF, including their sources (e.g., USEPA IRIS database or relevant literature). Record the corresponding toxicity parameters: RfD and OSF, including their sources (e.g., USEPA IRIS database or relevant literature). e.Calculate and record the HQ and its common Log10 HQ. If the RfD is missing, do not calculate HQ. Calculate and record the HQ and its common Log10 HQ. If the RfD is missing, do not calculate HQ. f.Calculate and record the ECR and its common Log10 ECR. If the OSF is missing, do not calculate ECR. Calculate and record the ECR and its common Log10 ECR. If the OSF is missing, do not calculate ECR. g.Assign risk interpretation: classify each species-chemical combination into one of five categories—(i) non-cancer hazard only, (ii) carcinogenic concern only, (iii) co-occurring risks, (iv) no detectable risk, or (v) data insufficient—based on threshold exceedance. Document all data deficiencies preventing risk calculation. Assign risk interpretation: classify each species-chemical combination into one of five categories—(i) non-cancer hazard only, (ii) carcinogenic concern only, (iii) co-occurring risks, (iv) no detectable risk, or (v) data insufficient—based on threshold exceedance. Document all data deficiencies preventing risk calculation. Note: Include but do not limit to unit conversions, model assumptions, reasons for missing data, sampling year, detection frequency, and any other information that may affect the interpretation of the results. Pause point: After compiling all risk results into the structured table, save the table for independent review before proceeding to map generation. This is a good point to verify data completeness and check for any missing toxicity benchmarks. Mapping of contamination and risk Timing: 5 days This section describes the method for visualising contamination scores (soils, wild plants, mammals) and the country-level spatial distributions of Log10 ECR and Log10 HQ. 16.Download spatial boundary dataa.Acquire vector boundary layers (administrative units, hydrological basins) from public geospatial repositories such as ArcGIS Hub or equivalent open-data platforms Download spatial boundary dataa.Acquire vector boundary layers (administrative units, hydrological basins) from public geospatial repositories such as ArcGIS Hub or equivalent open-data platforms a.Acquire vector boundary layers (administrative units, hydrological basins) from public geospatial repositories such as ArcGIS Hub or equivalent open-data platforms Acquire vector boundary layers (administrative units, hydrological basins) from public geospatial repositories such as ArcGIS Hub or equivalent open-data platforms Note: The choice of boundary should reflect the analysis scale. For the present study, national boundaries are used as the primary spatial unit for displaying contamination scores and risk indicators. 17.Import spatial boundary data into ArcGIS.a.Load the shapefile as a polygon layer using the “Add Data” function in ArcGIS. Import spatial boundary data into ArcGIS.a.Load the shapefile as a polygon layer using the “Add Data” function in ArcGIS. a.Load the shapefile as a polygon layer using the “Add Data” function in ArcGIS. Load the shapefile as a polygon layer using the “Add Data” function in ArcGIS. 18.Import contamination score or risk data tables.a.Contamination scores (country name, soil score Ssoil, wild plant score Splant, mammal score Sanimal).b.Risk indicators (country name, Log10 ECR and Log10 HQ derived from soil, plant, and animal tissue pathways). Import contamination score or risk data tables.a.Contamination scores (country name, soil score Ssoil, wild plant score Splant, mammal score Sanimal).b.Risk indicators (country name, Log10 ECR and Log10 HQ derived from soil, plant, and animal tissue pathways). a.Contamination scores (country name, soil score Ssoil, wild plant score Splant, mammal score Sanimal). Contamination scores (country name, soil score Ssoil, wild plant score Splant, mammal score Sanimal). b.Risk indicators (country name, Log10 ECR and Log10 HQ derived from soil, plant, and animal tissue pathways). Risk indicators (country name, Log10 ECR and Log10 HQ derived from soil, plant, and animal tissue pathways). Note: Ensure that the names of spatial units (e.g., country names) in the data tables exactly match those in the attribute table of the shapefile (including spelling, case, and abbreviations). Perform any necessary replacements before joining. 19.Join attribute data and generate thematic maps.a.Contamination score maps.i.Attribute join: Merge contamination indices to boundary polygons via country identifier.ii.Symbology: Use graduated colors (5 classes) for each score field (Ssoil, Splant, Sanimal) with an appropriate colour ramp (e.g., light green for low scores, dark red for high scores).iii.Output: Generate five separate maps: soil contamination scores, wild plant contamination scores, and mammals (carnivores, omnivores, and herbivores)contamination scores.b.Risk indicator maps (Log10 ECR and Log10 HQ)i.Attribute join: Merge tier-specific risk estimates to boundary polygons.ii.Symbology: Render Log10 ECR and Log10 HQ as separate thematic layers, differentiated by exposure pathway (soil, plant, tissue).iii.Threshold marking: Optionally, highlight regions exceeding the warning thresholds (Log10 ECR>−4 or Log HQ>0) using distinct colour intervals or hatching.iv.Deliverables: Generate six individual map sheets or three dual-panel composites, each matrix presenting both carcinogenic and non-carcinogenic risk surfaces. Join attribute data and generate thematic maps.a.Contamination score maps.i.Attribute join: Merge contamination indices to boundary polygons via country identifier.ii.Symbology: Use graduated colors (5 classes) for each score field (Ssoil, Splant, Sanimal) with an appropriate colour ramp (e.g., light green for low scores, dark red for high scores).iii.Output: Generate five separate maps: soil