Effects of Waste Disposal on Groundwater and Surface Water (Proceedings of the Kxeter Symposium, July 1982). IAIIS l'ubl. no. 139.

Use of dissolved oxygen modelling results in the management of river quality: case history of the Willamette River, Oregon

David A. RICKERT US Geological Survey, Reston, VA 22092, U.S.A. ABSTRACT In 1973 the U.S. Geological Survey initiated a study of the River Willamette, Oregon to determine: (a) the major causes of dissolved oxygen (DO) depletion, and (b) whether advanced treatment of municipal wastewaters was needed to achieve the DO standards. The study showed that rates of carbonaceous decay were low (k = 0.03-0.06 day- 1 ) and that point source BOD loadings accounted for less than one-third of the satisfied oxygen demand. Nitrification of industrially discharged ammonia was the dominant cause of DO depletion. The study led to the calibration and verification of a steady state DO model which was used to examine selected scenarios of BOD loading, ammonia loading and flow augmentation. In 1976, the modelling projections for the River Willamette were presented to resource managers and, by 1980, Oregon had (a) decided that neither tertiary nor advanced secondary treatment was needed for municipal wastewaters, (b) instituted an effluent standard on the major discharger of ammonia, and (c) acknowledged the need for flow augmentation to maintain DO standards during the summer.

THE DISSOLVED OXYGEN STUDY Introduction Historically, maintenance of high DO concentration has been the critical problem in the River Willamette. During summer low flow periods, the DO concentration in Portland Harbour was often zero, and for years, low DO levels inhibited the fall migration of salmon from the River Columbia. In recent years, summer DO levels have increased dramatically and autumn salmon runs have returned. The improvement has resulted largely from reductions in the loading of point source BOD, coupled with streamflow augmentation from storage reservoirs. Since 1972, all wastewater discharges in the basin have received secondary treatment. However, the Oregon Department of Environmental Quality (DEQ) still regards maintenance of high DO levels as the factor of 13

14 D.A,

Rickert

highest priority in managing the quality of the river. Physical Setting and Hydrology The River Willamette basin, a watershed of nearly 30,000 km 2 , is located in northwestern Oregon between the Cascade and Coast Ranges. Within the basin are the State's three largest cities: Portland, Salem and Eugene and approximately 1.7 million people, representing 70% of the State's population. The River Willamette main stem forms at the confluence of its Coast and Middle Forks, south of Eugene, and flows northward for 301 km through the Willamette Valley. The river is composed of three morphologic reaches (figure 1, Table 1 ) . Each reach has a unique hydraulic regime and, therefore, different velocities, sediment transport characteristics and patterns of biological activity. METRES 160 120 80

I

^ ^ - - _

Upstrea n Reach

40 EVÊL3 0 0

200

I ~

— 43 km

Newberg

Tidal

Pool

Reach

100

DISTANCE, IN RIVER KILOMETRES ABOVE MOUTH

FIG. 1 Diagram representing morphologic reaches of the River Willamette, Oregon. Most of the flow in the Willamette occurs in the November to March period as a result of persistent winter rainstorms and spring snowmelt. TABLE 1 Physical characteristics of the Willamette River (for discharge at Salem = 170 m 3 s - 1 ) .

Approximate bed slope (m km- 1 )

Representaative midchannel water depth (m)

Upstream

0.53

Newberg Pool

0.023

Tidal

0.02

Reach

(m s-1)

Approximate travel time in reach (days)

2

0.60

2.8

7.5

0.08

3.9

12

Average velocity

0.03

10

Each summer there is a naturally occurring low flow period; the timing, duration and magnitude of which are now largely controlled by reservoir releases. Since 1952, when large scale reservoir regulation began, low flows have been maintained at a minimum of

Oxygen modelling

in river

quality

15

about 170 m 3 s-1 (Salem gauge) by reservoir augmentation. In comparison, for the unusually dry year of 1973, computations indicate the naturally occurring low flow was 92 m 3 s- 1 . Water temperatures in the River Willamette and in all tributaries reach a maximum during the annual July through August low flow period. Temperatures during July average about 18°C in the Upstream Reach, 20°C in the Newberg Pool and 22QC in the Tidal Reach. Data Collection and Interpretation Review of existing data indicated that an appreciable DO deficit occurs in the Willamette only below river kilometre (RK) 139 and during the yearly low flow period of July through August. The DO data collection programme was developed to formulate a mathematical model for simulating conditions below RK 139 for the critical summer period. Emphasis was placed on intensive direct measurement of waste loads and model coefficients to avoid reliance on literature values and the estimation of coefficients by computerized curve fitting. The intensive data collection programme was conducted during the summers of 1973 and 1974, and the interpretive results were provided to the Oregon DEQ in 1976. DO Sources and Demands Figure 2 shows the predominant factors found to control DO concentrations in various segments of the River Willamette under summer low flow conditions. The profile represents average daily DO concentrations measured in 1973 and is constructed with time-of-travel on the horizontal axis to permit examination of the rate of change in measured DO levels.

