Grantee Research Project Results
Final Report: Community-level Management of Human Health Risks from Concentrated Animal Feeding Operations (CAFOs) with Defensive Natural Capital Investments
EPA Grant Number: R836942Title: Community-level Management of Human Health Risks from Concentrated Animal Feeding Operations (CAFOs) with Defensive Natural Capital Investments
Investigators: Hochard, Jacob P , Etheridge, Randall , Peralta, Ariane , Sims, Charles
Institution: East Carolina University , University of Tennessee
EPA Project Officer: Hahn, Intaek
Project Period: August 1, 2017 through July 31, 2019 (Extended to July 31, 2021)
Project Amount: $399,226
RFA: Integrating Human Health and Well-Being with Ecosystem Services (2016) RFA Text | Recipients Lists
Research Category: Human Health
Objective:
Concentrated Animal Feeding Operations (CAFOs) support local economies with jobs, tax revenue and community reinvestment but are associated with airborne ammonia, fecal particulate, hydrogen sulfide and phosphate emissions (Burkholder et al. 2007, Bunton et al. 2007) and increased nutrient levels, heavy metals and pharmaceuticals in surface water and groundwater (Burkholder et al. 2007, Celender 2008). North Carolina’s Duplin and Sampson counties are the nation’s leading swine producers with over 1,000 swine based CAFOs, a 4.4 million hog capacity and 1,600 waste lagoons. With more than 20% of individuals below the poverty line and nearly 40% employment in hog and pig farming, community livelihood is tied to an industry that likely exposes them to high levels of airborne and waterborne contaminant. Community-level management of pollution-related health risks requires (i) a local capacity to measure health effects and identify at-risk neighborhoods and (ii) an understanding of how community defensive investments into physical capital (e.g., expanding public water supplies, sewage services, storm water drains) relate, and potentially diminish the need for, community defensive investments into natural capital (e.g., restoring riparian buffers, installing vegetative buffers).
The health consequences of community exposure to CAFO-linked airborne and waterborne contaminants are poorly understood because the dispersion of contaminants, pathways to human exposure, and susceptibility of exposed individuals are difficult to trace (Cole et al. 2000). Further, the vulnerability of communities to upstream pollutants is mediated by local ecosystems (e.g., surrounding riparian buffers, forest cover, soil type) and access to public services (e.g. public water and sewage, storm water drains). Community-led mitigation of health risk must target at-risk neighborhoods by expanding public services or maintaining and reconstructing local sources of natural capital.
CAFO-linked contaminants and human health risks: Ammonia (NH3), hydrogen sulfide (H2S) and methane (CH4) airborne emissions are common byproducts of manure decomposition at swine CAFOs (Heederik et al. 2007). Schiffman et al. (2001) detected 331 fixed gases and volatile organic compounds at eastern North Carolina swine facilities. Airborne concentrations of these gasses, such as NH3, are increasing with closer proximity to CAFOs (mean concentrations of 29 ppb within one-half kilometer and 5ppb beyond 2km) exposing nearby residential neighborhoods and schools (Wilson and Serre 2007). Atmospheric reactions with these gases trigger gas-to-particulate conversion generating increased levels of fine particulate matter (PM2.5) while on-farm animal activity and manure transport generate increased levels of larger particulate matter (PM10) (National Research Council, 2003). The most significant concern of near-CAFO communities is odor (National Research Council, 2003), which can be detected in excess of 5 mi from the nearest swine CAFO (State Environ. Resource Center 2004).
Non-therapeutic use of antibiotics is applied to animal feed to sustain faster growth rates throughout the production process, which often results in administering steady doses spanning multiple years (Andersson and Hughes 2014). This increased antibiotic use in the United States is now considered a water contaminant of concern (Kummerer 2009, Michael et al. 2013, from Ibekwe et al. 2016). Animals excrete between 20-80% of antibiotics administered on farm, which results in a major source of unprocessed antibiotics in surrounding water bodies and soils (Thiele-Bruhn 2003, Mackie et al. 2006, Kummerer 2009, Taylor et al. 2011, Andersson and Hughes 2014). Specifically, such antibiotics may spread to surface and groundwater during manure spraying or from leaking waste lagoons (Sapkota et al. 2007). Surface and groundwater samples taken downstream from CAFOs report enterococci, fecal coliforms and Escherichia coli concentrations between 4 and 33 times higher than upstream samples (Sapkota et al. 2007), drug resistant genotypes of E. coli originating from swine houses (Ibekwe et al. 2016), and tetracycline residues were also detected in shallow downgradient groundwater sources (Mackie et al. 2006).
