Main

Wastewater treatment is energy intensive, consuming an estimated 0.8–4% of global electricity1,2. Meanwhile, it is a major source of CH4 and N2O emissions and contributes to approximately 5–6.5% of global non-CO2 greenhouse gas (GHG) emissions3,4. As energy-related CO2 emissions are expected to stabilize or decline with the transition to low-carbon systems5,6,7, the relative importance of non-CO2 GHGs is projected to rise in the coming decades. Their climate and co-pollutant impacts are profound: CH4 accelerates tropospheric ozone formation, threatening health and crops8, while N2O contributes to eutrophication and water quality degradation9. Together, these environmental and human health impacts create strong incentives for robust GHG accounting, which forms the foundation for credible climate baselines, effective mitigation strategies and progress towards net-zero targets10.

However, accurate quantification of fugitive GHG emissions from the wastewater sector remains challenging as they arise across a heterogeneous network of wastewater pathways11,12: from uncollected/untreated sanitation (that is, latrines, open defecation), decentralized systems (for example, septic tanks), to collection systems (flowing and stagnant sewers) and centralized treatment (treatment facilities, including sludge handling) and discharge of treated or untreated effluent, leakages and overflows. While the Intergovernmental Panel on Climate Change (IPCC) guidelines pave the fundamental methodologies to account for fugitive GHG emissions from those sources, not all wastewater systems/pathways are fully covered due to the fact that data to quantify their emissions are often insufficient or unavailable (Fig. 1a–c). Moreover, these pathways operate over diverse technologies, climates and operating regimes, producing emissions that can vary by orders of magnitude13,14,15,16. However, the widely adopted uniform emissions factors (EFs) recommended by the IPCC guidelines are derived from a limited number of measurement campaigns that fail to capture this high variability and may lead to systematic over- or underestimations of existing GHG inventories across all levels, from facility scale all the way up to national and global scales13,14,15,17. To address these limitations, an increasing number of field-measurement studies have been conducted to improve the understanding of GHG emissions variability from water resource recovery facilities (WRRFs). For example, field measurements at over 100 US WRRFs revealed facility-level CH4 EFs ranged from 0.005 to 0.05 g CH4 per gram of biochemical oxygen demand (BOD)14, whereas N2O EFs show much wider fluctuations, ranging from 0.00003% to 20.7% kg N2O-N per kilogram of total nitrogen (TN) (refs. 13,18). Gruber et al.19 carried out long-term measurement campaigns across a dozen WRRFs in Switzerland and proposed weighted EFs based on key facility characteristics (for example, carbon removal, nitrification only and full nitrogen removal) to better represent diverse N2O emissions profiles and reduce uncertainties associated with applying uniform EFs.

Fig. 1: Coverage of wastewater pathways in the IPCC guidelines.
figure 1

a, Complex wastewater treatment systems and discharge pathways and overall coverage of these pathways in the IPCC 2006 and 2019 guidelines. b,c, Detailed EFs suggested in the IPCC guidelines to estimate CH4 (b) and N2O (c) emissions. Multiple EFs under each pathway meaning different treatment systems and discharge pathways are considered. WRRF_sludge refers to emissions from onsite sludge disposals, for example, leakage from anaerobic digestion, composting, sludge-drying beds, flaring emissions and so on. DS denotes dry solid matter in treated sludge.

Source data

Full size image

In addition to limitations in methodological accuracy and completeness, the existing patchwork of GHG inventories also suffer from inconsistencies and insufficient transparency10. Although the IPCC guidelines form the basis of GHG inventories, high flexibility is given to entities in adopting country-specific assumptions, EFs, activity data, among others, resulting in a wide range of heterogeneity in coverage of wastewater pathway and accounting methodology20. Several global databases apply harmonized accounting frameworks to track anthropogenic GHG emissions and improve cross-country comparability, such as Global Water Intelligence (GWI)21, US Environmental Protection Agency (EPA) report3 and the Emissions Database of Global Atmospheric Research (EDGAR)4. But they rarely specify the wastewater pathways included or the detailed accounting methods, making it difficult to explain the large discrepancies in wastewater-sector GHG emissions estimates, which range from 410 to 768 million metric tons (MMT) of CO2-equivalent (CO2e) annually across these databases.

