ESBD_482241_1772049596706_RFQ 2026-001 Addendum No. 1.pdf

Purchasing Services Post Office Box 2570

Waco, Texas 76702-2570 Phone: 254-750-8098

Fax: 254-750-8063 melaniew@wacotx.gov

CITY OF WACO www.waco-texas.com

Date: 02/25/2026

RFQ No: 2026-001 Commodity: Waco Lake Study Purchasing Agent: Melanie Wade

Closing Time: 2:00 P.M. CST, Wednesday, March 11, 2026 Opening Time: 2:01 P.M. CST, Wednesday, March 11, 2026

RFQ Opening Location: Via Zoom Video and Dial In

Addendum No: 1

1. Questions and answers

2. SWAT modeling documentation

3. Water Quality Modeling assessment

The above-mentioned RFQ has been changed. Sign and return addendum to the Purchasing Office by the closing time and date with your RFQ response. Returning this page signed by your authorized agent will serve to acknowledge this change. All other requirements of the invitation remain unchanged. If you have any questions, please call or stop by the Purchasing Office at the above address.

Firm:____________________________________________________________________________________

Address___________________________________________________________________________________

Signature of Person Authorized to Sign Bid:______________________________________________________________________

Signor's Name and Title (print or type):_____________________________________________________________________________

E-mail Address:____________________________________________________________________________

Date:____________________Telephone:___________________________Fax:_________________________ mailto:melaniew@wacotx.gov file://NETAPP/HOMES/JodyC/0%20Purchasing/Utilities/2014/RFB%202014-012%20Generator%20Rebuild/www.waco-texas.com

Questions RFQ 2026-001 Waco Lake Study

1. Are 11x17 pages permitted, specifically for the project schedule? Yes

2. To confirm our understanding of the RFQ instructions for item (5), “Additional background information on no more than 15 pages,” could you please verify that the 15-page limit applies only to this subsection and does not apply to any other sections of the Statement of Qualifications? Yes, the 15-page limit applies to that subsection only and excludes the title page, table of contents, transmittal letter and required attachments such as the organizational chart.

3. Can you please provide us with copies of the CE-QUAL-W2 and 2001 SWAT models?

The attached documentation is for reference only and was obtained from an open-source website. The City of Waco does not have the original models.

4. Are we required to include both an org chart for our company as a whole and an org chart for the personnel who will be working on the Waco Lake Study? Or are we only required to include the Lake Study team org chart? The organizational chart will be for the portion of your business that will be involved with this project, complete with names and titles, also mark the individuals who will be involved in this project.

5. Does the City of Waco have an analytical lab, or should we include one in the qualification package? The City of Waco laboratory will not be available for this project;

submitters will need to provide an alternate laboratory for use during this project.

6. Does task 7 get included in the 18-month scope? Task 7 is not part of the initial 18-month timeline. If tasked an appropriate amount of time will be given to the selected firm to complete task 7.

7. Is there information available prior to 2000 for consideration in Task 1? The selected firm will be given access to the available historical records prior to developing a formal contract or proposal.

USDA: Lake Waco-Bosque River Initiative

Water Quality Modeling of Lake Waco Using CE-QUAL-W2 for Assessment of Phosphorus Control Strategies

Joan D. Flowers, Larry M. Hauck, and Richard L. Kiesling

TR0114

December 2001

Texas Institute for Applied Environmental Research Tarleton State University •Box T0410 •Tarleton Station •Stephenville, Texas 76402 V254.968.9567 • F254.968.9568

Water Quality Modeling of Lake Waco Using CE-QUAL-W2 Acknowledgments

The Texas Institute for Applied Environmental Research (TIAER) acknowledges the United States Department of Agriculture-Natural Resources Conservation Service (USDA-NRCS) for providing the funding for this research. Water quality modeling under this project was a collaborative effort between research scientists at TIAER and Texas A&M Blackland Research and Extension Center (BREC). Modeling of the Lake Waco-Bosque River watershed was performed by Chinnisamy Santhi under the direction of Jimmy Williams and William Dugas of BREC in Temple, Texas.

The authors would like to acknowledge the assistance of the developers of the CE-QUAL-W2 model, Thomas Cole, of the U.S. Army Corp of Engineers, Waterways Experiment Station and Edward Buchak, of J.E. Edinger and Associates, Inc. for their assistance and guidance in applying the model to Lake Waco.

TIAER would also like to recognize the input and assistance of numerous scientists and technical experts including Rajeev Jain of Edinger and Associates and members of the Technical Work Group (TWG) assembled for this project. The TWG, whose purpose was to provide technical oversight of the project, consisted of 19 members including: Clyde Bohmfalk (Chairman), Texas Natural Resource Conservation Commission (TNRCC); Jim Davenport (TNRCC); Larry Koenig (TNRCC); Allison Woodall (TNRCC); Mel Vargas (formerly of TNRCC); Larry Mitchell (TNRCC); Marie Knipfer (TNRCC); Hari Krishna (TNRCC); Charles Maddox (TNRCC); James Moore, Texas State Soil and Water Conservation Board (TSSWCB);

Beade Northcut (TSSWCB); Richard Garrett, City of Waco; Mike Jones, City of Waco; Allan Colwick, Natural Resources Conservation Service (NRCS); George Ward, University of Texas Center for Research in Water Resources (UT-CRWR); Mike Meadows, Brazos River Authority (BRA); Tom Conry (BRA); Joan Glass, Texas Parks and Wildlife Department (TPWD); and Shawneille Campbell, Environmental Protection Agency, Region VI (EPA).