contamination scores, wild plant contamination scores, and mammals (carnivores, omnivores, and herbivores)contamination scores.b.Risk indicator maps (Log10 ECR and Log10 HQ)i.Attribute join: Merge tier-specific risk estimates to boundary polygons.ii.Symbology: Render Log10 ECR and Log10 HQ as separate thematic layers, differentiated by exposure pathway (soil, plant, tissue).iii.Threshold marking: Optionally, highlight regions exceeding the warning thresholds (Log10 ECR>−4 or Log HQ>0) using distinct colour intervals or hatching.iv.Deliverables: Generate six individual map sheets or three dual-panel composites, each matrix presenting both carcinogenic and non-carcinogenic risk surfaces. a.Contamination score maps.i.Attribute join: Merge contamination indices to boundary polygons via country identifier.ii.Symbology: Use graduated colors (5 classes) for each score field (Ssoil, Splant, Sanimal) with an appropriate colour ramp (e.g., light green for low scores, dark red for high scores).iii.Output: Generate five separate maps: soil contamination scores, wild plant contamination scores, and mammals (carnivores, omnivores, and herbivores)contamination scores. Contamination score maps.i.Attribute join: Merge contamination indices to boundary polygons via country identifier.ii.Symbology: Use graduated colors (5 classes) for each score field (Ssoil, Splant, Sanimal) with an appropriate colour ramp (e.g., light green for low scores, dark red for high scores).iii.Output: Generate five separate maps: soil contamination scores, wild plant contamination scores, and mammals (carnivores, omnivores, and herbivores)contamination scores. i.Attribute join: Merge contamination indices to boundary polygons via country identifier. Attribute join: Merge contamination indices to boundary polygons via country identifier. ii.Symbology: Use graduated colors (5 classes) for each score field (Ssoil, Splant, Sanimal) with an appropriate colour ramp (e.g., light green for low scores, dark red for high scores). Symbology: Use graduated colors (5 classes) for each score field (Ssoil, Splant, Sanimal) with an appropriate colour ramp (e.g., light green for low scores, dark red for high scores). iii.Output: Generate five separate maps: soil contamination scores, wild plant contamination scores, and mammals (carnivores, omnivores, and herbivores)contamination scores. Output: Generate five separate maps: soil contamination scores, wild plant contamination scores, and mammals (carnivores, omnivores, and herbivores)contamination scores. b.Risk indicator maps (Log10 ECR and Log10 HQ)i.Attribute join: Merge tier-specific risk estimates to boundary polygons.ii.Symbology: Render Log10 ECR and Log10 HQ as separate thematic layers, differentiated by exposure pathway (soil, plant, tissue).iii.Threshold marking: Optionally, highlight regions exceeding the warning thresholds (Log10 ECR>−4 or Log HQ>0) using distinct colour intervals or hatching.iv.Deliverables: Generate six individual map sheets or three dual-panel composites, each matrix presenting both carcinogenic and non-carcinogenic risk surfaces. Risk indicator maps (Log10 ECR and Log10 HQ)i.Attribute join: Merge tier-specific risk estimates to boundary polygons.ii.Symbology: Render Log10 ECR and Log10 HQ as separate thematic layers, differentiated by exposure pathway (soil, plant, tissue).iii.Threshold marking: Optionally, highlight regions exceeding the warning thresholds (Log10 ECR>−4 or Log HQ>0) using distinct colour intervals or hatching.iv.Deliverables: Generate six individual map sheets or three dual-panel composites, each matrix presenting both carcinogenic and non-carcinogenic risk surfaces. i.Attribute join: Merge tier-specific risk estimates to boundary polygons. Attribute join: Merge tier-specific risk estimates to boundary polygons. ii.Symbology: Render Log10 ECR and Log10 HQ as separate thematic layers, differentiated by exposure pathway (soil, plant, tissue). Symbology: Render Log10 ECR and Log10 HQ as separate thematic layers, differentiated by exposure pathway (soil, plant, tissue). iii.Threshold marking: Optionally, highlight regions exceeding the warning thresholds (Log10 ECR>−4 or Log HQ>0) using distinct colour intervals or hatching. Threshold marking: Optionally, highlight regions exceeding the warning thresholds (Log10 ECR>−4 or Log HQ>0) using distinct colour intervals or hatching. iv.Deliverables: Generate six individual map sheets or three dual-panel composites, each matrix presenting both carcinogenic and non-carcinogenic risk surfaces. Deliverables: Generate six individual map sheets or three dual-panel composites, each matrix presenting both carcinogenic and non-carcinogenic risk surfaces. 20.Map finishing and exporta.Add map elements: North arrow, scale bar, legend, and a descriptive title (e.g., “Country-level Log10 ECR derived from soil”).b.Indicate missing data: For countries without data, set the fill colour to grey and label them as “No data (NA)” in the legend.c.Export: Save the final maps at high resolution (≥300 dpi) as TIFF/PNG or PDF files for inclusion in the manuscript or supplemental information. Map finishing and exporta.Add map elements: North arrow, scale bar, legend, and a descriptive title (e.g., “Country-level Log10 ECR derived from soil”).b.Indicate missing data: For countries without data, set the fill colour to grey and label them as “No data (NA)” in the legend.c.Export: Save the final maps at high resolution (≥300 dpi) as TIFF/PNG or PDF files for inclusion in the manuscript or supplemental information. a.Add map elements: North arrow, scale bar, legend, and a descriptive title (e.g., “Country-level Log10 ECR derived from soil”). Add map elements: North arrow, scale bar, legend, and a descriptive title (e.g., “Country-level Log10 ECR derived from soil”). b.Indicate missing data: For countries without data, set the fill colour to grey and label them as “No data (NA)” in the legend. Indicate