2010-2016 methane trends over Canada, the United States, and Mexico observed by the GOSAT satellite: contributions from different source sectors

We use seven years (2010-2016) of methane column observations from the Greenhouse Gases Observing Satellite (GOSAT) to examine trends in atmospheric methane concentrations over North America and infer trends in emissions. Local methane enhancements above background are diagnosed in the GOSAT data on a 0.5◦ × 0.5◦ grid by estimating the local background as the low (10th-25th) percentiles of the deseasonalized frequency distributions of the data for individual years. Trends in methane enhancements on the 0.5◦×0.5◦ grid are then aggregated nationally and for individual source sectors, using 5 information from state-of-science bottom-up inventories. We find that US methane emissions increased by 2.5± 1.4% a−1 (mean ± one standard deviation) over the seven-year period, with contributions from both oil/gas systems (possibly unconventional oil/gas production) and from livestock in the Midwest (possibly swine manure management). Mexican emissions show a decrease that can be attributed to a decreasing cattle population. Canadian emissions show year-to-year variability driven by wetlands emissions and correlated with wetland areal extent. The US emission trends inferred from the GOSAT data account 10 for about 20% of the observed increase in global methane over the 2010-2016 period.


Introduction
Methane is an important greenhouse gas with a calculated climate impact as important as carbon dioxide over a 10-year time horizon (Myhre et al., 2013;Etminan et al., 2016). Livestock, oil/gas, and waste are the leading anthropogenic sources.
Wetlands are the dominant natural source. Contributions from different source sectors and regions remain poorly quantified (Kirschke et al., 2013;Saunois et al., 2016). Atmospheric methane concentrations leveled off in the 1990s but have been 5 increasing again since 2007 (Dlugokencky et al., 2009). Interpretations of atmospheric observations from surface networks have reached conflicting conclusions as to the cause of the renewed increase, with attributions to (1) natural gas production based on correlation with ethane Hausmann et al., 2016;Helmig et al., 2016), (2) agriculture/wetlands based on isotopic information (Nisbet et al., 2016;Schaefer et al., 2016), (3) reduced biomass burning to reconcile the ethane and isotopic constraints (Worden et al., 2017), and (4) declining concentrations of the OH radical (the main methane sink) 10 based on the methylchloroform proxy (Rigby et al., 2017;. Satellite-based observations of atmospheric methane columns have been available from the TANSO-FTS instrument aboard the Greenhouse Gases Observing Satellite (GOSAT) continuously since May 2009 (Kuze et al., 2016). Cressot et al. (2016) found that the GOSAT data had limited success in detecting regional year-to-year trends for 2009 used GOSAT data from January 2010 to January 2014 to infer a 2.8% a −1 increase in methane emissions from the contiguous 15 United States (CONUS), based on the trend in the CONUS enhancement of methane relative to the Pacific Ocean taken as background. Bruhwiler et al. (2017) argued that this trend inference could have been biased by the brevity of the GOSAT record, atmospheric transport variability, seasonal bias in GOSAT sampling frequency, and the use of Pacific data as background. They pointed out that global inversions of the surface network data for 2000-2012 from the North American Carbon Program (NACP) reveal no significant CONUS emission trend. 20 Here we reexamine the trend in CONUS emissions implied by the GOSAT data by using a longer record (January 2010 -December 2016), an improved definition of the background that accounts for atmospheric transport variability, and sectoral source information from a new gridded version of the US Environmental Protection Agency (EPA) Greenhouse Gas Inventory . We relate the inferred trends to trends in the underlying activities, and evaluate consistency with trends in the surface network data. We also extend the trend analysis to Canada and Mexico. 25

Methods
GOSAT was launched in January 2009 in a Sun-synchronous low Earth orbit. It retrieves the atmospheric methane column by nadir measurements of solar back-scatter (1.65 µm absorption band). There has been no degradation of retrieval accuracy since the beginning of the record (Kuze et al., 2016). Observations in the standard mode are made at three circular pixels of 10 km diameter across the orbit track 260 km apart, separated by 260 km along the track. The same locations are sampled every 3 30 days, making for a temporally dense data set at those locations. The observations often switch from the standard mode to focus on targets and this affects the regularity of the sampling.
