Figure 1
Flow chart of steps to find the best analytical strategy for determination of δ13CH4 source signature, from mobile near-source measurements, based on controlled release experiment. The number in the top right corner corresponds to the methods subsection, where steps are explained in detail.
Table 1
Comparison of direct and averaged results using bag samples measured on IRMS and CRDS (treatment 1 averaging), and CRDS AirCore samples (raw data cluster, without C2H6 correction). The first three rows present comparison for individual releases, where sampling using two different measurement techniques was done simultaneously. Residual between reference IRMS strategy and other analytical strategies for atmospheric measurements. 1 standard error is used as an uncertainty.
| DATE OF SAMPLING | IRMS BAGS YORK FITTING | CRDS BAGS YORK FITTING | RESIDUAL (IRMS—CRDS BAGS) | CRDS AIRCORE YORK FITTING | RESIDUAL (IRMS—CRDS AirCore) |
|---|---|---|---|---|---|
| release 1.1 | –40.60 ± 0.23 | –41.3 ± 8.9 | 0.66 | – | – |
| release 1.2 | –40.56 ± 0.24 | –40.8 ± 3.2 | 0.21 | – | – |
| release 3.3 | –40.45 ± 0.54 | – | – | –40.4 ± 2.8 | –0.09 |
| all 26 releases | –40.23 ± 0.14 | –41.0 ± 6.7 | 0.79 | –41.0 ± 2.7 | 0.81 |
| direct sampling | –41.36 ± 0.17 |
Figure 2
Comparison of direct and averaged results for different linear fitting methods. Bag samples measured on IRMS and CRDS (treatment 1 averaging), and CRDS AirCore samples (raw, non-clustered data) from mobile near-source measurements conducted during controlled release experiment. Results using Miller-Tans with individual background subtracted (left) and Keeling (right) methods are compared. Black line represents IRMS reference value with its uncertainty (gray line). For averaged δ13CH4, uncertainties calculated as 1 standard error. The outlier result for MA linear fitting is not plotted. Compared linear fitting methods: OLS—Ordinary Least Squares, MA- major axis, BCES Y—Bivariate Correlated Errors and Intrinsic Scatter (Y|X) and BCES O—Bivariate Correlated Errors and Intrinsic Scatter Orthogonal.
Figure 3
Individual AirCore samples with and without a C2H6 on δ13CH4 correction from mobile near-source measurements conducted during controlled release experiment. Rejection criteria applied. For simplicity, only York fitting, and raw non-clustered data are presented, as the same trend is observed for all data processing strategies. Left: Miller-Tans with individual background subtracted method, Right: Keeling method. δ13CH4 calculated as 1 standard error. Black line represents IRMS reference value with its uncertainty (gray line).
Figure 4
Comparison of averaged CRDS AirCore samples for different cluster averaging from mobile near-source measurements conducted during controlled release experiment. C2H6 correction was not applied. Results using Miller-Tans with individual background subtracted (left) and Keeling (right) methods are compared. δ13CH4, uncertainties calculated as 1 standard error. Black line represents IRMS reference value with its uncertainty (gray line). Compared linear fitting methods: OLS—Ordinary Least Squares, MA- major axis, BCES Y—Bivariate Correlated Errors and Intrinsic Scatter (Y|X) and BCES O—Bivariate Correlated Errors and Intrinsic Scatter Orthogonal.
Table 2
CRDS AirCore samples for raw cluster data. nAirCore represents number of AirCore samples used to determine averaged δ13CH4 after applying rejection criteria. C2H6 on δ13CH4 correction not applied; one standard error is used as an uncertainty.
| LINEAR FITTING | δ13CH4 ± U(δ13CH4) (‰) | nAirCore KEELING METHOD | nAirCore MILLER-TANS 1 | nAirCore MILLER-TANS 2 | ||
|---|---|---|---|---|---|---|
| KEELING METHOD | MILLER-TANS METHOD 1 | MILLER-TANS METHOD 2 | ||||
| OLS | –41.1 ± 3.0 | –41.2 ± 1.5 | –41.2 ± 1.5 | 22 | 12 | 12 |
| MA | –24.2 ± 3.4 | –45.0 ± 1.7 | –45.0 ± 1.7 | 2 | 12 | 12 |
| York | –41.7 ± 2.8 | –41.0 ± 2.7 | –40.9 ± 2.1 | 21 | 12 | 12 |
| BCES Orthogonal | –46.5 ± 1.0 | –35.5 ± 2.2 | –39.8 ± 1.9 | 19 | 9 | 12 |
| BCES (Y|X) | –41.4 ± 2.8 | –41.2 ± 1.5 | –41.2 ± 1.5 | 25 | 12 | 12 |