missing data: For countries without data, set the fill colour to grey and label them as “No data (NA)” in the legend. c.Export: Save the final maps at high resolution (≥300 dpi) as TIFF/PNG or PDF files for inclusion in the manuscript or supplemental information. Export: Save the final maps at high resolution (≥300 dpi) as TIFF/PNG or PDF files for inclusion in the manuscript or supplemental information. Note: All maps are intended for descriptive visualisation only; no error bars or statistical tests are shown. Due to uneven data coverage, high-risk areas may reflect data availability rather than true contamination levels. This limitation should be clearly stated in the figure captions or main text. For large or environmentally diverse countries (e.g., China, United States), supplementary maps at sub-national scales (provinces, states, or watersheds) are recommended, while country-level maps serve as an overview. Uncertainty quantification and sensitivity analysis Timing: 5 days This section performs uncertainty analysis to assess the robustness of risk estimates. This step mainly includes Monte Carlo simulation and Sobol sensitivity analysis. 21.Monte Carlo Simulation for Uncertainty Analysis of HQ and ECR.a.Prepare input data and distribution assumptions.i.Retrieve previously calculated ADD point estimates stored by species, pesticide, and country.ii.Model each uncertain input (measured residues, Log Kow, half-life, intake rate, body weight, etc.) with a log-normal distribution.iii.Adopt a default relative standard deviation of 30% (CV = 0.3) if the literature does not report variability for a particular parameter.iv.Treat toxicity benchmarks (RfD and OSF) as fixed values in this simulation.Note: Environmental contamination data are usually right-skewed, so the log-normal distribution is appropriate. The default CV of 0.3 is consistent with the typical range of variability (20%–50%) observed for environmental contaminants and toxicokinetic parameters. Additional uncertainty may be introduced as a sensitivity test if explicitly stated.b.Set simulation parameters.i.Set the number of iterations to 10,000.ii.Fix the random seed (set.seed) to ensure reproducibility.c.Generate random samples.i.Draw 10,000 random values from the log-normal distribution for each parameter.ii.Use built-in random number generators in R.d.Compute output variables.i.For each iteration, calculate HQ and ECR.ii.Compute the 2.5th and 97.5th percentiles as the 95% confidence interval. Transform HQ and ECR to base-10 logarithms: Log10HQ and Log10ECR.iii.Repeat the above for Log10 HQ and Log10 ECR.e.Summarise simulation results.i.Compute the median of HQ and ECR.ii.Compute the 2.5th and 97.5th percentiles as the 95% confidence interval.iii.Repeat the above for Log10HQ and Log10ECR.f.Visualise and export results.i.Plot the probability density of the simulated values using a smooth kernel density estimate.ii.Overlay the density of the observed data (if available) to assess model fit.iii.Save medians, confidence intervals, and the corresponding log-transformed values as a CSV file.Note: This step is optional. Monte Carlo Simulation for Uncertainty Analysis of HQ and ECR.a.Prepare input data and distribution assumptions.i.Retrieve previously calculated ADD point estimates stored by species, pesticide, and country.ii.Model each uncertain input (measured residues, Log Kow, half-life, intake rate, body weight, etc.) with a log-normal distribution.iii.Adopt a default relative standard deviation of 30% (CV = 0.3) if the literature does not report variability for a particular parameter.iv.Treat toxicity benchmarks (RfD and OSF) as fixed values in this simulation.Note: Environmental contamination data are usually right-skewed, so the log-normal distribution is appropriate. The default CV of 0.3 is consistent with the typical range of variability (20%–50%) observed for environmental contaminants and toxicokinetic parameters. Additional uncertainty may be introduced as a sensitivity test if explicitly stated.b.Set simulation parameters.i.Set the number of iterations to 10,000.ii.Fix the random seed (set.seed) to ensure reproducibility.c.Generate random samples.i.Draw 10,000 random values from the log-normal distribution for each parameter.ii.Use built-in random number generators in R.d.Compute output variables.i.For each iteration, calculate HQ and ECR.ii.Compute the 2.5th and 97.5th percentiles as the 95% confidence interval. Transform HQ and ECR to base-10 logarithms: Log10HQ and Log10ECR.iii.Repeat the above for Log10 HQ and Log10 ECR.e.Summarise simulation results.i.Compute the median of HQ and ECR.ii.Compute the 2.5th and 97.5th percentiles as the 95% confidence interval.iii.Repeat the above for Log10HQ and Log10ECR.f.Visualise and export results.i.Plot the probability density of the simulated values using a smooth kernel density estimate.ii.Overlay the density of the observed data (if available) to assess model fit.iii.Save medians, confidence intervals, and the corresponding log-transformed values as a CSV file.Note: This step is optional. a.Prepare input data and distribution assumptions.i.Retrieve previously calculated ADD point estimates stored by species, pesticide, and country.ii.Model each uncertain input (measured residues, Log Kow, half-life, intake rate, body weight, etc.) with a log-normal distribution.iii.Adopt a default relative standard deviation of 30% (CV = 0.3) if the literature does not report variability for a particular parameter.iv.Treat toxicity benchmarks (RfD and OSF) as fixed values in this simulation.Note: Environmental contamination data are usually right-skewed, so the log-normal distribution is appropriate. The default CV of 0.3 is consistent with the typical range of variability (20%–50%) observed for environmental contaminants and toxicokinetic parameters. Additional uncertainty may be introduced as a sensitivity test if explicitly stated. Prepare input data and