Reallocation of Discretionary Diversion from Lake ...

1 downloads 0 Views 730KB Size Report
Consultant, Greenfield, WI 53221, USA, Email: [email protected] ..... the Central and Wisconsin Railroad and a small effect at Southwest Highway ...
Proceedings of 2013 IAHR Congress © 2013 Tsinghua University Press, Beijing

Reallocation of Discretionary Diversion from Lake Michigan to Improve Water Quality in the Chicago Area Waterways System Charles S. Melching Consultant, Greenfield, WI 53221, USA, Email: [email protected]

ABSTRACT: The Metropolitan Water Reclamation District of Greater Chicago is allowed to divert an annual average flow of 7.65 m3/s from Lake Michigan to improve water quality in the Chicago Area Waterways System (CAWS). A modeling study was done to determine if compliance with dissolved oxygen (DO) standards could be improved on the North Shore Channel (NSC) without adversely affecting compliance with the DO standards anywhere else in the CAWS while staying within the discretionary diversion limit. The DUFLOW modeling system developed in the Netherlands has been adapted to simulate water quality in the CAWS. This model was applied to evaluate the reallocation of discretionary diversion from Lake Michigan for three representative water years, a wet year (2008), dry year (2003), and medium year (2001). It was found that by reallocating the discretionary diversion within the CAWS full compliance with the DO standards could be achieved at Simpson Street and Main Street on the NSC during dry weather. Compliance over the entire year from 95.8 to 97.0% at Simpson Street and from 92.0 to 96.7% at Main Street could be achieved without negatively affecting compliance anywhere else in the CAWS, as opposed to 66.9 to 78.0% and 70.0 to 89.8% compliance at Simpson Street and Main Street, respectively, for the actual discretionary diversion taken in WYs 2001, 2003, and 2008. KEY WORDS: Water Quality Modeling, Dissolved Oxygen, Combined Sewer Overflows, Water Quality Management 1 INTRODUCTION The City of Chicago, Illinois, known for many years as America’s “Second City,” is located at the southern end of Lake Michigan, the fifth largest freshwater lake in the world (by surface area) that serves as the water supply for Chicago and surrounding communities. In the 1800s, Chicago built a network of combined sewers to drain stormwater and wastewater from the city to the Chicago River and then to Lake Michigan. During large storms the polluted combined sewer flows would extend far enough into Lake Michigan that they would enter the water supply intakes for Chicago. This contributed to very high levels of death by typhoid fever in Chicago, peaking at more than 170 per 100,000 residents in 1891 (Hill, 2000). In 1889, the Sanitary District of Chicago (now known as the Metropolitan Water Reclamation District of Greater Chicago, MWRDGC) was formed by the State of Illinois, and charged with building a canal that would carry flow from the polluted Chicago River away from Lake Michigan through the low continental divide west of Chicago to the Des Plaines River, Illinois River, and ultimately the Mississippi River (Lanyon, 2012). In 1892 construction began and in 1900 the Chicago Sanitary and Ship Canal

