Globally, aquatic systems contribute significantly to methane (CH₄) emissions, a powerful greenhouse gas. Small water bodies (<0.1 ha) emit disproportionately high amounts of CH₄ per unit area, accounting for a substantial portion of total lentic methane emissions despite covering a relatively small surface area. Many of these small systems are human-constructed for agricultural purposes, leading to a proliferation of anthropogenic water bodies that significantly impact global biogeochemical cycles. Agricultural ponds, also known as farm dams or dugouts, are small constructed water bodies (typically 0.01-1 ha) characterized by some of the highest per-area CH₄ emissions among freshwater ecosystems. These emissions are amplified by high concentrations of fertilizers and manure runoff, shallow depths leading to rapid warming, and increased organic matter, all contributing to ideal conditions for CH₄ production. Critically, these emissions are anthropogenic and should be included in national carbon inventories. While the Intergovernmental Panel on Climate Change (IPCC) guidelines now encourage the inclusion of agricultural ponds, a lack of data on their abundance and distribution hinders accurate incorporation into national greenhouse gas inventories. This study aims to address this knowledge gap by providing a first-order assessment of CH₄ emissions from agricultural ponds in the United States and Australia.

The literature reveals a significant knowledge gap regarding methane emissions from agricultural ponds. While some studies have highlighted the high per-area emissions from these small water bodies, the available data is insufficient to accurately incorporate them into national greenhouse gas inventories. Previous research has established a link between high nutrient concentrations from fertilizer and manure runoff and increased GHG emissions. Studies also indicate the importance of temperature in influencing CH₄ production and flux rates. The IPCC has recently updated guidelines to include constructed water bodies in national inventories, but there's a substantial need for better data on their distribution and emission rates. Existing estimates on the total emissions from these ponds are likely underestimated, due to their omission in many national inventories.

This study employed a multi-step approach to estimate methane emissions from agricultural ponds in the US and Australia. First, spatial datasets on agricultural pond distribution were obtained from AusDams.org for Australia (1.7 million ponds) and the National Hydrography Dataset for the US (7.8 million). These datasets were refined to include only ponds between 0.01 and 1 ha in area located within specific land-use types (crops, open forests, etc.). The researchers then conducted a meta-analysis of published data on methane emissions from agricultural ponds (N=286) to determine average emissions. This data was supplemented with new measurements from 11 temperate agricultural ponds in Victoria, Australia. To account for variations in temperature across different locations, the researchers standardized the methane emission rates to 15°C using the Boltzmann-Arrhenius relationship, incorporating temperature sensitivity data from Rosentreter et al. (2021) and the relative contributions of diffusive and ebullitive methane fluxes. Annual median temperatures were sourced from MODIS Terra Land Surface Temperature data. A model was built and calibrated to map temperature-adjusted methane emissions. Finally, the results were compared to figures reported in the latest national GHG inventories submitted to the UNFCCC for 2020. Uncertainty analysis was conducted using non-parametric bootstrapping to estimate overall uncertainty of the analysis, taking into account variability in methane flux, temperature sensitivity, and the proportion of ebullition flux. The study also acknowledges limitations, including the potential underestimation of carbon dioxide and nitrous oxide contributions and the simplification of seasonal variability in temperature and pond surface area.

The meta-analysis showed that total temperature-adjusted methane emissions (diffusion + ebullition) from agricultural ponds varied considerably, ranging from <1 to >10³ kg CH₄ ha⁻¹ year⁻¹. The average emission at 15 °C was predicted to be 204 kg CH₄ ha⁻¹ year⁻¹ (95% CI: 83–521; median: 157.7), which is within 12% of the IPCC emission factor for freshwater and brackish ponds (183 kg CH₄ ha⁻¹ year⁻¹). However, the IPCC emission factor is temperature-independent, underestimating emissions in warmer climates; at 30 °C the model predicted an average emission of 405 kg CH₄ ha⁻¹ year⁻¹ (95% CI: 164–1037; median: 314.2), which is twice the IPCC factor. In the US, 2.56 million agricultural ponds covering 420.9 kha were estimated to emit 95.8 kt CH₄ year⁻¹ (95% CI: 61–157). In Australia, 1.76 million ponds covering 291.2 kha emitted an estimated 75.1 kt CH₄ year⁻¹ (95% CI: 47–123). Converting this to CO₂-eq using a global warming potential of 28, these emissions equate to 4.79 Mt CO₂-eq year⁻¹ (95% CI: 3.01–7.86). Comparison with UNFCCC reports revealed significant underestimation of emissions; US reported emissions were 46% lower, and Australian emissions were 54% lower than the study's estimates for agricultural ponds between 0.01 and 1 ha. A significant portion of the discrepancy might stem from the UNFCCC reporting guidelines which allow separate accounting of emissions from agricultural ponds and manure contamination within those ponds. The researchers noted that agriculture contributes a large proportion of overall methane emissions in the US and Australia, and agricultural ponds could be a substantial, currently unaccounted-for portion of that contribution. Uncertainty analysis revealed that the average methane flux, temperature sensitivity, and contribution of methane ebullition had the most significant uncertainty (CV between 20% and 28%).

The study's findings highlight the substantial underestimation of methane emissions from agricultural ponds in national greenhouse gas inventories. The significant discrepancy between the estimated emissions and the reported values underscores the need for improved accounting of these emissions. The considerable uncertainty associated with certain parameters highlights the need for further research to refine estimates. This research emphasizes the critical role of small water bodies in the global carbon cycle, highlighting the need for inclusion of these systems in future carbon emission accounting and climate change mitigation strategies. Management interventions, such as fencing to exclude livestock, could reduce nutrient loads and significantly lower CH₄ emissions, alongside other environmental benefits.

This study provides a continental-scale assessment of methane emissions from agricultural ponds, revealing that current national greenhouse gas inventories significantly underestimate these emissions. The researchers' estimates suggest that agricultural ponds in the US and Australia emit substantially more CH₄ than previously reported. This underestimation necessitates improved accounting methods and highlights the importance of incorporating these emissions into national inventories. Future research should focus on refining emission estimates, particularly by addressing uncertainty in key parameters and investigating the contribution of other greenhouse gases. Furthermore, developing and implementing cost-effective management strategies to reduce emissions from agricultural ponds while maintaining water security and biodiversity are crucial.

The study acknowledges several limitations. The analysis primarily focused on methane emissions, neglecting the contribution of other greenhouse gases like CO₂ and N₂O. The model uses 10-year average temperatures, simplifying seasonal variability in pond surface area and temperature. The sample sizes for certain parameters in the meta-analysis were relatively small, introducing uncertainty in the estimates. The assumption of similar temperature-dependency of methane ebullition between lakes, reservoirs, and ponds may introduce additional uncertainty. Future research is necessary to address these limitations and provide even more precise and comprehensive estimates of emissions from agricultural ponds.