The authors would also like to recognize the assistance to Jim Rogers, computer programmer-analyst with TIAER for programming support and development of preprocessing, postprocessing and interface programs for the modeling effort. Although too numerous to mention by name, we also acknowledge the TIAER field crew, laboratory personnel, data analysts, and GIS specialists for their role in supplying the data necessary for the application and calibration of the lake model.

USDA2.2

Abstract

The Texas Institute for Applied Environmental Research (TIAER) conducted modeling assessments of alternative phosphorus (P) control strategies for the USDA Lake Waco-Bosque River Initiative. This report documents the results of CE-QUAL-W2 model calibration for Lake Waco and long-term (32-year) simulations assessing the impact of different scenarios on lake water quality.

A water quality target of 10 ppb PO4-P was recommended and adopted by the stakeholder committee for Lake Waco. The 10 ppb PO4-P target was established in terms of a mean concentration over the algal “growing” season defined as April through September. The lake target was defined for summer mean PO4-P concentrations within the surface water layer near the drinking water intake for the City of Waco. Alternative P control strategies were assessed based on long-term simulations of lake water quality relative to target attainment.

A two-phased approach was utilized for calibration of CE-QUAL-W2 for Lake Waco. CE- QUAL-W2 and SWAT were first calibrated independent of one another (i.e., the independent calibration phase). SWAT and CE-QUAL-W2 models were then integrated through an intermediate interface program, which translated stream flow and watershed loads predicted by SWAT into inflow boundary conditions required by CE-QUAL-W2. The second phase of calibration entailed fine-tuning of the integrated SWAT/CE-QUAL-W2 models to conditions observed in Lake Waco. Observational data collected over a two-year period from July 11, 1996 through July 29, 1998 were used in both the independent and integrated calibrations.

CE-QUAL-W2 calibration focused on the prediction of summer mean PO4-P concentrations in the surface layer of model segment 8, which corresponded to the location where the water quality target was defined. Inspection of daily time series output and summer mean concentrations predicted by CE-QUAL-W2 indicated acceptable agreement between observed and predicted data. The magnitude of temporal variations in response to seasonal changes and storm pulses were successfully reproduced by CE-QUAL-W2. Summer mean surface PO4-P concentrations predicted by CE-QUAL-W2 were within ± 21 of observed data.

Calibration statistics computed for surface and water column average concentrations over the entire calibration period were within acceptable tolerance ranges and depending on the constituent, indicated the quality of the calibration ranged from fair to very good. Results indicated that the calibration quality with respect to surface PO4-P concentrations in segment 8 was very good for both phases of calibration.

Alternative P controls were applied for the two major sources of P in the watershed— municipal wastewater treatment plants (WWTPs) and dairy manure application fields. P controls were assessed individually and as combination scenarios relative to target attainment. P controls which targeted nonpoint source loads from dairy manure application fields (haul off, P-rate, and P-diet scenarios) had the greatest impact on summer mean PO4-P concentrations in Lake Waco. Individual dairy BMPs were predicted to reduce summer mean PO4-P concentrations by 15 to 39 percent from future conditions depending upon BMP evaluated. Effluent limits imposed on WWTPs were predicted to reduce summer mean PO4-P in Lake Waco by less than 10 percent even under the strictest limit of 0.5 mg/L total phosphorus. Three combination scenarios were assessed relative to the lake target of 10 ppb.

Exceedence probability plots of combination scenarios indicated that even with the most aggressive implementation of P control measures in the watershed, the lake target would be exceeded roughly 35 to 40 percent of the years.

Contents

Chapter 1 Introduction

Chapter 2 Description of Study Area and Lake Waco Lake Waco-Bosque River Watershed Lake Waco Literature Review Target Development Identification of Phosphorus Sources in the Watershed

Chapter 3 Description of Lake Model Model Selection Modeling Approach Model Input Data

Physical Representation and Segmentation Scheme Model State Variables Collection of In-Lake Calibration Data Initial Conditions Boundary Conditions Hydraulic and Kinetic Parameters

Chapter 4 Model Calibration Water Balance Calibration

Independent Calibration Integrated Calibration

Hydrodynamic/DO/Temperature Calibration Water Quality Calibration Post-Audit of Effects of Conservation Pool Increase

Chapter 5 Evaluation of Phosphorus Control Strategies Description of Phosphorus Control Scenarios

Dairy Best Management Practices (BMPs) WWTP Phosphorus Control Strategies Combination Scenarios Effect of Individual Control Measures on Lake Water Quality

Effect of Combined Control Measures on Lake Water Quality & Target Attainment

Chapter 6 Summary and Conclusions Major Accomplishments of the Modeling Effort

References

Appendix A Model Calibration Results

Appendix B CE-QUAL-W2 Calibration Plots

Appendix C Procedure for Generating Exceedence Probability Charts

Tables

Table 1 Land uses within the Lake Waco-Bosque River watershed Table 2 Morphometric characteristics of Lake Waco at conservation pool level Table 3 Estimated percent contribution of phosphorus by source to the Bosque watershed Table 4 Water quality parameters Table 5 Final calibration values for temperature input parameters Table 6 Final calibration values for phytoplankton input parameters Table 7 Final calibration values for water quality input parameters