(CSSC) was opened to reverse the flow of the Chicago River, thus, diverting the wastewater and combined sewer overflows from Chicago away from Lake Michigan and toward the Mississippi River. Two additional channels were later opened to improve water quality in the Chicago area: (1) the North Shore Channel (NSC, completed 1910) to flush poor water quality in the North Branch Chicago River (NBCR) and (2) the Calumet-Sag Channel (completed 1922) to divert the Calumet River away from Lake Michigan. The NBCR, South Branch Chicago River, Chicago River main stem, and Little Calumet River (north) also have been widened, deepened, and straightened to efficiently carry treated wastewater away from Lake Michigan. The man-made canals and channels together with the modified natural watercourses compose a 122.8 km network known as the Chicago Area Waterways System (CAWS). The operation of the CAWS has been a great public health success for Chicago, but the CAWS still frequently experiences low dissolved oxygen (DO) concentrations that do not meet the standards for these waterways, especially in the NSC. The MWRDGC is allowed to divert an annual average flow of 7.65 m3/s from Lake Michigan to improve water quality in the CAWS. A modeling study was done to determine if the compliance with DO standards could be improved on the NSC without adversely affecting compliance with the DO standards anywhere else in the CAWS while staying within the discretionary diversion limit. This paper reports the results of the modeling study. 2 STUDY METHODS 2.1 Project Goals Evaluations of measured DO concentrations by the MWRDGC and Limnotech (Nemura, 2011) and of DO simulation results by Marquette University (Alp and Melching, 2006; Melching et al., 2010, 2013) and the MWRDGC (Zhang et al., 2007) have shown it is impractical to raise DO concentrations during storm periods above the DO standards via dilution from transferred, aerated effluent from the Terrence J. O’Brien Water Reclamation Plant [TJOWRP] (Alp and Melching, 2006; Melching et al., 2010, 2013) or increased discretionary diversion from Lake Michigan at Wilmette (Zhang et al., 2007). Low DO concentrations resulting from storms also cannot be easily remedied via supplemental aeration (Melching et al., 2010, 2013). Thus, the MWRDGC and Limnotech proposed a Wet Weather Limited Use (WWLU) DO standard to the Illinois Pollution Control Board (Nemura, 2011). Under the WWLU DO standard full compliance with the DO standards would be required during dry weather periods, whereas short variance periods during which the DO standards would not apply would be allowed following storms of sufficient size to affect DO concentrations in the CAWS. Comparison of daily rainfall data and the hourly DO concentrations measured by the MWRDGC’s continuous DO monitor (CDOM) system indicated that storms with average daily rainfall between 6.3 and 12.6 mm could substantially affect DO concentrations downstream in the CAWS for up to 2 days, storms with average daily rainfall between 12.7 and 25.3 mm could substantially affect DO concentrations for up to 4 days, and storms with average daily rainfall greater than 25.4 mm could substantially affect DO concentrations for up to 6 days (Nemura, 2011). For all of the CAWS except the CSSC the following DO concentration targets must be met or exceeded: 1) During the period of March through July, 5.0 mg/L at all times, 2) During the period of August through February: a. 4.0 mg/L as a daily minimum averaged over 7 days, and b. 3.5 mg/L at all times For the CSSC the following DO concentration targets must be met or exceeded: 1) 4.0 mg/L as a daily minimum averaged over 7 days, and 2) 3.5 mg/L at all times. For the CAWS the requirement to meet 3.5 mg/L at all times was more restrictive than the requirement of 4.0 mg/L as a daily minimum averaged over 7 days, and, thus, only the requirement to meet or exceed 3.5 mg/L was evaluated here. Given the past studies on methods to achieve compliance with the DO standards and the proposal of

