Water Treatment and Removal Efficiencies for Dissolved Organic Carbon and Disinfection By-Product Precursors


Water-treatment processes are designed to remove organic carbon, among other constituents, to minimize the formation of DBPs. The CRW and LO DWTPs coagulate and chlorinate raw source water simultaneously, which reduces DOC concentrations in finished water to about 0.7 mg/L (fig. 25). The loss of DOC during treatment was partially dependent on the source-water DOC concentration entering the treatment plant such that when DOC concentrations were elevated, the percentage of DOC removed increased (up to about 50 percent). Although these samples with high concentrations of DOC resulted in the highest percentage of DOC removed, they also had the highest DBP concentrations in finished water (fig. 11).


During the water treatment process, there was a 60–85 percent reduction in fluorescing material between source and finished water; 3 representative sample pairs are shown in fig. 26. There was also a marked decrease in SUVA and an increase in the FI values, suggesting the DOM remaining in finished water following coagulation and chlorination contained lower aromatic content and lower molecular-weight substances, as is commonly reported (Kitis and others, 2001; Sharp and others, 2006; Beggs and others, 2009). Compared with source water, finished water had consistently higher percentages of C1 relative to the other components, although the total fluorescence loading from C1 was lower indicating some removal (fig. 26). Because C1 was highest in samples from the tributary sites most heavily impacted by anthropogenic activities (fig. 20), there may be a connection between these land uses and the presence of DOM that is less amenable to removal by standard coagulation. Additional jar test experiments with similar fluorescence measurement might identify treatment methods that target removal of such “C1 carbon” that might lead to reduced DBP concentrations in finished water.


Results from the treatability jar-test experiments conducted on CRW DWTP source water during the four basin-wide surveys provided information on the amount of DOC and DBP precursors that could be removed by coagulation itself and in combination with PAC. Source-water DOC concentrations during the experiments ranged from 0.9 to 1.7 mg/L (fig. 27). Coagulation with optimum doses of alum and ACH coagulant reduced DOC concentrations 30 to 39 percent; the removal rate was highest when the DOC concentrations were highest during the October 2010 storm (table 12). Following coagulation, SUVA decreased and FI values increased, indicating preferential removal of the higher molecular-weight, aromatic carbon. 


Co-addition of PAC and coagulants led to slightly higher removal of DOC (4–10 percent higher) and lower DBPFP values compared to coagulant alone for all jar tests except the one conducted on November 9, 2011, which showed no difference (fig. 27). These decreases, however, only represented a further reduction in DOC concentration of about 0.1 to 0.2 mg/L. The amount of DOM removed by the laboratory jar tests was higher than that measured in finished water by about 20 percent, which was equivalent to about 0.2 mg/L DOC. The reduction in DOC concentration following coagulation led to a similar decrease in THM precursor concentrations; however, the reduction in HAA precursor concentrations was much greater. HAAFPs decreased about 70 percent during all four jar tests (table 12). 


The use of coagulation during drinking-water treatment removes TPC and DOC. Because DOC concentrations in the Clackamas River have been historically low, the CRW and LO DWTPs were designed primarily to target the removal of particles during treatment; thus, the coagulation and chlorination steps take place simultaneously. Removal of TPC and DOC prior to chlorination would, however, likely further reduce the formation of THMs and HAAs. Results from the jar tests showed coagulation with alum and ACH is particularly effective at removing HAA precursors. The preferential removal of HAA precursors over the bulk DOM pool and THM precursors agrees with the finding that the HAA precursor pool is associated with more aromatic, high molecular-weight material and that the pool of DOM is more amenable to removal by coagulation (Croué and others, 2000; Liang and Singer, 2003; Hong and others, 2008). Because the BQ values for HAAs in finished water are occasionally elevated, upgrades to treatment plants that allow for coagulation and filtration prior to chlorination may, therefore, be a design improvement. 


The CRW and LO DWTPs occasionally use PAC during water treatment (up to 5 mg/L at CRW and up to 50 mg/L at LO) to control for tastes and odors. In contrast to studies that have found substantial removal of HAAs and THMs with commercial charcoal-filtration water pitchers (Levesque and others, 2006), the addition of PAC provided only minor improvement in DOC and DBP precursor removal, based on the four jar tests conducted here. A full evaluation of this method to reduce DBP precursors may be warranted because these data only reflect conditions during the study period. The benefits of PAC may be more pronounced during periods of higher algal contributions when the DOM pool contains higher fractions of low-SUVA, high-FI material.


First posted February 11, 2013