Source: COLORADO STATE UNIVERSITY submitted to
ESTIMATING NUTRIENT LOADS FOR WATER QUALITY MANAGEMENT
Sponsoring Institution
National Institute of Food and Agriculture
Project Status
TERMINATED
Funding Source
Reporting Frequency
Annual
Accession No.
0200177
Grant No.
(N/A)
Project No.
COL00726
Proposal No.
(N/A)
Multistate No.
(N/A)
Program Code
(N/A)
Project Start Date
Jul 1, 2004
Project End Date
Jun 30, 2007
Grant Year
(N/A)
Project Director
Loftis, J.
Recipient Organization
COLORADO STATE UNIVERSITY
(N/A)
FORT COLLINS,CO 80523
Performing Department
CIVIL AND ENVIRONMENTAL ENGINEERING
Non Technical Summary
Most drinking water reservoirs along the Front Range of Colorado are experiencing degradation due to nutrient inputs that accelerate algae growth. The purpose of the research is to enhance the scientific basis for estimating nutrient loads for source water protection.
Animal Health Component
50%
Research Effort Categories
Basic
10%
Applied
50%
Developmental
40%
Classification

Knowledge Area (KA)Subject of Investigation (SOI)Field of Science (FOS)Percent
1120110205030%
1120210205030%
1120320205030%
1020210205010%
Goals / Objectives
The overall goal of the research is to enhance the scientific basis upon which water managers in Colorado and the West will estimate nutrient loads from multiple sources and land uses in their efforts to protect the quality of receiving waters, especially drinking water reservoirs. Obviously, this goal is far greater in scope than the proposed project would permit. Thus we focus on two more limited objectives. The first project objective is to refine estimates of nutrient (both nitrogen and phosphorus) losses from irrigated turf by completing the experimental studies begun in Colorado Experiment Station Project No. COL00705 Best Management Practices for Landscape Irrigation, which ends June 30, 2004. The second objective is to develop and demonstrate a cost-effective approach for estimating nutrient loads from multiple sources and land uses in watersheds along the Colorado Front Range. This objective will focus on phosphorus because it is the key nutrient for protecting Colorado drinking water reservoirs from eutrophication.
Project Methods
Research Approach, Objective #1: The first objective will be accomplished in the first year of the proposed project. The approach will be to continue both the field and lysimeter studies of N and P leaching and the greenhouse studies of N and P transport in surface and subsurface runoff that are reported in the current CRIS report for project COL 00705. Results thus far suggest that phosphorus in runoff from irrigated turf can occur at higher rates than those reported in other studies, but these results are not yet conclusive due to analytical difficulties associated with phosphorus in various forms: total, dissolved, ortho, organic, readily extractable. Thus a fourth year of experiments is needed. Research Approach, Objective #2: The second objective will use a case study approach, involving the Big Thompson Watershed. The project will include significant interaction with the Big Thompson Watershed Forum (BTWF). The general approach for this objective will consist of two components. The first component will be to establish a phosphorus mass balance for the entire watershed through estimation of annual phosphorus entering and leaving the watershed. The second component will be estimation of the contributions to those loads by external sources (trans-mountain diversion and precipitation) and by specific areas of the watershed and by source type within those areas. In both components seasonal and annual phosphorus loads at key nodes will be estimated using the BTW Forum monitoring data, possibly supplemented by additional data sources. Quantification of loads by source type will be more difficult, and thus less precise. In the area of point sources, we plan to work with the wastewater treatment plants to obtain reasonable estimates of P loads. At the sub-watershed scale, aggregated nonpoint contributions for a given segment will be estimated by differencing the load estimates for adjacent nodes and subtracting out the point source estimates. Within the subwatersheds, we shall use the watershed component of EUTROMOD, along with typical loading functions (for livestock, agricultural land, septic tanks, etc.) obtained from the literature. To validate those loading functions for local conditions, we shall directly monitor surface runoff from areas that contain significant nonpoint sources. We shall monitor phosphorus deposition in rain and snowfall over a range of elevations within the watershed.

