Sources of Organic Carbon that Contribute Disinfection By-Product Precursors


Many lines of evidence were considered to evaluate possible sources of DOM that form regulated DBPs in the Clackamas River basin. The individual samples provided some indication as to where and when samples had elevated DBPFP values, and longitudinal and temporal variations in the data were useful for identifying patterns in the mainstem, especially when combined with Data Grapher analyses of the continuous water-quality data from the four-station monitoring network (fig. 1).


Considering that STHMFP and SHAAFP values were not correlated (appendixes F5 and F6), these two classes of DBPs seem to have different fundamental sources, which may require different watershed-management strategies for control. The filtered STHMFP value from North Fork Reservoir during an algal bloom (see photograph 2a-b) was the single highest value, which points to this source as worthy of future monitoring should THM concentrations increase in the future. The highest specific, or carbon-normalized, DBPFP values (table 9 and fig. 16) suggest that carbon most prone to producing THMs came from the reservoir and lower main-stem Clackamas River in September 2011; carbon contributing to HAA formation was highest in the three largest lower-basin tributaries, Eagle, Clear, and Deep Creeks during the initial October 2010 storm (fig. 14). These streams drain basins containing large proportions of private timberland, rural residential, and some agricultural land, primarily Christmas tree plantations, nurseries, cane berries, and some row crops (fig. 3). The greatest amount of agricultural land in the basin, about 16 mi², is contained within the Deep Creek Basin where nurseries and Christmas tree farms are abundant (Carpenter, 2003). It is, however, unclear to what degree each of these potential sources contributes DBP precursors to the mainstem and downstream drinking-water intakes.


While lower-basin tributaries had the highest concentrations of organic carbon (DOC and TOC) and DBP precursors (fig. 17), because of the relatively low flows from these streams most of the time, the primary sources of carbon are located in the upper forested basin where most of the flow also originates. This finding is consistent with a study of the nearby McKenzie River that also found DBP precursors to be primarily derived from upper-basin terrestrial carbon (Kraus and others, 2010). Our results also show elevated THM- and HAA-precursor concentrations when DOC concentrations also were elevated (fig. 11), and DBPs were formed primarily from chlorination of dissolved, terrestrially-derived organic compounds such as humic and fulvic acids. This is consistent with results for the DBPFPs that showed the dissolved fraction dominating the precursor pool (fig. 14); although particulate carbon did contribute some DBP precursors, most were dissolved. 


Terrestrial Inputs


Multiple lines of evidence indicate terrestrial inputs are the dominant sources of carbon to the lower Clackamas River. First, most of the DBPFP loads in the lower Clackamas River at the CRW DWTP are already accounted for at Estacada (table 10), which drains a mostly forested basin. Second, the fluorescence data and PARAFAC model identified five dominant components, four of which are soil-related (fig. 21). Third, the co-dominant HAA in finished water was TCAA, an organic compound widely found in forest soils (McCulloch, 2002). 


TCAA is a chlorinated hydrocarbon with many sources. It is used for many industrial purposes, including the synthesis of other organic compounds, and as an herbicide, for example, but it is also a breakdown product of TCE and other solvents. TCAA is also formed during various chlorination processes, in wood-pulp processing and drinking-water treatment (McCulloch, 2002). TCAA is measureable in the atmosphere at concentrations ranging from less than 0.02 µg/L in Switzerland up to 20 µg/L in urban Berlin, Germany; global average concentrations are about 0.5 µg/L (McCulloch, 2002). Although TCAA is not volatile, it is highly soluble in water and can precipitate in rain; this explains its prevalence in forest soils, especially coniferous soils (McCulloch, 2002). While it is possible there is some background level of TCAA in the Clackamas River derived through this process, this has not been investigated.


Although unstudied here, the erosion of TCAA-containing soils might be a factor in the prevalence of TCAA in drinking water from the Clackamas River. In a previous study, Carpenter (2003) found that levels of total phosphorus (TP) in the forested tributaries of the Clackamas Basin were highly correlated with percentage of “non-forest upland”—mostly timber harvest areas (r = 0.96, p <0.001). TP also was correlated with silica concentrations (r = 0.84, p <0.001), which suggests soil erosion could be a source of phosphorus to the river. It is possible the loss of particulate and dissolved carbon from the forested watershed areas are lost in a similar fashion through erosion, although leaching of DOM from organic soil horizons also certainly contributes to riverine carbon. 


The flushing of decomposed organic matter by autumn and winter rains is a complex process governed by the types of organic matter and microbes present, thickness of the vadose zone and local water-table dynamics, temperature, and other factors. Decomposition of plant materials is a key process that successively transforms solid organic matter into fine particles, colloids, and DOM solutes that can leach into surface waters and form DBP precursors.