distribution assumptions.i.Retrieve previously calculated ADD point estimates stored by species, pesticide, and country.ii.Model each uncertain input (measured residues, Log Kow, half-life, intake rate, body weight, etc.) with a log-normal distribution.iii.Adopt a default relative standard deviation of 30% (CV = 0.3) if the literature does not report variability for a particular parameter.iv.Treat toxicity benchmarks (RfD and OSF) as fixed values in this simulation.Note: Environmental contamination data are usually right-skewed, so the log-normal distribution is appropriate. The default CV of 0.3 is consistent with the typical range of variability (20%–50%) observed for environmental contaminants and toxicokinetic parameters. Additional uncertainty may be introduced as a sensitivity test if explicitly stated. i.Retrieve previously calculated ADD point estimates stored by species, pesticide, and country. Retrieve previously calculated ADD point estimates stored by species, pesticide, and country. ii.Model each uncertain input (measured residues, Log Kow, half-life, intake rate, body weight, etc.) with a log-normal distribution. Model each uncertain input (measured residues, Log Kow, half-life, intake rate, body weight, etc.) with a log-normal distribution. iii.Adopt a default relative standard deviation of 30% (CV = 0.3) if the literature does not report variability for a particular parameter. Adopt a default relative standard deviation of 30% (CV = 0.3) if the literature does not report variability for a particular parameter. iv.Treat toxicity benchmarks (RfD and OSF) as fixed values in this simulation.Note: Environmental contamination data are usually right-skewed, so the log-normal distribution is appropriate. The default CV of 0.3 is consistent with the typical range of variability (20%–50%) observed for environmental contaminants and toxicokinetic parameters. Additional uncertainty may be introduced as a sensitivity test if explicitly stated. Treat toxicity benchmarks (RfD and OSF) as fixed values in this simulation.Note: Environmental contamination data are usually right-skewed, so the log-normal distribution is appropriate. The default CV of 0.3 is consistent with the typical range of variability (20%–50%) observed for environmental contaminants and toxicokinetic parameters. Additional uncertainty may be introduced as a sensitivity test if explicitly stated. Note: Environmental contamination data are usually right-skewed, so the log-normal distribution is appropriate. The default CV of 0.3 is consistent with the typical range of variability (20%–50%) observed for environmental contaminants and toxicokinetic parameters. Additional uncertainty may be introduced as a sensitivity test if explicitly stated. b.Set simulation parameters.i.Set the number of iterations to 10,000.ii.Fix the random seed (set.seed) to ensure reproducibility. Set simulation parameters.i.Set the number of iterations to 10,000.ii.Fix the random seed (set.seed) to ensure reproducibility. i.Set the number of iterations to 10,000. Set the number of iterations to 10,000. ii.Fix the random seed (set.seed) to ensure reproducibility. Fix the random seed (set.seed) to ensure reproducibility. c.Generate random samples.i.Draw 10,000 random values from the log-normal distribution for each parameter.ii.Use built-in random number generators in R. Generate random samples.i.Draw 10,000 random values from the log-normal distribution for each parameter.ii.Use built-in random number generators in R. i.Draw 10,000 random values from the log-normal distribution for each parameter. Draw 10,000 random values from the log-normal distribution for each parameter. ii.Use built-in random number generators in R. Use built-in random number generators in R. d.Compute output variables.i.For each iteration, calculate HQ and ECR.ii.Compute the 2.5th and 97.5th percentiles as the 95% confidence interval. Transform HQ and ECR to base-10 logarithms: Log10HQ and Log10ECR.iii.Repeat the above for Log10 HQ and Log10 ECR. Compute output variables.i.For each iteration, calculate HQ and ECR.ii.Compute the 2.5th and 97.5th percentiles as the 95% confidence interval. Transform HQ and ECR to base-10 logarithms: Log10HQ and Log10ECR.iii.Repeat the above for Log10 HQ and Log10 ECR. i.For each iteration, calculate HQ and ECR. For each iteration, calculate HQ and ECR. ii.Compute the 2.5th and 97.5th percentiles as the 95% confidence interval. Transform HQ and ECR to base-10 logarithms: Log10HQ and Log10ECR. Compute the 2.5th and 97.5th percentiles as the 95% confidence interval. Transform HQ and ECR to base-10 logarithms: Log10HQ and Log10ECR. iii.Repeat the above for Log10 HQ and Log10 ECR. Repeat the above for Log10 HQ and Log10 ECR. e.Summarise simulation results.i.Compute the median of HQ and ECR.ii.Compute the 2.5th and 97.5th percentiles as the 95% confidence interval.iii.Repeat the above for Log10HQ and Log10ECR. Summarise simulation results.i.Compute the median of HQ and ECR.ii.Compute the 2.5th and 97.5th percentiles as the 95% confidence interval.iii.Repeat the above for Log10HQ and Log10ECR. i.Compute the median of HQ and ECR. Compute the median of HQ and ECR. ii.Compute the 2.5th and 97.5th percentiles as the 95% confidence interval. Compute the 2.5th and 97.5th percentiles as the 95% confidence interval. iii.Repeat the above for Log10HQ and Log10ECR. Repeat the above for Log10HQ and Log10ECR. f.Visualise and export results.i.Plot the probability density of the simulated values using a smooth kernel density estimate.ii.Overlay the density of the observed data (if available) to assess model fit.iii.Save medians, confidence intervals, and the corresponding log-transformed values as a CSV file.Note: This step is optional. Visualise and export results.i.Plot the probability density of the simulated values using a smooth kernel density estimate.ii.Overlay the density of the observed data (if available) to assess model fit.iii.Save medians, confidence intervals, and the corresponding log-transformed values as a CSV file.Note: This step is optional. i.Plot the probability density of the simulated values using a smooth kernel density estimate. Plot the probability density of the simulated values using a smooth kernel density estimate. ii.Overlay the density of the observed data (if available) to assess model fit. Overlay the density of the observed data (if available) to assess model fit. iii.Save medians, confidence intervals, and the corresponding log-transformed values as a CSV file.Note: This step is optional. Save medians, confidence intervals, and the corresponding log-transformed values as a CSV file.Note: This step is optional. Note: This step is optional. 