The extent to which incomplete coverage of wastewater pathways and methodological discrepancies in GHG inventories compromise global emissions benchmarks remains unclear, highlighting a major knowledge gap in this field. Here we conduct a global synthesis of wastewater-sector GHG accounting methods adopted in national inventory reports (NIRs). We systematically compare and identify discrepancies in methodologies, coverage of wastewater pathways and adopted EFs and quantify the gaps these choices create in nationally reported emissions. NIRs are the best entry points for improving fugitive GHG reporting, given their broader coverage and widespread adoption among stakeholders. A total of 38 countries providing explicit pathway descriptions and/or disaggregated CH4 and N2O emissions data by pathways were analysed, including 30 Annex I and eight non-Annex I countries. These countries span five continents and various socio-economic contexts, collectively accounting for approximately 36% and 67% of sectoral emissions reported in EDGAR4 and GWI21, respectively.

Specifically, direct CH4 and N2O emissions from the following five main groups of wastewater pathways are reported in the wastewater sector according to IPCC guidelines: including (1) latrines; (2) septic systems or other decentralized systems (for example, constructed wetlands, small community-scale treatment facilities); (3) centralized WRRFs—both water and sludge (for example, dewatering and storage) lines and onsite sludge disposal (for example, incineration and composting); (4) effluent—discharge of treated wastewater and (5) untreated wastewater—discharge of untreated wastewater from scattered households, overflows or leakages and so on (Fig. 1a). By exposing where coverage is incomplete and where methods diverge from IPCC guidance, our analysis provides an evidence base for improving the comparability, completeness and transparency of wastewater GHG inventories. These insights can directly inform future updates to IPCC guidelines, refining NIR reporting practices, enabling more credible mitigation planning, resource allocation and target setting by governments, cities, utilities and industry stakeholders.

Inconsistent methodology and pathway coverage

For the latest NIRs of 38 countries studied, accounting methods and reporting timelines show substantial inconsistencies (Fig. 2a), highlighting the need for caution when using NIR-reported emissions to compare national total emissions across the countries or collectively sum up to provide global estimates of wastewater GHG emissions. Approximately 50% of the Annex I countries have adopted the IPCC 2019 Refinement, whereas only one non-Annex I country—Mexico—has applied this refined methodology. This observation makes sense, as Annex I and developed countries have historically contributed the most to global GHG emissions and are expected to take more stringent climate actions. Such efforts benefit from more accurate and comprehensive GHG inventories to inform policymaking. Notably, applications of the IPCC 2019 Refinement among the studied Annex I countries have increased from five countries in 2023 to 13 countries in 2025, which is likely to rise in the coming years. In terms of reporting timelines, all Annex I countries submitted emissions data through 2023 (except the United States which reported through 2022), reflecting the typical 2–3-year lag for data processing and review. In contrast, non-Annex I countries exhibit profound discrepancies with a longer time lag (Fig. 2a). For example, the most recent emissions data submitted in Egypt’s NIRs are approximately a decade out of date. Given the fast development of these emerging economies, an updated inventory would be urgently needed.

Fig. 2: Reported emissions and wastewater pathways being covered in the latest inventory year.
figure 2

a, NIR-reported emissions for domestic wastewater treatment. Countries labeled in the figure use the IPCC 2019 Refinement and/or report a latest inventory year other than 2023; unlabeled countries follow the IPCC 2006 Guidelines and report 2023 as the latest inventory year. Note that due to lack of granular data in China’s NIR, its reported emissions are synthesized from multiple peer-reviewed literatures quantifying national emissions by wastewater pathways, including latrines and septic tanks from ref. 45, collection systems from ref. 46, and centralized WRRFs and effluent from ref. 47. b, Wastewater pathways included in NIRs by country and gas. WRRF_sludge refers to emissions from onsite sludge disposals, for example, leakage from anaerobic digestion, composting, sludge-drying beds, flaring emissions and so on. c, Proportion of population connected to different wastewater pathways. The proportion of WRRF effluent is not shown, as all treated wastewater from WRRFs is assumed to be discharged and thus equal to the proportion of WRRFs.

Source data

Full size image

Additionally, coverage of wastewater pathways varies widely across countries, with CH4 inventories including one to four pathways and N2O inventories covering one to five pathways (Fig. 2b). For CH4 assessments, the most frequently included pathways are septic systems (37 of 38 countries) and WRRFs (33), followed by untreated wastewater (15), latrine (15) and WRRFs effluent (11). There are 13 countries also accounted for CH4 emissions from onsite sludge disposal, denoted as WRRF_sludge in Fig. 2b, which includes emissions from anaerobic digestion (AD) leakage, composting, sludge-drying beds or inefficient flaring. For N2O emissions, the most commonly covered pathway is WRRFs effluent (37 of 38 countries), followed by WRRFs (23), septic systems (6), untreated wastewater (4) and latrine (1). The exclusion of certain pathways is primarily attributed to two reasons. First, it aligns with the IPCC 2006 guidelines, which assume no CH4 emissions from well-managed, centralized aerobic WRRFs and omit N2O emissions from septic tanks and latrines. Second, certain pathways are classified as non-key categories and therefore omitted from reporting. For example, Indonesia excludes WRRFs from its inventory because of their limited contribution to national wastewater treatment (Fig. 2c). Generally, latrines and untreated wastewater are rarely included in NIRs. In countries where these pathways represent a minor fraction of total wastewater flows (for example, Denmark, Sweden and UK), their exclusion may have negligible impact on overall emission estimates. However, in many emerging and developing countries where these pathways remain widespread (for example, Egypt and Turkey), such omissions probably result in a large underestimation of wastewater GHG emissions.

