The Lamoille Watershed as a Model for Storm Water Nutrient Runoff from Dairy Farms in Vermont

The following was my environmental studies senior thesis.


            Storm water runoff from agricultural areas carries excess nutrients with it, which leads to dead zones in nearby lakes. Lake Champlain is the sixth largest lake in the United States, and is facing issues related to excess nutrients from farms. Research related to farms has focused on runoff related to manure spread on fields. One area not previously addressed is manure management on farms. There are two main types of manure pit styles on dairy farms; enclosed and, unenclosed. Three farms of each of these types were analyzed for nutrient runoff during storm events. Four key factors linked to nutrient runoff were analyzed through field sampling; phosphorus, dissolved oxygen, nitrate, and ammonium. The study took place in the Lamoille River Watershed in Vermont which feeds into Lake Champlain. Baseline data was collected before storms and samples were also collected after storms. The goal of this study was to identify a manure pit management style that minimizes nutrient runoff. While the data indicated that a manure pit that lacked a confining wall impacted phosphorus, ammonium, and nitrate, there was no significant impact on dissolved oxygen from such pits. However, manure pits had relativity low impact on the streams overall; other more important factors may be soil erosion, manure spreading, and septic tanks.


Dead zones are areas of reduced dissolved oxygen in bodies of water. They can occur naturally, or due to anthropogenic causes. Excess nutrient runoff causes algae to grow excessively, then sink and decompose. These nutrients include nitrate, nitrite, phosphorus, and ammonium. When algae decomposes, the decomposing bacteria consume large amounts of oxygen, leaving the water depleted of oxygen and killing marine life (National Oceanic and Atmospheric Administration 2013, Dybas 2005). In Lake Champlain, dead zones have begun to develop in the Missisquoi Bay, where the Missisquoi River flows into the lake, leading to regular summer fish kills (Vermont Public Radio 2012). The majority, 97%, of the nutrients causing these dead zones are from non-point source pollution (Federation of Vermont Lakes and Ponds 2013). Runoff from farms contributes up to 44% of this non-point source pollution (Federation of Vermont Lakes and Ponds 2013).

Lake Champlain is of vital importance to both Vermont and New York (New York State Department of Environmental Conservation 2013). Each year, the areas surrounding the lake contribute $4 billion dollars to the local economy (Lake Champlain Basin Program 2004). The lake provides drinking water for 188,000 people (New York State Department of Environmental Conservation 2013). Water quality issues in any of the tributaries flowing into Lake Champlain leads to water quality issues in the lake itself, including the development of dead zones. Lake Champlain’s nutrients come from farms, storm water, and wastewater treatment plants.

The Lamoille River watershed feeds into Lake Champlain (Galois 2002). The watershed contains 611 miles of rivers and streams. Agriculture covers 13% of the landscape. An estimated 100 miles of stream are at high or moderate risk for runoff pollution (Lamoille County Conservation District 2003). The largest waterway of the watershed, the Lamoille River, empties into Mallet’s Bay in Lake Champlain (Lamoille County Conservation District 2003). The majority of the Lamoille River does not meet Clean Water Act standards due to drawdown from hydroelectric plants, agricultural and storm water runoff, and atmospheric deposition of metals (Lamoille County Conservation District 2003). Since 1996, $7.7 million from public and private funds has gone to reducing nonpoint source pollution related to farms in the watershed. However, none of this investment went towards addressing the manure pit issues (Vermont Agency of Natural Resources 2009). The manure pits on dairy farms are a potentially significant nutrient loading issue in both the Lamoille Watershed and Lake Champlain.

Dairy farms were chosen as the focus of this research as they are the most common type of farm located near Lake Champlain (Vermont Agency of Natural Resources February 2009). Dairy farms have become an important part of Vermont as they attract tourists and are iconic landscape features (Lamoille County Conservation District 2003). It was not until the 20th century that dairy became the number 1 sector of agriculture in Vermont (Parsons 2010). In 1999, agriculture brought $252 million into the state (Saint Michael’s College 2014). Due to a recent decrease in milk prices, farmers make maximum use of their land, often farming up to the water’s edge (Vermont Agency of Natural Resources 2009). This increases the likelihood of nutrient runoff into rivers. The four critical chemical parameters associated with nutrient runoff from farms and the resulting degradation in water quality are phosphorus, dissolved oxygen, ammonium, and nitrate.

