Methods and Detailed Results: Tap Water for 500,000 Minnesotans Contaminated With Elevated Levels of Nitrate

Methods and Detailed Results: Tap Water for 500,000 Minnesotans Contaminated With Elevated Levels of Nitrate

By Sarah Porter, Senior Geospatial Analyst, and Anne Weir Schechinger, Senior Analyst of Economics

Nitrate is found in groundwater used as drinking water throughout the state of Minnesota. EWG’s report, “Tap Water for 500,000 Minnesotans Contaminated With Elevated Levels of Nitrate,”[1] analyzed nitrate levels in groundwater used by public water systems and private household wells as sources of drinking water. This summary provides descriptions of how that analysis was performed, as well as more-detailed results.

Public Water System Analysis

Public water systems are defined as those supplying drinking water for at least 15 people a year all year, or an average of at least 25 people for 60 days a year.[2] These systems can be publicly or privately owned and are regulated by the U.S. Environmental Protection Agency. However, the EPA commonly delegates regulatory authority to state agencies. In Minnesota, the Minnesota Department of Health, or MDH, directs the state’s Drinking Water Protection program.[3]

Public water systems include community water systems such as cities and towns, or what most people consider to be “municipal systems,” that serve their customers year-round and can serve up to millions of people. Non-community systems are also public water systems, but these systems include places like schools, gas stations, churches or campgrounds that tend to serve much smaller populations for shorter amounts of time.

Public water systems are required by the EPA to test their finished drinking water for nitrate. Testing frequency depends on how many customers each water system serves, and whether the system is a community or non-community system. In Minnesota, MDH monitors public water systems’ nitrate testing schedule and collects the results for the EPA. These tests can be found through the EPA’s Safe Drinking Water Information System, or SDWIS.[4]

EWG analyzed finished-water nitrate test results from all public water systems in Minnesota between 2009 and 2018. All data came from public records requests fulfilled by the Minnesota Department of Health. Data for 2009 through 2017 are included in EWG’s Tap Water Database.[5]

In analyzing the public water system data, we looked at every nitrate test that each system conducted between 2009 and 2018, specifically for SDWIS contamination code number 1040. The analysis looked only at tests for systems that are currently active according to the SDWIS database and that use groundwater as their main source of drinking water.

There are 6,626 active groundwater systems in the state. Of those, 6,566 tested for nitrate at least once between 2009 and 2018. Eighty-seven percent of the systems are non-community systems, but they serve just 16 percent of people (a little more than 575,000 people). Only 13 percent of the systems were community systems, but they serve 84 percent of people (more than 3 million).

Public Water System Results

The frequency of testing for nitrate varies by system type, but most systems tested every year. Of the 6,566 systems that tested for nitrate, 82 percent tested in all 10 years between 2009 and 2018 (Figure 1). However, community water systems tested more often than non-community systems – 87 percent of community systems tested every year, whereas 81 percent of non-community systems tested every year. Some systems also tested more often than once a year – 29 percent, or 1,911 systems, tested more than 10 times between 2009 and 2018.

Figure 1. Eighty-two Percent of Public Water Systems Tested Every Year Between 2009 and 2018.

EWG also looked at which public water systems had at least one test at or above 3, 5 or 10 milligrams per liter, or mg/L, between 2009 and 2018. We chose the 3 mg/L threshold because MDH considers that level of contamination an indication that groundwater is contaminated by human-generated sources and that contamination may increase.[6] Studies have found increased risk of different cancers with long-term ingestion of water with nitrate around 5 mg/L,[7] as well as birth defects in babies whose mothers consumed water with 5 mg/L of nitrate.[8] We also chose 10 mg/L because it is the legal limit. If public water systems are at or above this level, they are legally required to act to reduce levels.[9]

Of the 6,566 active groundwater systems that tested for nitrate, 11 percent, or 727 systems, had at least one test at or above 3 mg/L. Table 1 shows the number of systems with at least one test at or above 3, 5 or 10 mg/L, along with the population served. There is also an interactive map showing the spatial distribution of these systems here (hyperlink to map).

Table 1. Eleven Percent of Systems Tested Between 2009-2018 Had at Least One Test at or Above 3 mg/L for Nitrate.

