Reducing risks of antibiotics to crop production requires land system intensification within thresholds

The Anthropocene has witnessed significant socioeconomic development driven by human activities, but this has also created considerable environmental risks, including chemical pollution and threats to food security. Antibiotics, used globally for disease prevention and animal growth promotion, are a growing concern. Agricultural practices, particularly manure fertilization and wastewater irrigation, are major sources of antibiotic emissions into soils. This soil antibiotic pollution negatively impacts agricultural ecosystems, inhibiting plant growth and reducing crop production. With increasing antibiotic emissions, the risk to crop production is projected to rise, highlighting the critical need for managing antibiotic pollution risks to ensure sustainable crop production. While the antibiotic footprint is increasingly recognized, the consequences of antibiotic dispersion in soil are largely unknown. Assessing soil antibiotic pollution risks on a massive scale is challenging, hindering comprehensive analysis of risks to crop production. Existing studies often have restricted geographic extents and don't directly characterize risks to crop production. This study aims to address these gaps by investigating the scale-dependent effects of land system intensification on soil antibiotic pollution risk in China, a major consumer of antibiotics.

Land use significantly contributes to soil antibiotic pollution, but the precise causes remain complex, partly due to land system changes across diverse spatial scales. Land system intensification, including arable land expansion and increased agricultural inputs, aims to boost crop yields and economic returns but can exacerbate soil antibiotic pollution. High manure application rates increase antibiotic loading in soil, and larger arable land areas increase the likelihood of receiving agriculture-originated antibiotics. Understanding the relationship between antibiotic pollution and land system regimes, along with identifying thresholds for sustainable land system intensification, is crucial for risk management. Previous research suggests land systems distinctly affect antibiotic pollution risks from regional to global scales, emphasizing the need for sustainable land system intensification to control pollution at multiple scales. However, few studies have investigated how land system intensification across different spatial scales affects antibiotic pollution and its impact on crop production.

This study used a multi-faceted approach to analyze the impacts of land system intensification on soil antibiotic pollution risk in China. It involved the following steps: 1. **Data Collection:** The study compiled a comprehensive dataset of measured environmental concentrations (MECs) of nine frequently detected antibiotics in Chinese soil, including tetracycline, chlortetracycline, oxytetracycline, doxycycline, ofloxacin, norfloxacin, ciprofloxacin, enrofloxacin, and lomefloxacin. Additional soil samples were collected from Yunnan and Zhejiang provinces to supplement the existing data. Data on various anthropogenic factors (livestock density, human population density, GDP, chemical fertilization, pesticide use), climatic factors (temperature, precipitation), soil properties (clay content, organic carbon, bulk density, soil thickness, saturated hydraulic conductivity), and vegetation characteristics (groundwater table depth, normalized difference vegetation index) were also collected. All data were processed at a 1km spatial resolution. 2. **Risk Assessment Modeling:** The risk quotient (RQ) approach was used to assess the risks of the target antibiotics to crop growth. PNECs (predicted no-effect concentrations) were estimated using assessment factors based on toxicity data from the literature. RQs were calculated as the ratio of MEC to PNEC. Cumulative risks were determined by summing the highest RQs for each antibiotic. An ensemble random forest (RF) model was used to scale up the antibiotic risk assessment across China, using the collected data as predictors. A Monte Carlo approach was used for uncertainty analysis, creating 500 RF models to generate a distribution of model error. 3. **Scale-Dependent Analysis:** To analyze scale-dependent effects, four watershed levels were selected from the HydroSHEDS dataset to represent different spatial grains. Additionally, twenty datasets were generated by systematically excluding segments from the upper and lower ends of the dataset ranked by increasing human footprint, representing a gradient of human impacts. Linear regressions were used to evaluate relationships between antibiotic pollution risks and the set of explanatory variables. The contributions of land use and management to antibiotic pollution risks were quantified based on variance explained. 4. **Risk-Yield Tradeoff Analysis:** Generalized additive models (GAMs) were used to model the nonlinear relationships between antibiotic pollution risk and crop yields (maize, rice, wheat, vegetables). Risk-yield trade-offs were calculated by standardizing both variables and subtracting the standardized risk from the standardized yield. The moving window approach was used to analyze the effects of land system intensification on risk-yield trade-offs across different scales. Thresholds of land system intensification where the risk-yield trade-offs peaked were identified using the segmented package in R based on the development Kuznets curve hypothesis.

The study's key findings are: 1. **Widespread Antibiotic Pollution Risk:** The average cumulative RQ of antibiotics in Chinese soil was 6.1 ± 2.1, with high levels in central and eastern China. Approximately 11.4% of the land was contaminated by more than one antibiotic compound. Ofloxacin posed the highest risk. 2. **Nonlinear Impact on Crop Production:** The study found nonlinear relationships between soil antibiotic pollution risk and crop yields (maize, rice, wheat, vegetables). Crop yields significantly decreased when the cumulative RQ exceeded 8.30–9.98. 3. **Scale-Dependent Land System Effects:** Land systems (land use and management) were strongly correlated with antibiotic pollution risks across all scales, more so than population or economic indicators. The contributions of land use and management to antibiotic pollution varied with scale. At smaller scales, the contribution of arable land to antibiotic pollution was much higher (59.5 ± 12.7%) than at larger scales (12.9 ± 3.2%). 4. **Risk-Yield Tradeoffs and Thresholds:** The study revealed nonlinear relationships between land system intensification and risk-yield trade-offs, showing an inverted U-shaped curve. Thresholds of land system intensification where risk-yield tradeoffs peaked were identified. Vegetables and wheat showed higher thresholds for manure fertilization, irrigation, and arable land proportion compared to maize and rice, particularly at small scales.

The study's findings demonstrate the widespread risk of soil antibiotic pollution to crop production in China, highlighting the complex interplay between land system intensification and environmental consequences. The scale-dependent nature of the land system's effect on antibiotic pollution underscores the importance of considering spatial context in management strategies. The nonlinear risk-yield tradeoffs highlight the need to manage land systems intensification below identified thresholds to balance the benefits of increased crop production with the risk of antibiotic pollution. The higher thresholds for vegetables and wheat at smaller scales might reflect resilience in these crops to antibiotic pollution, but this also implies that managing these crops effectively requires careful attention to small-scale land management practices. The study's results demonstrate the need to integrate environmental considerations into strategies to increase agricultural productivity, and further research should focus on exploring interventions that enhance soil resilience to antibiotic pollution.

This study provides crucial insights into the scale-dependent relationship between land system intensification, soil antibiotic pollution, and crop production in China. The identification of thresholds for land system intensification, where the benefits of increased production outweigh the risks of pollution, offers valuable guidance for policymakers and land managers. Sustainable land system intensification strategies that remain below these thresholds are needed to mitigate the negative effects of antibiotic pollution while ensuring food security. Future research should focus on refining the risk assessment model, incorporating the spread of antibiotic resistance, and integrating biotic factors into the analysis to enhance the accuracy and predictive capabilities of the model.

Several limitations should be considered when interpreting the results. The antibiotic pollution risk might be overestimated due to the use of worst-case scenarios in risk assessment. The study primarily focused on the exposure risks to crops, neglecting the spread of antibiotic-induced resistance and overlooking the potential influence of biotic factors. The model assumed the absence of unreported antibiotics. Finally, the study relied on modeling results without experimental validation, lacking a mechanistic understanding of the processes involved.