Excess phosphate can lead to the excess growth of algae which can lead to development of dead zones. High levels of phosphate can be caused by fertilizer runoff and farm waste runoff. Phosphorus can also be naturally added into streams by the weathering of bedrock (Kaa 2006). However, phosphorus naturally occurs in low concentrations, so small changes in concentration dramatically affect water quality. This is why loading is a big concern (Environmental Protection Agency 2013). Concentrations of phosphate in streams are typically 20 to 200 µg/L in the United States (Environmental Protection Agency 2013). Approximately 85% of the streams near farms in the United States have concentrations of total phosphorus of 100 µg/L or higher (Environmental Protection Agency 2010).

Nitrogen is essential to the growth of plants. The main sources of nitrogen in streams are chemical fertilizers and animal waste, mostly in the form of nitrate (Vitousek 1997). Excess nitrate is not toxic but, like phosphate, can lead to the excessive growth of algae which can cause dead zones (Environmental Protection Agency 2010). Levels of nitrate naturally occurring nationwide in streams range from 0.12 to 2.2 mg/L (Environmental Protection Agency 2010). A recent EPA study reported that 60% of stream sites sampled near farms had nitrate concentrations of 2 mg/L or greater (Environmental Protection Agency 2010).

Ammonium is another compound containing nitrogen, and is linked to fecal matter from agricultural pollution (California Environmental Protection Agency 2014). Ammonium can be converted to both nitrite and nitrate by bacteria. Ammonium is highly toxic to fish even at low levels, particularly to cold water species, like salmonids found in Vermont and Lake Champlain (Environmental Protection Agency 2011).

Vermont has tried to deal with the problems of dairy farms and nutrient runoff in a variety of ways. It has invested $100 million in strategies to reduce the pollution problem across the state.  However, these efforts focused primarily on fixing point source pollution sources. These sources included waste water treatment plants and factories. All point sources totaled only 3% of the phosphorus ending up in Lake Champlain. The majority of the nutrient pollution ending up in the lake comes from manure and urban storm water (Bright Blue EcoMedia 2010).

Most research focuses on runoff from manure once it is spread onto fields (Kleinman 2002, Kleinman 2003). There has been little research done on how different manure pit types affect runoff. I evaluated how different dairy manure pit styles affect nutrient runoff from farms associated with storm events. Storm events are more toxic to the aquatic environment because they produce a high surge of pollutants. I measured concentrations of nitrate, ammonium, total phosphorus, and dissolved oxygen in streams before and after storm events alongside dairy farms using different styles of manure pits.


The study began on May 29th, 2014. Streams were visited both upstream and downstream of the manure pit of interest. Phosphorus, dissolved oxygen, nitrate, and ammonium samples where all measured. The process was repeated after the accuweather data collection center nearest each manure pit showed that it had rained. Samples were taken manually 24 hours before and after it rained.

Study area:

The definition of “farm” used in this study is land with more than 30 livestock animals. I identified farms on tributaries of the Lamoille River with similar flow rates using base map satellite imagery on ArcGIS. I classified them by two manure pit styles; unconfined and confined. Those that are confined have a wall and liner surrounding the pit. Those that are confined simply are pits dug into the ground (see appendix). I selected three farms of each type by assigning a number to each farm and then using a random number generator to select them.  Other factors such a nearby land usage, and nearby hazardous waste sites were taken into account when determining the suitability of sites. Sites within 500 meters of these factors were removed from the list of suitable sites.

Using satellite imagery to choose sites does not guarantee entirely appropriate selections. I used ground truthing to determine the suitability of the site chosen. Some selected sites were not suitable. I them randomly selected another site of the same manure pit style.