22.Variance Decomposition via Sobol Sensitivity Analysis.a.Select the case study.i.Illustrative application: Select a well-studied predator-contaminant pairing for method demonstration (consult the original study for a worked example).ii.Perform the analysis separately for Tier 1 (tissue-based, using measured liver concentrations) and Tier 3 (soil-based, using the soil-plant-food chain model).b.Define input parameters and their ranges.i.Tier 1 (tissue-based pathway): liver tissue concentration (Ctissue), BTF, BW, and oral slope factor (OSF).ii.Tier 3 (soil-based pathway): Csoil, plant bioaccumulation factor (BAFplant), rabbit intake rate (IRrabbit) and body weight (BWrabbit), rabbit fat and muscle BTFs, fox intake rate (IRfox), fox body weight (BWfox), and OSF.iii.For each parameter, define a uniform distribution range (e.g., minimum and maximum values taken from literature or ±20%–30% of the point estimate).c.Generate Saltelli sampling samples.i.Use the Saltelli algorithm to generate a parameter combination matrix.ii.Total model runs must strictly follow the original paper:Tier 1 (tissue-based): 2,000 runs;Tier 3 (soil-based): 10,000 runs.iii.Compute the base sample size N denotes total runs/(2D+2) (where D denotes number of parameters) and round to an integer. Fix the random seed for reproducibility.d.Run the model and calculate outputs.i.For each parameter set, compute the corresponding ECR value.ii.Store the results as a vector aligned with the parameter samples.e.Compute Sobol indices.i.Calculate the first-order index (S1) and total-effect index (ST) for each parameter.ii.Estimate 95% confidence intervals using bootstrap (200–500 resamples).f.Interpretation of results.i.S1 quantifies the direct contribution of one input to the output variance.ii.A high ST indicates that the parameter (including its interactions) has a large influence on output variance.iii.ST captures both direct and interaction effects.iv.When ST exceeds S1 by a wide margin, that parameter strongly interacts with others.g.Reporting.i.Present S1, ST, and their 95% confidence intervals in a table.ii.In the text, highlight the main drivers (e.g., body weight and intake rate as the largest contributors to ECR uncertainty, as reported in the original study). Variance Decomposition via Sobol Sensitivity Analysis.a.Select the case study.i.Illustrative application: Select a well-studied predator-contaminant pairing for method demonstration (consult the original study for a worked example).ii.Perform the analysis separately for Tier 1 (tissue-based, using measured liver concentrations) and Tier 3 (soil-based, using the soil-plant-food chain model).b.Define input parameters and their ranges.i.Tier 1 (tissue-based pathway): liver tissue concentration (Ctissue), BTF, BW, and oral slope factor (OSF).ii.Tier 3 (soil-based pathway): Csoil, plant bioaccumulation factor (BAFplant), rabbit intake rate (IRrabbit) and body weight (BWrabbit), rabbit fat and muscle BTFs, fox intake rate (IRfox), fox body weight (BWfox), and OSF.iii.For each parameter, define a uniform distribution range (e.g., minimum and maximum values taken from literature or ±20%–30% of the point estimate).c.Generate Saltelli sampling samples.i.Use the Saltelli algorithm to generate a parameter combination matrix.ii.Total model runs must strictly follow the original paper:Tier 1 (tissue-based): 2,000 runs;Tier 3 (soil-based): 10,000 runs.iii.Compute the base sample size N denotes total runs/(2D+2) (where D denotes number of parameters) and round to an integer. Fix the random seed for reproducibility.d.Run the model and calculate outputs.i.For each parameter set, compute the corresponding ECR value.ii.Store the results as a vector aligned with the parameter samples.e.Compute Sobol indices.i.Calculate the first-order index (S1) and total-effect index (ST) for each parameter.ii.Estimate 95% confidence intervals using bootstrap (200–500 resamples).f.Interpretation of results.i.S1 quantifies the direct contribution of one input to the output variance.ii.A high ST indicates that the parameter (including its interactions) has a large influence on output variance.iii.ST captures both direct and interaction effects.iv.When ST exceeds S1 by a wide margin, that parameter strongly interacts with others.g.Reporting.i.Present S1, ST, and their 95% confidence intervals in a table.ii.In the text, highlight the main drivers (e.g., body weight and intake rate as the largest contributors to ECR uncertainty, as reported in the original study). a.Select the case study.i.Illustrative application: Select a well-studied predator-contaminant pairing for method demonstration (consult the original study for a worked example).ii.Perform the analysis separately for Tier 1 (tissue-based, using measured liver concentrations) and Tier 3 (soil-based, using the soil-plant-food chain model). Select the case study.i.Illustrative application: Select a well-studied predator-contaminant pairing for method demonstration (consult the original study for a worked example).ii.Perform the analysis separately for Tier 1 (tissue-based, using measured liver concentrations) and Tier 3 (soil-based, using the soil-plant-food chain