The number of wastewater pathways included in NIRs does not necessarily reflect the extent of coverage, as it also depends on the total number of wastewater pathways in use within a country. For example, if a country covers three wastewater pathways in its NIR, the percentage of wastewater pathways being covered is probably varying between 60% to 100%, depending on whether five or three wastewater pathways are present. Among the 38 countries studied, the percentage of pathways being covered ranges from 25% to 100% for CH4 (Fig. 3a) and falls into 20% to 100% for N2O (Fig. 3b). Specifically, Switzerland is the only country that has achieved comprehensive coverage of all wastewater pathways for both CH4 and N2O. Japan has achieved full coverage for N2O. By comparing coverage across different economic groups, we found that average coverage of pathways for CH4 is comparable between developed and developing countries, whereas developing countries generally exhibit lower N2O coverage (Fig. 3c). These findings suggest major progress has been made in wastewater CH4 accounting, while opportunities exist for targeted improvements in N2O emissions quantification.

Fig. 3: Percentage of wastewater pathways being covered in NIRs.
figure 3

a,b, Coverage of wastewater pathways for CH4 (a) and N2O (b) emissions estimates in NIRs among the countries studied. Numbers in panels a and b indicate total pathways considered in NIRs, with actual total wastewater pathways in parentheses. c, Descriptive statistics of CH4 and N2O coverage percentage by economic levels. In c, dots indicate coverage of wastewater pathways in each country. Boxplots show 25th, 50th and 75th percentiles, and outlier bounds are based on 1.5× interquartile range (IQR; IQR equals to the 75th percentile minus the 25th percentile). Yellow dots represent the arithmetic means. Violins represent probability distributions using kernel density estimation. Coverage percentage is calculated as the ratio of the total pathways considered in NIRs to the total actual number of wastewater pathways, multiplying by 100. n is the sample size of coverage scores for each gas or the number of countries in each economic group because each country has one CH4 and one N2O coverage score. Basemaps in a and b from Natural Earth.

Source data

Full size image

Varying emissions factors adopted by different countries

Regarding EFs adopted in NIRs, they are typically sourced from three categories: IPCC default values, estimated EFs derived through expert judgements or tailored estimation to reflect country-specific conditions and measured EFs based on country-specific field measurements (Fig. 4a,b). Generally, measured and estimated EFs are more representative of local environmental and operational conditions and thus critical for improving inventory accuracy. However, such EFs remain limited—mostly for septic systems and WRRFs—while other pathways, such as latrine and discharge of treated or untreated wastewater, largely rely on default EFs (Fig. 4c,d).

Fig. 4: Emissions factors adopted in NIRs.
figure 4

a,b, Varying EFs used for CH4 (a) and N2O (b) emissions estimates across countries and wastewater pathways. Multiple EFs under each pathway meaning different treatment systems and discharge pathways are considered. c,d, Total number of distinct EFs applied in studied NIRs for CH4 (c) and N2O (d).

Source data

Full size image

Specifically, Japan is the only country that has measured both CH4 and N2O emissions from decentralized systems, reporting CH4 EFs of 0.0042–0.17 kg CH4 kg−1 BOD and N2O EFs of 0.007–0.02 kg N2O-N kg−1 TN22. But the widely used system in Japan is Johkasou rather than septic tank, serving 18.2% of the population in 2023. For septic tanks, measured CH4 EFs range from 0.113 (refs. 23,24) to 0.132 (ref. 25) kg CH4 kg−1 BOD according to field measurements conducted in Denmark, Sweden and the USA (Fig. 4a). These measured values equal to approximately half the IPCC default value of 0.3 kg CH4 kg−1 BOD, suggesting potential overestimation for countries adopting default EFs.

For WRRFs, six countries (that is, Germany, Japan, Mexico, Sweden, Switzerland and the UK) applied measured CH4 EFs (Fig. 4a). In particular, measured EFs for aerobic WRRFs ranging from 0.003 to 0.036 kg CH4 kg−1 BOD, indicating a substantial underestimation in those NIRs that adopted the IPCC 2006 guideline assumption that no CH4 emissions arise from well-managed aerobic facilities. This highlights the need for reporters to adopt IPCC 2019 default EFs (0.018 kg CH4 kg−1 BOD) or develop country-specific measurements that better reflect actual emissions.