Table 8 Calibration results for segment 8 (July 11, 1996 - July 29, 1998) Table 9 Phosphorus reduction scenarios Table 10 Percent change in predicted loading to Lake Waco for 1966 through 1997

Figures

Figure 1 Lake Waco-Bosque River watershed Figure 2 Lake Waco bathymetry Figure 3 Diagram of calibration process Figure 4 Lake Waco segmentation Figure 5 Lake Waco computational grid Figure 6 Location of Lake Waco sampling sites used in model calibration Figure 7 Main tributary gauging stations

Figure 8 Watershed delineations for Lake Waco model inputs Figure 9 Time-series plots of water balance calibration Figure 10 Time-series plots of predicted to observed water temperatures Figure 11 Time-series plots of predicted to observed dissolved oxygen concentrations Figure 12 Time-series plot of predicted and observed surface PO4-P data for segment 8 Figure 13 Time-series plot of predicted and observed surface chlorophyll-α data Figure 14 Calibration results for summer mean chl-α concentration for surface layer Figure 15 Calibration results for summer mean PO4-P concentration for surface layer Figure 16 Percent reductions in predicted mean summer PO4-P concentrations Figure 17 Exceedence probability chart for summer mean chlorophyll-α concentration Figure 18 Exceedence probability chart for summer mean PO4-P concentration

CHAPTER 1

Introduction

The Lake Waco-Bosque River Watershed Initiative (Initiative) was established in 1996 through a cooperative agreement between the Texas Institute for Applies Environmental Research (TIAER) and United States Department of Agriculture, Natural Resources Conservation Service (USDA-NRCS). The purpose of the Initiative was to develop a community based “bottom-up” approach for dealing with water quality problems associated with livestock which satisfies broad water quality objectives, on both a state and national level. Several components of the Initiative are pertinent to information contained in this report including the establishment of a water quality monitoring program in the basin for the assessment of current conditions, identification of pollutant sources, determination of suitable water quality targets, and finally the assessment of different management or control strategies in attaining desired water quality targets.

The Initiative employed a holistic, watershed approach to the entire Lake Waco drainage area and a watershed stakeholder process through the Bosque River Advisory Committee (BRAC) to address the specific issue of nutrient enrichment. The North Bosque River, which drains approximately three-quarters of the Lake Waco watershed, is currently listed on the State’s Section 303(d) list for nutrient enrichment (TNRCC, 2000). Although the lake itself is not listed by the State for impaired water quality, the general perception of local residents is that excessive nutrients are responsible for the unpleasant taste of their tap water and increased treatment costs to remedy the taste and odor problems.

This report documents the water quality modeling of Lake Waco that was performed for the Initiative. Although the bulk of this report documents the model calibration process and underlying assumptions, the more meaningful outcome of the modeling exercise is the assessment of various control strategies with respect to target attainment. The assessment results represented the culmination of the modeling effort and when condensed into exceedence probability plots were effective tools in conveying the “bottom-line” to members of the BRAC.

Water Quality Monitoring of Lake Waco Using CE-QUAL-W2

CHAPTER 2

Description of Study Area and Lake Waco

Lake Waco-Bosque River Watershed The Lake Waco-Bosque River watershed (Figure 1) encompasses 1,660 square miles, spanning portions of 6 counties in the Brazos River Basin. Four major tributaries are located in the watershed including the North, Middle and South Bosque Rivers and Hog Creek. Numerous water quality problems have been cited on the North Bosque including elevated nutrients and fecal coliform levels. The region’s most conspicuous nonpoint source contributor, the dairy industry, and municipal wastewater treatment plants have been targeted as the probable causes. The Middle and South Bosque Rivers, while receiving less attention, are not without problems. Elevated nitrogen levels associated with nonpoint source pollution from cropland areas threaten water quality in these watersheds (TNRCC, 1996).

Land use in the watershed (Table 1) is classified primarily as woodland or rangeland, comprising roughly 63 percent of the watershed area (McFarland and Hauck, 1999). Pasture and cropland areas comprise the next largest land use categories, covering 14 and 17 percent of the watershed, respectively. Pasture and cropland acreage receiving dairy waste were classified as a separate land use category. Dairy waste application fields and urban areas each comprised about two percent of the watershed.

Table 1 Land uses within the Lake Waco-Bosque River watershed.

Source: McFarland and Hauck (1999)

Land Use Category Percent of the Watershed

Wood/range 63%

Pasture 14%

Row crop 15%

Non-row crop 2%

Dairy waste application 2%

Urban 2%

Water < 1%

Other < 1%

Figure 1 Lake Waco-Bosque River watershed.

Lake Waco Lake Waco1 is a highly productive, polymictic reservoir (Kimmel, 1969) located in McLennan County. The reservoir which was impounded in 1965, is relatively young, but has experienced accelerated eutrophication, responsible for the shift in it’s trophic status from mesotrophic (Lind, 1979 & 1986) to eutrophic (Roark & Lind, 1988). Designated uses for the reservoir, classified as Segment 1225, include contact recreation, high aquatic life, and public water supply, serving an estimated 140,000 residents of Waco and nearby communities. The reservoir is also operated as a flood control reservoir by the United States Army Corp of Engineers (USACE), which shares water rights with the Brazos River Authority (BRA) for the conservation storage.

1 Although technically a misnomer, this water body is commonly referred to as “Lake Waco” (a.k.a.

Waco Reservoir) in much of the published literature and maps of the area (USGS, 1975) and its common usage is continued throughout this report.

Waco

0 32

Kilometers

Hamilton Co.