Here we use the version 7.0 proxy nadir retrievals of GOSAT methane data from Parker et al. (2011. The proxy method uses prior knowledge of carbon dioxide columns (based on the MACC-II inversion product (v13r2; Chevallier F. et al., 2010) accounting for seasonal and interannual variations) to infer methane column average dry mole fractions X CH4 (in ppb) from the ratio of retrieved methane and carbon dioxide columns. The proxy method takes advantage of the much larger variability in methane than in carbon dioxide mixing ratios (Frankenberg et al., 2006;. The resulting GOSAT X CH4 data 5 have been validated against the ground-based Total Carbon Column Observing Network (TCCON), and found to be of high quality with a single-scene precision of 0.7% (random error) and a systematic error of 4-6 ppb Buchwitz et al., 2015Buchwitz et al., , 2016. GOSAT observes in all seasons with near-uniform frequency south of 45 • N (CONUS and Mexico), but observations further north (Canada) are biased toward summer. The number of successful retrievals over Canada is 2-3 times less in winter than in summer (see Supplemental Material). 10 From a simple mass balance perspective, enhancements of column methane above the surrounding background in a strong source region can be linearly related to the emissions in that region Buchwitz et al., 2017).  estimated the CONUS background by using glint mode retrievals from GOSAT over the Pacific Ocean for the corresponding latitudes. Bruhwiler et al. (2017) pointed out that changes in large-scale meridional transport could alias trends in this background estimate onto trends in the emissions. 15 Here we define local background methane for a given CONUS location (0.5 • × 0.5 • grid cell, typically including a single repeated GOSAT measurement location) and for a given year as the low (10 th -25 th ) percentiles of the deseasonalized GOSAT methane observations within the given 0.5 • × 0.5 • grid cell, with seasonality removed using the seasonal-trend loess (STL) decomposition method (Cleveland et al., 1990). This approach assumes that the low percentiles of concentrations reflect meteorological conditions where local sources have relatively little effect on methane concentrations due to rapid ventilation. 20 Low percentiles are a standard approach for estimating the regional background at a measurement location (Goldstein et al., 1995). By choosing the 10 th -25 th percentile rather than a lower extreme we guard against the effect of measurement noise (random error). A permutation resampling test shows that GOSAT observations across North America are sufficiently precise that ≥10th percentiles are not affected by measurement noise (see Supplemental Material). We use the range defined by the 10 th -25 th percentile range as a measure of uncertainty in the background for purpose of determining the enhancement. 25 Systematic errors of 4-6 ppb in GOSAT observations (Buchwitz et al., 2016) do not affect the enhancement because the bias can be expected to similarly affect all percentiles of the methane observations. Local enhancements are inversely proportional to wind speed , but we find no significant trends in wind speeds over the 2010-2016 period that would contribute to our aggregated trends in methane enhancements (see Supplemental Material). Any trends in OH concentrations would also not affect the enhancement because the lifetime of methane against oxidation is 9-10 years (Prather et al., 2012;30 Kirschke et al., 2013), very long compared to the timescale for ventilation from the source region.
We examined the validity of our approach by comparing frequency distributions of GOSAT methane columns and related trends to continuous ground-based column observations available from the TCCON  network site at Lamont, Oklahoma (36.6 • N, 97.4 • W). Figure 1 shows the frequency distributions of the deseasonalized GOSAT and TCCON observations at Lamont. The GOSAT background defined by the 10 th -25 th percentiles is consistent with TCCON; we see that 35 the repeated observation strategy of GOSAT at its discrete sampling locations makes for a sufficiently dense data set for defining the 10 th -25 th percentiles with little effect from instrument noise. The local annual mean background increases between 2010 and 2015 in a consistent way in the GOSAT and TCCON data sets, reflecting the global increase in the methane background.
The enhancements above background also show comparable 2010-2016 trends between the two data sets, although the error standard deviations defined by the ranges of the 10 th -25 th percentiles are large and the trends at this single site are marginally 5 significant (p = 0.07). Below we will use enhancement statistics aggregated over a large number of sites in order to reduce that uncertainty and quantify trends.
To aggregate trends in methane enhancements for individual source sectors, we use bottom-up annual mean sectoral infor-  Bloom et al. (2017). From these inventories we select high-emitting grid cells at 0.5 • × 0.5 • resolution dominated by a particular source sector. The high-emitting grid cells are defined as having emissions larger than 0.5 tons h −1 , encompassing 80-90% of anthropogenic and wetland emissions in all three countries. A high-emitting grid cell is identified as dominated by a given source sector if that source sector accounts for more than 70% of the total emissions in the cell. This allows us to define grid cells dominated specifically by oil/gas, livestock, waste, and wetlands emissions. 20 Contributions from other sectors (up to 30%) may lead to some smoothing of results. Wetland-dominated areas determined by the WETCHIMP mean and WetCHARTs inventories differ significantly (see Supplemental Material). Using either of the two inventories alone may bias our results, and thus we conservatively require wetland-dominated areas to be determined as such in both inventories.
We define a total methane enhancement ∆ for a given year, source sector, and country as whereX CH4,i is the annual mean value of the deseasonalized column average dry mole fractions in 0.5 • × 0.5 • grid cell i for the given year, X CH4,b,i is the corresponding local background value, and the summation is over all high-emitting grid cells for that sector and country. We require grid cells to have at least eight valid retrievals for a given year, and about 70% of grid cells meet this requirement. To account for local background variation due to atmospheric transport, the summation 30 in Equation (1)  frequency distributions, shown in the lower left panel, are much broader than ours because they did not use annual averaging of the data. Their Pacific background distribution is similarly broader and is also lower than our local background, which is appropriately elevated by continental influences. Below we will use the aggregated enhancements by source sectors (Equation 20 1) to infer the trends and reduce the uncertainty.  The Canadian methane emissions show no significant seven-year trend but large year-to-year variability driven by wetlands.