Special attention should be paid to accurately account for CH4 leakages from facilities with ADs or onsite sludge disposal (for example, incineration, composting, sludge-drying lagoon26). Elevated CH4 emissions from facilities with ADs have been widely reported14,15. Among countries reporting AD-related emissions (Fig. 2b), three distinct methods are used: leakage rate, emissions per unit of dry solid (DS) and emissions per influent BOD. These values are derived from assumptions, national averages or field measurements, leading to wide variations. Specifically, reported leakage rate vary from 1.0% (USA)27 to 6.9% (Denmark)24. France, Netherlands and the UK estimate emissions per dry solid of sludge (0.002–0.018 kg CH4 kg−1 DS) (ref. 28), whereas others apply BOD-based EFs ranging from 0.003 to 0.48 kg CH4 kg−1 BOD.

N2O emissions from WRRFs vary considerably depending on treatment processes, operational conditions, wastewater characteristics, among other factors13,16. To capture this variability, several countries have adopted either national average EFs based on nationwide surveys covering diverse WRRF configurations or categorized WRRFs into distinct groups when developing country-specific accounting methods (Fig. 4b). Average EFs adopted in Austria, Denmark, Norway and Sweden range from 0.002 (ref. 29) to 0.009 (refs. 17,24,30) kg N2O-N kg−1 TN: at least five times larger than the IPCC 2006 default value (0.00032 kg N2O-N kg−1 TN) (ref. 31) but lower than the IPCC 2019 default value (0.016 kg N2O-N kg−1 TN) (ref. 12). Such high discrepancies highlight potential misestimation when using default EFs to estimate N2O emissions. Differentiating WRRFs into distinct groups, as reported in Japan and Switzerland, is considered a more refined approach to improve inventory accuracy. In Japan, extensive sampling surveys at 42 plants yielded four treatment-specific N2O EFs: anaerobic–anoxic–oxic reactor (1.6 × 10−4 kg N2O-N kg−1 TN), anaerobic–oxic reactor (3.8 × 10−4), activated sludge (1.8 × 10−3) and membrane bioreactor (1.7 × 10−5) (ref. 22). Similarly, Switzerland grouped WRRFs into carbon removal (0.04 kg N2O-N kg−1 TN), nitrification only (0.018) and year-round nitrogen removal (0.009) based on long-term measurements across over 14 facilities19. In practice, treatment technologies are more complex and vary widely across countries, highlighting the need for country- or facility-specific EFs. However, for countries in the absence of field measurements, adopting the IPCC 2019 default value or referring to measurements from neighbouring countries or global syntheses, may offer a more reasonable alternative than relying on the outdated IPCC 2006 default EFs.

Fig. 5: CH4 and N2O emissions gaps in the latest NIRs.
figure 5

a, Total emissions after adding CH4 and N2O emissions gaps to reported values. b, CH4 and N2O emissions gaps as a percentage of their respective reported values.

Source data

Full size image

Huge emissions gaps due to omitted pathways and outdated EFs

Given the widespread omission of wastewater pathways (Fig. 3) and the potential underestimation of WRRFs emissions in countries following the IPCC 2006 guidelines (Fig. 4), current NIRs are under-reporting emissions from wastewater sector. However, the extent to which these deficiencies compromise global benchmarks remains unknown. Clarifying this impact is important for improving the accuracy and comparability of national inventories, which form the basis of global stocktakes and mitigation planning. To quantify these emissions gaps arising from omitted pathways and outdated EFs, we applied EFs based on the latest understanding of emissions from each wastewater pathway (Table 1). Notably, onsite sludge disposal and stagnant/flowing sewers were excluded from this calculation due to lack of activity data.

Table 1 Methane and nitrous oxide emissions factors by wastewater pathways and corresponding assumptions
Full size table

Our estimates found an annual emissions gap of 73.2 MMT CO2e for the 38 countries studied, equivalent to 26.5% of total reported values in their latest NIRs for domestic wastewater treatment (Fig. 5a). This includes a CH4 emissions gap of 17.9 MMT CO2e (or 9% of the 201.8 MMT CO2e reported CH4 emissions) and a N2O emissions gap of 55.3 MMT CO2e (or 74% of the 74.2 MMT CO2e reported N2O emissions). Approximately 76% of these emissions gaps stem from the 12 emerging and developing countries, including Belarus, Russia, Turkey, Ukraine and all studied non-Annex I countries, which contribute to 10.0 and 45.3 MMT CO2e of the CH4 and N2O shortfalls, respectively.