I took water samples from along roadways, upstream and downstream from the farms from the center of the stream. The state owns 3.8 meters (12.75 feet) on either side of the road, therefore the sites were accessed using public lands (The Vermont Institute for Government 1994). The samples taken before the farm were taken downstream of the road, while the sample taken after the farm were taken upstream of the access road (figure 1). This insured that any pollution from the road was included in both samples. As a result, it canceled out as when the samples were analyzed they were analyzed using the difference between the upstream and downstream locations. Taking the sample upstream of the second road minimized runoff from that road entering into the samples.

When testing took place:

Each farm is unique in this study, so samples were only compared to the other samples from the same farm. A baseline was first established for each farm. Therefore, before and after storm event samples were collected. For this purpose, a storm was defined as an event when >0.635 cm (0.25 in) of rain falls within a 12-hour period. Weather data from Morrisville Stowe State Airport National Weather Service station, located in the center of the study area, was used in order to determine the exact amount of rainfall. In addition, samples were collected first from downstream and then from upstream.

Testing parameters:

Dissolved oxygen was measured in situ using a YSI Model 58 Dissolved Oxygen Meter. At each site three readings were taken and averaged. For this and subsequent field readings, each was conducted three times and then averaged. Temperature affects the solubility of dissolved oxygen. Therefore, I used the percentage of the maximum possible dissolved oxygen or percent saturation. This was calculated by the meter using temperature.

I collected samples to be analyzed for phosphate in 60 mL polyethylene bottles then preserved them by cooling to 4° C and acidified them with concentrated H2SO4 to lower the pH to less than 2. I analyzed the samples within 28 days of collection using a spectrophotometer at a wavelength of 650 nm in the lab at Alfred University. Environmental Protection Agency approved method number 365.2 was used to measure total phosphorus. Standards of 0 mg, 0.25 mg, 0.5 mg, 1 mg, and 10 mg were used. These standards were made from KH2PO4. I used field blanks, equipment blanks, and analytical replicates for phosphorus analysis. I used one equipment blank per day in the lab, and one field blank per day in the field. Five percent of the total samples analyzed per day were analytical replicates. Nitrates and ammonium was measured in situ using a Vernier Lab Quest meter with nitrate ion-selective electrode. Standards of 1 mg/L concentration and 100 mg/L concentration from Vernier were used to calibrate the probes.

Statistical analysis:

Pairing of upstream and downstream, as well as before and after storm data creates a situation where each variable is paired to two other variables. This was to be accounted for by using the double difference as the quantitative variable. This was created by taking the after storm values and subtracting upstream from downstream.  The same method was used to calculate the before stream value. Then the before storm value was subtracted from the after storm value. The resulting variable was the increase of contaminants added to a stream due to runoff. This method was repeated for all four factors.

Because the goal was to look at multiple pollutants at the same time I used a two-way MANOVA (multivariate analysis of variance) to analyze the final data.  The two categorical variables are manure pit type and a manure pit type and farm interaction term. Each individual measurement on each farm was sufficiently independent within each farm for this testing method. In addition, the interaction term accounted for variations between farms of the same manure pit type.


Rain events ranged from 0.55 cm to 7.44 cm. Both farms and rainstorm events increased nutrient levels, and decreased the level of dissolved oxygen (Table 1). The differences between the upstream values subtracted from the downstream values were statistically significant both before and after storm events.

The differences between before and after storm values, and between upstream and downstream values of all of the analytes were statistically significant (Table 2). This was expected since the difference between the upstream values subtracted from the downstream values was also significant.

Since visible erosion was a possible confounding factor, the sites with and without visible erosion were compared (Table 3). There was no measurable effect of visible erosion on the parameters tested.

Rainfall likely had an effect on NH4 levels but the type of pit did not. Due to strong outliers one site was removed from analysis. These outliers were likely from unexpected sources of runoff. Neither rainfall nor pit type had an effect on dissolved oxygen. Rainfall likely had an effect on NO3 levels while pit type did not. Rainfall had an effect on phosphorus, but the type of pit did not have an effect. There was no statistically significant effect on all of the values of the types of manure pits (Table 4).

