Characterization of Bed Sediment and Source Materials

Isotopic results for δ¹³C and δ¹⁵N as well as the C/N ratios for each sample are shown in table 6 (at back of report). The full datasets are shown in appendix A.

Bed Sediment

δ¹³C.—The δ¹³C of bed sediment for Tualatin River basin samples generally is between -29.4 and -26.9 per mil (fig. 4). No general trend with river mile was evident, although the lowest values of δ¹³C in the river tended to occur at RM 9.9. Among the tributaries, δ¹³C values for the Rock and Fanno Creek sites were lower than those for either the Gales or Dairy Creek sites. There were no statistically significant differences between tributary and river sites; however, the Rock and Fanno Creek results more closely resembled Tualatin River results from RM 8.7–9.9, whereas the Gales and Dairy Creek results were more similar to results from most upstream river sites.

The δ¹³C results for the late-summer-1999 samples were lower than those obtained at any other time. The lowest δ¹³C value measured for bed sediment, -30.6 per mil, was obtained during this sampling period. Median differences for pairwise site-matched comparisons between late-summer-1999 results and those from mid-summer-1999 and early-summer-2000 were -1.9 and -1.4 per mil, respectively, which were statistically different from zero (α=0.05, fig. 5). Statistical tests for differences with late-summer-1998 and with winter-2000 were not possible because the number of samples was too small.

Pairwise site-matched comparisons also showed that δ¹³C for mid-summer-1999 was greater than δ¹³C for early‑summer-2000. Because the difference was consistent among all samples (river and tributaries), it was highly significant (p<0.001), but it also was small (0.7 per mil). Although this difference may reflect a real seasonal effect (such as differences in source material), it is possible that this is a spurious correlation or an artifact related to laboratory performance that changed over time.

δ¹⁵N.—The δ¹⁵N of bed sediment for Tualatin River basin samples generally is between 1.7 and 6.3 per mil (fig. 4). No trends related to season or sampling period were observed.

Inspection of the δ¹⁵N results shows that the data divide neatly into two groups: Group A comprising lower river sites at and downstream of RM 26.9 plus the Rock Creek site, and Group B including river sites upstream of RM 26.9 plus the Gales, Dairy and Fanno Creek sites (fig. 4). The δ¹⁵N of Group A generally is between 4.9 and 6.3 per mil, whereas the δ¹⁵N of Group B generally is between 1.7 and 4.9 per mil. The difference in δ¹⁵N values between these groups is statistically significant. Two possibilities that could account for this difference are decomposition and WWTF effects. Increases in δ¹⁵N values of several per mil have been observed to occur with progressive decomposition (Bernasconi and others, 1997; Kendall and others, 2001). If the degree of decomposition explains the difference between Groups A and B, it could indicate that the sediments in the tributary and upper river sites were “fresher” than materials found in the lower river. This explanation is consistent with observed differences in measured SOD rates, which were statistically higher at tributary sites than at Tualatin River sites (Rounds and Doyle, 1997). Alternatively, if the dissolved nitrogen discharged from the WWTFs tends to be heavier (higher δ¹⁵N), then it is possible that vegetation in the riparian corridor downstream of the Rock Creek WWTF (RM 38.1) also might have a heavier nitrogen content. Transport of leaf litter with a higher δ¹⁵N from the near riparian area might help explain the higher δ¹⁵N of the bed sediment samples collected at and downstream of RM 26.9. Neither of these hypotheses, however, can be tested with the available data.

The δ¹⁵N results from late-summer-1998 bed sediment samples should be regarded with some skepticism. These samples were collected and processed using a different method than the other samples (cores versus surficial sediment grab samples, and drying at about 50 ºC versus freeze drying). It is not known if these collection and processing differences affected the results, but it would not be unreasonable to suspect that the higher temperature drying method caused additional nitrogen fractionation. Laboratory personnel remarked that the nitrogen in Tualatin River bed sediment samples was particularly labile. Because of the processing difference, these samples were omitted from further analysis.

C/N ratio.—The C/N atomic ratio of bed sediment samples from the Tualatin River basin generally falls within a narrow range of 12.5–17.7 (fig. 4). No trends related to river mile or sampling period were observed with the exception of the two winter-2000 samples which had much higher C/N ratios. Both samples from winter-2000 contained woody debris and although the larger pieces were removed from the samples during processing, it is likely that some wood remained in the sample at the time of analysis. Because of this confounding variable and the general lack of winter samples, it is impossible to draw conclusions about C/N ratios in bed sediment during the winter season. Two samples obtained in early-summer-2000 had particularly low C/N ratios which cannot be attributed to any known cause.