Adverse health outcomes among swine operation workers are well documented (Heederik 2007) while studies have also found similar health impacts in communities neighboring CAFOs. Increased mood disturbances linked to odor detection (Schiffman et al. 1995), elevated rates of headaches, runny nose, sore throat, excessive coughing, diarrhea, burning eyes (Wing and Wolf, 2000), increased blood pressure (Wing et al. 2013) and incidence of asthmatic symptoms (Mirabelli et al. 2006) have all been recorded in eastern North Carolina’s CAFO-neighboring communities. More broadly, early and in utero exposure to PM10, total suspended particulates (TSP) and contaminated water has been linked [beyond a mere correlative association] to increasedcommunities is odor (National Research Council, 2003), which can be detected in excess of 5 mi from the nearest swine CAFO (State Environ. Resource Center 2004).
Non-therapeutic use of antibiotics is applied to animal feed to sustain faster growth rates throughout the production process, which often results in administering steady doses spanning multiple years (Andersson and Hughes 2014). This increased antibiotic use in the United States is now considered a water contaminant of concern (Kummerer 2009, Michael et al. 2013, from Ibekwe et al. 2016). Animals excrete between 20-80% of antibiotics administered on farm, which results in a major source of unprocessed antibiotics in surrounding water bodies and soils (Thiele-Bruhn 2003, Mackie et al. 2006, Kummerer 2009, Taylor et al. 2011, Andersson and Hughes 2014). Specifically, such antibiotics may spread to surface and groundwater during manure spraying or from leaking waste lagoons (Sapkota et al. 2007). Surface and groundwater samples taken downstream from CAFOs report enterococci, fecal coliforms and Escherichia coli concentrations between 4 and 33 times higher than upstream samples (Sapkota et al. 2007), drug resistant genotypes of E. coli originating from swine houses (Ibekwe et al. 2016), and tetracycline residues were also detected in shallow downgradient groundwater sources (Mackie et al. 2006).
Adverse health outcomes among swine operation workers are well documented (Heederik 2007) while studies have also found similar health impacts in communities neighboring CAFOs. Increased mood disturbances linked to odor detection (Schiffman et al. 1995), elevated rates of headaches, runny nose, sore throat, excessive coughing, diarrhea, burning eyes (Wing and Wolf, 2000), increased blood pressure (Wing et al. 2013) and incidence of asthmatic symptoms (Mirabelli et al. 2006) have all been recorded in eastern North Carolina’s CAFO-neighboring communities. More broadly, early and in utero exposure to PM10, total suspended particulates (TSP) and contaminated water has been linked [beyond a mere correlative association] to increased infant mortality, shorter gestation periods and lower birth weights (Chay and Greenstone 2003, Currie et al. 2009, Almond and Currie 2011, Currie et al. 2013, Aizer and Currie 2014).
Abiotic and biotic mediating factors: The vulnerability of downstream communities to upstream pollutants is mediated by biotic and abiotic factors. Community exposure to airborne CAFO contaminants relates to CAFO animal capacity, proximity, wind direction, wind speed and temperature (Wilson and Serre 2007, Wing et al. 2008). Shelterbelts, or vegetative buffers, limit the dispersion of CAFO odors and airborne pollutants to nearby neighborhoods by intercepting low-lying (20-30 feet) particulate matter (Tyndall and Colletti 2007). Nitrate (NO3) and nitrite are the two forms of nitrogen (N) in the National Primary Drinking Water Regulations and both are linked to blue-baby syndrome (U.S. EPA, 2016). The high amounts of organic and inorganic N contained in hog waste will many times be converted to NO3 through biological processes in the soil. This waste, whether land applied or leaching out of a lagoon, has the potential to reach private groundwater wells used for human consumption. The maximum NO3 concentrations in surface waters near animal feeding operations in eastern North Carolina are above the 10 mg L-1 (measured as N) that is the maximum contaminant level (Harden, 2015, U.S. EPA, 2016). These elevated N concentrations may be due to direct surface runoff, linked to groundwater contributions, or a combination of both. One of the challenges faced by CAFO operators is managing lagoon waste during periods of heavy precipitation. At the extreme, heavy precipitation may cause local stream or river levels to rise high enough that they overtop lagoon embankments resulting in increased contaminants in local waters (Heaney et al. 2015). A more likely scenario would occur after an extended wet or rainy period. Wet fields prevent waste from lagoons from being land applied. The combination of being unable to remove waste from the lagoons and rainfall entering the lagoons can result in waste leaving lagoons through spillovers.