Generally, CH4 and N2O emissions gaps are largely due to omitted or underestimated emissions from WRRFs, with additional contributions in some countries from unaccounted emissions from discharge of treated effluent (CH4) and septic systems (N2O), depending on national reporting practices (Fig. 5b). Whereas absolute emissions gaps are lower in developed countries, their relative impact is substantial and may alter current ranking of total emissions. In some cases, under-reported emissions could increase national wastewater GHG baselines by up to 450% for CH4 emissions, as observed in Spain and by up to 550% for N2O emissions as shown in Finland (Fig. 5b). It should be noted that when such numbers are extrapolated globally, the annual gap is approximately 150 MMT CO2e (or 46% of current reported emissions from domestic wastewater treatment), by incorporating CH4 emissions from discharge of treated effluent, N2O emissions from septic tank and updated WRRF EFs in the remaining unstudied countries with the assumption that unstudied countries follow the IPCC 2006 guidelines as recommended by the United Nations Framework Convention on Climate Change (UNFCCC).

Given the large variability in WRRF emissions, we also estimated such gaps using a relatively conservative approach, assuming a CH4 EF of 0.018 kg CH4 kg−1 BOD (IPCC 2019 default) and N2O EF of 0.008 kg N2O-N kg−1 TN (mean of measured EF in NIRs), while keeping EFs for other pathways unchanged. Under this condition, the annual emissions gap is 52 MMT CO2e across the 38 countries studied and about 94 MMT CO2e at a global scale. The precise quantification remains challenging due to the lack of national inventories and field data for these countries.

Best practices for accurate and comparable GHG inventory

The results show reporting biases due to limited coverage and methodological constraints make wastewater GHG baselines unreliable and hinder meaningful comparisons between countries. An estimated 19–27% annual emissions gap of currently reported wastewater emissions arises from widespread omission of key pathways and the use of outdated EFs. Closing this gap calls for an update to the IPCC guidelines to enable systematic accounting of emissions from all wastewater pathways, alongside the development of more representative EFs. Our dataset identifies countries with more adequate NIRs and those employing country-specific EFs based on field measurements. These countries can serve as noteworthy exemplars. For instance, Switzerland recently achieved full coverage for both CH4 and N2O by expanding its inventory to include decentralized systems that were previously overlooked. Although onsite sanitation serves only 1.3% of the Swiss population, it contributes an estimated 7% of total wastewater GHG emissions32, underscoring the importance of incorporating all relevant pathways. Japan also stands out for its full coverage for N2O emissions and its deliberate effort to account for emissions from sludge treatment, including sludge line in WRRFs, facilities treating human waste from decentralized systems and ocean dumping of sludge. To enhance methodological rigour, both countries have classified WRRFs into distinct categories based on measured EFs to better capture emissions variations that enhance both the accuracy and policy relevance of their national inventories. However, such progress depends on the availability of country-specific measurements and institutional capacity for data collection and interpretation.

Improving global wastewater GHG estimates require more robust activity data, along with expanded data-sharing platforms, harmonized data translation protocols and strengthened dialogue across technical and policy communities. Whereas several global wastewater datasets exist (for example, HydroWASTE33, JMP and other country-level estimates34), distinct inconsistencies across sources are common. More frequent updates are also needed, given the rapid evolution of wastewater treatment practices and corresponding GHG dynamics. Particularly, data on wastewater leakage remains sparse. One study quantified sewage leakage and identified approximately 56% (53%–59%) of domestic total nitrogen loads leak out without treatment35. Thus, an improvement of the leakage dataset not only enhances GHG inventory accuracy but also yields co-benefits for public health and environmental protection by effectively pinpointing mitigation strategies.

Considering the consensus reached at the IPCC Expert Meeting and aware of the broad scientific community’s responsibility to provide clearly defined data to policymakers36, there is a pressing need to ensure that wastewater GHG estimates are accurate, policy-relevant and comparable across different communities. For the IPCC community, there is a need to develop refined EFs with advanced understanding of emissions dynamics from the wastewater sector. Although the IPCC Emission Factor Database aims to serve as a centralized repository for EFs and associated parameters37, it currently heavily compiles default values. Incorporating more measured EFs from NIRs and peer-reviewed studies would substantially enhance its adoption. Additionally, clearer guidance on onsite sludge disposal is also needed, ideally consolidated within the same chapter rather than spread across multiple sectors. Strengthening enforcement of best practices is equally critical. These improvements should be integrated across the currently Seventh Assessment Report, including emissions pathways (Working Group III), climate system changes (Working Group I) and broader societal impacts (Working Group II).