 

With at Least 1 test >= 3 mg/L

With at Least 1 test >= 5 mg/L

With at Least 1 test >= 10 mg/L

System Type

Systems

Population

Systems

Population

Systems

Population

Community

95

405,386

55

258,985

20

146,202

Non-Community

632

67,597

358

38,251

104

8,448

Total

727

472,983

413

297,236

124

154,650

Vulnerability

EWG also analyzed which public water systems were near groundwater that is highly vulnerable to contamination. We used the Minnesota Water Table Aquifer Vulnerability shapefile from the Minnesota Geospatial Commons to determine which areas in the state were vulnerable to contamination, and we looked only at areas that were considered to have “high vulnerability.”[10] We then looked at how many public water systems that tested at or above 3, 5 or 10 mg/L for nitrate were also within one mile of one of these highly vulnerable groundwater areas.

Most public water systems that tested at or above 3, 5 or 10 mg/L were within one mile of a highly vulnerable groundwater area. Eighty-nine percent of systems with at least one test at or above 3 mg/L were near a highly vulnerable groundwater area, whereas 92 percent of systems with at least one test at or above 5 mg/L and 91 percent of systems with at least one test at or above 10 mg/L were near a highly vulnerable groundwater area. Table 2 contains the number of systems and their populations with at least one test at or above 3, 5 or 10 mg/L that were within one mile of a highly vulnerable groundwater area.

Table 2. Nitrate Levels in Public Water Systems Within One Mile of a Highly Vulnerable Groundwater Area.

 

With at Least 1 Test >= 3 mg/L

With at Least 1 Test >= 5 mg/L

With at Least 1 Test >= 10 mg/L

System Type

Systems

Population

Systems

Population

Systems

Population

Community

87

383,225

51

257,063

20

146,202

Non-Community

559

60,193

327

34,888

93

7,833

Total

646

443,418

378

291,951

113

154,035

Private Well Analysis

The Minnesota Department of Health calculates the number of Minnesotans who rely on private wells by subtracting the number of people served by a community water system from the total state population.[11] According to community water system population information from the EPA, almost 4.4 million people in Minnesota are served by a community water system, which is 78 percent of the total state population. Just under 1.3 million people, or 22 percent of the state population, get their drinking water from private wells.

Nitrate results were collected from four different state programs in Minnesota focused on nitrate testing of private wells. Each program, described below, has a different geographic focus, testing frequency and scale at which the data were provided. All test results collected between 2009 and 2018 were analyzed and aggregated to the township level using the Minnesota City, Township, and Unorganized Territory layer provided on the Minnesota Geospatial Commons.[12] Within each township, the total number of tests collected from each program and the number of tests at or above 3, 5 and 10 mg/L nitrate was recorded. An interactive map showing the spatial distribution of test results by township can be found here (hyperlink to map).

Township Testing Program

The Minnesota Department of Agriculture, or MDA, Township Testing Program began in 2013 with the goal of assessing nitrate-nitrogen concentrations in private wells at the township scale.[13] The program is intended to support the 2015 revised Nitrogen Fertilizer Management Plan, or NFMP, and focuses on townships across the state that have been identified as vulnerable to groundwater contamination and have significant row crop production. As of February 2019, 306 vulnerable townships from 42 counties have participated in the program.

All available township-level results were assembled from online PDFs. Initial test results were used for all townships, except two in Washington County and three in Morrison County, for which only final test results were provided. Wells with construction issues or nearby potential point sources of nitrogen were removed from final test results, which are intended to include only wells potentially impacted by applied commercial agricultural fertilizer. Townships that had completed final testing reported the number of wells with test results at or above a 3, 5 and 10 mg/L nitrate level. This was about two-thirds of the total number of wells tested through the program. The other one-third of wells were located in townships that had not yet completed final testing and reported only the number of wells at or above the legal limit of 10 mg/L. For those townships, it was assumed that tests at or above 10 mg/L were also above 3 and 5 mg/L. In total, results for 30,656 unique wells were assembled. 