Soil as a natural filter: The ecosystems and biogeochemical processes that naturally alter and control the delivery of potential contaminants to ground and surface waters vary spatially and temporally. Natural sources of freshwater come from groundwater aquifers. The deeper the aquifer, the longer the opportunity for water to move downward and be filtered by soil as precipitation gets deposited on Earth's surface. Soils have physical and chemical properties that make them natural filters. The characteristics of soil considered to be natural capital are texture, mineralogy and soil organic matter (Palm et al. 2007). This definition was expanded upon by Robinson et al. (2009) highlighting more interacting and complex soil properties involving ‘matter, energy and organization’. Soils provide important provisioning and regulating ecosystem services across scales. For example, soils provide services such as habitat for microbes responsible for water purification (micron scale), flood mitigation (landscape scale), and air purification (global scale) (Dominati et al. 2010).
Soil provision services such as filtering are important for interception of CAFO-specific contaminants through storing, intercepting, and processing of solutes resulting in water purification. Soil structure (i.e. arrangement of soil particles) and texture (i.e. percent composition of sand, silt, clay) impact soil’s ability to absorb and retain nutrients or contaminants through physical and chemical processes (Dominati et al. 2010, Robinson et al. 2013). Soil structure and texture also regulate flooding via retention and storage of water. Finer texture, clay-dominated soils are characterized by low hydraulic conductivity at the surface, which can prevent substantial infiltration and will result in most of the contaminants running off into surface waters. In contrast, sandier, coarser textured soils have increased hydraulic conductivity, which will allow the contaminants to infiltrate leading to potential contaminant removal, further lateral movement into groundwater, or horizontal movement to surface waters. Soil permeability or hydraulic conductivity at the surface and deeper horizons also plays a role in the delivery of contaminants to different water bodies.
Microbial processing of nutrients: Microbially-mediated recycling of nutrients and detoxing wastes are biological processes (Dominati et al. 2010). Riparian buffers and wetlands can reduce the amount of nitrogen reaching other systems through uptake by vegetation and microbially facilitated reduction of reactive nitrogen species (e.g. nitrate [NO3-]) to unreactive dinitrogen gas (N2). A greater area of wetlands in watersheds containing CAFOs has been linked to improved water quality compared to those with fewer wetlands (Harden, 2016). To offset nutrient runoff, best management practices (BMPs), such as implementation and management of riparian buffers, are used to intercept and process nonpoint nutrient pollution from agricultural fields (Mayer et al. 2007). Nitrogen loadings can be removed from surface runoff via multiple processes. First, by infiltration, although various forms of N are highly mobile below the subsurface, specifically NO3-, the loading is no longer at the surface. Infiltration through soil or engineered media can separate or trap particulate forms of N, although decomposition can slowly release organic nitrogen fractions. Ammonium (NH4+) can be adsorbed to negatively charged soil or media, or volatized under high pH (>8) conditions. Plant assimilation of nitrogen can occur through the uptake NH4+ and NO3-. Through multiple processes, microorganisms can transform various species of N into dinitrogen gas or nitrous oxide (N2 or N2O). Under aerobic conditions, organic N can be converted to NH4+ via ammonification. Nitrifying bacteria can convert NH4+ to NO3- via nitrification. Denitrifying bacteria can convert NO3- to gaseous forms of N2 or N2O, which volatize into the atmosphere. Microorganisms play a key role in functions essential for improving water quality by controlling virtually all nitrogen transformations in these systems (reviewed by Thamdrup 2012).
State and Federal Regulations on CAFOs: CAFOs are regulated federally under the Clean Water Act. Animal feeding operations (AFO) that have at least 2,500 swine each weighing 55 pounds or more and 10,000 swine weighing less than 55 pounds are considered large CAFOs. Medium CAFOs have at least 750 swine each weighing 55 pounds or more, 3,000 swine each weighing less than 55 pounds and a waste lagoon (or manure pipe) or direct animal contact with surface water. CAFOs are federally required to submit nutrient management plans that detail manure storage capacity, take measures to separate animals from surface water, install diversions of drinking water from the production area, and conduct soil tests and appropriate manure spraying and nutrient management practices. In addition, large CAFOs are required to design the production area to contain all facility manure plus the runoff from a 25-year 24-hour rainfall event (a large storm). If design specifications are met, manure overflows are permitted during such storm events. Within land application areas, CAFOs are required to analyze the soil for phosphorus every 5 years, to avoid applying manure within 100 feet of surface water and “from time to time” inspect land applications for leaks.