RIVER KILOMETRES ABOVE MOUTH

301

SALEM NEWBERG 138 80

WILLAMETTE FALLS

PORTLAND HARBOR -

16

9.6

Dischar ge at Salem • 170 m 3 s "

DO recovery

^ B e n t h i c demand —H-**j ry - N a t u r a l plus augmented streamfiow; atmospheric

i

I

oration; carbonaceous deoxygena' 15

TIME OF TRAVEL (DAYS!

FIG. 2 Average daily DO concentrations of the River Willamette under observed 1973 low flow conditions.

16

D.A.

Ri

chart

Along the course of the River Willamette the observed DO profile is the net balance of oxygen demands exerted and oxygen resources contributed. Nitrificiation, carbonaceous deoxygenation and a benthic demand are the oxygen demands; whereas natural and augmented streamflow, tributary inflows, atmospheric reaeration, reaeration at the Willamette Falls and Columbia River water are the oxygen sources. In the modelled segment of river, the composite DO profile shows a rapid decrease of DO from RK 139 to 84, a DO "plateau" in the Newberg Pool (see figure 1 and Table 1 ) , a DO increase over Willamette Falls, a gradual decline in DO between RK's 40 to 20, a sharp decrease in DO between RK's 20 to 11 and recovery of DO below RK 6. The DO decrease between RK's 139-84 amounts to 30% of DO saturation, a drop of approximately 2.5 mg l-1 at the prevailing water temperature. The decrease results from a large industrial load of ammonia that induces the growth of nitrifying bacteria on the shallow, cobbly, diatom encrusted river bottom. The nitrification proceeds at a high rate (k = 0 . 7 day-1 [log ] ) , outpacing the capacity of the river to replenish DO through atmospheric reaeration. Water entering the newberg Pool at RK 84 contains a residual level of ammonia, but owing to the dramatic change in morphology (Table 1 ) , no further nitrification is observed from analysis of nitrogen species data. The DO plateau in the Newberg Pool (RK's 84-43) represents an approximate balance between losses from carbonaceous deoxygenation and inputs from atmospheric reaeration. The in-river rate of carbonaceous deoxygenation is low (k = 0.03 day-1 [log ]) owing to (a) the fact that all wastewaters entering the river receive secondary treatment, (b) low non-point source (NPS) loadings of BOD . , and (c) the low surface area to volume ratio of the river in this reach. The same rate prevails in the Tidal Reach, whereas in the shallow Upstream Reach, k is about 0.06 day- 1 . Below RK 43 there are abrupt increases in DO resulting from aeration at Willamette Falls and inflow of high DO, low temperature water from the Clackamas River. Between RK's 40-20 the gradual decline in DO results from slow travel times (Table 1) which allow carbonaceous deoxygenation to exceed atmospheric reaeration. This DO loss is augmented by a benthic demand downstream of RK 20. Below RK 11 the profile flattens, and farther downstream, DO levels begin to recover. The flattening and recovery result from the mixing of Willamette water with upstream flow of cold, high DO water from the River Columbia. WIRQAS DO Model The model chosen for the study was developed and used for more than 30 years by C.J. Velz (1970). The basic model is applicable to conditions of steady (invariable), non-uniform (changing cross sectional geometry), plug (non-dispersive) flow. The computer program as formulated for the study is named the WIRQAS (Willamette Intensive River Quality Assessment Study) DO Model. The WIRQAS Model has been applied only for summer low flow, high temperature,

Oxygen modelling

in river

quality

17

steady state conditions. McKenzie et al. (1978) described the calibration and verification of the model. Explanation of Base Condition To make effective use of the steady state model, it was necessary to define a streamflow and waste loads that represented average summer conditions for 1973 and 1974. The associated DO profile could then be used as a base for comparing DO profiles simulated for anticipated future ranges of streamflow and waste loads. Streamflow The selected base condition flow was 170 m 3 s- x at the Salem gauge. This flow was selected as the comparative base because (a) the U.S. Army Corps of Engineers attempts to maintain this level as the minimum summer low flow needed for navigational purposes (and did so during most the summers of 1973 and 1974), and (b) observations by the Oregon DEQ and the U.S. Geological Survey indicated that this discharge is needed as a summertime minimum for achievement of State DO standards. Water temperatures used for the base condition were those measured during the 1973 verification period when the average streamflow was 170 m 3 s- 1 . For all curves, DO saturation at the model boundary point (RK 139) was set at the 1973-1974 average of 92%. Waste loadings In addition to a base condition streamflow, it was also necessary to select values from loading data measured during 1973 and 1974 to obtain an average summertime base condition. Table 2 summarizes the selected base condition loadings of BOD and NOD for both point and non-point sources (NPS) between RK's 139 and 8, plus the in-river loadings at the model boundary point (RK 139). All loadings were computed using the base condition flow of 170 m 3 s- 1 . Detailed data and the methods used to compute the values in Table 2 were reported by Mines et al. (1978). TABLE 2 Summary of base condition BOD , and NOD loadings to the modelled segment (RK's 139-8) of the River Willamette BOD

Sources

NOD

Total Loading — • (kg day-1) Percent (kg day-1) Percent (kg day-1) Percent

Nonpoint

28,900

43

2,500

5

31,400

28

Point

37,800

57

43,900

95

81,700

72

Total

66,700

100

46,400

100

113,100

100

To evaluate the significance of data in Table 2, it is important to note that only 57% of the total BOD , is derived from point sources. Moreover, Mines et al. (1978) reported that 41% of the total point source BOD ^ loading which enters the entire River River is from municipal wastewater treatment plants, whereas 59% is from industrial discharges. Regarding NOD, there is an overwhelming contribution of point source (95%) relative to NPS (5%) loading,

18

I). A.

Richert

with most of the point source ammonia being discharged from a Boise Cascade pulp and paper mill at RK 137. Oxygen Demands Exerted Using base condition inputs, the WIRQAS Model was run to determine the percentage by source of satisfied oxygen demands, between RK's 139-8, for the summer, low flow conditions of 1973 and 1974. The total demand exerted by all sources was 80,700 kg day-1 and the percentage contributions were as follows: Nitrification of point source ammonia 32 Point source carbonaceous deoxygenation 28 NPS carbonaceous deoxygenation 22 Benthic-oxygen demand 16 Nitrification of NPS ammonia 2 Total 100 The results show that oxidation of point source ammonia represents the largest oxygen demand even though it occurs exclusively between RK's 137-89. at the comparative point source loadings (BOD = 37,800 kg day- 1 , NOD = 43,900 kg day- 1 ; Table 2 ) , this finding results primarily from the much greater rate at which NOD is exerted (k = 0.7 day- 1 , k = 0.06 or 0.03 day- 1 ). In evaluating the impact of carbonaceous deoxygenation, it is important to consider the relative proportion of the point and NPS demands (Table 2 ) . The importance arises because during summer low flow conditions, NPS BOD . loading to the River Willamette represents essentially natural background demands that seemingly cannot be modified. This means, for base conditions, that only slightly more than one-half of the materials causing carbonaceous deoxygenation are point source derived and, therefore, possibly amenable to reduction through wastewater treatment. Implications of Modelling Results As an aid to river quality management, the WIRWAS Model was used to test planning alternatives concerning (a) BOD loading, (b) ammonia loading, (C) low flow augmentation, and (d) the effects of possible removal or reduction of the benthic oxygen demand in Portland Harbour. Figure 3 illustrates the type of information developed by presenting a single factor diagram for reservoir low flow augmentation. To aid the reader, the base condition DO profile is included in figure 3. Also included are the State DO standards, which change segment-to-segment as follows: RK's 0-43, 5 mg l- 1 , to protect anadromous fish passage and population by resident fish; RK's 43-81, 6 mg I- 1 , to protect anadromous fish passage and salmonid rearing; RK's 81-137, 7 mg l- 1 , to protect salmonid fish rearing and spawning; and above RK 137, 90% of saturation, to protect salmonid spawning. The range of flows from 92-283 m 3 s - 1 in figure 3 was selected because it represents an envelope around the summer low flows that are likely to occur in the future. The profile for 92 m 3 s - 1 requires some mention. This discharge is near the lowest minimum monthly flow ever recorded for July under natural (non-augmented)

Oxygen modelling

in river

quality

19

conditions. Moreover, predictions from a flow model indicate that without flow augmentation, natural streamflow would have approximated this level during July of the unusually dry year of 1973. Figure 3 shows the marked impact that flow augmentation has on river DO levels. At the 92 m 3 s- x discharge, the base condition BOD and ammonia loadings would cause violations of State DO standards by wide margins at most locations. The estimated DO saturation levels at 92 m 3 s- 1 are about 25% less than the base condition (flow augmented to 170 m 3 s- 1 ) at RK 89, and about 20% less at RK's 43 and 8.

SALEM 138

WILLAMETTE FALLS 43

NEWBERG 80

27

PORTLAND HARBOR

EXPLANATION Staie DO standards (mg r ' ) of percent saturation

150

120

90

60

30

0

RIVER KILOMETRES ABOVE MOUTH

FIG. 3 DO profiles for selected levels of streamflow at Salem. Point source BOD _ and NOD loadings held . . ult constant. At a flow of 142 m 3 s- 1 , the State DO standards would be violated between RK's 113-80 by as much as 10% of DO saturation. Moreover, at this flow and these base condition loadings, the DO standards would barely be met in the Newberg Pool and in the lower end of Portland Harbour. In evaluating increased flow augmentation, it should be recognized that improvements in quality are less dramatic as the DO profile is raised toward the level of saturation. Thus, increasing flow from 170 to 227 m 3 s- 1 causes DO improvements of only slightly greater magnitude than that achieved through an increase from 142 to 170 m 3 s- 1 . At higher levels, an increase from 227 to 283 m 3 s- 1 results in only about half of the DO improvement obtained by moving from 170 to 227 m 3 s- 1 . For the base condition loadings, it appears that augmentation of flows to above 198 mSs-1 (computed but not shown) would not provide enough improvement in DO levels to warrant taking the water from competetive uses.

20

D.A.

Rickcrt

CONCLUSIONS PRESENTED TO DEQ IN 1976 The results and implications of the DO study were presented to DEQ in 1976, together with a detailed series of one, two and three factor diagrams that depicted simulated DO levels for selected scenerios of future flow and wastewater loadings. The implications of the study were subsequently reported by Rickert & Hines (1978), and use of the diagrams for resource planning was described by Rickert et al^ (1980). The conclusions of the overall study encompassed five basic points: (a) Future achievement of DO standards in the River Willamette will require continued augmentation of summer low flows in addition to pollution control. Minimum summer flows of at least 170 m 3 s - 1 (Salem gauge) are needed to meet the standards at all locations below Salem. (b) Point source loading of ammonia is the major cause of oxygen depletion in the modelled segment of river (RK's 139-8). Because most of the ammonia comes from one industrial source, reduction of ammonia loading offers a relatively simple alternative for achieving a large improvement in summertime DO. (c) Although DO levels in Portland Harbour are currently above the State standard, removal or partial reduction of a localized benthic demand would improve summer DO concentrations between RK's 16-8. However, the feasibility of reducing the demand has yet to be determined. (d) BOD loading from municipal wastewater treatment plants presently exerts a relatively small impact on DO. However, increased efficiency of BOD removal at the largest municipal plants and at selected industrial plants might be desirable in the future. The benefits to be gained from this alternative would best be determined after the industrial ammonia loading has been reduced to a lower level. (e) For the foreseeable future, there is no need for municipal, advanced waste treatment to protect DO levels in the Willamette River. Maintenance of a minimum low flow of 170 m 3 s - 1 and reduction of industrial ammonia loading are the important management needs.

MANAGEMENT ACTION Since completion of the study, the Oregon DEQ and other agencies have made considerable use of the data and modelling results. This concluding section describes how the information has been used or has influenced management perceptions. New Schedule for Applying Effluent BOD Standards During the study period, the Oregon DEQ had firm plans to place, by 1980, a BOD effluent standard of 10 mg l- 1 on all municipal, wastewater treatment plants (WTP's) that discharged directly into the River Willamette. Such a standard was to be attained by upgrading treatment at every municipal WTP to high level secondary treatment.

Oxygen modelling

in river

quality

21

The WIRQAS study computed that implementation of the standard would decrease the municipal summertime loading of BOD to the Willamette from about 13,800 to 8,000 kg day- 1 . The largest individual decrease would be about 3,200 kg day-1 at the Salem WTP (RK 125.3). Using the WIRQAS model, the Geological Survey determined that, during the critical summer period, the reduced loading would produce only minor improvements in DO saturation, the maximum increase being 2% at RK 8. Further evaluation indicated the virtual lack of improvement stemmed from (a) the small reduction attainable in the total loading of BOD (about 9%), (b) the low rates at which BOD is exerted in the river (k = 0.06 or 0.03 day- 1 ), and (c) widely dispersed locations within the basin of the larger municipal WTP 1 s. Based on the modelling results, the Oregon DEQ dropped the 1980 achievement date and instead proposed that the standard be implemented only when each WTP reaches its hydraulic capacity and must be upgraded. This new approach is certainly more acceptable to municipalities. However, the projection still stands that implementation of the standard will achieve only a very slight improvement in the critical summertime DO. Therefore, it will be interesting to see if DEQ really pursues this issue as more pressing water quality problems arise. A First Time Ammonia Effluent Limitation Prior to the WIRQAS results, there were vague impressions among water quality specialists that nitrification might be occurring in the River Willamette. As previously noted, the study found that oxidation of point source ammonia represented the largest oxygen demand in the modelled segment of river. Moreover, most of the ammonia was discharged from one source, a Boise Cascade pulp and paper mill at Salem (RK 137). During the summers of 1973 and 1974, the maximum summertime daily discharge of ammonia-nitrogen from the mill was nearly 10,000 kg day- 1 . However, review of records indicated that substantially lower values were more representative of average summertime conditions. Accordingly, a value of 7,400 kg day-1 ammonia-nitrogen was used as the base condition input. DEQ conducted follow-up studies in 1976 which confirmed the WIRQAS results. Then, early in 1978, time came for renewal of the Boise Cascade discharge permit. Previously, the permit did not include an ammonia discharge limit. During the 1978 negotiations, an effluent limit of 3,409 kg day-1 ammonia-nitrogen was set for the critical DO period of 1 June - 31 October. For the remainder of the year, the limit was set at 4,546 kg day-1 ammonia-nitrogen. Acknowledged Need to Maintain Flow Augmentation Since initiation of significant flow regulation in 1952, the Corps of Engineers has attempted to maintain a minimum summertime level of 170 m 3 s - 1 at Salem. This low flow level was set primarily to enable efficient navigation in the Willamette, but, over the years, many

22 D.A.

Rickert

resource managers came to realize that the added water had a marked beneficial effect on summertime DO levels. In 1976, the WIRQAS results documented the significance of flow augmentation to the summertime quality of the lower Willamette River. For conditions existing during the mid-1970's, flow augmentation was more important to the attainment of DO standards than any increased amount of wastewater treatment. The Federal funds which built the Willamette Basin reservoir system were provided for flood control, navigation, power and irrigation. There was no mention of water quality enhancement. Therefore, it is now difficult, if not impossible, for the operating guidelines to be rewritten to acknowledge officially water quality enhancement as an objective of the system. However, the DEQ, the Corps of Engineers and other agencies now fully recognize the importance of flow augmentation. Since presentation of the WIRQAS results, there has been increased interagency communication and cooperation. There is now an unofficial but strong commitment that flow augmentation will be provided to aid the attainment of DO standards. REFERENCES Hines, W.G., McKenzie, S.W., Rickert, D.A. & Rinella, F.A. (1978) Dissolved oxygen regime of the Willamette River, Oregon, under conditions of basinwide secondary treatment. USGS Circ., 715-1. McKenzie, S.W., Hines, W.G., Rickert, D.A. & Rinella, F.A. (1978) Calibration and verification of an applied dissolved oxygen model of the Willamette River, Oregon. USGS Circ., 715-J. Rickert, D.A. & Hines, W.G. (1978) River-quality assessment: implications of a prototype project. Science, 200(4346), 1113-1118. Rickert, D.A., Rinella, F.A., Hines, W.G. & McKenzie, S.W. (1980) Evaluation of alternatives for achieving desirable dissolvedoxygen concentrations in the Willamette River, Oregon. USGS Circ., 715-K. Velz, C.J. (1970) Applied Stream Sanitation, John Wiley & Sons, Inc. New York.