UNC-BMElab ongoing research topics
BMElab director: Marc Serre
DEFINITION: Research topics are active research efforts that are ongoing at the UNC-BMElab research group and designed to provide a framework upon which new dynamic and innovative students can define their Honor project (BS students), technical report (MSPH and MSEE students), thesis (MS students), and dissertation (PhD students). They each correspond to an area in which some research has already been completed, is ongoing, or is being proposed for future works.
GOALS: At the UNC-BMElab, research topics generally aim to: (i) map environmental water and air quality, (ii) find drivers of environmental pollutions and adverse health outcomes, and (iii) solve environmental and public heath injustices.
DISCIPLINES: To reach these aims, students at the UNC-BMElab delve in the following three research disciplines:
1. The first is the field of geostatistics, which is used to map the spatial distribution of environmental and health processes across space and time. This consists in modeling the spatial or space-time variability of environmental and health processes using space/time covariance models, and in estimating process values at unsampled/unobserved points using the kriging and Bayesian Maximum Entropy (BME) methods of geostatistics.
2. The second is the field of spatial statistical regression, which is used to study associations between an environmental/health outcome and suspected spatial predictors. In particular, students become expert in the use of Land Use Regression (LUR) of environmental quality data to discover what are drivers explaining the variability of water or air quality, and inform policy decision makers tasked with reducing the source of water or air pollution. Students also become knowledgeable about spatial regression techniques used in environmental epidemiology to discover environmental determinants of adverse health outcomes.
3. The third is risk assessment, which is used to calculate what is the population disease burden attributable to a specific water or air pollutant of interest. This consists in using geostatistics to estimate environmental exposure, in interpreting epidemiologic results to model dose response functions, and in using uncertainty propagation to estimate a disease or mortality rate attributable to the environmental pollution of interest.
LIST: Following is a list of ongoing research topics:
· Modeling of SARS-CoV-2 in wastewater treatment plants and mapping COVID19 incidence across North Carolina: a large and still on-going multi institutional project on tracking SARS-CoV-2 in the wastewater across a range of North Carolina Municipalities, with many threads including sewershed delineation, inflow and infiltration modeling, COVID19 case geocoding, disease mapping, and SARS-CoV-2 sewer modeling.
· The effect of land applied sludge on the environment and public health in North Carolina: The treatment of wastewater results in sludge, or biosolid, loaded with heavy metals and potentially pathogenic and carcinogenic contaminants. These sludge are applied in agricultural fields, however little is known about the effect it has on the environment and on the health of people living nearby. This environmental injustice research topic involves the use of Geographic Information Systems (GIS) to create the first high-resolution spatial database of all sludge applications in North Carolina, and the use of Land Use Regression to study the association between the land application of sludge and environmental contamination.
· The effect of confined Animal Feed Operations (CAFOs) on the environment and public health in North Carolina: CAFOs hold large numbers of animals such as pigs and chicken in confined barns. They produce large amounts of largely untreated waste that is stored in open lagoons and spread on land and in the air shared with people living nearby. This environmental injustice research topic consists in the study of the effect that CAFOS have on water quality deterioration in nearby streams, and on microbial contamination of nearby homes. This work involves community participatory research and fieldwork to collect samples in homes near CAFOs, the quantification of microbial loads and antimicrobial resistance, the development of the first high-resolution spatial database of CAFOs’ barns and lagoons, and the development and implementation of spatial regression methods to quantify the effect that CAFOs have on the environment and people’s health.
· The space/time mapping of groundwater quality across North Carolina: Many people in North Carolina rely of private wells as their drinking water. Groundwater has been found to contain inorganic and organic contaminants (Arsenic, Radon, Nitrate, pesticides, etc. ) at elevated levels, but observational data is sparse and knowledge of areas with unsafe levels is limited. To address this public health issue we combine Land Use Regression (LUR) and Bayesian Maximum (BME) methods to obtain LUR/BME estimates of groundwater contaminants and map levels across North Carolina. Ongoing research is continuously refining the maps using additional data from multiple sources such as land use and geology, and addressing new pollutants that have not been mapped yet.
· Space/time river geostatistics to map surface water quality along all streams and rivers in the United States: Rivers are a unique ecosystem that is critical to the health and well-being of society as it supports aquatic life, recreational activities, and is a major source of drinking water. However estimating surface water quality is an intriguing statistical problem because of the role that river geomorphology and flow play. Classical geostatistical approaches use an Euclidean (i.e. straight line) distance metric, which does not account for river distance. To address this issue we have made substantial methodological contributions by developing a novel river geostatistical framework. This framework is being used to map all water quality pollutants along all river miles across the United States.