It is hypothesized here that increased deposition of nitrogen, from atmospheric and other sources, has accelerated decomposition rates of organic matter in portions of the watershed including deforested and previously harvested regrowth forests. Because of the large store of bulk carbon in the forest and the extensive timber harvesting in the basin (see numerous U.S. Forest Service “watershed analyses” reports referenced in Carpenter [2003]), decomposition of organic matter and burned slash left over from previous harvest operations might be an important source of carbon to the river that may have increased over time and may help explain the increase in DBPs observed over the last three decades (fig. 4). Although Federal forestland in the Pacific Northwest, including most of the Clackamas Basin, has been recovering since the Northwest Forest Plan was enacted in 1993, timber harvesting was extensive in the late 1980s and early 1990s, which produced a large reservoir of decaying organic matter that may have subsequently leached DOM. Previous studies cited in Turner and others (2011) suggest that estimates of forest carbon stocks in the region covered by the Northwest Forest Plan had declined substantially during 1953–87 from high rates of timber harvesting. In recent years, however, regrowth of many of these forests has increased net ecosystem productivity, thereby possibly shifting the balance toward storing carbon through sequestration (Turner and others, 2011). Whether or how this may change export of DBP precursors is, however, not known but could be examined with further study.


Algae


Even though algae fix carbon through photosynthesis and contribute to organic-matter pools, the degree to which benthic and phytoplankton algae contribute DBP precursors in the Clackamas River remains an important unanswered question. Although it is well-established that algae can be a source of DOM-containing DBP precursors, there is conflicting evidence regarding whether DOM produced by algae is more or less reactive per unit carbon compared with terrestrial sources (Jack and others, 2002; Nguyen and others, 2005; Kraus and others, 2011). Differences in DBP-precursor content of algal-derived DOM likely arise from a combination of factors, including algal species, growth stage, release of extracellular material, and environmental processing of algal-derived material and compounds (Jack and others, 2002; Nguyen and others, 2005). 


To initially address this question, correlations were examined between chlorophyll-a and DBPFP (fig. 29). This approach focuses on the viable particulate algal cells that fluoresce and did not include non-fluorescing particles, such as dead or decaying algal filaments. Although the correlations between chlorophyll-a and THMFP were weak, the relation with HAAFP was significant (p <0.001) for filtered (r = 0.56) and unfiltered (r = 0.49) samples (fig. 29). 


Other lines of evidence suggest algae are having an influence over carbon amount and quality in the Clackamas River. While water-column chlorophyll-a concentrations were atypically low in the lower mainstem (fig. 8), there were several observations from the data that not only demonstrate how algae can affect carbon amounts and DOM quality in the river, but also provide some evidence algae is contributing THM precursors. 


The relatively high STHMFP in the September 2011 sample from the release depth within North Fork Reservoir (table 9), for example, indicated the DOM present during the blue-green algae bloom was reactive and formed THMs. The amount of reactive organic matter produced by any individual bloom (and the production of algal toxins or taste and odor compounds) could depend on a number of factors, including the overall size and health of the bloom. Future studies of Timothy Lake or North Fork Reservoir could begin to characterize these processes. The concentrated “algae grab” sample with a high abundance of Anabaena flos‑aquae (fig. 20) shows a strong signal in the lower ultraviolet “protein-like” region, which was interpreted in this system to be indicative of the more labile, freshly produced organic matter expected in such a sample (fig. 21). This signal, represented by PARAFAC component C5, was also apparent in the full suite of reservoir samples (fig. 22). Organic matter recently contributed by these blooms is expected to contain a greater protein-like signal, particularly because Anabaena flos-aquae is a nitrogen-fixer and a common member of the phytoplankton assemblage in North Fork and Timothy Lake during summer. 


As discussed above, algae may also contribute HAA precursors, as the positive correlations with chlorophyll-a suggest (fig. 29). Further, the highest HAAFP values occurred during the October 2010 storm (fig. 16), at the end of the growing season, when concentrations of chlorophyll-a due to algal sloughing were highest.


Phytoplankton Blooms


Blooms of blue-green algae occurred in Timothy Lake and North Fork Reservoir during this study, prompting the Oregon Health Authority to issue a human-health recreational advisory for both water bodies. Although these blooms were not as severe as in years past, elevated STHMFP values in North Fork Reservoir in September 2011 (table 9) indicate a high degree of reactivity for this type of carbon, a finding that is consistent with other studies (Jack and others, 2002; Kraus and others, 2011). Because the phytoplankton blooms in 2010 and 2011 were seemingly small, the high degree of reactivity shown by the high STHMFP suggest phytoplankton populations in the reservoirs could become important sources of THM precursors, especially if larger blooms occur in the future.