Comanche Co.

Bosque Co.

Coryell Co.

McLennan Co.

Somervell Co.

Erath Co.

Hico

Lake Waco

Stephenville

Dublin

Iredell

Meridian

Clifton

Cranfils Gap

Valley Mills

Crawford

McGregor

Hog Creek

North Bosque

D uffau C reek

Green Creek

N orth Bosque

N orth Bosque

So ut h Bo sq ue

Middle Bosque

Neils Creek

Meridian Creek

Chapter 2 Description of Study Area and Lake Waco Lake Waco is relatively shallow, ranging in depth from approximately 3 meters in the riverine portions of the arms to 23 meters near the dam (Figure 2). A recent volumetric survey by Texas Water Development Board (Sullivan et al., 1995) determined that the volume had declined from its original capacity of 152,500 acre-feet in 1965 to 144,830 acre-feet in 1995. The TWDB survey reported a surface area of 7,194 acres at conservation pool elevation and a maximum depth of 76 feet near the dam. A sediment accumulation rate of 164.8 acre-feet per year was estimated for the 25 year period between 1970 to 1995. Table 2 lists some of the morphometric characteristics of Lake Waco.

Figure 2 Lake Waco bathymetry.

Table 2 Morphometric characteristics of Lake Waco at conservation pool level.

(adapted from Sullivan et al., 1995)

Conservation pool elevation 138.7 m (455 ft.)

Surface area 2911 ha (7194 Ac) Volume 179x106 m3 (144,830 Ac-ft.)

Maximum depth 23 m (76 ft.)

Mean depth 6 m (20 ft.)

Length of shoreline 91 km (60 mi)

0 5 10

Kilometers

0-3 3-6 6-9 9-12 12-15 15-18 >18

Islands

Depth below conservation pool (meters)

One of the more prominent features of Lake Waco’s morphometry is due to its location at the confluence of the North Bosque River with the Middle and South Bosque Rivers resulting in a “L” shape with two branches oriented at almost right angles to each other. Inflows from the North Bosque River enter the northern most arm of the lake and converge with inflows from the Middle and South Bosque Rivers and Hog Creek, which form the southern arm.

A second feature of Lake Waco important in the understanding of reservoir behavior is the uneven distribution of flows and nutrients entering through each arm. This uneven distribution is due to marked differences in the size of the drainage basins along with differences in their soil and land use characteristics. As a result, the north arm of the reservoir dominates the inflow to the reservoir. For the period of Lake Waco monitoring used in this report (July 11, 1996 to July 29, 1998), approximately three-quarters of the flow entered through the north arm. The north arm is the major supplier of phosphorus to the lake ecosystem contributing an estimated 78 percent of the soluble phosphorus and 73 percent of the total phosphorus load to Lake Waco (McFarland and Hauck, 1999). The south arm contributes a significant portion of the nitrogen load to the lake due to the influence of row-crop agriculture. Row-crop agriculture was estimated to be the largest contributing source of total nitrogen (TN), suppling approximately 49 percent of the TN load. Land use estimates indicated that 93 percent of the row-crop agriculture in the watershed drains into the south arm of Lake Waco (McFarland and Hauck, 1999).

Literature Review A review of historical data and published literature pertaining to Lake Waco was performed to identify important limnologic processes to be modeled. Available literature indicated four dominant factors governing hydrodynamic and trophic response of the lake: wind driven mixing, light extinction by organic and inorganic turbidity, adsorption/desorption, equilibrium of clay-bound phosphorus, and water residence time. Previous studies of Lake Waco were also helpful in characterizing the phytoplankton community for specification of broad taxonomic or functional groupings within the model that would adequately mimic the real world community response. Published literature pertinent to parameterization of the lake model is summarized in the following section.

Historical data suggest a strong and persistent presence of bloom-forming diatoms and blue-green algae within the phytoplankton community of Lake Waco. Kimmel (1969) observed continual diatom dominance through November following a diatom bloom in August 1968.

Species composition observed in May 1977 consisted mainly of diatoms and green algae (Wyrick, 1978). Lind (1979) noted a spring diatom bloom in mid-May 1977. Algal assays conducted at the time of the bloom indicated no nutrient limitations for the main body of the reservoir near the dam (Wyrick,1978). Numerical dominance of blue-green algae was observed in August 1981, with numbers exceeding 40 percent of the total (USGS, 1982).

To adequately represent algal dynamics, a minimum of three phytoplankton groups, (green algae, blue-green algae, and diatoms) would need to be simulated. The inclusion of blue-green algae in the model was further supported by growing concerns about taste and odor problems in the City of Waco’s water supply that may be caused by blooms of certain blue-green algae species. It is interesting to note that taste and odor problems were experienced in Lake Waco within the first few years following impoundment which led to the installation of artificial aerators in problem areas including one in the deepest region near the outlet works (Biederman and Fulton, 1971). Thermal stratification persisting from May to October was observed prior to installation of the aerators in 1970 (Biederman and Fulton, 1971). Following installation, stratification is occasionally observed during calm periods but is generally short-

Chapter 2 Description of Study Area and Lake Waco lived (Kimmel and Lind, 1972). The causative agent responsible for taste and odor problems in Lake Waco is unknown but the blue-green algae or cyanobacteria which are commonly blamed for these problems may only be indirectly responsible due to their ability to fix atmospheric nitrogen. Concomitant increases in actinomycetes following cyanobacteria blooms may produce the metabolites responsible for the problems (Lind and Katzif, 1988).