· Space/time BME data fusion to map air pollutants nationally and globally: Air pollution concentrations change rapidly over time, with sharp spatial gradients over space, and it is therefore difficult if not impossible to map air pollution accurately and at high resolution over large regions. Data characterizing air pollution come from monitoring stations providing exact observations (i.e. hard data) at sparse locations, and air quality chemical transport models providing approximate predictions (i.e. soft data) at fine scale resolution. There is a need to leverage these two data sources to obtain the best estimate of air pollution. The space/time Bayesian Maximum Entropy (BME) of modern geostatistics is an ideal framework for this task. We are continuously refining CTM model outputs and using BME for the data fusion of observations and model outputs to create the most accurate estimate available of air pollution at the finest resolution possible. Ongoing efforts consist in refining estimates of criteria air pollutions (PM2.5, ozone, etc) over the US and globally, as well as tackling new air pollutants.
· Epidemiology of air pollutants for novel health outcomes: The UNC-BMElab is collaborating in several epidemiological studies to provide air quality exposure estimates and quantify the association between air pollution and novel health outcomes. These studies have participants that are generally located in the US, for which air pollution exposure has to be calculated using the BME method as specific locations, times, and for specific time windows of exposure. One proposed study will be looking at the association between yearly Styrene concentration in the air and neurological symptoms amongst participants residing in Louisiana and surrounding Gulf region. Another study will be looking at the association between yearly PM2.5 and ozone concentration and cardiometabolic risks for participants in the same region. Another study will investigate the effect of weekly ozone concentration on revascularization outcomes amongst participants residing in New England.
· Global burden of disease attributable to ozone: Air pollution is one of the primary cause of deaths and diseases for humans, and it is preventable. In order to inform policies and treaties that would abate air pollution globally and save thousands of lives, it is necessary to be able to estimate accurately air pollution globally and assess the morbidity and mortality attributable to air pollution. However, modeling ozone across the globe presents a challenge due to the sparsity of observations available in some regions of the world, and the lack of reliable remote sensing data for ozone. In this research topic, we rely on the first global database of ozone monitoring data, and on global ozone chemical transport models, to obtain the most accurate maps available of ozone across the globe. Finally, we combine this information with known dose response functions to evaluate the burden of disease attributable to ozone globally.
· Short term air quality health effects and financial risks associated with wildfires: Recent wildfires in California have had significant adverse effects, including substantial increases in air pollution and destruction of land and infrastructures. However, little is known about the effect that increased air pollution has on human health, and on the financial risk associated with fire destruction. In this research topic we use a BME geostatistical analysis of air quality monitoring data to map air pollution during fire events and we assess the short term health impact that is attributable to this air pollution. We also study risks of future fires and quantify the financial losses associated with these risks.
· Space/time mapping and GIS analysis of monitoring and evaluation data on Water, Sanitation, and Hygiene (WASH) projects: Water, sanitation, and hygiene (WASH) programs often lack data and evidence adequate to increase efficiency and impact. To address this need, the Water Institute at UNC has implemented a monitoring, evaluation and learning focus area, through which data are being collected to assess the state of a system or the need of an intervention. For example, these data may consist in water quality observations at homes, at water taps, etc. There is a need to interpolate these sparse observations across space and time, in order to create maps showing how water quality changes across space, and evolves over time. In this research topic, students will work in collaboration with the Water Institute to perform a space/time mapping analysis of monitoring and evaluation data, and use GIS to create maps of water quality. These maps allow practitioners to learn about the success or challenge faced by a WASH program, and enable them to act in a way that increases efficiency and impact.
· Disease mapping of sexually transmitted infections and the opioid epidemic in North Carolina: Several disease and mortality outcomes, such as those associated with sexually transmitted diseases or opioid overdoses, result in outbreaks and epidemics. Mapping the disease and mortality rates associated with these health outcomes provide critical information about the underlying epidemics and are crucial in taking public health measures to reduce or eliminate these epidemics. However; while epidemiologic surveillance data exist, it is difficult to interpolate these data and create disease maps because of the small number problem, i.e. the problem associated with incidence rates with a small number denominator. In this research topic, we address this issue by using BME for the space/time interpolation of disease data with small denominator, and we create the best available maps of disease. Past work and ongoing work focuses on mapping sexually transmitted across North Carolina. Future work is looking at mapping opioid rates across the state.