At the national and subnational levels, greater attention to the wastewater sector is needed through collaboration with domestic stakeholders to improve data collection and by enhancing transparency around data limitations and methodological assumptions in inventories. Future efforts should specifically consider adding pathways frequently omitted such as decentralized systems and discharge of treated or untreated wastewater, especially where they represent a substantial share of wastewater flows. Facilitating cross- and intra-country learning and providing targeted technical support, especially for non-Annex I and developing countries, can help bridge data and capacity gaps. The IPCC is well positioned to lead these efforts, given its responsibility for both assessment reports and inventory guidelines. Taken together, these practices can improve the robustness of global wastewater GHG inventories, inform future revisions to IPCC guidelines and support more effective climate policies and mitigation strategies.

Methods

This study examines limitations of fugitive GHG (mainly CH4 and N2O) emissions reported for the wastewater sector in NIRs. Direct CO2 emissions are excluded because they are predominately biogenic and do not result in long-term net carbon additions. Although the IPCC notes that 4–15% of wastewater-related CO2 emissions derive from fossil sources (for example, soaps and detergents)48,49, no consensus method exists to quantify these non-biogenic CO2 emissions; hence, CO2 emissions from wastewater are not generally reported to NIRs.

System boundary

Five main groups of wastewater treatment systems and pathways contribute to CH4 and N2O emissions and are reported under the wastewater sector according to IPCC guidelines and sanitation service delivered, including (1) latrines which provide basic/limited sanitation services for households not collected to public sewers, (2) septic systems or other decentralized systems that treat uncollected wastewater, (3) centralized WRRFs—including both water and sludge lines and onsite sludge disposal; (4) WRRF effluent and (5) untreated wastewater discharged from scattered households, overflows or leakages. Onsite sludge disposal (for example, incineration and composting) is included within the WRRF category because it is an integral component of WRRF operations, and its emissions are reported under the wastewater sector according to the IPCC 2019 Refinement. Sludge disposed offsite is not considered in this study, as these emissions are attributed to other sectors. For example, CH4 emissions from disposal of sludge in a landfill are included in the solid waste sector, and N2O emissions from land application are reported under the agriculture sector.

Beyond the five main wastewater pathways, sewer collection systems are probably a major but under-characterized emissions source. The IPCC guidelines provide CH4 EFs only for open/stagnant systems (Fig. 1b). Although IPCC acknowledges flowing sewers as potential sources of CH4 and N2O, their emissions are assumed negligible due to insufficient data. Among all the NIRs available, only Switzerland reports emissions from flowing sewers by applying a mean value of 0.00255 kg CH4 kg−1 COD (ref. 50) according to peer-reviewed measurements51,52. This study does not address this topic given such limited consideration in current accounting practices. With more data on sewer types, operational conditions and monitored emissions, flowing sewers may be incorporated into future IPCC guidelines.

IPCC GHG accounting methodologies

Given the IPCC guidelines form the basis of fugitive GHG accounting methodologies, we compiled default EFs by wastewater pathways suggested in three main IPCC documents, including the IPCC 2006 Guideline, the IPCC 2013 Supplement and the IPCC 2019 Refinement (Fig. 1b,c and Supplementary Dataset). Wastewater pathways covered and EFs suggested in the successive IPCC guidelines are evolving with more field measurements and enhanced understanding of the magnitude and variations of GHG emissions from each pathway. For example, the CH4 EFs are mainly derived by multiplying the maximum potential methane emissions (Bâ‚€) and modifying methane correction factors. In the IPCC 2006 guideline, although it covers most wastewater pathways for CH4 emissions estimations (Fig. 1b), the methane correction factors are derived from expert judgement, and it is assumed that there are no CH4 emissions from well-managed, aerobic WRRFs and discharge of treated effluent. Methods for N2O emissions are only provided for treated effluent discharge and WRRFs with controlled nitrification and denitrification, based on one study of a single plant in the USA53. In comparison, the IPCC 2019 Refinement made substantial updates. Generally, it incorporated CH4 and N2O EFs for constructed wetlands through cross-referencing the IPCC 2013 Supplement, updated CH4 and N2O EFs for centralized WRRFs and septic tank based on dozens of peer-reviewed literatures and differentiated discharge-related EFs by receiving waterbody types to enable application of high-tier approaches (Fig. 1b,c).

