|Figure 1. Globally averaged carbon dioxide (CO2) mole fraction based on observations from stations contributing to the GAW Programme (Source: WMO, 2018)|
The effects of climate change are becoming more evident. Governments are addressing the challenge of climate change through international accords such as the Paris Agreement in 2015. To evaluate progress towards climate targets, governments have adopted a process of national greenhouse gas (GHG) emissions reporting following agreed protocols. These protocols were established by the Intergovernmental Panel on Climate Change (IPCC) and described in the 2006 IPCC Guidelines for National Greenhouse Gas Inventories (IPCC, 2006). In May 2019, the IPCC Plenary approved a Refinement to the Guidelines (in press), which outlines the important role of atmospheric observations and analysis to improve estimations of national GHG emissions. The 2019 Refinement describes key components and steps to be applied when using atmospheric measurements and inverse models for comparison with inventory estimates (Chapter 6, Quality Assurance/Quality Control and Verification).
Accurate and precise atmospheric measurements of major GHGs have revealed an increase in their concentrations globally (Figure 1). However, the concentrations of ozone-depleting substances, which are also GHGs, have been decreasing in the past decade, in response to the Montreal Protocol. The global network of GHG observations, coordinated by the WMO Global Atmosphere Watch (GAW), provides us with alerts of dangerous changes in the climate system.
Atmospheric GHG concentrations are a result of the balance between sources and sinks (see Figure 2) and are influenced by transport and mixing processes. To limit global warming, it is important to quantify the sources, as these can be controlled. However, estimating emissions from atmospheric observations is not a trivial task, as a relationship between the concentrations at a given observation point and the sources upstream needs to be derived. This relationship is determined by atmospheric transport and can be simulated by an atmospheric transport model. Doing this accurately places high demand on the model performance.
First attempts to use atmospheric observations to estimate emissions of GHGs date back to the 1980s (e.g. CFC-11, Fraser et al., 1983). These early studies largely addressed global to continental scales relying on coarse-resolution models and observations from a sparse global network, primarily from the National Oceanic and Atmospheric Administration (NOAA) global flask sampling programme. Quantifying emissions at large scales is essential, but it only describes the net global emissions that determine the increase of long-lived GHGs in the atmosphere. However, these estimates provide little information on individual sources and processes required by policymakers, who need to take action at the national, subnational or regional levels. With the expansion of the observational network, especially in developed countries, and the increasing capability and resolution of atmospheric transport models, estimating emissions at smaller national scales has become possible.
|Figure 2. Atmospheric concentrations of GHGs are a balance between the sources (what comes into the atmosphere) and sinks (what is taken out from it). The cumulative contributions to the global carbon budget since 1870. The carbon imbalance represents the gap in our current understanding of sources and sinks. (Source: Global Carbon Budget 2018, Global Carbon Project)|
The atmospheric, carbon cycle and climate change scientific communities have produced a number of studies on the potential for atmospheric GHG concentration measurements and model analyses to evaluate and help to inform improved estimates of GHG emissions (for example, National Research Council (2010), Ciais et al. (2010), IPCC (2010)). These studies concluded that a realization of this approach would require additional investment in research, increasing the density of well-calibrated atmospheric GHG measurements and improving atmospheric transport modelling and data assimilation capabilities.
The need for harmonization and documentation of the methodologies for emission estimation from atmospheric observations, as well as sharing of good practices, led to the establishment of the Integrated Global Greenhouse Gas Information System IG3IS (WMO Bulletin 66 (1), 2017) at the Seventeenth World Meteorological Congress in 2015.
Emission estimates to support national inventories
The United Nations Framework Convention on Climate Change (UNFCCC), which entered into force in 1994, is now a nearly universal agreement with 197 participating countries. Parties to the Convention were asked to “periodically update, publish and make available (...) national inventories of anthropogenic emissions by sources and removals by sinks of all greenhouse gases not controlled by the Montreal Protocol.” Methodologies were not well defined at the beginning, and countries were asked only to report to the extent their capabilities permitted. Nevertheless, many countries started to collect and report information on GHG emissions, though at variable levels of detail and frequency.