Progress 07/01/04 to 06/30/07

Outputs
This project first completed a small-scale experimental study of phosphorus transport from irrigated turf and then investigated alternative modeling approaches for predicting phosphorus loads from western mountain watersheds. The primary contribution of the experimental study was to suggest that phosphorus transport in both surface and subsurface drainage from irrigated turf might be more substantial than is generally reported and could be a significant source from a source water protection perspective. The mobility of phosphorus was greatly increased by the addition of ammonium sulfate fertilizer, even when phosphorus fertilizer was not added at the same time. The major finding of the model evaluation was that export coefficient models appear to be the most appropriate for use by watershed forums and similar organizations with limited modeling expertise. These models are less data intensive, more straightforward, and more quickly adaptable for complex water networks than more complex watershed models. Export coefficients found in literature were derived in flatter, more humid conditions than those present in Colorado's Front Range. Therefore new export coefficients, appropriate for application in more mountainous and arid terrain, were developed in this study. This involved a regression approach to define the relationship between land cover type areas and observed phosphorus loads. Data were obtained from the following watersheds in Colorado: Cache la Poudre, Big Thompson, St. Vrain, Clear Creek, Blue River, South Platte headwaters. Since phosphorus loads were found to be lognormally distributed, the natural logs of the loads were used as the response variables in regression. Areas of each land cover type were calculated using 10-m DEMs and GIRAS land cover data. These areas were also estimated for basins limited to a specified distance from the stream channel. These buffer widths included 120, 240, 480, and 3000-m distances. Stepwise regression was performed in Minitab for whole basins as well as each buffer width. The 480-m buffer produced the best regression model with the highest adjusted R2 (72.08) and the lowest PRESS statistic (30.9067). The land cover export coefficients in this model included developed, barren, forest, agriculture, and wetland and were found to be 0.00434, -0.0076, 0.00058, 0.00075, -0.0084 kg/ha/yr respectively. Area-weighted discharge had a coefficient of 3.02 kg/yr. Several problems with the regression model were identified, and the regression results should be viewed as preliminary. The non-negligible constant (2.244) suggests that there are factors influencing phosphorus export which are not accounted for by the regression model. Furthermore, strong correlations were found between a number of predictor variables. These correlation issues are being addressed through refinements to the regression analysis. In addition, phosphorus and discharge data from 8-digit HUCs for the entire ecoregion will be used in future regression analyses. The larger dataset and more uniform basin sizes should improve the regression models and produce better export coefficients.

Impacts
The results of this project will be used by watershed organizations, such as the Big Thompson Watershed Forum, and by water suppliers, such as Front Range cities, to develop and improve their water quality assessments and management plans for protecting water sources from eutrophication. The work will provide a low-cost, first-cut approach to identifying and prioritizing phosphorous sources for further study and application of Best Management Practices. This will be especially critical in the early stages of any future development of phosphorus TMDLs. Non-profit watershed organizations in particular, rely heavily on volunteers, donations, and grants. Improving their ability to assess and convey watershed information will directly influence their ability to solicit community involvement and funding, and therefore increase their resources for implementing watershed protection and improvement projects.

Publications

  • No publications reported this period


Progress 01/01/06 to 12/31/06

Outputs
In this second year of the project, the focus has shifted entirely to estimation of phosphorus loads from multiple sources and land uses in mountain watersheds. The project is designed to meet the needs of watershed organizations like the Big Thompson Watershed Forum, in Loveland, Colorado. Such organizations need to determine the scope of a particular water quality problem, estimate how conditions might change over time, and evaluate alternative management options. They typically have very limited financial resources and moderate technical expertise. Integrated modeling systems have the potential to meet their needs by providing a means for managing and interpreting data, simulating various conditions and management options, and generating information-rich reports. Selection of an appropriate watershed model is central to this approach, and the extent of data available to drive the model is a key factor in selection. Model complexity must be weighed against the resource limitations of the user organization. Watershed characteristics must be considered and will be a significant limiting factor. Mountain watersheds, with highly variable climate and terrain, present a special challenge that is of particular importance to Colorado and has not been adequately met in previous research. Models considered here can be divided into two types: mechanistic (process-based) and statistical (data-based). Mechanistic models can be subdivided into three categories based on level of complexity: export coefficient models, loading function models, and chemical simulation models. Mechanistic models are designed to simulate watershed and chemical transport processes using mathematical functions. Statistical models do not attempt to simulate hydrologic or chemical transport processes, but instead identify statistical relationships between observed water quality data and various predictors. This project is currently finishing an in-depth look at all three levels of mechanistic models and their applicability to a mountainous Colorado watershed. There is little information published regarding appropriate export coefficients for climate and terrain typical to Colorado. Therefore export coefficients are being calculated from existing Colorado water quality data and corresponding land cover information using both frequentist and Bayesian techniques. These calculated export coefficients will be used in a simple spreadsheet model and the results compared to those of more complex models like EUTROMOD and HSPF in BASINS. Next, this project will evaluate the ability of a statistical model to predict and assess phosphorus loading. Such a model seeks statistical relationships between combinations of various watershed characteristics (ie landscape metrics) and observed phosphorus loads. These relationships can then be used in a predictive capacity to assess water quality under changing conditions and new scenarios. The performance of a statistical model applied to a mountainous watershed will then be compared to that of a mechanistic model to determine the most appropriate model for use in the final integrated modeling system.