Turbulent mixing caused by steady southwesterly or northwesterly winds (Lind, 1971) contributes to both a shallow photic zone due to resuspension of inorganic sediments and a deep mixed layer, which within the main body of the reservoir often includes the entire water column. This high mixing depth to photic depth ratio can contribute to a condition termed “morphometric oligotrophy” (Kimmel & Lind, 1972).

The influence of wind mixing on Lake Waco is responsible for considerable resuspension of clays (Lind, 1986). Additionally, these clays were found to be the primary removal mechanism for orthophosphate through adsorption processes. The dynamic equilibrium established between soluble and solid-bound fractions of phosphorus is also capable of maintaining a relatively constant amount of soluble orthophosphate under well mixed conditions (Kimmel and Lind, 1970).

Light extinction by organic and inorganic turbidity was found to be most limiting of algal growth (Kimmel, 1969). High turbidity due to suspension of clays originating from tributary inflows along with strong wind action was responsible for limiting algal production. High turbidity within the south arm of the reservoir was due to inflows of the South Bosque River, which according to Kimmel (1969) is “the principal silt contributor to Waco Reservoir.” Marl and clays of the Black Prairie physiographic province within the South Bosque drainage basin are more readily transported in surface runoff than the Grand Prairie soils within the North Bosque watershed (Kimmel and Lind, 1972). As noted by Lind (1979), light demonstrated a greater governing role than did nutrients in spring phytoplankton populations. High inorganic turbidity in the lake was cited as a possible explanation for differences in the relationship between total phosphorus and chlorophyll-α concentration observed by Lind (1979), through either a reduction in light penetration or sorption of phosphorus onto clay particles. Preliminary testing of a phytoplankton production model suggests that the combined influences of physical factors outweighs that of chemical factors three to one (Lind, 1979).

Lind (1971) observed a relatively consistent relationship between soluble and total organic matter within the four main tributaries draining the Lake Waco watershed. This relationship indicated that the amount of organic matter in the dissolved form exceeded that in the particulate form roughly three to one. Periods of high flow, when daily discharge exceeded the mean daily discharge for the month by more that five times, were excluded in developing this relationship. It was noted that during these flood dates, a significantly greater proportion of the organic matter was present in the particulate form.

Water residence time also plays an important role in controlling phytoplankton productivity.

Chlorophyll-α may be negatively related to reservoir inflows (Heiskary, 1995) particularly at high flows due to shorter residence times and cell washout (Walker, 1987). A poor relationship between phytoplankton production and chlorophyll-α in Lake Waco was observed during periods of severe drought (i.e., long residence times) such as that observed by Lind in 1978, which was characterized by low inflows and minimal external nutrient supply. During this time, the reservoir level dropped two meters and algal “production was more dependent on physical factors and recycling of internal nutrient stores” (Lind, 1979). During 1977, which was characterized by more normal rainfall patterns (both amounts and frequency), seasonal patterns in algal production tended to parallel chlorophyll concentrations (Lind, 1979).

Artificial aeration at two sites in Lake Waco was initiated in 1970 to discourage chemical stratification and hypolimnetic anoxia. One aerator is located in the deepest portion of the lake near the outlet works and the other near the confluence of the North Bosque River. A study conducted by Roark and Lind (1988) concluded that the effects of the aerator were minor and limited to within 50 meters of the aerator.

Target Development Target development for Lake Waco centered on identifying the nutrient most limiting to the phytoplankton community in the lacustrine portion of the lake and determining a range or threshold value for the limiting nutrient conducive of a healthy algal community. From bioassay studies conducted by Dr. Owen Lind, phosphorus was determined to be the nutrient most limiting to algal communities (Davalos-Lind and Lind, 1998). Relationships between chlorophyll-α and orthophosphate concentrations developed from TIAER monitoring data revealed a threshold effect with a Monod-type growth response (Kiesling et al., 2001).

Relationships were established based on near-surface summer mean concentrations of orthophosphate and chlorophyll-α for stations located in model segment 8. The summer growing season for phytoplankton, defined as April through September, provided the strongest relationship between the response variable (chlorophyll-α) and the stimulus (PO4- P). Based on these relationships and recommendations by the Technical Work Group (TWG), a target range between 8 and 14 parts per billion (ppb) was proposed with a preliminary target for mean summer (April-September) orthophosphate concentration set at 10 ppb.

Identification of Phosphorus Sources in the Watershed Percent contribution of soluble (PO4-P) and total (TP) phosphorus in the Bosque watershed (Table 3) were estimated by McFarland and Hauck (1999). While comprising about 2 percent of the watershed area, dairy waste application fields were estimated to contribute 35 percent of the PO4-P and 21 percent of the TP within the watershed for the period examined. The wood and range land use category was also estimated to contribute a large percentage of the phosphorus load (22 percent of the PO4-P and 30 percent of the TP) but comprises roughly 63 percent of the watershed area. Other sources identified in the watershed include pasture, cropland, urban areas, and point source discharges from municipal wastewater treatment plants (WWTPs).