With the Kyoto Protocol entering into force in 2005, the communication of annual National Inventory Reports (NIRs) became mandatory for all Annex 1 (developed) countries. Details were specified in the 2006 IPCC Guidelines, which proposed a tiered approach. The simplest, Tier 1, relies on default emission factors (EFs), while in the more detailed, Tiers 2 and 3, country-specific methods, data and models can be incorporated. Another step forward was made when the Paris Agreement introduced the provision of NIRs by all signatory countries on a biennial basis. Preparation and submission of a NIR poses a great challenge for developing countries that have yet to learn the use of the official reporting process.
Developing a GHG emission inventory is challenging because a wide range of information has to be used depending on the types of source. Socio-economic, and other, statistical data may not be available in sufficient detail and in a timely manner. National emission inventories provide fine-grained information on individual sources, which allow policymakers to assess their relative share and design effective emission reduction measures. However, the quality of these inventories can be assessed only by checking for completeness and compliance with recommended procedures. The most relevant number in terms of climate change effects, ie., specifically the total emissions per country, cannot be assessed by independent means.
As atmospheric concentrations respond to the sum of all emissions, observation-based estimates can provide invaluable constraints on the total emissions of a country. However, they are less suited to provide information on individual source categories since multiple sources and sinks interplay. In that sense, inventories and observation-based estimates are complementary and should be used together to improve and build trust in national emission estimates. With dense observation networks and measurements of auxiliary parameters such as isotopic composition of greenhouse gases or concentrations of co-emitted gases, additional source-specific information can be gained to support the validation of national emission inventories beyond country totals. Observation-based estimates can be particularly valuable for trace gases with large uncertainties in their emissions.
Improvements in science
Emission estimates of national totals based on inverse modelling techniques have greatly improved over the last 20 years. Implementation of this approach includes a combination of atmospheric observations and modelling, and comprises four key components, all of which have seen significant development over this time frame:
- Atmospheric observations became much more accurate, and more robust instruments are now available. Furthermore, the frequency of measurements and the number of observation sites have greatly increased, with extended networks being developed in many countries such as Australia, China, Germany, India, Switzerland, United Kingdom of Great Britain and Northern Ireland, and United States of America. Furthermore, satellite remote-sensing of GHGs has advanced remarkably since the first measurements of CO2 and methane (CH4) total columns by SCIAMACHY in 2002. Today, satellites such as GOSAT, OCO-2 and TROPOMI are providing observations with accuracies that are sufficient to constrain emissions at large regional scales. However, further improvements in coverage, resolution, gases observed and precision will be required to push the limits towards the scale of individual countries and emission hotspots. CEOS coordinates activities related to Earth system observations from satellite and develop the long-term strategies for their evolution. In particular, they recently defined a global architecture for monitoring atmospheric CO2 and CH4 that includes the current state of satellite-based GHG measurements and refers to IG3IS as a common framework.
- Time-evolving three-dimensional meteorology has seen significant advances through improved data assimilation, increased computing power, better representation of atmospheric processes and higher spatial resolution. For example, the horizontal resolution of operational global weather forecast models has evolved from 80 km 20 years ago down to between 9 km and 20 km now, with similar improvements in the vertical dimension.
- Transport models, which are driven by these three-dimensional meteorological fields have matured considerably through applications in a multitude of research projects. They are now able to use more meteorological parameters at higher spatial and temporal resolution than previously. Furthermore, online integrated models, which compute the meteorology and the transport of GHGs simultaneously and consistently within the same model, have been developed.
- Inverse models, which integrate the information from the observations and the atmospheric transport models, have seen important changes with improved use of advanced algorithms optimally combining information from atmospheric observations with knowledge on distribution of emissions.
|Figure 3. UK emission estimates of HFC-134a. Inventory reported values: purple 2013 submission; black, 2019 submission. Inverse modelling (InTEM) estimates: blue, using one observation site; orange, using three observation sites; green, using four observation sites.|
To assess the quality of the emissions reported in the national inventory, the UK uses a completely independent method (described in Arnold et al., 2018) for deriving its GHG emission estimates that relies upon a combination of atmospheric observations and inverse modelling. The results are reported annually in the UK NIR to UNFCCC. The UK uses the significant differences in the emissions inventory and the observation-based estimates to identify areas of the inventory worthy of further investigation. Its network of observation sites, called the UK DECC (Deriving Emissions related to Climate Change) (Stanley et al., 2018), consists primarily of tall tower telecommunication masts equipped with state-of-the-art observation equipment. These measure CO2, CH4, nitrous oxide (N2O), hydrofluorocarbons (HFCs), perfluorocarbons (PFCs), sulphur hexafluoride (SF6) and nitrogen trifluoride (NF3) at high frequency, precision and accuracy.