Impacts
The results of this project will be used by watershed organizations, such as the Big Thompson Watershed Forum, and by water suppliers, such as Front Range cities, to develop and improve their water quality assessments and management plans for protecting water sources from eutrophication. The work will provide a low-cost, first-cut approach to identifying and prioritizing phosphorous sources for further study and application of Best Management Practices. This will be especially critical in the early stages of any future development of phosphorus TMDLs. Non-profit watershed organizations in particular, rely heavily on volunteers, donations, and grants. Improving their ability to assess and convey watershed information will directly influence their ability to solicit community involvement and funding, and therefore increase their resources for implementing watershed protection and improvement projects.

Publications

  • No publications reported this period


Progress 01/01/05 to 12/31/05

Outputs
The first year of project COL00726 revealed a correlation between ammonium sulfate fertilizer application and phosphate concentration in surface runoff. Therefore, in 2005 the fertilizer treatments for the greenhouse surface runoff experiments were modified to further examine this relationship. Although P export was the focus of this second set of greenhouse experiments, nitrate/nitrate concentrations continued to be measured. As in the 2004 greenhouse experiments, the turf-covered boxes were irrigated weekly with a hand sprinkler until runoff occurred. Surface runoff and leachate from the boxes were collected separately and analyzed for phosphate and nitrate/nitrite. No fertilizer was applied to the boxes during the first five weeks of the experiment. During the sixth week, ammonium sulfate fertilizer was applied in equal amounts (25.7 kg N/ha) to each grassed box to assess variability among the boxes. At the end of the ninth week, superphosphate fertilizer was applied to Box 1 and Box 3 (55 kg P/ha) and ammonium sulfate fertilizer was applied to Box 1 and Box 2 (77kg N/ha). This fertilizer scheme simulated four different treatments: ammonium sulfate and superphosphate fertilizer applied simultaneously, ammonium sulfate fertilizer applied alone, superphosphate fertilizer applied alone, and no fertilizer. Phosphate concentrations in surface runoff did not appear to be influenced by the application of ammonium sulfate fertilizer at either application rate. However application of superphosphate fertilizer during the ninth week of the experiment corresponded to a large increase in surface runoff phosphate concentrations in from Box 1 and Box 3. Concentrations were observed to increase from approximately 2 mg/L to 25 mg/L during the first runoff event. Concentrations fell substantially in subsequent runoff events and reached levels similar to those measured prior to fertilizer treatment by the second event. Surface runoff losses of N in the greenhouse study generally increased with increasing nitrogen application. The most important findings of this research are that P losses from irrigated turf may be larger than previously reported, and P mobility is strongly influenced by ammonium sulfate fertilizer application. Superphosphate fertilizer, applied alone, did not influence phosphorus leaching rates observed during the 2004 lysimeter experiments. When superphosphate was applied in conjunction with ammonium sulfate fertilizer, leachate phosphate concentrations were observed to increase. This may have been the result of temporary acidification of the soil solution produced by the application of ammonium sulfate fertilizer. The second portion of the project, dealing with estimation of phosphorus loads from multiple sources and land uses in mountain watersheds has just begun with a review of information needs of watershed organizations and of available watershed modeling tools. Modeling may enable watershed organizations to efficiently determine the scope of a water quality problem, estimate how conditions might change over time, and evaluate alternative management options.

