Water Quality Modeling in the Barr Lake and Milton Reservoir Watershed


by George Patten, Associate Scientist, Integral Consulting

Figure 1.
Overview map showing Barr/Milton “datashed” project boundary. From the 2008 Barr Lake and Milton Reservoir Watershed Management Plan. Available online at www.barr-milton.org.

Watershed Overview
Barr Lake and Milton Reservoir are two warm-water reservoirs located on the high plains northeast of Denver. Water quality impacts in the lakes associated with high levels of nutrients, primarily as nitrogen and phosphorus, led to their inclusion on Colorado Department of Health and Environment’s (CDPHE) 303(d) list of impaired waters in 2004. The 303(d) listing is medium priority for exceedance of the upper limit pH standard of 9.0, and requires completion of a Total Maximum Daily Load (TMDL). In May of 2005, the Barr Lake and Milton Reservoir Watershed (BMW) Association formed with the mission “to encourage cooperation, involvement, and awareness by interested parties in collaborative efforts to improve the water quality of Barr Lake and Milton Reservoir.” In order to fulfill this mission, and to begin technical work in support of TMDL development, the BMW Association worked closely with consultants at Integral Consulting, who served as coordinators and technical advisors, and AECOM, Inc. who conducted a water quality modeling investigation of Barr Lake, Milton Reservoir and the surrounding watershed. The objective of this modeling work was to develop an accurate water quality predictor in the two lakes to help inform proper management scenarios for limiting pH exeedances in the lakes.

The BMW Association is a consensus-driven group dedicated to improving water quality through collaborative efforts. The group includes representatives from cities and towns, major wastewater treatment facilities, irrigation companies, drinking water providers, agricultural water users, and recreational users. The group formed as a 501(c)(6) organization in 2005 and is currently governed by a Board of Directors.

Figure 1 presents an overview of the Barr-Milton project boundary. The project boundary encompasses a large area, over 850 square miles. Land uses include mostly developed residential, commercial or industrial areas (38 percent) and agricultural areas (55 percent).

The South Platte River represents the mainstem of the watershed and provides flow for both lakes. Barr is filled via the Burlington canal, which transports surface water diverted from the South Platte as well as treated effluent from Denver Metro Wastewater Reclamation District. Milton is filled through a South Platte surface water diversion at the Platte Valley headgate, which feeds the Platte Valley canal. Milton is also fed by the Beebe Draw, a canal fed by seepage from Barr Lake. Barr Lake and Milton Reservoir are classified as able to support warm-water, class 2 aquatic life. Currently, the water in Barr and Milton is stored and used for downstream irrigation purposes, as well as secondary recreational uses. In 2004, both lakes were also designated as potential sites for drinking water supply for the Denver Metro region.

Water Quality Issues
Barr and Milton both experience excessive nutrient loading, from point and nonpoint sources, which contributes to high pH and eutrophic conditions during certain times of the year. Low summertime clarity, high total phosphours and nitrogen, and chlorophyll-a concentrations, and severe algal blooms from July to October are all evidence of water quality problems in the reservoirs. The state of Colorado does not currently have water quality standards for nitrogen and phorphorus,which are the root cause of the impairment in Barr and Milton, but is in the process of developing standards for nutrients.

The BMW Association currently focuses solely on the nonattainment of the pH standard at the two lakes. The goal in addressing lake pH is to work with the State to develop and implement a pH TMDL, which will contain allowable load measurement and a plan for achieving load reductions. The BMW Association recognizes that high pH levels in the lakes are caused by excessive in-lake primary productivity (i.e., algae growth) which results from high nutrient loads.

Water Quality Modeling
Numerical models were developed by AECOM, in partnership with the BMW Association, to provide a predictive platform for phosphorus management scenarios and to support pH TMDL development for Barr Lake and Milton Reservoir. The model results will guide the BMW Association and regulators in the process of determining allowable levels of nutrients that will result in pH attainment. AECOM and the BMW Association released the final modeling report, titled Watershed and Lake Modeling for a TMDL Evaluation of Barr Lake and Milton Reservoir, in August 2009.