2

a WWLU DO standard, the goals of increasing discretionary diversion from Lake Michigan at Wilmette are full compliance with the DO standards during dry weather and provision of some remediation of low DO concentrations during wet weather on the NSC. The CDOM monitoring locations at Simpson Street and Main Street are used as the evaluation points. A small increase in discretionary diversion can substantially shorten the time of non-compliance with the DO standards during wet weather periods, but even a large increase cannot offset the high pollutant loads resulting from storm flows and result in compliance with the proposed DO standards during the periods of highest DO demand during storm runoff. Thus, small increases in discretionary diversion during wet weather were examined, and the value beyond which only marginal increases in compliance with DO standards resulted was set as the best value of discretionary diversion during wet weather. The increase in discretionary diversion needed to achieve the foregoing goals was determined for WYs 2001, 2003, and 2008. These years were chosen as representative wet (2008), dry (2003), and medium (2001) years. The discretionary diversion at the Chicago River Controlling Works (CRCW) and O’Brien Lock and Dam (O’Brien) then was evaluated to determine how the required increase in diversion at Wilmette can be obtained without substantially affecting compliance with the DO standards downstream from these other points of diversion from Lake Michigan. The increases of discretionary diversion at Wilmette and the decreases in discretionary diversion at CRCW and O’Brien presented here yield great improvements in compliance with the DO standards on the NSC without adversely affecting compliance elsewhere in the CAWS, but are not an optimum allocation of discretionary diversion but rather a “proof of concept” that the discretionary diversion can be used in a more beneficial manner. Developing an optimal approach to the use of the available discretionary diversion is the subject of later phases of the work reported here. 2.2 Water-Quality Model In anticipation of various water-quality management needs the MWRDGC began an intensive sampling of hourly DO and temperature throughout the CAWS in 1998, and entered into an agreement with Marquette University in 2000 to develop a water-quality model for the CAWS that was suitable for simulating constituent concentrations during unsteady-flow conditions resulting from the more than 200 combined sewer overflow points discharging to the CAWS. The DUFLOW (2000) water-quality model developed in the Netherlands was selected to simulate flow and water quality in the CAWS. In particular, the DUFLOW water-quality simulation option that adds the DiToro and Fitzpatrick (1993) sediment flux model to the Water Quality Analysis Simulation Program (WASP4) (Ambrose et al., 1988) model of constituent interactions in the water column is applied. DUFLOW distinguishes among transported material that flows with water, bottom materials that are not transported with the water flow, and pore water in bottom materials that are not transported but that can be subject to similar water-quality interactions to those for the water column. Flow movement and constituent transport and transformation are simulated within DUFLOW and constituent transport is defined by advection and dispersion. The flow simulation in DUFLOW is based on the 1-D partial differential equations that describe unsteady flow in open channels (de Saint-Venant equations). These equations are the mathematical translation of the laws of conservation of mass and momentum. The DUFLOW model was calibrated, verified, and applied over entire Water Years (WYs) (i.e. the period from October 1 to September 30 designated by the year it ends). By studying rainfall data in the Chicago area from 1951 to 2010 and CSO Pumping Station volumes from 1992 to 2010, WYs 2001, 2003, and 2008 were selected as representative medium, dry, and wet years, respectively, during the period in which the MWRDGC had established the extensive hourly DO and temperature monitoring network. An extensive data set including hourly in-stream DO data at 25 locations, monthly in-stream water-quality measurements at 18 locations, daily composite treatment plant effluent measurements, daily solar radiation data, and detailed hydraulic data were used to calibrate and verify the water-quality model at a 1-hour output time step. All water quality parameters including DO were measured by the MWRDGC. There are up to 8760 measured hourly DO data at each location within each of the calibration and verification periods. Throughout the calibration process it was aimed to match hourly measured and simulated DO concentrations as much as possible. On the other hand, as Harremoës et al. (1996)

3

mentioned, it is almost impossible to match all the measured hourly data if there are a large number of data to be fitted to. It was particularly hard to match measured DO concentrations over the entire simulation period at certain locations that are dominated by CSO flows, such as the NSC. Thus, model calibration was done manually via a conservative approach, in which the goal was to better match the lower DO concentrations resulting from CSOs and produce similar probability of exceedence for different DO concentrations. Using this approach, the simulations of any management alternative that can bring DO concentrations to desired levels can also work well in the actual situation. Melching et al. (2010, 2013) provide tables that list the percentages of the simulated and measured DO concentrations higher than 3, 4, 5, and 6 mg/L target DO levels in WYs 2001 and 2003, respectively (the results for WY 2008 will be published later in 2013). Especially for the lower DO concentrations, the DUFLOW water-quality model predicted DO concentrations with relatively high accuracy. Thus, it was concluded that, in general, the DUFLOW model represents water-quality processes in the CAWS well enough to be a useful tool for determining the allocation of discretionary diversion from Lake Michigan that can meet the goals of this study. Details of the calibration procedure and results are given in Melching et al. (2010). 2.3 Procedure to Reallocate Discretionary Diversion from Lake Michigan A procedure was developed for increasing the discretionary diversion from Lake Michigan at Wilmette. In this procedure, once the DO concentration at either Simpson Street or Main Street drops within a tolerance of the applicable DO standard the increased discretionary diversion would begin and it would end at the time when the DO concentration in the original simulation again exceeds the DO standard plus the tolerance. Initially the tolerance was set as 0.5 mg/L, but this was found through several DO simulation trials to be too conservative and the tolerance was reduced to 0.3 mg/L. That is, in August-February when the DO concentration dropped below 3.8 mg/L or in March-July when the DO concentration dropped below 5.3 mg/L the increased discretionary diversion would begin and it would end when the DO concentrations exceeded these values in the original simulation. This approach is conservative in terms of the amount of increased discretionary diversion because in actual operations once the increased diversion begins, this increase in flow will result in the DO concentration rising above the DO target concentration sooner than in the original simulation. However, varying the start and end times of increased diversion with each increase in diversion would require a prohibitive number of simulations to define a precise operation plan for this “proof of concept” level study. After identifying periods in the original simulation that had DO concentrations below the target DO concentrations of 3.8 mg/L for August-February or 5.3 mg/L for March-July, constant increased discretionary diversion values of 0.283, 0.424, and 0.566 m3/s were applied to each of these periods, and the value of 0.424 m3/s was found to be most effective in improving the percentage compliance with the DO standards. That is, 0.566 m3/s yielded only marginal improvements relative to 0.424 m3/s, and, thus, the simulation results from an increased discretionary diversion of 0.424 m3/s became the base for focusing additional discretionary diversion at new sub-periods with low DO concentrations. After increasing the discretionary diversion at Wilmette by 0.424 m3/s for the periods identified previously, a second group of periods were identified in which the DO concentration was less than the DO standard plus 0.2 mg/L (3.7 mg/L for August-February, and 5.2 mg/L for March-July) and which either were dry weather periods or wet weather periods with DO concentrations only slightly below the DO standard. Constant additional increased discretionary diversion values of 0.283 and 0.424 m3/s were applied to each of these periods (i.e. total increased diversion for these periods of 0.708 and 0.850 m3/s, respectively), and the value of 0.424 m3/s was found to be most effective in improving the percentage compliance with the DO concentrations. After increasing the additional discretionary diversion at Wilmette by 0.424 m3/s for the periods identified in the second level of discretionary diversion augmentation, periods were identified in which the DO concentration was less than the DO standard plus 0.1 mg/L (3.6 mg/L for August-February, and 5.1 mg/L for March-July). A constant additional increased diversion of whatever amount is needed to meet the DO standard then was determined by trial and error for the remaining dry weather periods that do not meet the DO standard. Because of the travel time from Wilmette to Simpson Street or Main Street, in some cases (less than 15% of the periods), it was necessary to begin the increases in