Table 1. Paired t-test comparison of downstream and upstream values of nutrients, as well as before and after storms values, Lamoille watershed, VT, May- August 2014.

  Mean change in downstream values before and after storms (mg/L) Mean change in upstream values before and after storms (mg/L) Downstream vs. upstream value before and after storms (P-values)
Ammonium   0.20   0.70 0.014
Nitrate   0.50   1.10 0.009
Phosphorus   0.23   0.53 0.004
Dissolved oxygen – 0.33 – 0.77 0.007

Table 2. Means of nutrient values found upstream and downstream of manure pits. Lamoille watershed, VT, May- August 2014.

  Upstream mean (mg/L) Downstream mean (mg/L) P-value
Ammonium 0.62 0.86 0.006
Nitrate 2.88 3.40 0.010
Phosphorus 0.23 0.33 0.005
Dissolved oxygen 9.19 8.95 0.002

Table 3. Nutrient levels found in streams with and without visible erosion, p-values from t-tests, Lamoille watershed, VT, May- August 2014.

  Mean for sites with visible erosion (mg/L) Mean for sites without visible erosion (mg/L) P-values between sites with and without visible erosion
Ammonium 0.73 0.79 0.523
Nitrate 2.89 3.14 0.790
Phosphorus 0.22 0.28 0.907
Dissolved oxygen 9.02 9.07 0.632

Table 4. Results from a generalized linear model comparing downstream values subtracted from upstream values before and after storm events. Lamoille watershed, VT, May- August 2014.

  Mean for confined manure pits (mg/L) Mean for unconfined manure pits P-value between types. P-value for amount of rainfall
Ammonium 0.81 0.79 0.953 0.016
Nitrate 2.18 2.22 0.763 0.011
Phosphorus 0.34 0.29 0.725 0.006
Dissolved oxygen 8.95 8.63 0.863 0.003


The goal of this study was to identify a manure pit management style that minimizes nutrient runoff. While the data indicated that not having a confining wall impacted phosphorus, ammonium, and nitrate, there was no significant change with dissolved oxygen. The lack of change in dissolved oxygen indicates that stream and rivers were not the areas affected by the high nutrient levels. Likely, lakes and ponds were most impacted as these tend to have high primary productivity than rivers and streams.

Manure pits had relativity low impact on the streams; other more important factors may be soil erosion, manure spreading, and septic tanks. Nonpoint source pollution continues to be an issue within the Champlain watershed. Even though the number of farms has decreased over the years, the high nutrient levels have not decreased. More studies are needed in order to identify which management practice will be successful. Adaptive resource management is ideal for this situation as the impacts of nutrient loading are large within the watershed. Adaptive resource management is when management decisions are made without having all the data necessary and then changed as more data becomes available.

From these results it is clear that current management practices are not effective in the state of Vermont. While storm water nutrient runoff is a significant issue within the state the impacts that using confining walls would have are questionable. The benefit they provide does not warrant the financial investment. Other methods, such as increasing the quality of riparian habitats, and reducing erosion by increasing soil stability, appear to be more cost effective.


1998 Clean Water Act Status. Environmental Protection Agency, Office of Water, Washington DC, USA.

Bright Blue EcoMedia. November 29, 2010. Bloom. <;

California Environmental Protection Agency. 2014. Ammonia. <;

Doolan, B. The Geology of Vermont. 1996. Department of Environmental Conservation, Division of Geology and Mineral Resources, Burlington, Vermont, USA.

Dybas C. L. 2005. Dead Zones Spreading in World Oceans. BioScience, 55(7), 552-557.

Environmental Protection Agency. July 8, 2013. Water Quality Standards Handbook – Chapter 3: Water Quality Criteria (40 CFR 131.11). <;

Environmental Protection Agency. June 27, 2012. Common Manure Handling Systems. <;

Environmental Protection Agency. March 6, 2012. Dissolved Oxygen and Biochemical Oxygen Demand. <;

Environmental Protection Agency. June 2, 2011. What’s the Problem?. <;

Environmental Protection Agency. February 17, 2010. Nitrogen and Phosphorus in Streams in Agricultural Watersheds. <;

Federation of Vermont Lakes and Ponds. 2013. Annual meeting minutes.