As part of the Township Testing Program, a follow-up nitrate sample was offered to all homeowners who had a detectable level of nitrate in their initial test. Location data and test results were provided by MDA for those households that participated in follow-up testing. Follow-up test results were assigned to the township in which the well was located, and the number of tests at or above 3, 5 and 10 mg/L within each township was recorded. In total, 4,665 nitrate test results were provided for 4,282 wells as part of the follow-up testing.

Southeast Voluntary Nitrate Monitoring Network

The MDA Southeast Voluntary Nitrate Monitoring Network initially began in 2006 as a coordination among nine southeast Minnesota counties.[14] The goal of the program was to monitor long-term trends of nitrate concentrations in private drinking water wells in southeast Minnesota, as karst geology makes this region vulnerable to groundwater contamination. The first sampling took place in 2008. In 2014, MDA coordinated with county water planners and the Southeast Minnesota Water Resources Board to ensure nitrate sampling continued on an annual basis and that the original well network was kept intact. The program attempts to sample the same wells – around 650 unique wells – on an annual basis. Wells are tested once annually, in August.

MDA provided EWG with Southeast test results in the form of 2-mile-diameter buffers around the actual location of the one or more sampling wells located within the buffer. In total, 4,392 nitrate test results were provided for 651 wells since 2009. Data was aggregated to the township scale by assigning each buffer to the township in which its centroid was located, and the number of tests at or above 3, 5 and 10 mg/L within each township was recorded.

Central Sands Private Well Network

The MDA Central Sands Private Well Network is a voluntary testing program focused on 14 counties in the Central Sands portion of Minnesota.[15] The program began in 2011 in response to concerns about high nitrate levels in private drinking water wells. The goal of the program is to determine nitrate trends through an effort to sample the same wells, around 550 unique wells, at least once annually. Wells are tested once a year, in March.

MDA provided EWG with Central Sands test results in the form of 1.5-mile diameter buffers, with a single sampling well located somewhere inside each buffer. Well locations represent the long-term sampling network of the Central Sands program. In total, 3,463 nitrate test results were provided for 551 wells since 2011. Data was aggregated to the township scale by assigning each buffer to the township in which its centroid was located, and the number of tests at or above 3, 5 and 10 mg/L within each township was recorded.

New Domestic Well Nitrate Testing

The Minnesota Department of Health maintains a database of nitrate tests collected during construction of private drinking water wells across the state.[16] The agency has been collecting data since the early 1990s, although only samples collected between 2009 and 2018 were used for this analysis. MDH provided EWG with location data and test results for each of the 45,598 wells sampled during the past 10 years. Test results were assigned to the township in which the well was located, and the number of tests at or above 3, 5 and 10 mg/L within each township was recorded.

Private Well Results

The large majority (91 percent) of nitrate tests of private wells in Minnesota were collected as part of the MDH New Domestic Well or the MDA Township Testing Program (Figure 2). Both programs primarily collected only one sample from each well. Exceptions to this are wells that participated in the Township Testing Follow Up Program, for which one or more repeated tests were collected. The frequency of testing for the Central Sands and Southeast Monitoring Programs, which are focused on establishing a long-term monitoring network, is once per year. Private wells in the Central Sands monitoring network have an average test frequency of 6.3 tests per well since 2013, whereas private wells in the Southeast monitoring network have an average test frequency of 6.7 tests per well since 2009.

Figure 2. Number of Private Well Nitrate Tests Collected by Program

The number of tests collected, and the number and percentage of tests at or above 3, 5 or 10 mg/L from each domestic well program, are listed in Table 3. Not surprisingly, the Township Testing Follow Up program, which tests only wells that had already been found to have elevated nitrate, reported the highest percentage of contaminated wells. The Southeast monitoring program reported the second-highest percentage of contaminated wells, followed by initial results from the Township Testing program and the Central Sands program. The MDH New Domestic Well program had the lowest percentage of tests with elevated nitrate, with only four percent of tests at or above 3 mg/L and one percent of tests at or above 10 mg/L. The MDH New Domestic Well program is also the only dataset analyzed in which sampling was not specifically targeted to vulnerable groundwater areas of the state.