4

discretionary diversion before the DO concentrations in the original simulation were below the target concentrations. Once the discretionary diversion was increased at Wilmette such that full compliance with the DO standards was obtained during dry weather, the total amount of additional discretionary diversion at Wilmette was computed as an average over the year. Traditionally, the MWRDGC begins taking discretionary diversion in May or June, takes it most intensely in July and August, but still retains a substantial amount in case of extreme conditions in September and October. Thus, typically a large amount of discretionary diversion has been taken in October that may not really be needed because of the cooler temperatures and the reduced DO standard in October. Thus, it was felt that October would be a good time to reduce the discretionary diversion at CRCW and/or O’Brien without affecting the compliance with the DO standards downstream from these points. This procedure was successfully applied to each of the demonstration years. 3 REALLOCATION RESULTS In this section detailed results are presented for WY 2001, but only summary results are presented for WYs 2003 and 2008 due to space limitations. 3.1 Reallocation for Water Year 2001 Following the procedure described in Section 2.3, 43 periods were found in the original simulation for WY 2001 that had DO concentrations below the target DO concentrations of 3.8 mg/L for August-February or 5.3 mg/L for March-July. In the course of testing the three constant increased discretionary diversion values of 0.283, 0.424, and 0.566 m3/s it became clear that there were 11 wet periods totaling 360 hours at Simpson Street and 681 hours at Main Street that could not be brought into compliance with the DO standards through increased discretionary diversion at Wilmette. Initially, 3 of the periods were thought to be close enough to the DO standard that increased discretionary diversion might help them achieve full compliance with the DO standard, but subsequent increases in discretionary diversion for these periods yielded no improvement in compliance. After increasing the discretionary diversion at Wilmette by 0.424 m3/s for the 43 periods described previously, 25 periods were identified in which the DO concentration was less than the DO standard plus 0.2 mg/L (3.7 mg/L for August-February, and 5.2 mg/L for March-July) and which either were dry weather periods or wet weather periods with DO concentrations only slightly below the DO standards. These periods ranged in duration from as short as 4 hours to as long as 9 days and 8 hours. After increasing the additional discretionary diversion at Wilmette to 0.850 m3/s for the 25 periods described previously, 9 periods were identified in which the DO concentration was less than the DO standard plus 0.1 mg/L (3.6 mg/L for August-February, and 5.1 mg/L for March-July). Constant additional increased discretionary diversion values of 0.283, 0.566, and 0.850 m3/s were applied to each of these periods (i.e. total increased diversion for these periods of 1.133, 1.416, and 1.699 m3/s, respectively). One period was brought into full compliance with the DO standards with a total increased discretionary diversion of 1.133 m3/s. Three periods were brought into full compliance with the DO standards with a total increased discretionary diversion of 1.416 m3/s. Two periods were brought into full compliance with the DO standards with a total increased discretionary diversion of 1.699 m3/s. Obtaining full compliance with the DO standards for the final three periods was a little more complex. For the period of 0:00 on August 8th to 9:00 on August 9th a total increased diversion of 1.699 m3/s can bring this period into full compliance with the DO standards, but this increased diversion needs to start 3 hours before the DO concentration drops below 3.6 mg/L. For the period of 8:00 on April 7th to 13:00 on April 8th a total increased diversion of 1.699 m3/s can bring this period into full compliance with the DO standards, but this increased diversion needs to start 6 hours before the DO concentration drops below 3.6 mg/L. Thus, these periods do not fall into a systematic approach to increasing discretionary diversion as found for all the other periods. For the dry weather period of 0:00 on September 25th to 4:00 on September 27th the low DO concentrations are a product of the September 17-24 storm period. Thus, in order to meet the DO standard throughout the dry period (as per the goals of this study), the further increase in discretionary 5

diversion from the initial 0.424 m3/s needs to begin 24 hours earlier (i.e. 0:00 on September 24th). To achieve full compliance with the DO standard, a total increase in discretionary diversion at Wilmette of 2.124 m3/s must be applied between 0:00 on September 24th and 14:00 on September 25th, and a total increase in discretionary diversion at Wilmette of 1.416 m3/s must be applied between 14:00 on September 25th and 4:00 on September 27th. Figure 1 shows the simulated DO concentrations at Simpson Street and Main Street for the original simulation and the simulation for the final allocation of increased discretionary diversion at Wilmette. In total, the increases in discretionary diversion result in 95.8% and 92.0% compliance with the DO standards at Simpson Street and Main Street, respectively, as opposed to 71.2% and 70.0% compliance, respectively, in the original simulations.

Figure 1 Simulated dissolved oxygen (DO) concentrations compared with the DO standards at Simpson Street (top) and Main Street (bottom) on the North Shore Channel for the original (calibrated) simulation and the simulation for the final allocation of increased discretionary diversion from Lake Michigan at Wilmette. Figure 2 shows the actual (original) discretionary diversion and total increased discretionary diversion from Lake Michigan at Wilmette needed to achieve the results shown in Figure 1 for WY 2001. In total over the year the increase in the discretionary diversion from Lake Michigan at Wilmette amounts

6

to 0.258 m3/s or 3.38% of the total annual allowed discretionary diversion of 7.65 m3/s. Applied over a 31 day period this amounts to 3.04 m3/s. The average daily discretionary diversion for October 2000 at CRCW and O’Brien was 2.58 and 2.82 m3/s, respectively. Thus, partially reducing the October diversion at these locations could easily provide 3.04 m3/s of increased discretionary diversion at Wilmette. Initially, the discretionary diversion at CRCW was set to zero for the entire month of October. However, it was found that the period of noncompliance during wet weather between October 4 and 7, 2000 increased from 41 to 66 hours at Loomis Street and from 32 to 38 hours at Cicero Avenue relative to the results for the actual discretionary diversion at CRCW for October 2000. Thus, it was necessary to restore the actual discretionary diversion on October 3-7, 2000, to achieve the same compliance at Loomis Street and slightly improved compliance (28 vs. 32 hours of noncompliance) at Cicero Avenue. Figure 3 shows the change in simulated DO concentrations at Clark Street of the Chicago River main stem and Loomis Street on the South Branch Chicago River. Discretionary diversion at CRCW stopped on October 19th, thus, the simulated concentrations converge after the 19th. The shift in diversion from CRCW to Wilmette has lowered DO concentrations at these two locations, but the simulated DO concentrations still are far above the DO standard of 3.5 mg/L for the month of October except for the aforementioned storm period.

Figure 2 Original discretionary diversion from Lake Michigan at Wilmette and the total increased discretionary diversion from Lake Michigan at Wilmette needed to achieve the compliance with the dissolved oxygen standards indicated in Figure 1.

Figure 3. Change in dissolved oxygen concentrations in October 2000 resulting from the reduction in discretionary diversion at the Chicago River Controlling Works The discretionary diversion at CRCW for the month of October 2000, thus, was reduced by 2.03 7

m3/s (from 2.58 to 0.55 m3/s). Therefore, another 1.01 m3/s had to be taken from the discretionary diversion at O’Brien to be withdrawn at Wilmette. For the month of October 2000, full compliance with the DO standards was achieved at all monitoring locations along the Little Calumet River (north), Calumet-Sag Channel (Cal-Sag), and the CSSC downstream from the junction with the Cal-Sag. To try to maintain this full compliance, the discretionary diversion at O’Brien was reduced 36.1% on all days in October to provide the needed flow increase at Wilmette. The results of this simulation indicated that full compliance with the DO standards was maintained at all monitoring locations downstream from O’Brien. Figure 4 shows the change in simulated DO concentrations at the Central and Wisconsin Railroad on the Little Calumet River (north) and Southwest Highway on the Calumet-Sag Channel. The shift in diversion from O’Brien to Wilmette has had very little effect on simulated DO concentrations at the Central and Wisconsin Railroad and a small effect at Southwest Highway (and other locations on the Calumet waterway). The reason for this is that on many days in October 2000 the 36.06% reduction in discretionary diversion is a fairly small amount of the total flow from Lake Michigan on that day.

Figure 4 Change in dissolved oxygen concentrations in October and November 2000 resulting from the reduction in discretionary diversion at the O’Brien Lock and Dam for October 2000 3.2 Reallocation for WYs 2003 and 2008 The same procedures were applied to reallocate the discretionary diversion to Wilmette from CRCW and O’Brien for WYs 2003 and 2008 as were used for WY 2001. For WY 2003, the proposed increases in discretionary diversion result in 97.0% and 95.5% compliance with the DO standards at Simpson Street and Main Street, respectively, as opposed to 78.0% and 81.0% compliance, respectively, in the original simulations. In total over WY 2003 the increase in the discretionary diversion from Lake Michigan at Wilmette amounts to 0.199 m3/s or 2.61% of the total annual allowed discretionary diversion of 7.65 m3/s. Applied over a 31 day period this amounts to 2.35 m3/s. The average daily discretionary diversion for October 2002 at CRCW and O’Brien was 3.66 and 2.24 m3/s, respectively. Thus, partially reducing the October diversion at these locations could easily provide 2.35 m3/s of increased discretionary diversion at Wilmette. Initially, it was tried to apply the reduction uniformly (64.1% reduction) across all days with discretionary diversion at CRCW in October 2002. However, this resulted in increases in non-compliance with the DO standards during storm periods in October 2002 of 17, 4, and 5 hours at Loomis Street, Cicero Avenue, and Lockport, respectively. These non-complying storm periods stretched from 0:00 on October 3rd to 11:00 on October 7th. Thus, the discretionary diversion for 0:00 on October 1st to 12:00 on October 7th was restored, and a 77.19% reduction in discretionary diversion was applied to the remainder of October. For WY 2008, the proposed increases in discretionary diversion result in 97.0% and 96.7% compliance with the DO standards at Simpson Street and Main Street, respectively, as opposed to 66.9% and 89.8% compliance, respectively, in the original simulations. In total over WY 2008 the increase in the discretionary diversion from Lake Michigan at Wilmette amounts to 0.174 m3/s or 2.27% of the total annual allowed discretionary diversion of 7.65 m3/s. Applied over a 31 day period this amounts to 2.05 m3/s.

8

The average daily discretionary diversion for October 2007 at CRCW and O’Brien was 4.04 and 2.54 m3/s, respectively. Thus, partially reducing the October diversion at these locations could easily provide 2.05 m3/s of increased discretionary diversion at Wilmette. Initially, it was tried to apply the reduction uniformly (50.7% reduction) across all days with discretionary diversion at CRCW in October 2007. This reduction resulted in no decrease in compliance with the DO standards at any location on the South Branch Chicago River or CSSC in October 2007. As an alternative reallocation of discretionary diversion, the discretionary diversion at O’Brien was reduced by 2.05 m3/s in October 2007 by uniformly reducing the discretionary diversion by 80.4%. This reallocation did not decrease the compliance with the DO standards at any point downstream of O’Brien on the Little Calumet River (north) or Calumet-Sag Channel in October 2007. 4 CHALLENGES TO THE DEVELOPMENT OF OPERATIONAL PROCEDURES TO OPTIMALLY USE THE DISCRETIONARY DIVERSION The simulations done in this study have revealed some challenges and information that will affect the development and usefulness of an approach for more optimal use of the available discretionary diversion to achieve high levels of compliance with the DO standards. It will be difficult to develop a procedure that will result in full compliance with the DO standards at all times during dry weather because of the following challenges. 1) The periods during which DO concentrations do not meet the DO standards can be substantially different at Simpson Street and Main Street. Simpson Street is more prone to low DO concentrations during periods when no discretionary diversion is taken at Wilmette (i.e. November through mid-April). During these periods the primary source of additional oxygen in the upper NSC is the back up of TJOWRP effluent into the upper NSC. Thus, the stagnant waters at Simpson Street and points upstream get little added oxygen from late fall to early spring and are prone to extended periods of DO concentrations below the DO standards. Whereas Main Street tends to experience longer periods of low DO concentrations than Simpson Street in the late spring through early fall when discretionary diversion is taken at Wilmette. This may be the result of the lingering effects of combined sewer overflows and the hydraulic dam resulting from the TJOWRP outfall. Thus, it may be difficult to find a single monitoring point on the NSC that can adequately identify periods of low DO concentrations at Simpson Street and Main Street so that the discretionary diversion at Wilmette can be triggered to counteract all periods of low DO concentrations throughout the NSC. 2) The study initially identified 93 periods among all three water years in which the DO standards were not met. Within these were 27 wet weather sub-periods that could not be brought into compliance with the DO standards by increasing the discretionary diversion from Lake Michigan at Wilmette. The majority of the dry weather periods could be brought into compliance with the DO standards by increasing the discretionary diversion once the DO concentrations got within 0.3 mg/L of the DO standards (i.e. less than 3.8 mg/L in August-February, and less than 5.3 mg/L in March-July). However, there were 12 periods for which the increase in discretionary diversion needed to start 3-6 hours earlier than the onset of low DO concentrations at the monitoring points. In these cases, the travel-time from Wilmette to Simpson Street and/or Main Street required the high DO Lake Michigan water to already be on the way to head off periods of low DO. This operation can be done in hind sight for a modeling study, but in practical operations 10-20% of the dry weather events might experience short periods of non-compliance with the DO standards until the high DO Lake Michigan water can spread through the entire NSC. 3) Even though the majority of periods could be brought into compliance by initiating the increased discretionary diversion when the DO concentration went below the “threshold” (3.8 mg/L in August-February, and 5.3 mg/L in March-July), the rate at which the discretionary diversion increases still needs some consideration. That is, some periods were brought into compliance with an increase of 0.424 m3/s, others required an additional 0.283 or 0.424 m3/s if DO concentrations continued to drop after the first increase, however, others required another 0.283 to 2.83 m3/s if DO concentrations continued to drop after the first and second increases.