Galois, P., Léveillé, M., Bouthillier, L., Daigle, C., and Parren S. 2002. Movement Patterns, Activity, and Home Range of the Eastern Spiny Softshell Turtle (Apalone spinifera) in Northern Lake Champlain, Québec, Vermont. Journal of Herpetology, 36(3), 402-411.

Great Lakes Fishery Commission. April 16, 2004. TOP:009.2 Procedures for Measurement of Ammonia in Stream Water. <;

Kaa, J., and Kopacek, J. 2006. Impact of soil sorption characteristics and bedrock composition on phosphorus concentrations in two bohemian forest lakes. Water, Air, & Soil Pollution, 173(1-4), 243-259.

Kleinman, P. J. A., and Sharpley, A. N. 2003. Effect of broadcast manure on runoff phosphorus concentrations over successive rainfall events. Journal of Environmental Quality, 32(3), 1072-81.

Kleinman, P. J., Sharpley, A. N., Moyer, B. G., and Elwinger, G. F. 2002. Effect of mineral and manure phosphorus sources on runoff phosphorus. Journal of Environmental Quality, 31(6), 2026-33.

Lake Champlain Basin Program. July 2004. Landforms of the Basin. <>

Lamoille County Conservation District. 2003. Lamoille River Watershed Agriculture and River Resource Guide. <;

National Oceanic and Atmospheric Administration. “Dead zone” is a more common term for hypoxia, which refers to a reduced level of oxygen in the water. January 11, 2013. <;

Northeast Organic Farming Association of Vermont. 2012. Vermont Certified Organic Farms. <;

New York State, Department of Environmental Conservation. 2013. Lake Champlain Watershed Water Quality Management Planning. <>

Parsons, B. 2010. Vermont’s Dairy Sector: Is There a Sustainable Future for the 800 lb. Gorilla? Food System Research Collaborative. 1(4): 1-12.

Saint Michael’s College. 2014. Agriculture in Vermont. <;

The Vermont Institute for Government. 1994. The public right of way and you. <;

Vermont Agency of Natural Resources. February 2009. LAMOILLE RIVER BASIN Water Quality Management Plan- DRAFT. <;

Vermont Agency of Natural Resources Watershed Management Division. March 2012. Cyanobacteria in Vermont–Basic Information. <;

Vermont Department of Environmental Conservation.  Vermont Watershed Grants 2007-2012. <;

Vermont Public Radio. August 27, 2012. Lake Champlain Water Quality Gets Worse As Summer Winds Down. <>

Vitousek, P. M., Aber, D. J., Howarth, R. W., Likens, G. E., Matson, P. A., Schindler, D. W., Schlesinger W. H., and Tilman D. G. 1997. Technical Report: Human Alteration of the Global Nitrogen Cycle: Sources and Consequences. Ecological Applications, 7(3),737-750.

Zhi-Jun L., D. Weller, D. Correll, and T. Jordan. 2000. Effects of land cover and geology on stream chemistry in watersheds of Chesapeake Bay. Journal of the American Water Resources Association 36(6): 1349-1363.







Table 5. GIS analysis of the suitability of the sampling sites. Lamoille watershed, VT, 2014.

Site Number Distance from manure pit to stream (meters) Impervious surfaces Stream Slope insures runoff flow into stream? Scale at which other possible pollution appears on map (in feet)
 9 107 Roads and homes Browns River Yes 1:10,000
21 130 Roads and airports Kentfield Brook Yes 1: 6,000
63  45 Roads and homes Seymour River Yes 1: 8,000
86 242 Roads and homes Stones Brook Yes 1: 8,000
91 115 Roads and homes Wilkins Brook Yes 1:12,500
97 275 Roads and fields Polly Brook Yes 1:10,000


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s