Leaf Litter and Soil

Coniferous litter, deciduous litter, and woody litter show similar values of δ¹³C, δ¹⁵N, and C/N ratios (fig. 6). The only statistically significant difference among these sample data is that δ¹⁵N is slightly higher for woody material than for coniferous litter. Different plants utilize different biochemical pathways during photosynthesis, and the pathway affects the isotopic composition. The ranges of δ¹³C, δ¹⁵N, and C/N values measured in this study are similar to those reported elsewhere for plants that use the C3 pathway of CO₂ uptake (table 4). Vegetation in the riparian areas of the Tualatin Basin is almost exclusively trees and shrubs that utilize the C3 pathway, rather than plants that use the C4 pathway (corn, bamboo and tropical grasses) or the CAM pathway (cacti and desert plants).

The range for C/N is wide, with the highest values associated with cones, seed pods and woody material. Wide ranges of C/N are not uncommon for terrestrial material in various stages of decay (Kendall and others, 2001).

As materials decompose, the carbon and nitrogen content both decrease. The rates of decay are such that the C/N ratio generally decreases as the material becomes progressively more decomposed (Middelburg and Nieuwenhuize, 1998; Kendall and others, 2001). In addition, δ¹⁵N tends to shift toward higher values as the lighter isotopes are removed during decomposition (Bernasconi and others, 1997; Middelburg and Nieuwenhuize, 1998; Kendall and others, 2001). Reported changes in δ¹³C with decomposition, however, show less agreement. Middelburg and Nieuwenhuize (1998) reported that the δ¹³C was 2-3 per mil lower for more degraded material, but Kendall and others (2001) reported that δ¹³C usually increased with decomposition, although the changes were small.

Two samples of decomposed terrestrial detritus were compared to plant litter in this study. Because only two samples of decomposed terrestrial detritus were collected and the same material was not tracked through the decomposition process, caution in interpretation is warranted. The δ¹³C values for detritus are within the range for plant litter, suggesting that δ¹³C changed little during decomposition for these samples. The detrital δ¹⁵N values, however, are at the high end of the range for litter, suggesting that some fractionation of nitrogen might have occurred during decomposition. Furthermore, the C/N ratios for the detritus samples are at the low end of the range for litter, supporting the hypothesis that the C/N ratio decreases with increasing decomposition.

Soils collected in the Tualatin River basin resemble plant litter for δ¹³C and δ¹⁵N, but have lower C/N values that are similar to those of the decomposed terrestrial detritus (fig. 6). This change in C/N ratio is what would be expected as soil organic matter derived from plant material ages and becomes more decomposed.

Macrophytes

Duckweed is a common macrophyte in the Tualatin River. Duckweed mats on the order of 1 m² or larger are not unusual during low flow in the summer, particularly where they collect on the upstream side of a stationary floating object such as a log. In addition to duckweed, these mats collect small cones and twigs. Compared to terrestrial plant litter, sorted duckweed (only duckweed leaves and roots) had slightly lower δ¹³C (-29.8 to -28.5 per mil), higher δ¹⁵N (6.8 to 10.5 per mil), and lower C/N (11.2 to 14.6) values (fig. 6); the differences for δ¹⁵N and C/N were statistically significant. The δ¹³C is less than typical values reported in the literature for aquatic macrophytes (-27 to -20 per mil; Finlay and Kendall, 2007). The C/N range is particularly narrow for duckweed compared to terrestrial plant litter. Compared to sorted duckweed, the unsorted material had δ¹⁵N and C/N values closer to leaf litter, showing the influence of cones and twigs. The decomposed duckweed sample had values similar to the unsorted material.

Two large WWTFs discharge treated municipal effluent into the Tualatin River at RMs 38.1 and 9.3. The δ¹³C values for effluent particulate (>0.45 μm) are similar to those for terrestrial plant litter and soils (fig. 6). The δ¹⁵N values are significantly greater than those measured for the terrestrial plants or soils that were sampled. δ¹⁵N results for sorted duckweed and WWTF effluent particulate samples are similar (7.2 to 11.6 per mil). This similarity may be a coincidence, but it is not unreasonable to conclude that duckweed downstream of RM 38.1 obtain a substantial amount of nitrogen from WWTF effluent. Bioavailable nitrogen is in abundant supply due to the WWTF discharges, and uptake in such circumstances typically does not alter the isotopic composition of the nitrogen source (Finlay and Kendall, 2007). Because only WWTF effluent particulate were analyzed in this study, it is not known if the δ¹⁵N of dissolved nitrogen in effluent is similar. Compared to other sources, the C/N values for effluent particulate are very low (5.3 to 8.0).

Phytoplankton and Periphyton

The reservoir reach (RM 30–3.4) of the Tualatin River is known for algal production. Because of the river depth and turbidity in this reach, little light reaches the river bottom except at the edges. As a result, most algae in this reach are phytoplankton (free-floating algae); however, periphyton (attached algae) are common at a few riffle sections. To sample phytoplankton, seston samples were collected during the summer from RMs 26.9 to 5.5. As previously described in the Sample Processing section, two size fractions were prepared. Figure 7 shows that the δ¹³C, δ¹⁵N, and C/N values for the two size fractions of seston (>80 μm and 80–202 μm) are nearly identical, indicating that the presence of zooplankton or macroscopic debris in the >80 μm sample did not have a significant effect. Only results for the 80–202 μm size fraction are included in subsequent data analyses.

Seston samples showed distinctive and statistically significant (α=0.05) trends with river mile for δ¹³C, δ¹⁵N and C/N (fig. 8). The δ¹³C decreased from an average of -28.8 per mil at RM 26.9 to an average of -38.0 per mil at RM 5.5. Similarly, C/N values decreased from 16.5 to 6.1. The δ¹⁵N showed a similarly strong, but opposite trend; the average δ¹⁵N increased from 5.8 to 13.4 per mil. These changes reflect the increasing presence of phytoplankton in the lower (more downstream) reaches of the river as evidenced by increasing chlorophyll-a concentrations. Samples that contain moderate to large amounts of phytoplankton (>15 μg/L chlorophyll-a), therefore, are characterized by low δ¹³C (< -32 per mil), high δ¹⁵N (>8 per mil) and low C/N (<12). These effects of location and chlorophyll-a level are clearly evident when the δ¹⁵N and C/N are plotted against δ¹³C in figure 9 with annotations showing the chlorophyll-a concentrations.

Hagg Lake seston showed a pattern that was different from seston in the Tualatin River (fig. 8). Visual examination of all Hagg Lake seston samples showed appreciable amounts of phytoplankton. The C/N ratios of these samples are similar to the samples from the lower river that contain the most phytoplankton, but the δ¹³C and δ¹⁵N values indicated much less fractionation. The low concentrations of nutrients in Hagg Lake compared to the lower Tualatin River may account for the differences between the seston samples from these locations. Algae have been reported to exhibit less isotopic fractionation when carbon and nitrogen are in short supply relative to the algal growth rate (Finlay and Kendall, 2007).

Periphyton samples from this study show a wide range of δ¹³C, δ¹⁵N, and C/N values (fig. 8). Although δ¹³C values for most periphyton samples are similar to those for phytoplankton-rich seston (<-32 per mil), two samples (one‑quarter of the total) had anomalously high δ¹³C values (near -18 per mil). No other samples collected during this study had δ¹³C values that high. The δ¹⁵N values for periphyton generally were less than those for seston samples with high levels of chlorophyll-a (<8 per mil compared to >8 per mil); the C/N values for phytoplankton-rich seston and periphyton were similar (<12).

Suspended Sediment

Suspended sediment samples show trends with river mile that resemble those of seston samples, but the trends are not as robust: δ¹³C and C/N decrease with increasing river mile and δ¹⁵N increases (fig. 10). Suspended sediment samples with low δ¹³C (<-32 per mil), high δ¹⁵N (>7 per mil) and low C/N (<10) were from the same sections of the river where phytoplankton were most prevalent. This result is not surprising because the suspended sediment sampling did not attempt to exclude phytoplankton. As a result, the filterable material from the lower river reaches includes a substantial amount of plankton during the summertime growing period. The main difference between the suspended sediment and seston samples is that the former includes particles both smaller than 80 μm and larger than 202 μm. The particles smaller than 80 μm are composed of plankton as well as colloidal materials from soil and streambank materials. Particles larger than 202 μm include some zooplankton as well as larger sediment and detrital materials. The more robust trend in the seston samples, therefore, is consistent with the fact that the seston samples are composed predominantly of plankton whereas the suspended sediment samples are a mixture of biological and other materials.

The decreasing trend in δ¹³C with river mile persists somewhat for the suspended sediment samples collected in winter, despite the fact that phytoplankton are not prevalent at that time of year due to high flow, limited light, and cold temperatures. The general lack of phytoplankton in winter is consistent with the absence of any trends with river mile for δ¹⁵N and C/N during the winter sampling. Consequently, the trend in δ¹³C during the winter is not directly attributable to downstream trends associated with plankton populations.

The δ¹⁵N and δ¹³C results show that the suspended sediment samples can be separated into two groups: samples collected from the Tualatin River at and downstream of RM 26.9 have lower δ¹³C and higher δ¹⁵N than samples collected from the upper river and tributaries (α=0.05). The primary difference, as discussed above, may be caused by the presence of more phytoplankton in the lower river samples.

First posted August 17, 2010