A recent example of how the observation-based emission estimates have been used to inform the traditional inventory estimate is demonstrated in Figure 3. In the UK Inventory Report of 2013, the annual emission estimates for HFC-134a (purple bars), a gas predominantly used in mobile air conditioning and, to a lesser extent, as an aerosol propellant, were consistently higher by more than 50% than the observation based estimates (blue and orange lines) from 1998 onward. This result motivated the UK to commission an industry expert to review the UK HFC-134a inventory estimates. As a result, the inventory estimate has been revised and moved nearer to the inverse modelling estimates, as seen by the black bars in the figure. There is still work to be done to close the gap entirely.
|Figure 4. Switzerland’s CarboCount CH measurement network, established in 2012 to measure CH4 and CO2. At Beromünster, additional instruments were installed in 2016 and 2019 to measure N2O and synthetic GHGs, respectively. (Source: Oney et al. 2015)|
In 2012, Switzerland set up a GHG observation network with one tall tower and additional sites on hilltops and smaller towers (see Figure 4, Oney et al., 2015). Together with the long record of observations at the high Alpine site Jungfraujoch, these measurements were used to estimate Switzerland’s CH4 emissions, which were found to be consisten with the national inventory (Henne et al., 2016). Since 2016, these estimates have been reported annually in an Annex of the Swiss NIR to UNFCCC. The consistency has been confirmed for all years investigated so far (2013–2017).
Additional measurements of N2O were established in 2016, and a first estimate derived from these measurements was published in the 2019 Swiss NIR. It suggested that N2O emissions, in contrast to CH4, may have been underestimated by up to 30%. However, due to the large uncertainties in the inventory and the observation-based estimates, the difference was not statistically significant, and more years of observations will be needed to corroborate these differences.
Since observations of HFCs and other synthetic GHGs are available only at Jungfraujoch, a simpler interspecies correlation method, not requiring any atmospheric transport modelling, is employed to estimate HFC emissions. These estimates have also been reported in the Swiss NIR since 2016 and have shown broad consistency with the traditional inventory numbers for most species. To obtain more robust estimates for synthetic gases, additional measurements at the tall tower site Beromünster started in August 2019 (Figure 4).
Australia incorporated atmospheric verification into its annual NIR in 2009. South-eastern Australian synthetic GHG (HFCs, PFCs and SF6) emissions are estimated by the Commonwealth Scientific and Industrial Research Organization (CSIRO) and the Met Office in the UK from atmospheric observations obtained at Cape Grim, Tasmania, using inverse modelling and interspecies correlation techniques. South-eastern Australian emissions of these synthetic GHGs are scaled to Australian emissions on a population or activity basis.
|Cape Grim station, Tasmania (Source: Bureau of Meteorology)|
Comparison of emissions estimates based on atmospheric observations with the emissions in the traditional Australian inventory showed significant differences for individual HFCs, PFCs and SF6, but aggregated emissions of these synthetic GHGs from both estimations were in general agreement.
Following the IPCC recommendation, the annual Australian HFC EFs from 2006 onward and SF6 EFs from 2010 onward have been adjusted in line with HFC and SF6 emissions estimated from atmospheric concentrations and trends measured at Cape Grim. In addition to the calibration of annual EFs, HFC fluctuations observed at Cape Grim are also used to vary gas speciation in the HFC emissions model used in the inventory. PFC EFs in the inventory have not been adjusted to date to reflect PFC emissions derived from atmospheric data. In the future, Australia plans to use GHG observations from a variety of sites (for example, Aspendale, Victoria) and platforms like research vessels, with better targeted inverse modelling and interspecies correlation techniques to improve the accuracy of the observation based estimates of regional and national emissions.
As the 2006 IPCC Guidelines do not provide any advice on the direct use of inverse modelled emission estimates, Australia has opted to use the fluctuations in the modelled estimates to adjust the annual HFC and SF6 leakage rates. This ensures the trends in the atmospheric observations are replicated in the inventory. The strength of this approach is that it enables the inventory emission estimates to better reflect improvements in industry practice in terms of gas handling, equipment maintenance and decommissioning.
Role of IG3IS in developing observation-based emission estimates
As a document addressing inventory compilers, the 2019 Refinement of the IPCC guidelines does not provide detailed guidance on implementing the national atmospheric measurement and modelling system. Instead, it refers to country examples, WMO GAW recommendations on observation techniques and the IG3IS Science Implementation Plan for further guidance. The 2006 IPCC Guidelines and the 2019 Refinement promote the use of emission estimates based on atmospheric measurements but remain cautious about potential difficulties implementing such an approach.
When it comes to finding the most efficient way to use atmospheric measurements for emission estimates, the IG3IS Science Implementation Plan outlines a number of techniques that are available for building new national systems and improving existing ones. The planned or already tested techniques provide recommendations for the type of inverse modelling algorithms, the atmospheric transport models, the choice of observation sites, and the type of measurement devices and what optional parameters might be measured." There are also new, country-specific challenges faced by those who implement the estimates based on atmospheric measurements. For instance, existing working examples of national systems are built in countries isolated from neighbours by oceans or mountain ridges. Operating atmospheric inversions in a country located downwind of strong GHG sources will present different challenges.
IG3IS plays an important role in providing a common framework for development of harmonized methodologies and benchmarking. As a community of experts, IG3IS is well positioned to support the development and evaluation of necessary expertise, and to provide guidance to overcome technical difficulties based on up-to-date science and experience of national teams that have already established working systems. Those implementing an observation-based approach in their country as support to the GHG inventory are welcome to contact the IG3IS team through its website (ig3is.wmo.int). They will receive targeted advice that considers the specific circumstances and conditions in their country.
Arnold, T. et al., 2018: Inverse modelling of CF4 and NF3 emissions in East Asia. Atmospheric Chemistry and Physics, 18:13305–13320.
Ciais, P. et al., 2010: Geo Carbon Strategy. Geneva, GEO Secretariat / Rome, Food and Agriculture Organization of the United Nations.
Fraser, P.J. et al., 1983: Global distribution and southern hemispheric trends of atmospheric CCl3F. Nature, 302:692–695.
Henne, S. et al., 2016: Validation of the Swiss methane emission inventory by atmospheric observations and inverse modelling. Atmospheric Chemistry and Physics, 16(6):3683–3710.
Intergovernmental Panel on Climate Change (IPCC), 2006: 2006 IPCC Guidelines for National Greenhouse Gas Inventories. Prepared by the National Greenhouse Gas Inventories Programme (H.S. Eggleston, L. Buendia, K. Miwa, T. Ngara and K. Tanabe, eds). Hayama, Japan, Institute for Global Environmental Strategies.
———2010: Expert Meeting on Uncertainty and Validation of Emission Inventories (H.S. Eggleston, J. Baasansuren, K. Tanabe and N. Srivastava, eds.). Utrecht, the Netherlands, 23–25 March 2010.
———2019 Refinement to the 2006 IPCC Guidelines for National Greenhouse Gas Inventories (advance publication version, accepted by IPCC).
National Research Council, 2010: Verifying Greenhouse Gas Emissions: Methods to Support International Climate Agreements. Washington, DC, The National Academies Press.
Oney, B. et al., 2015: The CarboCount CH sites: Characterization of a dense greenhouse gas observation network. Atmospheric Chemistry and Physics, 15(19):11147–11164.
Stanley, K.M. et al., 2018: Greenhouse gas measurements from a UK network of tall towers: technical description and first results. Atmospheric Measurement Techniques, 11(3):1437–1458.
World Meteorological Organization, 2018: WMO Greenhouse Gas Bulletin, No. 14, 22 November 2018. Geneva.
Shamil Maksyutov, National Institute for Environmental Studies, Japan
Dominik Brunner, Empa, Swiss Federal Laboratories for Materials Science and Technology
Alistair Manning, Met Office, UK
Paul Fraser, CSIRO Oceans and Atmosphere, Australia
Oksana Tarasova, WMO Secretariat
Claudia Volosciuk, WMO Secretariat