Table 3 Estimated percent contribution of phosphorus by source to the Bosque watershed for November 1, 1995 through March 30, 1998. Source: McFarland and Hauck (1999)

Source PO4-P (%) TP (%) Land Use (%)

Dairy waste application 35 21 2

Row crop 11 17 15

Non-row crop 2 2 2

Pasture 10 16 14

Wood/range 22 30 63

WWTP 9 4 not applicable

Urban 11 10 2

CHAPTER 3

Description of Lake Model

Model Selection The two-dimensional, laterally averaged, numerical model, CE-QUAL-W2, was selected for simulation of the hydrodynamics and water quality of Lake Waco and to assess the impact of alternate allocation scenarios on selected water quality constituents. CE-QUAL-W2, developed and supported by the U.S. Army Corp of Engineers Waterways Experiment Station (USACE-WES), is recognized as a state-of-the-art hydrodynamic/water quality model (Cole and Buchak, 1995). In addition to the ability to simulate 21 water quality state variables, CE- QUAL-W2 builds on the proven hydrodynamic solution techniques of its predecessors, the Laterally Averaged Reservoir Model (LARM) and the Generalized Longitudinal and Vertical Hydrodynamics and Transport (GLVHT) model. The capability of the model to predict changes in water quality in two dimensions (longitudinally and vertically) for multiple branches were among the characteristics which made CE-QUAL-W2 particularly suited to simulation of Lake Waco. The assumption of well-mixed conditions in the lateral direction is usually suitable to long, narrow reservoirs dominated by longitudinal flow through the water body. The configuration of Lake Waco with two relatively narrow branches that exhibit strong longitudinal gradients in flow and nutrient concentrations support the use of lateral averaging and the need for greater longitudinal resolution.

The model’s hydrodynamic sophistication, allowing for the addition or subtraction of layers and segments, and the use of an auto-stepping algorithm for greater numerical stability were suitable for modeling the changes in water level due to temporary storage and routing of flood flows through the reservoir. The emphasis placed on both hydrodynamic and water quality capabilities, favored the selection of CE-QUAL-W2 over simpler hydrodynamic approaches with comparable water quality capabilities such as the Water Quality Analysis Package (WASP). The model’s implicit calculation of vertical and horizontal velocities by layer was also an important consideration in the selection of CE-QUAL-W2 over WASP.

In addition to simulation of hydrodynamics and water temperature, the water quality constituents necessary for this application were associated with nutrient forms transported from the watershed and associated changes in the phytoplankton community. As discussed earlier, at least three algal groups would need to be specified in the model to adequately represent algal community dynamics of interest in Lake Waco. Since the currently released version of the model (version two) was limited to a single algal group, a prerelease of version three was obtained from USACE, which allowed for multiple algal groups. Initial testing of the beta version of CE-QUAL-W2 (V3) demonstrated that this version possessed the capabilities required for simulation of the unique hydrodynamics and water quality conditions in Lake Waco. Although a published manual for the prerelease version (V3) was not available, Cole and Tillman (1999) document the application of a similar prerelease version to Lake Monroe and discuss the capabilities and enhancements to the model since release of version two.

Water Quality Modeling of Wake Waco Using CE-QUAL-W2 Modeling Approach A complicating factor in the calibration process for this project was the need to integrate the CE-QUAL-W2 model with the Soil and Water Assessment Tool (SWAT) model as a means to predict the effects of various phosphorus control strategies in the watershed. This integration necessitated independent calibration of both SWAT and CE-QUAL-W2 followed by a calibration check of the integrated model output as depicted in Figure 3. A two-phase calibration approach was adopted for this project. The first phase consisted of calibration of SWAT and CE-QUAL-W2 independent of one another. During the independent calibration phase, data from TIAER’s stream monitoring network served dual purposes of calibration data for SWAT and input to CE-QUAL-W2. Following the initial calibration (referred to as independent calibration hereafter), the watershed and lake models were integrated through an intermediate interface program, which allowed the predicted stream quality output from SWAT to be used as input to CE-QUAL-W2. The second phase of calibration consisted of fine-tuning the integrated SWAT/CE-QUAL-W2 model (referred to as integrated calibration hereafter). The rational for using a two-phase calibration approach stemmed from the desire to account for the uncertainties associated with simulation of the landscape and receiving water body separately before lumping the uncertainties in the integrated modeling system.

The watershed loading model, SWAT, developed by the United States Department of Agriculture-Agriculture Research Service (USDA-ARS), was applied by Blackland Research and Extension Center (BREC) to simulate watershed processes. The SWAT model is physically based, uses readily available inputs, and is computationally efficient to operate on large basins in a reasonable time. SWAT is capable of simulating long periods for computing the effects of management changes. Major components of the model include hydrology, weather, erosion, soil temperature, crop growth, nutrients, pesticides and agriculture management. A complete description of SWAT model components may be found in Arnold et al. (1998). Santhi et al.

(2001) provides documentation of SWAT calibration and application to the Bosque River watershed.

The approach adopted by the TWG for evaluation of alternative phosphorus (P) control strategies involved the use of long-term simulations to minimize the stochastic uncertainty associated with model predictions. Based on the availability of input data for CE-QUAL-W2, a 32-year simulation period was selected, which utilized historic weather data from 1966 through 1997 as representative climate data for simulations of future conditions. A 20-year planning horizon was adopted for projection of watershed conditions to future (year 2020) conditions. Future conditions in the watershed were estimated based on projected changes in land-use patterns, population, urban area and dairy herd size. For those sectors controlled by permit limitations, future conditions were represented by either the projected growth from current levels to year 2020 or fully permitted levels, whichever was greater. The efficacy of P control measures in improving water quality in Lake Waco would be assessed relative to target attainment under future conditions.

