Herbicide leakage into seawater impacts primary productivity and zooplankton globally

Millions of tons of herbicides are applied annually, with 70% eventually reaching the ocean. Many herbicides inhibit photosynthesis, posing a significant threat to marine phytoplankton, which contribute approximately 50% of global primary productivity and are crucial for carbon cycling. Previous research demonstrated the sensitivity of some dominant offshore phytoplankton species to triazine herbicides, even at ambient concentrations. While individual studies have shown herbicide toxicity to microalgae, understanding the *in situ* ecological effects on phytoplankton communities and their impact on primary productivity requires a broader approach. This study addresses the gap in knowledge by assessing the global distribution of herbicides in coastal waters and quantifying their impacts on marine primary productivity and the food web.

Extensive research exists on the toxicity of herbicides to microalgae, demonstrating significant effects on photosynthetic physiology, nutrient uptake, and gene expression. However, extrapolating these findings to *in situ* ecological effects is challenging. Marine primary productivity is a complex system involving various phytoplankton groups with different growth rates, substrate preferences, and ecological roles. The impact of herbicides cannot be fully understood by examining single species or population responses alone. The community structure and particle size distribution of phytoplankton are key indicators of primary productivity; the effects of herbicides on these aspects are crucial to determine. Existing research on herbicide toxicity often focuses on single herbicides in small regions, lacking a comprehensive understanding of global coastal water pollution. Assessing the cumulative toxicity of various herbicides at environmentally relevant concentrations, given the complex composition of herbicides in seawater and additive toxicity among herbicides with the same mode of action, presents a major challenge.

This study involved multiple steps: 1. **Data Collection:** Spatiotemporal distribution data of herbicides (1990-2022) were collected from published surveys at 661 bay and gulf stations worldwide. Triazine herbicides were the focus due to their widespread distribution and known effects on phytoplankton. 2. **Toxicity Equivalent Conversion:** A toxicity equivalence database was established for 12 triazine herbicides using *Phaeodactylum tricornutum* (a model diatom) to determine their relative toxicities. The concentration-response curves were fitted to logistic or Weibull equations. Concentrations of each herbicide were then converted into atrazine toxic equivalent quantities (TEQs) using a concentration addition model. This normalization enabled a comparison of herbicide effects across different locations. 3. **Microcosm Experiments:** A controlled microcosm system was established using natural seawater with four atrazine concentrations (0, 0.5, 5, and 50 nmol L⁻¹). Changes in phytoplankton community structure, particle size distribution (pico-, nano-, micro-phytoplankton), chlorophyll *a* concentration, growth rates (intrinsic and net), and zooplankton grazing rates were assessed using high-throughput sequencing, microscopic observations, and classical dilution experiments. 4. **Herbicide Risk Assessment:** A global geographic distribution map of herbicide risk was created using data from PEST-CHEMGRIDSv1 and FAOSTAT databases. Risk scores were calculated based on herbicide application rates, crop-specific absorption, soil coverage, and residue persistence. Areas with high-risk scores were identified. 5. **Correlation Analysis:** Herbicide residues in waters adjacent to high-risk agricultural areas (Bohai and Yellow Seas) were measured, and correlations between herbicide use, residues, and phytoplankton primary productivity inhibition were investigated.

The study found that triazine herbicides were widely detected across the sampled sites, with high concentrations in the Gulf of Mexico, East Asia, and Vilaine Bay. Triazine herbicide concentrations showed an increasing trend over time in certain regions. The toxicity of the 12 triazine herbicides to *P. tricornutum* varied significantly. The TEQ analysis revealed that high equivalent atrazine concentrations were found in Camps Bay and Maputo Bay (South Africa), the Yellow Sea and Bohai Sea (Asia), and the Gulf of Mexico. High-throughput sequencing revealed that atrazine exposure significantly altered phytoplankton community structure, shifting dominance from Bacillariophyta (particularly *Chaetoceros*) to Dinophyceae, and from nano- to micro-phytoplankton. Chlorophyll *a* concentrations (a proxy for primary productivity) decreased significantly with increasing atrazine concentration, with the most substantial decrease observed in nano-phytoplankton. Atrazine also reduced the intrinsic growth rates of all phytoplankton size classes. Zooplankton community composition shifted from copepod larvae to smaller ciliates with increasing atrazine concentration, possibly due to food web disruption rather than direct toxicity. The study quantified the inhibition of phytoplankton primary productivity at various locations, finding that more than 5% of sites exhibited primary productivity inhibition greater than 5%, with 10% showing >10% inhibition. A strong correlation was observed between herbicide usage in agricultural areas and herbicide residues in adjacent coastal waters.

The findings demonstrate that herbicide runoff into coastal waters has significant, widespread impacts on marine primary productivity and the food web. The shift in phytoplankton community structure toward more resistant, but less productive species, coupled with altered particle size distribution and reduced growth rates, contributes to the overall decline in primary productivity. The observed changes in zooplankton communities further highlight the cascading effects of herbicide pollution. The results underscore the need for comprehensive approaches to assessing the cumulative toxicity of complex herbicide mixtures and considering the indirect effects on higher trophic levels. The strong correlation between agricultural herbicide use and coastal water pollution emphasizes the importance of sustainable agricultural practices to mitigate this environmental problem.

This study provides the first global-scale assessment of herbicide impacts on marine primary productivity, demonstrating significant and widespread negative effects at environmentally relevant concentrations. The altered phytoplankton community structure, particle size distribution, and zooplankton communities underscore the complex, cascading impacts of herbicide pollution. Future research should focus on exploring the long-term consequences of these impacts, investigating the interactions between herbicides and other environmental stressors (e.g., nutrient pollution, climate change), and developing effective strategies for reducing herbicide runoff into coastal waters.

While the study used a large dataset and controlled microcosm experiments, some limitations exist. The toxicity equivalence conversion, while useful, may not fully capture the complexities of interactions among different herbicides. The microcosm experiments may not perfectly replicate the diverse conditions found in natural ecosystems. The study primarily focused on triazine herbicides; other classes of herbicides may have different effects. The short-term nature of the microcosm experiments might not fully capture long-term impacts on marine ecosystems.