NIR data collection

The most recent NIRs for the 38 studied countries, submitted by 15 July 2025, were downloaded from the UNFCCC portal (https://unfccc.int/ghg-inventories-annex-i-parties/2025). NIRs in French (N = 1), Spanish (N = 2), Portugal (N = 1) and Russian (N = 2) were translated into English using open-source software (Google Translate and ChatGPT). We reviewed the domestic wastewater treatment and discharge section in detail and collected data of interest, including wastewater pathways considered, methodologies (for example, EFs, population served by each wastewater pathway, organic inputs in wastewater and other activity data), reported CH4 and N2O emissions in kiloton (kt), emissions contribution of each pathway if data are available, among others. For comparison, EF units for CH4 and N2O are converted to kg−1 CH4 kg−1 BOD and kg−1 N2O-N kg−1 TN, respectively.

For China, its NIRs do not provide granular data on emissions from domestic wastewater treatments, the specific pathways included and detailed methodologies. We therefore supplement these data with national estimates from peer-reviewed literature. Given substantial differences in system boundaries, input data, EFs and methodological assumptions across studies, we solely compiled emissions data by wastewater pathways rather than comparing input data and EF discrepancies or uncertainties among these studies. Specifically, latrine and septic tank emissions are obtained from Cheng et al.45; WRRFs and effluent emissions from Wang et al.47 and wastewater collection system emissions from Gao et al.46. Data extracted were cross checked by team members to validate and enhance the accuracy of the information extracted. The collected data points were documented into a structured dataset as provided in the Supplementary Dataset.

In addition to the 38 countries analysed in this study, many other countries also submit NIRs to UNFCCC that probably contain methodological details relevant to wastewater GHG accounting. However, our objective is not exhaustively reviewing all NIRs but to identify their discrepancies in coverage of wastewater pathways and accounting methodologies among countries. The 38 selective countries, including 30 Annex I and eight non-Annex I countries, are considered representative that collectively contribute to approximately 36–67% of reported emissions for the wastewater sector. They span five continents and cover a wide range of economic contexts: developing and emerging economies (n = 12; eight non-Annex I countries plus four Annex I countries including Belarus, Russia, Turkey and Ukraine), the G7 developed economies (n = 7; all Annex I countries: Canada, France, Germany, Italy, Japan, UK and USA) and non-G7 developed economies (n = 19; the remaining Annex I).

Ratio of population served by various wastewater treatment pathways

The ratio of the population served by each pathway was obtained from the World Health Organization and United Nations Children’s Fund (WHO/UNICEF) Joint Monitoring Programme (JMP), which tracks the global percentage of the population using safely managed sanitation services over time (2000–2022; accessible from https://data.unicef.org/resources/dataset/drinking-water-sanitation-hygiene-database/). Household surveys and censuses are the primary source of JMP on the different types of facility used by the population. Given the rapid development of wastewater industry especially in developing countries, the latest JMP data in 2022 were used for the following analysis to capture the updated wastewater pathways in each country.

The WHO/UNICEF JMP classified sanitation services into four main groups: (1) improved sanitation facilities (that is, latrines and other, septic tanks, sewer connections, wastewater treated); (2) basic (improved facilities that are not shared with other households); (3) limited (improved facilities shared with other households) and (4) unimproved (use of pit latrines without a slab or platform) and open defecation. Although groups of these sanitation services differ from wastewater pathways defined in this study, underlining linkages exist and the ratio of population served by each pathway is calculated as follows.

$${r}_{\mathrm{latrine}}={\mathrm{JMP}}_{\mathrm{latrines}}$$
(1)
$${r}_{\mathrm{septic}}={\mathrm{JMP}}_{\mathrm{septic}\,\mathrm{tanks}}$$
(2)
$${r}_{\mathrm{WRRF}}={\mathrm{JMP}}_{\mathrm{wastewater}\,\mathrm{treated}}$$
(3)
$${r}_{\mathrm{effluent}}={\mathrm{JMP}}_{\mathrm{wastewater}\,\mathrm{treated}}$$
(4)
$${r}_{\mathrm{untreated}}={\mathrm{JMP}}_{\mathrm{sewer}\,\mathrm{connections}}-{\mathrm{JMP}}_{\mathrm{wastewater}\,\mathrm{treated}}+{\mathrm{JMP}}_{\mathrm{open}\,\mathrm{defecation}}$$
(5)

This calculation and categorization were double checked and validated through the following three steps. First, we summed the ratio of population served by each pathway. Most countries reached a total of 100%, while a few slightly lower but still exceeded 99.1%. Second, we compared JMP data with those reported in NIRs and identified several inconsistencies. For example, the JMP dataset reports no population using decentralized systems in the Netherlands and Switzerland, which contradicts the NIRs. To resolve this, we aligned our data with NIRs where possible. Third, we identified 11 countries that reported CH4 or N2O emissions from latrines or untreated wastewater in their NIRs, but ten of them had no corresponding population recorded for these pathways in the JMP dataset. In such cases, we manually adjusted the population share for latrines and untreated wastewater to match data reported in NIR if data were available; otherwise, we estimated them as the remainder after subtracting the sum of all other pathway shares from 100%.

