Potential Sources of Organic Matter to Tualatin River Sediments

To assess the degree to which a range of sampled materials (soil, leaf litter, detritus, plankton, duckweed, WWTF effluent) might be important sources of organic material to Tualatin River bed sediment, the carbon and nitrogen isotopic data and C/N ratios of these sources were compared to those of bed sediment. Because the organic material in bed sediment has decomposed over time, the isotopic and C/N values have likely changed relative to their original sources. Microbial decomposition increases the δ¹⁵N by a few per mil and can substantially decrease the C/N ratio (Middelburg and Nieuwenhuize, 1998; Kendall and others, 2001). Finlay and Kendall (2007) report that the δ¹⁵N increases associated with decomposition are most pronounced in streams that contain ample nitrate, a condition that is true for the Tualatin River. Decomposition also may alter the δ¹³C, but the effect is less and tends to be small (Kendall and others, 2001).

Several substances (bryozoan, duckweed, periphyton, and WWTF effluent particulate) can be dismissed as primary sources of organic matter to bed sediment because the patterns in their isotopic and C/N ratios do not match those of bed sediment, even if trends associated with decomposition are taken into account (fig. 11). The δ¹³C values for bryozoan and periphyton are substantially different from those of bed sediment. The δ¹⁵N values of bryozoan, duckweed, and WWTF effluent particulate are greater than those for bed sediment. The C/N values for bryozoan, periphyton and WWTF effluent particulate are less than those for bed sediment. As these substances decompose, their δ¹⁵N values can be expected to increase and their C/N ratios to decrease, making the differences compared to bed sediment even greater.

A comparison of seston and suspended sediment to bed sediment shows that the seston and suspended sediment samples that contain the most phytoplankton (those at and downstream of RM 16.2) do not resemble bed sediment for δ¹³C, δ¹⁵N or C/N (fig. 12). The expected changes associated with decomposition (increased δ¹⁵N, decreased C/N) will further amplify these differences. Seston and suspended sediment obtained from tributaries and in the more upstream reaches of the Tualatin River compare more favorably with bed sediment. It is difficult to know whether this suspended particulate organic material is a source to the bed sediment or vice versa. Bed sediment samples collected in late‑summer-1999 have lower δ¹³C and higher δ¹⁵N than other bed sediment samples, which might indicate some contribution from phytoplankton. If the contribution of phytoplankton to these samples were significant, however, the C/N ratio of the late-summer-1999 bed sediment samples should have been lower. For all of these reasons, phytoplankton probably is not contributing significant amounts of organic matter to Tualatin River bed sediment. This conclusion is consistent with field measurements of SOD rates in the Tualatin River, which indicated that the river reaches with the largest algal populations did not have significantly higher SOD rates, except at RM 5.5 where a slightly higher rate might be attributed to some deposition of algal detritus (Rounds and Doyle, 1997).

A comparison of terrestrial source materials to bed sediment shows considerable overlap for δ¹³C and δ¹⁵N for deciduous litter, woody material, and soil (fig. 13). The range of δ¹³C values for plant litter and woody material encompasses the δ¹³C range for bed sediment. Although δ¹⁵N values for many of the deciduous litter samples are lower than those of bed sediment, increasing these values by 2–3 per mil (within the expected effect of decomposition) places them directly in the range of δ¹⁵N values for bed sediment. The C/N ratios for all litter and woody material samples are much higher than those for bed sediment, but these ratios are expected to decrease with decomposition. In fact, the two samples of decomposed terrestrial detritus show C/N ratios that are much more similar to those of bed sediment. The δ¹³C, δ¹⁵N, and C/N data from soil samples also show substantial overlap with the bed sediment data. Although this could indicate that eroded soil is a source of bed sediment, it is equally consistent with plant litter and woody material being the primary source of organic matter to soil and Tualatin River bed sediment.

Principal Components Analysis

Principal components analysis (PCA) is a statistical technique used to reduced the dimensionality of datasets that contain multiple variables. In PCA, new axes (principal components) are created that are linear combinations of the original variables. The greatest variability within the dataset is captured by the first principal component (PC1). The second principal component (PC2) is orthogonal to PC1 and captures the greatest remaining variability. PCA was applied to this dataset as a method to visualize the three variables δ¹³C, δ¹⁵N, and C/N in a two dimensional plot. As part of PCA, the data were normalized (by subtracting the mean and dividing by the standard deviation of the dataset) to account for the different scales of the three variables (δ¹³C, δ¹⁵N, and C/N).

The results of the PCA are summarized in table 5. Together, PC1 and PC2 capture 86 percent of the variability in the dataset. The principal components plot is shown in figure 14. The black vectors (arrows) in that figure originate at the mean value for the entire dataset and point in the direction associated with each of the individual variables, δ¹³C, δ¹⁵N, and C/N. Note that the δ¹³C vector is pointing toward increasing negative values (decreasing δ¹³C ) because that direction is most pertinent to this discussion.

Examination of the principal components graph (fig. 14) leads to the following observations, several of which were noted earlier:

• Bryozoan, duckweed, and WWTF effluent particulate do not plot at locations similar to bed sediments and, therefore, probably are not major sources of organic matter to bed sediment.
• Several soil samples plot in locations similar to those of bed sediment, suggesting that soil may contribute organic material to bed sediment or that the organic material in soil and bed sediments have a common source.
• Terrestrial litter (coniferous, deciduous, or woody) have principal component values characterized by higher C/N and lower δ¹⁵N compared to bed sediment. Decomposition of this material results in decreased C/N and increased δ¹⁵N. Assuming a 25 percent decrease in C/N coupled with a 2 per mil increase in δ¹⁵N results in a trajectory that leads directly to the region where bed sediments plot. Examination of data reported by Middelburg and Nieuwenhuize (1998) for suspended matter in progressive stages of decomposition showed an approximate 10 percent decrease in C/N with each 1 per mil increase in δ¹⁵N, which is consistent with the calculated trajectory from this study.
• Decomposition generally leads to increasing δ¹⁵N and decreasing C/N. Although the slope is unknown, the typical direction is toward the upper left (lower PC1 and higher PC2). Any shift in this direction will make seston, suspended sediment, WWTF effluent particulate, periphyton, duckweed, and bryozoan less similar to bed sediments.
• Increasing amounts of phytoplankton are associated with increasingly negative δ¹³C and positive δ¹⁵N and, therefore, cannot be a major source to bed sediment. The bed sediment samples from late-summer-1999 from downstream of RM 26.9 shift slightly toward increasing phytoplankton, indicating that phytoplankton probably contributes a small amount of organic material to bed sediment during certain times in the lower reaches of the river.
• Suspended sediment and seston have similar characteristics, with seston showing an increased influence from phytoplankton. Some suspended sediment samples are similar to bed sediment and soil, but others are clearly affected by the presence of phytoplankton.

The principal components plot also enables a qualitative examination of the effect of mixing several sources. For example, a mixture of WWTF effluent particulate and terrestrial plant litter mathematically could lead to the observed bed sediment pattern. This mixture, however, is not realistic because not all bed sediments were collected downstream of WWTFs. In general, the bed sediment samples plot in a region that makes it unlikely that extensive mixtures of the potential source substances could account for the δ¹³C, δ¹⁵N, and C/N values of bed sediment. Bed sediment samples plot farther to the upper right than most of the potential source materials except soil and two unusual periphyton samples. Erosional and transport processes throughout the drainage basin make soil a likely potential source of organic matter in bed sediment. The possibility that some unusual types of periphyton could be present in sufficient quantity to create a mixture resembling bed sediment, however, is remote.

First posted August 17, 2010