The Barr/Milton watershed model consists of two linked public domain models which were slightly modified based on site-specific characteristics to better describe the complexities of water flow pathways and nutrient/pH dynamics. AECOM modeled watershed nutrient loads with a modified version of the Surface Water Assessment Tool (SWAT) model developed by the USDA for surface water operated through an ArcSWAT interface using ERSI’s ArcGIS software. A second numeric model, EPA’s Water Quality Analysis Simulation Program (WASP), was used to predict conditions within the reservoirs. The WASP model includes multiple nutrient interaction equations and phytoplankton growth simulations, as well as nutrient settling rate simulations. These models are designed to examine indicator parameters selected for pH compliance (e.g. chlorophyll-a [chl-a], transparency, nitrogen, and phosphorus) and to predict water quality outcomes for various nutrient load levels that enter the lakes.

Prior to model development, AECOM and the BMW Association compiled an extensive set of watershed and water quality data. Geo-spatial data were acquired from the BMW Association Technical Committee, State and Federal agencies and other sources, while meteorological data were gained through the National Oceanographic and Atmospheric Administration (NOAA). Comparative water quality and flow data were acquired from AMEC (formerly Hydrosphere) databases and from the State of Colorado. The simulation period of 2003-2004 was chosen according to data availability and requirements of the SWAT model, while interpretations and assumptions were made to fill any data gaps.

The initial step in the model development process was to develop a detailed and accurate conceptual understanding of the complex system of water transfers in the project study area. The next critical piece in the model development process was model setup and calibration. The purpose of calibration was to gain the most accurate agreement between predicted and observed values of flow and water quality parameters, with the focus being two primary inflow locations at Barr Lake and Milton Reservoir. Model calibration required extensive model adjustments to create the most realistic platform for estimating observed conditions in the project area.

Model sensitivity was tested for four test scenarios:1) wastewater controls 2) reservoir management 3) land use alterations and 4) in-lake modifications. The results of these model runs show that reductions in phosphorus concentrations in Barr and Milton can be achieved through reductions in phosphorus loads coming from the largest wastewater treatment plants located in the project boundary. The sensitivity analysis also showed that internal loading of phosphorus currently stored in lake bottom sediments is significant, and that reductions in point sources of phosphorus may have diminishing returns due to high internal loading. The effect of nonpoint sources were shown to have relatively little effect on total phosphorus concentrations in the reservoirs.

A second set of four scenarios were run through the model to test water management options. These scenarios included 5) detailed wastewater controls 6) reductions from permitted areas 7) reductions from permitted and non-permitted areas and 8) modified reduction to internal loads. The results from these scenarios had similar implications as the first four model runs, showing that the water quality problems in Barr and Milton are largely a result of point source discharges from WWTPs. Reductions in phosphorus loads from the two largest WWTPs is predicted to have the greatest effect on concentrations in the reservoirs, but smaller WWTPs could also be important once overall phosphorus concentrations are reduced. Thus, the model indicates that reductions in phosphorus loads from the two largest WWTPs in the watershed represent the most likely target for approaching the goal of pH compliance in the reservoirs.

These model results will inform pH TMDL development for Barr Lake and Milton Reservoir. While the model indicates that controls on WWTP discharges will likely result in the greatest reduction in phosphorus concentrations and pH levels, the BMW Association recognizes that phosphorus treatment of wastewater is expensive to implement and would require extensive facility expansion. Additional management practices may be the form of multiple watershed controls and in-lake treatment options. The Watershed and Lake Modeling for a TMDL Evaluation of Barr Lake and Milton Reservoir report outlines many of these in-lake treatments including biomanipulation, aeration, mixing, phosphorus inactivation and algaecides. A final version of the pH TMDL for Barr and Milton is expected by mid-summer 2010. The BMW Association will also continue to conduct public outreach campaigns to educate communities about the history of the Barr-Milton watershed project, current pH TMDL process and future efforts to manage water quality in the watershed.

In summary, the combination of the modified SWAT and WASP platforms produced a linked watershed-lake model capable of predicting final concentrations of phosphorus and nitrogen compounds, chl-a, and pH in Barr Lake and Milton Reservoir over time. The models are a useful tool for estimating existing conditions in the watershed and reservoirs and for predicting outcomes of changes to the system that might affect water quality. Model results show that pH reductions in Barr and Milton can be achieved through reductions in phosphorus loads coming from the wastewater treatment plants located in the project boundary. The sensitivity analysis also showed that internal loading of phosphorus currently stored in lake bottom sediments is significant. Based on model results, the BMW Association is considering both point source and in-lake management controls to bring lake pH within attainment of water quality criteria.

For more information about this project and to view the modeling report, please visit the Barr Lake and Milton Reservoir Watershed Association website at: