The global methane budget requires accurate knowledge of the emissions and sinks of this greenhouse gas to assess whether policies aimed at mitigating emissions are working. Emissions evaluation is a challenging task because the methane lifetime in the atmosphere allows gases to disperse across long distances and thus are difficult to observe directly. Long-term ground-based networks can provide accurate hemispheric averages and decadal variations, but such systems must be combined with space-based observations to capture seasonal variability in order to fully characterize methane trends.
The inverse modeling approach converts atmospheric abundance to emissions by solving for revised emissions based on observed changes in atmospheric concentrations and first-guess estimates of emissions (or “priors”). This can be done using a variety of models, from simple mass balance approaches to data assimilation techniques similar to those used in numerical weather forecasting. Inverse modeling is particularly useful for attributing the contribution of specific sources to changes in atmospheric concentrations. This can help inform bottom-up emission inventories of methane sources that may not have the necessary accuracy to account for all of the emissions associated with a particular source category. For example, changes in atmospheric abundance due to increased emissions from Arctic wetlands and permafrost are expected to be greater than the current EDGAR-based estimate of total Arctic methane emissions, and inverse modelling could provide information on how this change might be best mitigated by future policy interventions.
Bottom-up estimates of methane emissions https://theinscribermag.com/how-tara-oilfield-services-delivers-customized-solutions-in-production-testing/ rely on activity data (e.g., number of livestock or natural gas operations) multiplied by emission factors to determine total emissions from a given source. This approach can be limited by uncertainty in the activity data, the emission factor, and atmospheric transport, but it can also benefit from direct measurements of source emissions and inverse modeling. For example, comparing model-based versus independent measurements of emissions at 29 landfills across six continents resulted in an unbiased correlation coefficient (Willmott index) > 0.76 for the modeled vs. measured comparisons (Spokas et al., 2015).
Inverse modeling can be used to evaluate bottom-up inventory estimates of methane, as well as other air pollutants. It is especially useful for assessing how accurately the emissions from specific sources are being attributable to changes in atmospheric abundances and in quantifying spatial variability.
Inverse modeling studies based on satellite measurements have been conducted at both global and regional scales. Inversions based on time-averaged satellite retrievals from SCIAMACHY and GOSAT appear to produce results similar to those from inversions based on surface measurements, even after accounting for uncertainties such as water vapor bias correction (Houweling et al., 2016). Inversions based on a combination of surface and space-based observations do not seem to significantly differ from inversions based on only in situ observations, as evidenced by the fact that both Wecht et al. (2014a) and Turner et al. (2015) used a global modeling framework and in situ observations to solve for U.S. emissions, and both obtained results within the error range of previous top-down estimates from EDGAR and CARB.