Table 3. Distribution of Nitrate Tests at or Above 3, 5 and 10 mg/L

Program

Number of Tests Collected

Number of Tests >= 3 mg/L

Number of Tests >= 5 mg/L

Number of Tests >= 10 mg/L

Township Testing Initial

30,656

5,494 (18%)

4,449 (15%)

2,773 (9%)

Township Testing Follow-Up

4,665

2,799 (60%)

2,197 (47%)

1,175 (25%)

Central Sands

3,463

373 (11%)

232 (7%)

115 (3%)

Southeast

4,392

1,397 (32%)

1,050 (24%)

423 (10%)

MDH New Domestic Well

45,598

1,843 (4%)

1,121 (2.5%)

450 (1%)

Total

88,774

11,906 (13%)

9,049 (10%)

4,936 (6%)

Private Well Household Analysis

Only unique wells with elevated nitrate reported from the Township Testing (initial results), Central Sands, Southeast and MDH new domestic well programs were used to estimate the number of households with elevated nitrate in their drinking water. Follow-up results from the township testing program were excluded, because these households are already captured in the initial test results. There may be some duplication of households between the MDH New Domestic Well program and the other three programs. However, the lack of exact well locations in those three programs made it impossible to identify any duplication. Table 4 lists the number and percentage of households from each program with at least one test at or above 3, 5 or 10 mg/L.

Table 4. Distribution of Households With at Least One Test at or Above 3, 5 or 10 mg/L Nitrate

Program

Number of Households Sampled

Number of Households With at Least 1 Test >= 3 mg/L

Number of Households With at Least 1 Test >= 5 mg/L

Number of Households With at Least 1 Test >= 10 mg/L

Township Testing Initial

30,656

5,494 (18%)

4,449 (15%)

2,773 (9%)

Central Sands

551

88 (16%)

62 (11%)

40 (7%)

Southeast

651

232 (36%)

193 (30%)

101 (16%)

MDH New Domestic Well

45,598

1,843 (4%)

1,121 (2.5%)

450 (1%)

Total

77,456

7,657 (10%)

5,825 (7.5%)

3,364 (4%)

Vulnerability

EWG also determined the number of households with elevated nitrate that were near groundwater that is highly vulnerable to contamination. We classified each township by the proportion of land area occupied by a “high vulnerability” classification. Townships with at least 25 percent of their land area occupied by a high vulnerability classification were considered vulnerable to groundwater contamination. 890 of 2,696 townships in Minnesota fell under this vulnerable classification. 

Sixty-four percent (56,627 of 88,774) of all nitrate tests of private wells in Minnesota in the past 10 years were collected from a vulnerable township. Eighty-eight percent (10,478) of tests at or above 3 mg/L were located in a vulnerable township, whereas 89 percent (8,060) of tests at or above 5 mg/L, and 90 percent (4,445) of tests at or above 10 mg/L, were located in a vulnerable township. Table 5 lists the number of tests with elevated nitrate that are located within a vulnerable township from each program.

Table 5. Most Private Well Tests With Elevated Nitrate Are in a Vulnerable Township

Program

Number of Tests Collected From a Vulnerable Township

Number of Tests >= 3 mg/L

Number of Tests >= 5 mg/L

Number of Tests >= 10 mg/L

Township Testing Initial

27,347

5,164

4,179

2,609

Township Testing Follow-Up

4,221

2,573

2,017

1,076

Central Sands

1,933

252

149

75

Southeast

3,304

1,280

974

406

MDH New Domestic Well

19,822

1,209

741

279

Total

56,627

10,478

8,060

4,445

The proportion of households with elevated nitrate that were located within a vulnerable township was also analyzed (Table 6). Township Testing Follow Up results were again excluded to prevent duplication of households. Of the 7,657 household wells with at least one test at or above 3 mg/L, 6,638 wells, or 87 percent, were located in a vulnerable township. Eighty-eight percent, or 5,132, of the wells testing at or above 5 mg/L, and 89 percent, or 3,005, of wells at or above 10 mg/L were located in a vulnerable township.

