An estimate of anthropogenic CO 2 inventory from decadal changes in oceanic carbon content

Tanhua et al. 10.1073/pnas.0606574104.

Supporting Information

Files in this Data Supplement:

SI Figure 7
SI Figure 8
SI Figure 9
SI Figure 10
SI Text
SI Table 2
SI Figure 11
SI Figure 12
SI Figure 13
SI Figure 14
SI Figure 15
SI Figure 16
SI Figure 17
SI Figure 18

SI Figure 7

Fig. 7. Comparison between D DIC as calculated with the direct comparison method (D DIC_direct) and with the eMLR method (D DIC_eMLR). (Left) The geometric mean linear regression plotted in gray. (Right) The difference D DIC_direct - D DIC_eMLR (d D DIC) plotted vs. depth for the whole data set.

SI Figure 8

Fig. 8. (Left) The C_ant vs. time for surface waters with different temperature and hence buffer capacity, assuming constant air-sea disequilibrium. (Right) The ratio vs. time, where The surface equilibrium is calculated with CO₂SYS, with constant nutrient, salinity and alkalinity but with temporally increasing pCO₂ values.

SI Figure 9

Fig. 9. The atmospheric CO₂ concentration vs. time. (Left) The measured atmospheric CO₂ concentrations (Mauna Loa record from 1959 and the 75 year smoothed Law Dome record before that time, downloaded from CDIC, http://cdiac.esd.ornl.gov/). (Right) The exponential nature of anthropogenic CO₂, where the natural, background, level of CO₂ is set to 275 ppm, the red line is a linear fit of the data between years 1750 and 2000. The red circle indicate year 1850 when the CO₂ concentration was 285 ppm, i.e., only about 10 ppm higher than the preindustrial concentration. From this point in time the atmospheric increase in CO₂ is nearly exponential.

SI Figure 10

Fig. 10. For a given TTD we can calculate C_ant from equation (2) that can be regarded as the "true" answer. From this we can also determine the D C_ant, i.e., the change in DIC between 1981 and 2004, and then use (5) to calculate C_ant from D C_ant. The results of this calculation is shown here, where the estimated C_ant by the extrapolation (dashed line) is compared to the calculated C_ant for mean age of the TTD varying from 25 to 400 years (solid line). For this calculation we have assumed an inverse Gaussian TTD and that the width (D)of the TTD is equal to the mean age (G). The agreement between the two ways of estimating C_ant is excellent, i.e., the difference between the two curves is much smaller than errors in estimating D C_ant from measurements. The difference between the two curves is solely due to the not perfect exponential atmospheric increase of anthropogenic carbon.

SI Figure 11

Fig. 11. The mean residual (eMLR based estimate: "true" D DIC) (black dots), and standard deviation of the residuals (red dots) of the 2s perturbation experiments in the eastern basin. See supporting text for more information on the four experiments.

SI Figure 12

Fig. 12. The mean residual (eMLR based estimate: "true" D DIC) (black dots), and standard deviation of the residuals (red dots) of the 4s perturbation experiments in the western basin. See supporting text for more information on the four experiments.

SI Figure 13

Fig. 13. The section across the mid-Latitude North Atlantic (see Fig. 1) of the difference (eMLR based estimate: "true" D DIC). The red line mark the division between the western and eastern basin in our calculations.

SI Figure 14

Fig. 14. The eMLR based estimates of D DIC in the western basin (black dots) and the DIC perturbation forced on the data (red line).

SI Figure 15

Fig. 15. Results from experiment 5 with the mean residual (black dots) and standard deviation of the residuals (red dots); upper panels from the eastern basin, lower panels from the western basins. (Left) The result with 2s perturbation of the data (i.e., experiment 1); the middle panels show the case where the modern data are perturbed with 20s, and the old data with 2s (experiment 5/1). (Right) The case in which the historic data has been perturbed with 20s, and the moderns data with 2s (experiment 5/2). Note that the scale of the x axis is different for the middle panels.

SI Figure 16

Fig. 16. C_ant calculated with the three different methods; eMLR (dark gray crosses), D C* (light gray diamonds) and TTD (black dots) for the stations surrounding the crossover with cruise A05R, south of point "A" in Fig. 1. (A) C_ant vs. depth (m mol kg^-1) for the three methods with the insert showing a blown-up view of the layer between 500 and 2,000 m. (B) Cumulative C_ant inventory (mol m^-2) calculated for the three methods, the light gray dashed line marks the highest column inventory for the DC* method.

SI Figure 17

Fig. 17. C_ant calculated with the three different methods; eMLR (dark gray crosses), DC* (light gray diamonds) and TTD (black dots) for the stations surrounding the crossover cruise A20 at point "C" in Fig. 1. (A) C_ant vs. depth (m mol kg^-1) for the three methods with the insert showing a blown-up view of the layer between 500 and 2000 m. (B) Cumulative C_ant inventory (mol m^-2) calculated for the three methods, the light gray dashed line marks the highest column inventory for the DC* method.

SI Figure 18

Fig. 18. C_ant calculated with the three different methods; eMLR (dark gray crosses), DC* (light gray diamonds) and TTD (black dots) for the stations surrounding the crossover cruise A16 at point "F" in Fig. 1. (A) C_ant vs. depth (m mol kg^-1) for the three methods with the insert showing a blown-up view of the layer between 500 and 2,000 m. (B) cumulative C_ant inventory (mol m^-2) calculated for the three methods, the light gray dashed line marks the highest column inventory for the DC* method.