The state of North Carolina issues 5-year permits for AFOs and requires annual inspections by state officials. Similar to federal regulations, each facility must have a Certified Animal Waste Management Plan developed by a Certified Technical Specialist (North Carolina Department of Environmental Quality, 2016). The state permitting process also outlines 31 operating and maintenance requirements and 27 monitoring and reporting requirements.1 In 1997, the state of North Carolina imposed a moratorium on “new and expanded” swine farms, which became permanent in 2007 for those operations “using anaerobic waste lagoons as primary waste treatment” (North Carolina Department of Environmental Quality, 2016). Under state regulations, waste lagoons must be designed with 1 to 2 feet of freeboard as not to be inundated by a 100-year flood. The state does not have direct regulations on odor but maintains general setback requirements. Swine CAFOs must “typically be 1,500 feet from any occupied residence; 2,500 feet from schools, churches, and hospitals; and 500 feet from property lines.” Additional permits on land application of manure are not required if “waste is applied at agronomic rates and a vegetation buffer of at least 25 feet is maintained from perennial waters”.
Averting behavior of at-risk communities: In the presence of airborne and waterborne pollutants, individuals exhibit the capacity to make defensive investments to avoid exposure. Direct measurement of health outcomes without regard to such “compensatory behavior” risks understating the social costs from pollutants by ignoring those expenses individuals incur to self-protect. Zivin et al. (2011) find that households respond to water contaminant violations, under the Environmental Protection Agency’s Safe Drinking Water Act (SDWA), by purchasing bottled water locally. Residents increased bottle water purchases by 22 percent in response to microorganism violations and 17 percent in response to chemical violations revealing an estimated nationwide avoidance cost of $60 million for all 2005 violations.
Neidell (2009) finds that smog alerts in Southern California deter individuals from partaking in outdoor events, which reduces asthma-related hospitalizations, but that such responses to alerts may be short-lived (Zivin and Neidell 2009). During smoggy conditions increased use of air conditioners have also been recorded in addition to reduced time spent outdoors (Bresnahan et al. 1997). Aizer and Currie (2014) find that lower infant birth weights among black high school dropout mothers are associated primarily with a higher propensity of white college educated mothers to relocate away from exposure in between births. Averting behavior has also been recorded in a developing country context where on-river Indonesian households are less likely to ingest trash-polluted river water but are less reactive to the presence of upstream bathing activity (Garg et al. 2016). Averting specific types of pollutants in this way may suggest the visibility, or salience, of exposure is an important motivating determinant for making defensive investments in averting behavior by way of public services or ecosystem services. Unlike these studies that examine averting behavior arising from private defensive investments, we expand this literature by recognizing that communities also enable averting behavior with investments into public goods such as community infrastructure and the reconstruction and maintenance of local ecosystems.
Summary/Accomplishments (Outputs/Outcomes):
Between 2017 and 2021, the research team collected and analyzed data on the following human health and environmental outcomes:
- 710,186 North Carolina household-level georeferenced birth outcomes, medical risk factors and congenital anomalies occurring between 2006 and 2012 of which a portion of these births experienced in utero exposure to Hurricane Irene’s August 27, 2011, landfall (NC Department of Health and Human Services Vital Statistics Office).
- 49,588 North Carolina household-level georeferenced total coliform and 48,893 North Carolina household-level georeferenced E. coli private well water samples collected and analyzed from 2013 to 2018 (NC State Laboratory of Public Health).
- 12,140 household-level georeferenced private well water nitrate samples collected from North Carolina’s mountains, piedmont, and coastal plain regions and analyzed from 2013 to 2018 (NC State Laboratory of Public Health).
- 759 private drinking water well construction logs georeferenced at the household level and constructed after July 1, 2008, and in the North Carolina counties of Bladen, Duplin, Sampson, and Pender (collected and transcribed manually from the respective public health offices in each county).