model). i.Illustrative application: Select a well-studied predator-contaminant pairing for method demonstration (consult the original study for a worked example). Illustrative application: Select a well-studied predator-contaminant pairing for method demonstration (consult the original study for a worked example). ii.Perform the analysis separately for Tier 1 (tissue-based, using measured liver concentrations) and Tier 3 (soil-based, using the soil-plant-food chain model). Perform the analysis separately for Tier 1 (tissue-based, using measured liver concentrations) and Tier 3 (soil-based, using the soil-plant-food chain model). b.Define input parameters and their ranges.i.Tier 1 (tissue-based pathway): liver tissue concentration (Ctissue), BTF, BW, and oral slope factor (OSF).ii.Tier 3 (soil-based pathway): Csoil, plant bioaccumulation factor (BAFplant), rabbit intake rate (IRrabbit) and body weight (BWrabbit), rabbit fat and muscle BTFs, fox intake rate (IRfox), fox body weight (BWfox), and OSF.iii.For each parameter, define a uniform distribution range (e.g., minimum and maximum values taken from literature or ±20%–30% of the point estimate). Define input parameters and their ranges.i.Tier 1 (tissue-based pathway): liver tissue concentration (Ctissue), BTF, BW, and oral slope factor (OSF).ii.Tier 3 (soil-based pathway): Csoil, plant bioaccumulation factor (BAFplant), rabbit intake rate (IRrabbit) and body weight (BWrabbit), rabbit fat and muscle BTFs, fox intake rate (IRfox), fox body weight (BWfox), and OSF.iii.For each parameter, define a uniform distribution range (e.g., minimum and maximum values taken from literature or ±20%–30% of the point estimate). i.Tier 1 (tissue-based pathway): liver tissue concentration (Ctissue), BTF, BW, and oral slope factor (OSF). Tier 1 (tissue-based pathway): liver tissue concentration (Ctissue), BTF, BW, and oral slope factor (OSF). ii.Tier 3 (soil-based pathway): Csoil, plant bioaccumulation factor (BAFplant), rabbit intake rate (IRrabbit) and body weight (BWrabbit), rabbit fat and muscle BTFs, fox intake rate (IRfox), fox body weight (BWfox), and OSF. Tier 3 (soil-based pathway): Csoil, plant bioaccumulation factor (BAFplant), rabbit intake rate (IRrabbit) and body weight (BWrabbit), rabbit fat and muscle BTFs, fox intake rate (IRfox), fox body weight (BWfox), and OSF. iii.For each parameter, define a uniform distribution range (e.g., minimum and maximum values taken from literature or ±20%–30% of the point estimate). For each parameter, define a uniform distribution range (e.g., minimum and maximum values taken from literature or ±20%–30% of the point estimate). c.Generate Saltelli sampling samples.i.Use the Saltelli algorithm to generate a parameter combination matrix.ii.Total model runs must strictly follow the original paper:Tier 1 (tissue-based): 2,000 runs;Tier 3 (soil-based): 10,000 runs.iii.Compute the base sample size N denotes total runs/(2D+2) (where D denotes number of parameters) and round to an integer. Fix the random seed for reproducibility. Generate Saltelli sampling samples.i.Use the Saltelli algorithm to generate a parameter combination matrix.ii.Total model runs must strictly follow the original paper:Tier 1 (tissue-based): 2,000 runs;Tier 3 (soil-based): 10,000 runs.iii.Compute the base sample size N denotes total runs/(2D+2) (where D denotes number of parameters) and round to an integer. Fix the random seed for reproducibility. i.Use the Saltelli algorithm to generate a parameter combination matrix. Use the Saltelli algorithm to generate a parameter combination matrix. ii.Total model runs must strictly follow the original paper:Tier 1 (tissue-based): 2,000 runs;Tier 3 (soil-based): 10,000 runs. Total model runs must strictly follow the original paper:Tier 1 (tissue-based): 2,000 runs;Tier 3 (soil-based): 10,000 runs. iii.Compute the base sample size N denotes total runs/(2D+2) (where D denotes number of parameters) and round to an integer. Fix the random seed for reproducibility. Compute the base sample size N denotes total runs/(2D+2) (where D denotes number of parameters) and round to an integer. Fix the random seed for reproducibility. d.Run the model and calculate outputs.i.For each parameter set, compute the corresponding ECR value.ii.Store the results as a vector aligned with the parameter samples. Run the model and calculate outputs.i.For each parameter set, compute the corresponding ECR value.ii.Store the results as a vector aligned with the parameter samples. i.For each parameter set, compute the corresponding ECR value. For each parameter set, compute the corresponding ECR value. ii.Store the results as a vector aligned with the parameter samples. Store the results as a vector aligned with the parameter samples. e.Compute Sobol indices.i.Calculate the first-order index (S1) and total-effect index (ST) for each parameter.ii.Estimate 95% confidence intervals using bootstrap (200–500 resamples). Compute Sobol indices.i.Calculate the first-order index (S1) and total-effect index (ST) for each parameter.ii.Estimate 95% confidence intervals using bootstrap (200–500 resamples). i.Calculate the first-order index (S1) and total-effect index (ST) for each parameter. Calculate the first-order index (S1) and total-effect index (ST) for each parameter. ii.Estimate 95% confidence intervals using bootstrap (200–500 resamples). Estimate 95% confidence intervals using bootstrap (200–500 resamples). f.Interpretation of results.i.S1 quantifies the direct contribution of one input to the output variance.ii.A high ST indicates that the parameter (including its interactions) has a large influence on output variance.iii.ST captures both direct and interaction effects.iv.When ST exceeds S1 by a wide margin, that parameter strongly interacts with others. Interpretation of results.i.S1 quantifies the direct contribution of one input to the output variance.ii.A high ST indicates that the parameter (including