CNL Report: Historic Comparison of Gravimetric and Reconstructed Mass March 19, 2001

Author: Robert Eldred

Summary

Reconstructed mass excluding nitrate is compared with gravimetric mass for each year of the IMPROVE network. Two parameters were used to investigate the trends. The first is the slope of the regression line through the origin, assuming gravimetric mass is independent. The second is the ratio of the two means. The slope emphasizes points with large mass, while the ratio reflects the average samples. Both are useful in looking at the relationship.

Figure 1 shows that the slope at Western sites is similar throughout the history of the network, except for a batch of defective Teflon filters in 1990, when there was extra gravimetric mass not accounted for by reconstructed mass or by elemental hydrogen. The Western sites exclude California sites with elevated nitrate. At the Eastern sites, the ratio is much lower than those for Western sites for the first two years, higher for 1990, and similar to the Western sites for 1991-99. The higher gravimetric mass for the first two years does not reflect laboratory relative humidity, since the RH during the early years of the network was actually lower than that at the later years. The dip in the ratio at Western sites in 1990 was not observed at Eastern sites. There has been a small increase in the ratio for both West and East since 1992.

Figure 1. Annual slope through the origin of reconstructed vs. gravimetric mass for Western and Eastern sites. This emphasizes the points with large mass.

Figure 2 shows the same trends for the ratio of means, except for an increase in 1998 for Western sites. This is associated with a discrepancy between organic mass measured on the quartz and Teflon filters for mid-range concentrations. The difference was largest in fall 1998. No major increase was observed at the Eastern sites.

Figure 2. Annual ratio of mean reconstructed mass divided by mean gravimetric mass for Western and Eastern sites. This emphasizes the points with average mass.

Discussion

Reconstructed mass (RCMC) is the sum of components of sulfate (4.125 S), organic mass (1.4 OC), light-absorbing (elemental) carbon, soil, NaCl (2.5 Na). Organic mass and light-absorbing carbon are determined from the quartz filter. All other parameters are from the Teflon filter. Nitrate is excluded because the gravimetric mass on the Teflon filter will have lost some of the nitrate during sampling. For the annual plots, including nitrate increased the slope about 6% for Western sites and 4% for Eastern sites.

Western Sites. Table 1 and Figures 3 and 4 show annual comparisons for Western sites without major nitrate. (San Gorgonio, Sequoia, Point Reyes, Pinnacles, and Redwood are excluded.) In 1990, the comparison differed from all other years, with reconstructed mass accounting for 12% less of the gravimetric mass. All samples were part of a batch (Batch D) that had been rejected because of pin holes and debris on the filter. However, because we were unable to obtain a replacement batch in time, it was necessary to use these filters for approximately five months. For Batch D, whatever was producing the mass artifact was not reflected in the measured variables. That is, using H and absorption did not result in a better fit.

The replacement batch (Batch E) was actually worse, with a organic artifact. For Batch E, we were able to reconstruct the mass using H and absorption from the Teflon filter instead the quartz values. That is, for batch E, there was a significant in organic mass estimated from hydrogen and carbon for approximately 7% of the samples. Because the gravimetric mass and hydrogen concentrations were invalidated whenever there was a significant difference between the two estimates of organic mass, this plots do not show the gravimetric mass artifact.

A much smaller discrepancy was observed in samples collected in fall 1998. Comparison of organic carbon on the quartz filter and organic hydrogen on the Teflon filter indicates that the problem seems to be associated with organic mass. Either organic was lost from the Teflon or gained on the quartz. The difference is observed on many, but not all, samples. The reconstructed mass using hydrogen gave a slightly better fit to the gravimetric mass than did reconstructed mass using carbon. For this quarter, there were more secondary quartz filters with abnormally high carbon concentrations than for other quarters. We are investigating whether some of the quartz lots had higher artifact.

Eastern Sites. Table 2 and Figures 5 and 6 show annual comparisons for Eastern sites. The Eastern sites differ from the Western sites in three cases. (1) The slope and ratios for the first two years were much less than at Western sites. (2) The increased gravimetric mass observed at Western sites in 1990 is not seen at Eastern sites. (3) The decreased gravimetric mass relative to reconstructed mass observed at Western sites in 1998 is not seen at Eastern sites. In both cases, the higher mass loadings at Eastern sites may have masked the effect.

Table 1: Comparison of Gravimetric and Reconstructed Mass by Year for Western sites without major nitrate. (San Gorgonio, Sequoia, Point Reyes, Pinnacles, and Redwood are excluded.) Years begin in March. The column “ratio” gives the ratio of mean reconstructed to mean gravimetric mass.

year	r2	ratio	slope	intercept	slope through origin
1988	0.88	0.87	0.81	0.24	0.85
1989	0.94	0.83	0.82	0.03	0.83
1990	0.74	0.75	0.62	0.62	0.71
1991	0.90	0.80	0.75	0.20	0.79
1992	0.92	0.85	0.79	0.23	0.83
1993	0.92	0.84	0.81	0.12	0.83
1994	0.92	0.83	0.79	0.16	0.81
1995	0.94	0.88	0.80	0.31	0.85
1996	0.93	0.88	0.82	0.23	0.86
1997	0.86	0.82	0.78	0.15	0.81
1998	0.93	0.94	0.84	0.37	0.85
1999	0.94	0.87	0.83	0.13	0.85
average	0.90	0.85	0.79	0.23	0.82

Table 2: Comparison of Gravimetric and Reconstructed Mass by Year for Eastern sites. The definitions are the same as in Table 1.

year	r2	ratio	slope	intercept	slope through origin
1988	0.89	0.78	0.61	1.62	0.71
1989	0.93	0.81	0.69	1.21	0.77
1990	0.91	0.82	0.75	0.60	0.79
1991	0.96	0.80	0.74	0.55	0.78
1992	0.97	0.85	0.79	0.61	0.83
1993	0.94	0.84	0.73	1.19	0.81
1994	0.94	0.83	0.72	1.18	0.79
1995	0.95	0.88	0.80	0.77	0.85
1996	0.93	0.89	0.85	0.41	0.88
1997	0.93	0.83	0.82	0.03	0.83
1998	0.93	0.87	0.82	0.55	0.86
1999	0.95	0.85	0.83	0.28	0.84
average	0.93	0.84	0.76	0.75	0.81