As described above, phytoplankton in the two Clackamas reservoirs could be an important source of DBP precursors for the Clackamas River. Blooms of blue-green algae and diatoms regularly occur in these reservoirs (see Carpenter, 2003), although the severity and duration vary widely, likely responding to the specific growing conditions and other factors that affect bloom development. Although sustained flow during summer in 2010 and 2011 may have limited the residence time in North Fork Reservoir, blooms were observed in the reservoir both years, but biomass was relatively low—less than 3 µg/L chlorophyll-a at the log boom. The horizontal and vertical mobility of these blooms makes it challenging to accurately characterize conditions with just one sampling. Considering longitudinal patterns, the STHMFP measurement in North Fork Reservoir was nearly two times higher compared with Carter Bridge in late summer, possibly reflecting the importance of the reservoir-derived carbon in forming THMs and possibly also contributing to taste and odor issues. The DBPFP measurements also suggested phytoplankton or their exudates produce DBP precursors. The highest STHMFP value for filtered samples, again, was from the release depth within North Fork Reservoir (table 9) and suggests the reservoir was a source of THM precursors to downstream sites.


Previous studies have examined fluxes of DBP precursors in reservoirs and lakes and found them to act as either a source or sink for DBP precursors (Stepczuk and others, 1998; Nguyen and others, 2002; Bukaveckas and others, 2007). In a recent study examining changes in DOM amount and composition in San Luis Reservoir, California, it was found that DBP precursors (particularly HAAs) increased during the summer months because of high phytoplankton activity; however, the reservoir was a sink for DBP precursors during the winter months when decomposition processes predominated (Kraus and others, 2011). Results from the Clackamas River study also suggest that information about the composition of organic matter within reservoirs, as well as its reactivity with respect to THM and HAA formation, can help explain downstream changes in source-water quality.


Benthic Algae and Periphyton Sloughing Events


Although conditions were not particularly favorable for algal growth in 2010–11, nuisance levels of benthic algae (greater than 100–150 mg/m²; Welch and others, 1988) developed in some main-stem and tributary locations during this study (table 7). Large masses of stalked diatoms, filaments of green algae, and ears of blue-green algae (see photographs 1a-e) were common among other types of algae in the mainstem.


On the basis of water-column chlorophyll-a concentrations in the lower mainstem, which were nearly always relatively low (less than 2 µg/L), it appears on initial inspection that benthic algae were not important contributors to the water column during the study. While large algal “sloughing” events did not occur, higher-than-normal streamflow and other possible factors such as high rates of benthic invertebrate grazing or other factors may have limited or moderated water-column concentrations of chlorophyll-a during the study, as discussed above. Regardless of this finding, recognizing the unusually high streamflow, the high algal biomass observed in the river (table 7), along with the knowledge of past algal sloughing events (U.S. Geological Survey, 2012), it is suspected that sloughed benthic algae might, at times, be a significant source of DBP precursors to the lower river.


As described previously, sloughing of benthic algae is a common feature of the Clackamas River and some other Cascade Range rivers such as the North Umpqua and Rogue Rivers. Such events do not necessarily take place every year, but in some years (fig. 30), concentrations can be high for a system dominated by periphyton. Such sloughed algae impacts water clarity and fishing conditions in the Clackamas River and has clogged DWTP intake screens in the past. This demonstrates that, at times, benthic algae are a source of TOC in the river and, thus, likely a source of DBP precursors.


These events are sometimes associated with increased streamflow resulting from storms, the annual drawdown of Timothy Lake (which increases flow in the mainstem by 100 ft³/s or more), and (or) changes in reservoir releases. For example, chlorophyll-a concentrations in the tributaries reached moderate levels (5–8 µg/L) during the October 2010 storm, which contributed to observed increases in the mainstem. The algal biomass in the mainstem more than doubled between Barton and Carver from tributary inputs—especially Deep Creek—and possibly also from benthic algae sloughed from main-stem locations downstream from Barton. Although DBP concentrations were not particularly high in finished water at the time, it raises the question about the degree to which algae may be contributing DBP precursors because chlorophyll-a levels can become elevated in the river, as this storm demonstrated. 


Hydroelectric project operations, including flow ramping for power production and the annual drawdown of Timothy Lake near the end of summer starting around Labor Day, can increase the amount of algal cells, DOC, and TOC in the upper river (Carpenter, 2003). The drawdown releases phytoplankton produced through the summer growing season, and increased flows—about 10 percent higher flow in the mainstem—can scour and suspend benthic algae into the water column. Even though, as a condition of the new Federal Energy Regulatory Commission operating license, flow ramping in the upper river at the Three Lynx powerhouse, upstream from Three Lynx Creek (fig. 1), is not as great as in years past, hydroelectric project operations still cause abrupt changes in flow that can scour algae from the riverbed and increase carbon concentrations in the river.


Although not explicitly measured here, the Timothy Lake water also may contribute DBP precursors, although it did not appear to increase DBPFP substantially at Carter Bridge in this study during the drawdown (fig. 14). A comparison of the longitudinal conditions in the mainstem near the beginning of the September 2011 drawdown period to those later during the drawdown, near the peak in flow, shows the TPC and THMFP did increase in samples from Estacada downstream to the LO DWTP intake (fig. 31A). The unfiltered formation potentials were 22–81 percent higher than the filtered THMFP (not shown), suggesting a particulate source.


Whether these September 2011 increases were due to sloughing of periphyton, or from displacement of bottom water from North Fork Reservoir (fig. 32) is not known. This “displacement” mechanism was first proposed by the USGS during the August 2003 taste and odor event at the CRW DWTP, when decreases in water temperature and increases in turbidity were noted at the Estacada water-quality monitor downstream from the North Fork Reservoir at the time of the drawdown. Taste and odor problems eventually developed in early October 2011, prompting the CRW DWTP to begin using powdered activated carbon (PAC). The cause of the event was not identified, but in August 2003, taste and odor problems were traced to geosmin, a known taste and odor compound (Graham and others, 2010), which was detected at concentrations between 0.024 and 0.044 µg/L in North Fork Reservoir at depths of 80 and 30 ft, respectively (data furnished by CRW). Given that a bloom of blue-green algae did occur in 2011, it is certainly possible that this contributed to the observed taste and odor problem. While water displacement from North Fork Reservoir may have contributed to the 2011 taste and odor event, the cooler temperatures and higher flows in 2010–11 probably minimized this process because residence times were likely shortened and thermal temperature stratification weaker compared with 2003. Drawdown also may have enhanced the export of blue-green algae cells to the outflow in 2011 by disrupting population “layering” at depth within the reservoir, which has been noted previously for this reservoir (Carpenter, 2003).


The September 2011 event also provided an opportunity to examine how water-column chlorophyll-a levels and turbidity respond to the increased flow in the river during the drawdown (fig. 31B). Even though the drawdown was one of the main differences between these dates, there was a jump in flow of 190 ft³/s in the lower river that did not cause much of a response from the chlorophyll-a sensor. This was not totally unexpected because it is probably the more senescent algae that would detach during the drawdown (if it were benthic algae). If blue-green algae cells were released from the reservoirs, those also would not cause a large response from the chlorophyll-a probe; phycocyanin probes are much better suited for detecting blue-green algae. Although the full time-series data, which has considerable within-day variation, does not show any apparent increase in chlorophyll-a during drawdown, the daily median values (48 daily measurements) do show a very small increase (fig. 31C) that may have been caused by algal sloughing. These increases were not large, however, but the small jumps in streamflow that occurred at the Oregon City streamflow gage do seem to line up, more or less, with these small increases in chlorophyll-a, possibly indicative of scouring.This process has been noted previously in May 2003, for example (U.S. Geological Survey, 2012). The increases in chlorophyll-a in 2011 were, however, quite small and probably within instrument error. While this does add to the weight of the evidence that algae are a source of DBP precursors and provides a possible mechanism that could enrich source waters, these data are not sufficient to prove or disprove the hypothesis regarding algal sloughing.


During this possible sloughing event, the THMs in finished-water samples from the CRW DWTP actually decreased, while those from the LO DWTP increased. The lower THM4 values at CRW were not, however, attributable to the use of PAC, because it was not used during the active data-collection phase of the study but was employed later in October 2011 as taste and odor problems were developing.


Wastewater


Although the Clackamas River indirectly or directly receives wastewater effluents from three wastewater treatment plants (fig. 1) and effluents from possibly thousands of septic tanks in the basin, the effluents did not produce a carbon signature that could be construed as being definitively related to wastewater. Although sampled less frequently (n = 3), there was also no definitive wastewater signal from Deep Creek, which receives wastewater-treatment-plant discharges from the city of Sandy and town of Boring through Tickle Creek and North Fork Deep Creek, respectively. 


First posted February 11, 2013