its interactions) has a large influence on output variance.iii.ST captures both direct and interaction effects.iv.When ST exceeds S1 by a wide margin, that parameter strongly interacts with others. i.S1 quantifies the direct contribution of one input to the output variance. S1 quantifies the direct contribution of one input to the output variance. ii.A high ST indicates that the parameter (including its interactions) has a large influence on output variance. A high ST indicates that the parameter (including its interactions) has a large influence on output variance. iii.ST captures both direct and interaction effects. ST captures both direct and interaction effects. iv.When ST exceeds S1 by a wide margin, that parameter strongly interacts with others. When ST exceeds S1 by a wide margin, that parameter strongly interacts with others. g.Reporting.i.Present S1, ST, and their 95% confidence intervals in a table.ii.In the text, highlight the main drivers (e.g., body weight and intake rate as the largest contributors to ECR uncertainty, as reported in the original study). Reporting.i.Present S1, ST, and their 95% confidence intervals in a table.ii.In the text, highlight the main drivers (e.g., body weight and intake rate as the largest contributors to ECR uncertainty, as reported in the original study). i.Present S1, ST, and their 95% confidence intervals in a table. Present S1, ST, and their 95% confidence intervals in a table. ii.In the text, highlight the main drivers (e.g., body weight and intake rate as the largest contributors to ECR uncertainty, as reported in the original study). In the text, highlight the main drivers (e.g., body weight and intake rate as the largest contributors to ECR uncertainty, as reported in the original study). Expected outcomes This protocol generates three primary outputs. First, it produces region-specific contamination scores (Ssoil, Splant, Sanimal) that rank pesticide contamination levels across countries or spatial units relative to the global mean. Second, it yields tier-specific average daily dose (ADD) estimates for herbivores, carnivores, and omnivores, derived from either measured tissue concentrations (Tier 1), measured plant concentrations (Tier 2), or soil-to-plant modeled concentrations (Tier 3). Third, it calculates non-carcinogenic hazard quotients (HQ) and carcinogenic ecological risks (ECR) for each species-pesticide combination, with associated 95% confidence intervals from Monte Carlo simulations. All results are compiled into structured tables and visualized as spatial maps, enabling comparative risk assessment across regions and trophic levels. Users can expect to identify high-risk regions and data gaps, and to prioritize future monitoring efforts. Limitations First, the toxicity benchmarks are human-health oriented. The RfD and OSF values employed in this protocol were primarily obtained from the USEPA IRIS database, which does not account for the physiological characteristics, metabolic capacities, or species-specific sensitivities of different wildlife taxa. Also, the external intake is back-calculated from limited biomonitoring data using a steady-state PBK model. In many cases, the measured pesticide concentration represents only a single point in time and may not reflect the full internal exposure history of the animal. Because pesticide residues in wildlife tissues can decrease over time through elimination, metabolism, growth dilution, or other dissipation processes, the steady-state assumption may lead to uncertainty or bias in the estimated intake. Future studies should consider incorporating toxicokinetic dissipation parameters, such as elimination rates or biological half-lives, or collecting repeated biomonitoring measurements over time to better characterize temporal variation in internal exposure. Consequently, the resulting hazard quotients and ecological carcinogenic risks should be interpreted as screening-level indicators rather than definitive risk determinations. Second, extrapolating PBK model parameters introduces uncertainty. The BTF and other species-specific input parameters were largely derived from laboratory or domesticated animals. Applying these values to free-ranging wildlife inevitably adds uncertainty. Although allometric scaling and species-specific data were used wherever possible, unmeasured interspecific differences in metabolic pathways, tissue partitioning, and elimination rates may bias the predictions. Sensitivity analyses confirmed that body weight and intake rates are major drivers of uncertainty, underscoring the need for improved species-specific inputs and independent validation. Third, limitations exist in plant exposure modeling. In the absence of measured pesticide concentrations in wild plants, this protocol applies a plant PBK transfer model originally designed for fruit trees, supplemented by the Trapp model. This dual-model approach helps fill data gaps but has not been fully validated for wild herbaceous species, and the associated uncertainty requires further evaluation. Fourth, geographic data coverage is uneven. The 100 studies compiled in the original application of this protocol covered only 30 countries, with sampling heavily concentrated in North America, Europe, and parts of Asia, while Africa, Oceania, and Central Asia remain severely under-represented. Therefore, the identification of high-risk regions largely reflects the distribution of available literature rather than the true global picture of pesticide contamination. The elevated risk estimates observed in ecologically protected areas also reveal critical gaps in current monitoring systems, particularly regarding wildlife tissues and remote ecosystems. Finally, the contamination score is relative by nature. Contamination score is a comparative index derived from global mean concentrations and does not correspond to any toxicological safety threshold. A high contamination score value