Table 6. Most Private Well Households With Elevated Nitrate Are in a Vulnerable Township

Program

Number of Households Tested in a Vulnerable Township

Number of Households With at Least 1 Test >= 3 mg/L

Number of Households With at Least 1 Test >= 5 mg/L

Number of Households With at Least 1 Test >= 10 mg/L

Township Testing Initial

27,347

5,164

4,179

2,609

Central Sands

313

58

38

23

Southeast

475

207

174

94

MDH New Domestic Well

19,822

1,209

741

279

Total

47,957

6,638

5,132

3,005

Demographic Analysis

To find out who is being impacted the most by nitrate contamination of groundwater in Minnesota, we looked at data from the U.S. Census for both public water systems and private wells. We found that most public systems and private wells that had elevated levels of nitrate are in rural areas, and that many are in areas with median household incomes below the state’s income.

Median Household Income

We used data from the 2017 American Community Survey to determine which public water systems and private wells were in areas with median household incomes below Minnesota’s median household income.[17]

Specifically, we found the 2013-2017 five-year median household income for every census block group in the state. Census block groups are the smallest census unit and provide the most-detailed information possible.

For the public water systems, we assigned the median household income to each public water system based on which census block group the systems were located in. For the private well township-level data, we assigned the median household income to each township based on which census block group contained the center of the township. In both cases, a public system or township was considered to have a median income below the state’s income if their income was less than $65,559, the 2013-2017 median household income for the state of Minnesota.

More than half of public water systems and private well townships are in census block groups with median household incomes below the state’s income. Table 7 gives the number of public water systems and private well townships where median household income is below the state average, as well as their percentages compared to all public water systems and all private well townships with at least one nitrate test at or above 3, 5 and 10 mg/L.

Table 7. Over Half of Public Water Systems and Private Well Townships With at Least One Nitrate Test at or Above 3 mg/L Had Median Household Incomes Below the State’s Average

 

With at Least 1 Test >= 3 mg/L

With at Least 1 Test >= 5 mg/L

With at Least 1 Test >= 10 mg/L

 

Count Below State Income

Percent Below State Income

Count Below State Income

Percent Below State Income

Count Below State Income

Percent Below State Income

Public Water Systems

392

54%

203

49%

63

51%

Townships With Private Wells

377

51%

299

48%

209

47%

Many of the public water systems and private well townships that are near vulnerable areas are also in census block groups with median income below the state’s income. Out of the 646 public water systems that had at least one test at or above 3 mg/L and were within one mile of highly vulnerable groundwater, 347 had a median income below the state’s income. And out of the 427 townships that had private wells with at least one test at or above 3 mg/L, and had 25 percent of their land area in the high vulnerability area, 232 had a median income below the state’s income.

Rural Versus Urban

To find out whether the public water systems and private well townships with elevated nitrate were in rural or urban areas, we used the 2010 Census Urban and Rural

Classification.[18]

Using specific population density numbers, the census delineates which areas in the country are urban. Places outside of these urban areas are delineated as rural.

If a public water system is located in one of these census-delineated urban areas, it was considered to be an urban system, and if not, it was labeled rural. For the townships that contain private well tests, if 50 percent of their area is within an urban area, they were classified as urban. If not, they were considered rural. 

Almost all public water systems and private well townships with at least one nitrate test at or above 3 mg/L are in rural areas. Table 8 provides the number of public water systems and private well townships that are in a rural area, as well as their percentages compared to all public water systems and all private well townships with at least one test at or above 3, 5 and 10 mg/L.

Table 8. Most Public Water Systems and Private Well townships With at Least 1 test at or Above 3 mg/L Are in a Rural Area.

 

With at Least 1 Test >= 3 mg/L

With at Least 1 Test >= 5 mg/L

With at Least 1 Test >= 10 mg/L

 

Rural Count

Rural Percent

Rural Count

Rural Percent

Rural Count

Rural Percent

Public Water Systems

597

82%

352

85%

108

86%

Townships With Private Wells

710

96%

602

98%

438

99%

Most of the public water systems and private well townships that are near vulnerable areas are also rural. Out of the 646 public water systems that had at least one test at or above 3 mg/L and were within one mile of highly vulnerable groundwater, 525 were in a rural area. And out of the 427 townships that had private wells with at least one test at or above 3 mg/L and had 25 percent of their land area in the high vulnerability area, 404 were rural. 


[1] link to wrapper