We also cross checked the amount of treated wastewater by comparing the WHO/UNICEF dataset (calculated as treated wastewater per population equivalent (PE) × total PE × ratio of population served by WRRF) with HydroWASTE33 and Jones et al.’s dataset34, resulting in ± 40% differences primarily because of high uncertainties of rough estimates using treated wastewater per PE and percentage of population using WRRFs. Despite huge inconsistencies, we adopted the WHO/UNICEF JMP dataset for the following assessment as it is the most comprehensive wastewater dataset with granular data of all wastewater pathways. Further validation and uncertainty analysis are needed once a more accurate wastewater dataset becomes available.

Estimation of emissions gaps

Emissions gaps caused by both omitted pathways and underestimated CH4 and N2O emissions from WRRFs by following the IPCC 2006 guidelines were estimated in this study. These emissions gaps were estimated using updated EFs for septic tanks and WRRFs based on large amount of field measurements published in literature and IPCC 2019 default EFs for other pathways (Table 1), which reflect the latest understanding on emissions from each wastewater pathway. Given a lack of detailed activity data for each country such as treatment technologies and waterbodies that untreated and treated wastewater discharged, we did not consider these variations, and corresponding assumptions are listed in Table 1.

Field measurements have been widely shown that the IPCC 2006 default CH4 EF (0 kg CH4 kg−1 BOD) and N2O EF (3.2 g N2O PE−1 year−1) for well-managed aerobic WRRFs underestimate actual WRRF emissions. This suggests a systematic underestimation in national inventories that rely on these default values. We identified several countries using the above default CH4 EFs for aerobic WRRFs, including Austria, France, Spain, Turkey, Ukraine; and others applying the above default N2O EFs, including Belgium, Italy, South Africa, Turkey, Ukraine. For these countries, beyond omitted wastewater pathways, we also estimated the net emissions change resulting from updating WRRF EFs.

The annual CH4 and N2O emissions (in kg) from latrine, septic tank, WRRF and untreated wastewater discharge, for any given country are calculated using equations (6) and (7), respectively

$${\mathrm{CH}}_{4,i,j}={P}_{i}\times {r}_{i,j}\times {\mathrm{BOD}}_{i}\times 0.001\times 365\times {\mathrm{EF}}_{\mathrm{CH}4,j}$$
(6)
$${{\rm{N}}}_{2}{{\rm{O}}}_{i,j}={P}_{i}\times {r}_{i,j}\times {\mathrm{TN}}_{i}\times {\mathrm{EF}}_{{\rm{N}}2{\rm{O}},j}$$
(7)

where i and j indicate country i and wastewater pathway j, respectively; Pi is population of country i; ri,j is percentage of population severed by pathway j in country i; BODi is country-specific per capita biochemical oxygen demand, g BOD PE−1 day−1; TNi is country-specific total nitrogen in wastewater, kg TN PE−1 year−1; EFCH4,j is and EFN2O,j are CH4 and N2O EFs for wastewater pathway j, in kg CH4 kg−1 BOD and kg N2O-N kg−1 TN, respectively.

The annual CH4 and N2O emissions (in kg) from discharge of treated wastewater are further considered BOD and TN removal rate in WRRFs as listed in equations (8) and (9), respectively

$${\mathrm{CH}}_{4,i,j}={P}_{i}\times {r}_{i,j}\times {\mathrm{BOD}}_{i}\times {(1-R}_{i})\times 0.001\times 365\times {\mathrm{EF}}_{\mathrm{CH}4,j}$$
(8)
$${{\rm{N}}}_{2}{{\rm{O}}}_{i,j}={P}_{i}\times {r}_{i,j}\times {\mathrm{TN}}_{i}\times {(1-R}_{i})\times {\mathrm{EF}}_{i,j}$$
(9)

where Ri represents country-specific BOD removal rate in WRRFs, unitless.

TNi is calculated by multiplying country-specific protein consumption per capita per year and a list of constant parameters without considering country-specific variations, including fraction of nitrogen to protein (0.16), factor for additional nitrogen from household products (1.1), factor for non-consumed protein (1.1) and factor for industrial and commercial co-discharged protein (1.25). Country-specific protein consumption and BOD per capita were obtained from NIRs and supplemented with data from the Food and Agriculture Organization (FAO) Statistics Division (https://www.fao.org/faostat/en/#data/FBS/visualize).

CH4 and N2O emissions (in kg) were further converted to million metric tons (MMT) of CO2e by multiplying global warming potential of 27 and 273, respectively. Detailed data are available in Supplementary Dataset.