The Harvard Sustainability Plan advocates for greater collaboration and integrative communication across disciplines for “more powerful and effective solutions to our most pressing problems.” Below is a short summary of my experiences in a research seminar in the School of Engineering and Applied Sciences.

I entered the semester having “Engineering Sciences 91r” on my study card, knowing with certainty that I did want to delve more into the world of research. I had only a faint of idea of what form that research might take, however. In the class, students work, for just the semester, on a research project with a faculty member. Looking far and wide at summaries of faculty in the departments at SEAS, I knew, surely, that I wanted to engage in a project with some focus on environmental science, engineering, or health, and I scheduled meetings with professors to speak about how an undergraduate might be able to contribute to their ongoing projects for the semester. My search had ended when I met with Elsie Sunderland, a professor in both SEAS and the School of Public Health, and Cindy Hu, a doctoral student in Environmental Health, who were both excited to tell me about work the Sunderland Lab had done on the biogeochemistry of methylmercury. The lab also had done recent work on perfluorinated alkylated substances (PFASs), a set of emerging pollutants.

With the help of Professor Sunderland and Cindy, I pinpointed my research project on human exposure due to perfluorooctanoate acid (PFOA), a PFAS, via drinking water. The compound is extensively used as an industrial surfactant in the synthesis of ingredients for fire-fighting foams and water and oil repellents for fabrics and leather, among many other current and historic uses. Highly insoluble in water, PFOA displays high mobility in soil and sediment, and is very resistant to degradation in air. A key study on the exposure of PFOA to human subjects in 2001 found that observable levels of PFOA were present in the blood of 96 percent of children tested in over 23 states. In combination with a slew of toxicity studies that concluded that PFOA elicits developmental, reproductive, and carcinogenic effects in animals, this has necessitated some concern as to how PFOA might harm humans. I chose New Jersey as the geographical constraint to my study because, in 2007, the state issued a guidance level for public drinking wells that was ten times lower than the advisory later set by the EPA in 2009.

First, intrigued by the mathematical rationales used by the EPA and the New Jersey Department of Environmental Protection (NJ DEP) for their advisory and guidance levels, I compared the protectiveness—essentially, the overall safety and conservatism—of the assumptions made. Overall, I found that the calculations by the NJ DEP and the EPA, though quite different, were slightly muddled in terms of choosing the most protective values for different variables. Where explanation was needed for a mathematical choice, the two governmental bodies often lacked. I did, however, give them some leeway: the NJ DEP’s guidance value and the EPA’s health advisory were created with a dearth of research on how PFOA, an emerging and highly undocumented compound, interacts with the human body.

Next, intrigued about how PFOA concentrations distribute throughout the state statistically, I conducted an exploratory spatial analysis using a few computer programs, and found that a few counties with high concentrations clustered in the New York metropolitan area. I was also curious about how several different point sources might cluster in counties with high average drinking water concentrations. Although a vast array of results were mapped and calculated, there was a telling impasse on the quality of data. Because PFOA is an emerging pollutant, few data were available at the spatial level, so many of the results were, from the start, statistically weak.

Percentile map of PFOA drinking water concentrationsPercentile map of PFOA drinking water concentrations

Next, I began a series of simple calculations to estimate how drinking water concentrations of PFOA measured across the state could contribute to concentrations in the blood of sample adult and child humans. Here, “sample” means that these individuals were given parameters (bodyweight, water consumption, etc.) that were average values found in literature; therefore, they were not necessarily accurate, but rather for my own educational purposes. Although the estimated concentration factors I calculated—the ratios between water and blood concentrations—were in the neighborhood of the factor reported by the New Jersey Department of Environmental Protection (NJ DEP), my advisors and I were unsatisfied. These calculations were simple and non-iterative, and the results were highly skewed by the most outlying concentrations reported in drinking water wells.

More was certainly needed on this front, so, to alleviate this issue, we conducted series of Monte Carlo simulations, which, rather than single-step calculations, use repeated, random sampling to calculate results. To calculate a single blood concentration, for example, the program chooses values randomly from the distributions it is given for each variable; for my project, this was done separately for adult men and women, and three groups of children based on age. Once these simulations were conducted, and later refined, I found the youngest group children, age 3-6, reported the highest average blood concentrations, reflecting earlier findings that children generally display higher concentrations. Later, once I compared the simulation results with a value found to yield immunotoxic effects in children, these results became a more significant cause of worry. All sets of individuals reported sizable numbers at risk for these health effects, with children at the most risk. Additionally, every simulation reported drinking water as the most significant contributor to blood concentrations—consistently greater than all other variables. In my opinion, these results warrant some concern regarding exposure to PFOA as well as more reports by the NJ DEP to monitor drinking water concentrations, even in light of the guidance level.

Although the final conclusion of my research project was an alarming one to make, the experience itself—interacting with faculty members, reading research papers, and immersing myself in a specific topic—was extremely rewarding. Throughout the course, I learned invaluable skills in probabilistic modeling, presenting research, and writing a formal scientific report, and became much more versed in the terminology of environmental health studies. I hope to invoke these skills, as well as the connections I made with faculty and graduate students, in my career as a student at Harvard and beyond.

I learned invaluable skills in probabilistic modeling, presenting research, and writing a formal scientific report, and became much more versed in the terminology of environmental health studies. I hope to invoke these skills, as well as the connections I made with faculty and graduate students, in my career as a student at Harvard and beyond.