- 189,096 North Carolina household-level georeferenced birth outcomes occurring within three miles of an animal feeding operation (AFO) between 1996 and 2017 (NC Department of Health and Human Services Vital Statistics Office).
Human health and environmental outcomes were examined using a variety of statistical techniques including quasiexperimental analyses that leverage a series of natural experiments (e.g., Hurricane Irene’s storm path) and external stressors that create plausibly exogenous sources of variation in outcome variables of interest (e.g., variation in wind direction, precipitation, hydrological flow, and soil compositions that determine environmental exposure risks across space and time). As such, some of the findings are interpreted to be plausibly causal while others point to correlative or suggestive relationships of interest that require further study.
Figure 1: A conceptual diagram linking example upstream CAFO-related contaminants with example downstream community health outcomes.
Conclusions:
- Across North Carolina, in utero exposure to Hurricane Irene decreased birth weights and gestation lengths while increasing the likelihood of preterm and low birth weight outcomes. For this specific storm event, groundwater contamination and potential exposure from private well drinking water cannot explain the observed birth impacts. Rather, evidence suggests that storm anticipation disrupted healthcare services that led to the permanent cancellation and delay of prenatal care appointments.
- Swine lagoon that are located hydrologically upstream of private drinking water wells cause a decisive spike in the detection rate of total coliform presence and E. coli presence when air temperatures exceeded a critical threshold of 32.2°C (90°F). We estimate that “cold weather sampling” – the collection of private well water samples during cooler air temperatures – systematically underestimates annual risk of total coliform and E. coli well water contamination by 25% and 103%, respectively. Federal guidance for private well water monitoring should be revised to increase “hot weather sampling” for communities that are located downstream of swine lagoons, are prone to high air temperatures and are dependent on private well water systems.
- Pregnant women located within three miles of an AFO experienced decreased birth weights, decreased gestation lengths and increased rates of preterm births when they spent more days of their pregnancy downwind of the nearest AFO. These effects increase in magnitude for pregnant women located within two miles of an AFO and the driving mechanism of downwind days of exposure is centered empirically on the summer months when sprayfields are likely to be active.
Secondary Findings
- After controlling for soil profiles and socioeconomic factors that may determine private well construction specifications, such as income-based geographic sorting away from surface sources of pollution, households within 3km of concentrated animal feeding operations (CAFOs) increase well casing depth by 30 feet for each km increase in proximity. This observed well drilling behavior may represent a form of private adaptation to perceived risks of groundwater contamination that is expected to cost approximately $406,000 per year (or $88 million total) for our rural North Carolina study area that includes Bladen, Duplin, Sampson, and Pender counties, which together are approximated to contain 70,000 private wells.
- Temporal changes in rainfall and temperature alone are not enough to predict nitrate contamination using machine learning models in most areas of North Carolina but show potential in the coastal plain region where our study area is located.
Journal Articles on this Report : 4 Displayed | Download in RIS Format
Other project views: | All 22 publications | 5 publications in selected types | All 5 journal articles |
---|
Type | Citation | ||
---|---|---|---|
|
Ethridge R, Pascual-Gonzales J, Hochard J, Peralta AL, Vogel TJ. Predicting nitrate exposure from groundwater wells using machine learning and meteorological conditions. AUTHOREA 2021. |
R836942 (Final) |
Exit |
|
Hochard J, Sbashidze N, Bawa R, Ethridge R, Li Y, Peralta A, Sims C, Vogel T. Air temperature spikes increase bacteria presence in drinking water wells downstream of hog lagoons. Science of the Total Environment 2023;867, doi: 10.1016/j.scitotenv.2023.161426. |
R836942 (Final) R840181 (2021) |
Exit |
|
Hochard J, Li Y, Abashidze N. Associations of hurricane exposure and forecasting with impaired birth outcomes. Nature Communications 2022;13(1):6746. |
R836942 (Final) |
|
|
Hochard J, Abashidze N, Bawa R, Carr G, Kirkland B, Li Y, Matlock K, Siu WY. Predicting groundwater contamination to protect the storm-exposed vulnerable. Climate Risk Management 2023;40:100499. |
R836942 (Final) R840181 (2022) |
Exit |
Progress and Final Reports:
Original AbstractThe perspectives, information and conclusions conveyed in research project abstracts, progress reports, final reports, journal abstracts and journal publications convey the viewpoints of the principal investigator and may not represent the views and policies of ORD and EPA. Conclusions drawn by the principal investigators have not been reviewed by the Agency.