The 2014 maximum can be explained by a maximum of wetland areal extent (Bloom et al., 2017) (See Fig. S6 in Supplemental Material). Observations in the oil/gas dominated region of Canada (mainly natural gas in Alberta) are too sparse for inferring a significant oil/gas emission trend and are not shown here.
Mexican national emissions (excluding oil/gas offshore emissions) show a 5-10% decrease over the 2010 to 2016 period 5 that appears to be largely driven by livestock. The decrease of livestock emissions (4.0±1.6% a −1 ) is consistent with the 17% decrease in the Mexican cattle population over that period as reported by the Foreign Agriculture Service of the US Department of Agriculture (2015) and shown in Figure 5. The slight increase in Mexican emissions from 2012 on suggests an increasing source to compensate for the declining livestock emissions but GOSAT observations are too sparse to identify that source.
The CONUS data imply a significant increase in methane emissions from 2010 to 2016, with a trend of 2.5±1.4% a −1 10 derived from linear regression that is consistent with our previously calculated mean trend of 2.3% a −1 averaged over the 0.5 • × 0.5 • gridded trends in Figure 2. Breakdown by sector suggests that US oil/gas emissions increased at a marginally significant level (2.9% a −1 , p = 0.03) from 2010 to 2015. Oil and unconventional (hydraulic fracturing) gas production grew by 15% a −1 and 19% a −1 , respectively during that period ( Figure 5), though production rate is not necessarily a predictor of emissions (Peischl et al., 2015). 15 The US livestock emissions show a 3.5±1.8% a −1 increase in our analysis, largely reflecting the agricultural Midwest where high-emitting grid cells are concentrated (Figure 3). These grid cells emit 0.95 Tg CH4 a −1 from enteric fermentation (cattle) and 0.55 Tg CH4 a −1 from manure management (swine) according to the gridded EPA inventory .
The cattle population in that region does not show a significant trend ( Figure 5)  (2017) suggests that this year-to-year variability is related to wetland areal extent, same as for Canada (See Fig. S6 in Supplemental Material), though the definition of wetland areal extent may vary significantly (Poulter et al., 2017). Here the WetCHARTs extended ensemble used GLOBCOVER land cover data (Bontemps et al., 2011) and the Global Lakes and Wetlands Database (GLWD; Lehner and Dölla, 2004) to represent spatial wetland extent, and ERA-interim precipitation to account for temporal wetland extent (Bloom et al., 2017). 35 Inverse analyses of methane concentrations in surface air measured as part of the North American Carbon Program (NACP; Wofsy and Harris, 2002) for 2010-2014 reveal no significant trends in US emissions over that period (Benmergui et al., 2015).
We examined whether the trends inferred from this work (significant trends after 2012) are consistent with the information provided by NACP surface data. For this purpose, we examined the residuals (observed minus simulated methane concentrations) of the CarbonTracker-Lagrange (CT-L) methane transport model (Benmergui et al., 2015) driven with two sets of emissions (1) These sites are strongly influenced by large livestock/wetlands, livestock, and oil/gas sources, respectively (Benmergui et al., 10 2015). There is no significant trend in the residuals of the CT-L simulation driven by either our GOSAT-inferred emission trends or CT-L posterior emissions, and the two sets of residuals are statistically indistinguishable. We find similar results for other NACP sites that are less sensitive to source regions. This implies that the trends found in this work are compatible with the constraints provided by NACP data. This also suggests that the surface data may be spatially too sparse to adequately infer trends of the magnitude detected by GOSAT. Mexican decreasing trend appears to be due to a declining cattle population. Canada shows no significant long-term trend but large year-to-year variability associated with wetlands and correlated with variations in wetland areal extent, though this trend 20 is weighted toward summer because of the seasonal bias in observation frequency (less observations in winter). The US trend is +2.5 ± 1.4% a −1 for the period and appears to reflect contributions from both oil/gas and livestock. Assuming 38-55 Tg CH4 a −1 for the CONUS emissions, including 29-40 Tg CH4 a −1 from anthropogenic sources (Miller et al., 2013;Wecht et al., 2014;Turner et al., 2015; and 9-15 Tg CH4 a −1 from wetlands (Melton et al., 2013;Bloom et al., 2017), we deduce an increasing emission trend of 0.9-1.3 Tg CH4 a −1 over the 2010-2016 period, which would account 25 for about 20% of the global increase in atmospheric methane (Rigby et al., 2017). Our analysis is mainly limited by the length of GOSAT record, and a longer record can provide more reliable results. The definition of local background may also not fully account for the variation in atmospheric transport. Our trend analysis should be compared to trends inferred from inverse modeling (Bruhwiler et al., 2017), which better account for the role of atmospheric transport but have their own errors notably in the prior assumptions of emission patterns . Future inversions combining GOSAT and surface 30 network data with improved bottom-up estimates are needed to provide more robust trend analyses.