for a given region indicates that contamination levels are above the global average but does not directly imply actual ecological risk. The level of risk must be assessed jointly with the ADD and toxicity benchmarks. Troubleshooting Problem 1 Extreme concentration values bias the contamination score (S). [See Step 6]. Extremely high or low pesticide concentrations at a few sampling sites may strongly influence the log-transformed deviation from the global mean, leading to overestimation or underestimation of the regional contamination score. Potential solution Examine pesticide-level distributions and identify outliers using boxplot or Grubbs’ test. Perform sensitivity analysis by recalculating contamination scores after removing values above the 99th percentile or below the 1st percentile. Report both the original and trimmed scores to show the influence of extreme values. Problem 2 Missing BTF for a target wildlife species. [See Step 8]. The required BTF for a specific species and tissue (e.g., fox liver) is not available in the supplementary tables or published literature, preventing direct back-calculation of ADD. Potential solution Use allometric scaling based on body weight and metabolic rate from a closely related species with known BTF. Alternatively, adopt a default BTF from a surrogate species (e.g., use rat BTF with a safety factor of 2) and explicitly document the extrapolation method. Perform a sensitivity analysis by varying the BTF by ±50% to assess its impact on the final HQ/ECR. Problem 3 Inconsistent results between Tier 2 and Tier 3 exposure estimates for the same species and region. [See Step 9 and Step 10]. When both measured plant concentrations (Tier 2) and measured soil concentrations (Tier 3) are available for the same country, the calculated ADD values for herbivores or carnivores may differ substantially, making it difficult to decide which estimate to report. Potential solution Always prioritise Tier 2 over Tier 3 because Tier 2 uses measured plant data, which is one step closer to the actual exposure pathway. If the discrepancy is large (e.g., more than one order of magnitude), investigate possible causes: differences in sampling sites, sampling years, or unit conversions. As a sensitivity measure, report both estimates in supplemental information and discuss the range in the main text, using the more conservative (higher) estimate for risk screening. Problem 4 Unavailability of environmental or physiological parameters required for the Trapp soil-plant transfer model (Tier 3). [See Step 10]. The transpiration rate model requires air temperature, relative humidity, leaf surface area, stomatal resistance, and other parameters. These data may not be available for many sampling sites, especially in remote or under-studied regions. Potential solution Use default values provided in Table S1 for parameters such as leaf mass per area, conductance, and water content. For meteorological parameters (temperature, humidity), use long-term averages from the nearest WorldClim or ERA5 grid cell. If multiple parameters are missing, consider a simpler empirical BAF (e.g., from literature) and treat the resulting plant concentration as a high-uncertainty estimate, flagged in the results. Problem 5 Monte Carlo simulation does not converge or requires excessive computation time when many input parameters are sampled. [See Step 21]. The standard 10,000 iterations may be insufficient if the output distribution has a long tail, or conversely, may be too slow when the model is complex (e.g., Tier 3 with many parameters). Potential solution Increase the number of iterations to 50,000 or 100,000 and check convergence by tracking the running mean of the median and 95th percentile. For speed, use parallel processing (e.g., foreach in R) or reduce the number of parameters by fixing those with very small variance (CV < 10%). Alternatively, replace full Monte Carlo with a Latin Hypercube sample of 2,000–5,000 runs, which often achieves similar accuracy with fewer iterations. Resource availability Lead contact Further information and requests for resources should be directed to and will be fulfilled by the lead contact, Zijian Li (lizijian3@mail.sysu.edu.cn). Technical contact Technical questions on executing this protocol should be directed to and will be answered by the technical contact, Zijian Li (lizijian3@mail.sysu.edu.cn). Materials availability This study did not generate new unique reagents. Data and code availability The article includes the code needed for the protocol, and any further questions can be directed to the lead contact. Acknowledgments This work was financially supported by the National Natural Science Foundation of China (32472598) and the Shenzhen Science and Technology Program (JCYJ20250604174437049). Author contributions S.C., writing – review and editing, writing – original draft, methodology, and data curation. Z.L., writing – review and editing, writing – original draft, methodology, funding acquisition, data curation, and conceptualization. Declaration of interests The authors declare no competing interests. Declaration of generative AI and AI-assisted technologies in the writing process During the preparation of this work, the authors used DeepSeek and ChatGPT in order to improve readability and language. After using this tool/service, the authors reviewed and edited the content as needed and take full responsibility for the content of the publication. Footnotes Supplemental information can be found online at https://doi.org/10.1016/j.xpro.2026.104824. Supplemental information Document S1. Tables S1–S3 Document S2. Article plus supplemental information References Associated Data Supplementary Materials Document S1. Tables S1–S3 Document S2. Article plus supplemental information Data Availability Statement The article includes the code needed for the protocol, and any further questions can be directed to the lead contact.