Future projections through the year 2020 also considered physical changes in the normal operating depth of Lake Waco and potential water quality changes that could result. In anticipation of increased water supply demand by the City of Waco, the USACE (1982) approved the reallocation of a portion of Lake Waco’s storage capacity from flood control to water supply. Dam alteration for the proposed reallocation was scheduled to begin in 1998 and upon completion would increase the conservation pool elevation by 2.1 meters (7 feet).

Chapter 3 Description of Lake Model Figure 3 Diagram of calibration process.

The increased pool elevation could potentially impact water quality. Greater water depths could lead to greater anoxia in the bottom waters, both in extent and duration of exposure.

Under anoxic conditions, PO4-P and NH3-N may be released from the sediments into the overlying water column further accelerating the eutrophication process. The rise in conservation pool elevation will also inundate new areas along the lake’s periphery.

Terrestrial vegetation such as grasses, trees and shrubs once submerged will soon die and decay. The newly inundated areas along the lake’s periphery will serve as an additional organic nutrient source and exert an oxygen demand as it decays.

The TWG recommended, however, that the modeling evaluations be performed for current lake conditions and a post-audit model run be performed with the increased conservation pool elevation to examine model’s response to increased depths. Further, the TWG recommended that this post-audit be performed with the calibrated values, ignoring the effects of decaying vegetation in newly inundated areas. Based on the guidance of the TWG, except for a post-audit model exercise, all CE-QUAL-W2 calibration and subsequent applications occurred for lake conditions existing at the time of the study.

Model Input Data Input data for calibration of CE-QUAL-W2 were assembled from numerous sources which varied based on whether CE-QUAL-W2 was being calibrated with observed inflows or with inflows predicted by SWAT. Input data required for model setup may be divided into four broad categories, 1) physical and spatial characteristics of the water body defining the computational grid, 2) time-varying boundary conditions describing the meteorologic and hydrologic influences on the water body, 3) initial conditions, and 4) kinetic and hydrodynamic data characterizing the physical, chemical, and biological processes.

Physical Representation and Segmentation Scheme Bathymetric data were required for specification of the computation grid defining the longitudinal and vertical segmentation of the water body and cell widths. The computational grid defined for Lake Waco was based on the desired longitudinal and vertical resolution, the branch configurations, and the reservoir bathymetry.

The reservoir was discretized into a finite-difference grid consisting of three branches. The main body of the reservoir was designated as branch 1, the North Bosque Arm (North Arm) as branch 2, and the Middle and South Bosque Arm (South Arm) as branch 3. Each branch was divided into segments approximately 500 meters in length. Segments were divided into 1-meter thick vertical layers. The segmentation of Lake Waco and the associated computational grid are depicted in Figures 4 and 5, respectively. The computational grid consisted of 47 segments and 37 layers with 311 active cells at the conservation pool elevation of 455 feet (139 meters). The computational grid was defined up to the maximum design water surface elevation (elevation of 505 ft or 154 m), which is well above lake levels observed during the lake’s history.

The CE-QUAL-W2 model requires specification of boundary segments at the upstream and downstream boundaries of each branch. Segments 1, 9, 10, 28, 29, and 47 are boundary segments. The model also requires boundary layers at the surface and bottom of the reservoir.

Layer 1 is the surface boundary layer. The bottom boundary layer varies by segment depending on depth. Boundary grid cells are depicted in Figure 5.

Bathymetric data for the lake were used to assign volumes to each cell in the numerical grid.

Reservoir bathymetry was obtained in digital form from Baylor University’s Department of Geology from the most recent volumetric survey of Lake Waco (Sullivan et al., 1995). These data were used to determine grid characteristic below the conservation pool elevation (layers 17-36). Topographic data from the USGS 7.5 minute Digital Elevation Model (DEM) were used to determine volumes for cells above the conservation pool elevation (layers 2-16). Layer 1 is a boundary layer required by the model. Boundary cells are required at the top and bottom and at the upstream and downstream boundaries.

Two branch inflows and five tributary inflows were defined in the computational grid (Figure 4). The North Bosque River enters segment 11 at the uppermost segment of the north arm and constitutes the inflow to branch 2. The Middle Bosque River enters at segment 30 and forms the inflow to the south arm (branch 3). Two main tributaries, the South Bosque River and Hog Creek enter branch 3 at segments 32 and 35, respectively. Three small ungauged areas are located adjacent to the lake and enter at segments 27, 39, and 45.

Figure 4 Lake Waco segmentation.

Fi gu re

L ak e W ac o co m pu ta ti on al g ri d

1 2 3

4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37

C on se rv at io n Po ol B ou nd ar y

C el ls

S eg m en ts L a y e r

B ra nc h

B ra nc h

B ra nc h

Model State Variables The variables modeled (or state variables) in this application include:

1. Hydrodynamics (water level, horizontal velocity, and vertical velocity)

2. Conservative tracer

3. Water temperature

4. Dissolved oxygen

5. Four organic matter pools

6. Ammonium

7. Nitrate+nitrite

8. Orthophosphorus

9. Three phytoplankton groups

Organic matter (OM) in the model is partitioned into four pools based on classification of physical state (dissolved versus particulate) and into one of two classes (labile or refractory) characterizing the mineralization (decay) rate of the organic compound to inorganic nutrients.

Labile OM is more readily mineralized (i.e., faster decay rates), whereas refractory OM is less readily mineralized (i.e., slower decay rates). The four OM pools are labile dissolved organic matter (LDOM), labile particulate organic matter (LPOM), refractory dissolved organic matter (RDOM), and refractory particulate organic matter (RPOM).