Impacts
Results of the first and second years of this project will impact the development and refinement of Best Management Practices for nutrient management for turf-type grasses in Colorado landscapes and other applications. The results will also be important for estimating the contributions of turf to nutrient budgets for source-water protection plans. Results of the second and third years will be used by watershed organizations, such as the Big Thompson Watershed Forum, and by water suppliers, such as Front Range cities, to develop and improve their water quality assessments and management plans for protecting water sources from eutrophication. The work will provide a low-cost, first-cut approach to identifying and prioritizing phosphorous sources for further study and application of Best Management Practices. This will be especially critical in the early stages of any future development of phosphorus TMDLs.

Publications

  • Morgan, Jennifer, E. 2005. Nitrogen and Phosphorus Transport from Managed Turf Grass. M.S. Thesis. Colorado State University, Department of Civil Engineering, Fort Collins, CO.


Progress 07/01/04 to 12/31/04

Outputs
The first year of this new project continues the work of project COL 00705 which involved three years of both field drainage lysimeter and greenhouse experiments to evaluate losses of both nitrogen and phosphorus under alternative water and fertilizer application rates. Experimental results for nitrogen generally confirmed those reported in the literature. Nitrogen leaching generally increased with both increasing water application and increasing nitrogen fertilizer application rates. Surface runoff losses of N in the greenhouse study generally increased with increasing nitrogen application. In the case of phosphorus, previous research has suggested that phosphorus export through both subsurface leaching and surface runoff is negligible. However results from COL 00705 differed from those findings, with significant P losses observed in both surface runoff and subsurface drainage. Therefore, under this new project in 2004, work focused on P, though N measurements continued. No additional P was applied to the greenhouse boxes in 2004. The boxes were irrigated weekly with a hand sprinkler until runoff occurred. Soil samples were collected prior to the initial irrigation and at the end of the study and analyzed for both Olsen P and total P. Leachate and surface runoff were collected separately and analyzed for nitrate/nitrite and dissolved phosphorus. During the last few weeks of the study, ammonium sulfate fertilizer was added to improve the health of the turf. Upon ammonium sulfate application, N levels in runoff increased from an average 0.7 mg/L to 1.3 mg/L. P levels increased much more significantly, rising from an average 1.1 mg/L to 5.8 mg/L. The 2004 lysimeter experiments were designed to more closely examine the effect of ammonium sulfate on phosphorus mobility in the subsurface. The 24 12-inch diameter drainage lysimeters (described in reports for COL 00705) were reseeded, and soil cores were taken at the start of the 2004 study. Soil samples were analyzed for water extractable phosphorus to estimate the P fraction that was readily available for leaching. Water was applied by sprinklers at application rates ranging from 0.5 to 2.5 times ET. After each irrigation, leachate volume was recorded and samples analyzed for nitrate/nitrite and dissolved P. Fertilizer was applied twice: first ammonium sulfate only at 3 rates, 0.75 lb per 100 ft2, 0.25 lb per 100 ft2, and zero. In the second application, fertilizer was not applied to 8 of the lysimeters. 8 lysimeters received triple super phosphate at 2 lb/100 ft2, and 8 lysimeters received the same P application plus 0.75 lb per 100 ft2 ammonium sulfate. As in previous years higher irrigation rates resulted in increased leaching. Ammonium sulfate application increased not only the nitrogen concentration in the soil leachate, but increased the phosphorus concentration as well. Mobilized by ammonium sulfate, leachate phosphorus concentrations were observed in excess of 12 mg/L, much higher than typically reported in the literature. To sum up, P losses from irrigated turf may be larger than previously reported, and P mobility is strongly influenced by ammonium sulfate fertilizer application.

Impacts
Results of the first year of the project will impact the development and refinement of Best Management Practices for nutrient management for turf-type grasses in Colorado landscapes and other applications. The results will also be important for estimating the contributions of turf to nutrient budgets for source-water protection plans. Results of the second and third years will be used by watershed organizations, such as the Big Thompson Watershed Forum, and by water suppliers, such as Front Range cities, to develop and improve their water quality assessments and management plans for protecting water sources from eutrophication. The work will provide a low-cost, first-cut approach to identifying and prioritizing phosphorous sources for further study and application of Best Management Practices. This will be especially critical in the early stages of any future development of phosphorus TMDLs.

Publications

  • No publications reported this period