9

Development of an effective and implementable system of increases requires further study. Findings of this study that inform the development of a practical operation procedure to more optimally take discretionary diversion at Wilmette include the following. 1) A small discretionary diversion of 0.283 m3/s should be taken even during low DO periods resulting from wet weather in order to shorten the period of non-compliance with the DO standards during wet weather periods. 2) If the DO concentration during the last day of a wet weather period is greatly below the DO standard and no rain is forecast so that tomorrow is likely to be a dry period, the discretionary diversion should be greatly increased early on the last day of the wet weather period (ideally at 0:00 on that day), so that compliance with the DO standard can be achieved during the subsequent dry weather period. 3) A possible operation rule for taking discretionary diversion at Wilmette might be that if DO concentrations on the NSC are close to the March DO standard on the last day of February, increased discretionary diversion should start at noon (12:00) on that day to avoid DO standard violations in early March. 4) The results of this study have shown that discretionary diversion can be reduced in October at either CRCW or O’Brien or both without adversely affecting compliance with the DO standards at downstream locations. A careful re-evaluation of when discretionary diversion is withdrawn at these locations may identify other times when DO concentrations at other locations throughout the CAWS can be brought into compliance with the DO standards without adversely affecting compliance at any other location in the CAWS. 4 CONCLUSIONS The simulations done in this study indicate discretionary diversion could be shifted from CRCW and O’Brien in October 2000 to Wilmette at various times throughout WY 2001 so that full compliance with the DO standards during dry weather and greatly increased compliance during wet weather could be achieved on the NSC in WY 2001 without decreasing compliance anywhere else in the CAWS. For WY 2003, discretionary diversion could be shifted from CRCW in October 2002 to Wilmette at various times throughout WY 2003 so that full compliance with the DO standards during dry weather and greatly increased compliance during wet weather could be achieved on the NSC in WY 2003 without decreasing compliance anywhere else in the CAWS. For WY 2008, discretionary diversion could be shifted from CRCW or O’Brien in October 2007 to Wilmette at various times throughout WY 2008 so that full compliance with the DO standards during dry weather and greatly increased compliance during wet weather could be achieved on the NSC in WY 2008 without decreasing compliance anywhere else in the CAWS. Overall, the compliance over the entire year with the DO standards ranged from 95.8 to 97.0% at Simpson Street and from 92.0 to 96.5% at Main Street for the reallocations determined in this study. These reallocations are not the optimal utilization of discretionary diversion throughout the CAWS just major improvements in compliance along the NSC, and an improvement in compliance at many other points downstream of the NSC. These cases illustrate that the MWRDGC can substantially improve compliance with the DO standards just by making better use of the allowable discretionary diversion from Lake Michigan. Later phases of this study will develop a more optimal approach to the use of the available discretionary diversion and aeration station operations to achieve full compliance with the DO standards at all times during dry weather and improved compliance during wet weather. ACKNOWLEDGEMENT The study is supported by the Metropolitan Water Reclamation District of Greater Chicago through Purchase Order No. 3073684. References Alp, E. and Melching, C.S., 2006, Calibration of a Model for Simulation of Water Quality During Unsteady Flow in the Chicago Waterway System and Application to Evaluate Use Attainability Analysis Remedial Actions, Institute for Urban Environmental Risk Management Technical Report No. 18, Marquette University, Milwaukee, Wis. and

10

Research and Development Department Report No. 2006-84, Metropolitan Water Reclamation District of Greater Chicago, Chicago, Ill. Ambrose, R.B., Wool, T.A., Connolly, J.P., Schanz, R.W., 1988. WASP4, A Hydrodynamic and Water Quality Model—Model Theory, User’s Manual, and Programmer’s Guide, U.S. Environmental Protection Agency, EPA/600/3-87-039, Athens, Ga. Di Toro, D.M. and Fitzpatrick, J., 1993. Chesapeake Bay Sediment Flux Model. Contract Rep. No. EL-93-2, Prepared for U.S. Army Engineer Waterway Experiment Station. Vicksburg, Miss., Hydro Qual Inc., Mahwah, N.J. DUFLOW, 2000. DUFLOW for Windows V3.3: DUFLOW Modelling Studio: User’s Guide, Reference Guide DUFLOW, and Reference Guide RAM, EDS/STOWA, Utrecht, The Netherlands. Harremoёs, P., Napstjert, L., Rye, C., Larsen, H.O., 1996. Impact of rain runoff on oxygen in an urban river. Water Science and Technology, 34(12), 41-48. Hill, L., 2000. The Chicago River – A Natural and Unnatural History, Lake Claremont Press, Chicago, Ill. Lanyon, R., 2012. Building the Canal to Save Chicago, Xlibris Corporation, Chicago, Ill. Melching, C.S., Alp, E., and Ao, Y., 2010. Development of Integrated Strategies to Meet Proposed Dissolved Oxygen Standards for the Chicago Waterway System, Institute for Urban Environmental Risk Management Technical Report No. 20, Marquette University, Milwaukee, Wis. Melching, C.S., Ao, Y., and Alp, E., 2013. Modeling evaluation of integrated strategies to meet proposed dissolved oxygen standards for the Chicago Waterway System, Journal of Environmental Management, 116(2013), 145-155. Nemura, A. 2011. Pre-Filed Testimony of Adrienne D. Nemura—Wet Weather Limited Use for Aquatic Life in the Chicago Area Waterway System, Limnotech, Ann Arbor, Mich. Zhang, H., Bernstein, D., Kozak, J., and Jain, J. 2007. Model Simulations for Evaluating the Impact of Various Options of Discretionary Lake Diversion to the Chicago Waterway System, Monitoring and Research Department Unpublished Internal Report, Metropolitan Water Reclamation District of Greater Chicago, Chicago, Ill.

11