Organic matter (OM), as specified in CE-QUAL-W2, was not directly measured by TIAER and had to be estimated based on organic phosphorus (Org-P) measurements. Org-P was computed as total phosphorus (TP) minus orthophosphate phosphorus (PO4-P). Two CE- QUAL-W2 input parameters, ORGN and ORGP, govern the breakdown of OM into Organic N (Org-N) and Organic P (Org-P) fractions. By definition, ORGN is the fraction of N in OM and ORGP is the fraction of P in OM used by CE-QUAL-W2. Published literature and model default values suggested a typical value of ORGN equal to 0.08 (8 percent of OM is Org-N).

The fraction of Org-P in OM (ORGP) was estimated relative to ORGN assuming that the ratio of ORGN to ORGP was similar to the mass ratio of Org-N to Org-P loads to the Lake Waco (equation 1).

1)

Monitoring data from stream stations on major tributaries indicated that Org-N to Org-P ratios were variable both temporally and spatially. Restrictions in the model allowed only a single value of ORGN and ORGP to be used within a simulation; therefore, cumulative organic loads to the lake over the period of interest were computed and used to estimate the value of ORGP. This method would ensure that the total organic N and P load to the lake would be realized. Time-series generated for OM would reflect the temporal variation in Org- P; however, Org-N loads would be under or overestimated on a daily basis but the cumulative load would be correct. Considering the importance of P as the nutrient most limiting of algal production and the focus of water quality targets selected for Lake Waco, it was important to accurately reflect the Org-P loads.

Cumulative organic N and organic P loads to Lake Waco for the calibration period were estimated based on measured stream flow data to be 2,548,045 kg and 507,670 kg, respectively.

Rearranging equation 1 and assuming ORGN = 0.08, yields an ORGP value of 0.016 (equation 2), which was used in CE-QUAL-W2.

ORGN

ORGP

ΣOrg-N Load ΣOrg-P Load

2)

Consequently, OM was back computed from Org-P timeseries data (equation 3).

3)

The same procedure was used for both the observational data and SWAT predictions. An ORGP of 0.016 was used for calibration and for long-term simulations of P control scenarios.

The ORGN values was set at 0.08 for the independent calibration. For the integrated model calibration and long-term simulations, ORGN was adjusted based on the cumulative loads predicted by SWAT. Estimation of organic matter was performed for initial conditions and specification of inflow boundary conditions.

The phytoplankton community of Lake Waco was represented in CE-QUAL-W2 by three algal groups based on temperature and light preferences and competitive ability under nutrient limited conditions. Algal group one was loosely defined as diatoms, although other types of algae such as flagellates and cryptomonads may also be included in this group. Algae in this group were considered to be spring and fall dominants with optimal temperatures below that of the other two groups. Alga of this group were also characterized by slightly higher growth rates and considerably higher settling rates compared to the others groups due to their siliceous frustrules. It should be noted at this point that silica was not modeled in this application. There is some evidence that under bloom conditions, silica may limit diatom growth in Lake Waco; however, the primary focus being on phosphorus along with the lack of data characterizing inflowing silica concentrations and the inability of SWAT to predict these concentrations, precluded the simulation of dissolved silica concentrations in Lake Waco.

Specification of the other two algal groups were based primarily on their ability to compete for limited resources under nutrient limited conditions. Algal group two, broadly defined as green algae, would prevail under phosphorus limited conditions, whereas, group three, characterized by blue-green algae, would prevail when nitrogen was limiting.

Collection of In-Lake Calibration Data Water quality data collected at eight sites within Lake Waco (Figure 6) were used for the purpose of model calibration. Monthly monitoring began in June 1996 except for biweekly sampling, which was initiated for the summer 1998 sampling period to better characterize algal community changes. The duration of algal bloom conditions are generally short, lasting days to weeks before available resources are depleted. Residual water quality effects indicating the occurrence of an algal bloom may only be apparent for a several weeks following the bloom and could go undetected based on a monthly sampling frequency.

Data collected through July 29, 1998 were used in model calibration. Comparisons of model predictions to observed data focused on model segment 8, which is the area of the lake where the phosphorus target was defined. Four sampling sites were established within segment 8 to characterize water quality conditions near important features such as the confluence of Landon Branch, the reservoir outlet structure (riser), and spillway. Water quality data collected at these four sites were averaged to provide observed data comparable to the lateral average predicted by CE-QUAL-W2. During the first portion of the model calibration period (through March 1998), observational data for segment 8 were represented as a mean values with error bars representing the range of values observed during that sampling date. Latter in

ORGP 0.08 507670( )

2548045

--------------------------------- 0.016==

OM Org-P 0.016⁄= the calibration period (April 1998 to July 1998), the number of sampling sites in segment 8 was reduced to a single site (LW013) to accommodate an increase in sampling frequency from monthly to biweekly monitoring. Four additional sites (LW070, LW012, LW015, and LW060) were used for model comparisons with respect to longitudinal gradients along the north and south arms of Lake Waco.

The water quality parameters listed in Table 4 were determined either directly or indirectly from TIAER sampling data. At each lake monitoring site, water temperature and dissolved oxygen (DO) profile measurements were taken at one-meter increments from the surface through the water column. Organic and inorganic nutrients and total suspended solids concentrations were determined for surface, mid-depth and bottom samples collected at each site.