Select Page
Melanie Edwards, PStat.
Senior Solution Architect

Melanie Edwards, PStat.

Senior Solution Architect

Ms. Melanie Edwards is an accredited statistician who routinely provides insight on the interpretation of analyses and presentation of findings for experts from a variety of backgrounds, including environmental science, ecology, toxicology, and engineering, and for non-scientists. She has provided litigation support in the form of depositions, trial testimony, and expert reports. Her work includes investigations of multichemical and multimedia interactions, environmental forensics, spatial and temporal trend analysis, modeling exposure and effect relationships, probabilistic modeling, and study design considerations. With more than 20 years of experience, Ms. Edwards has provided statisti...

Ms. Melanie Edwards is an accredited statistician who routinely provides insight on the interpretation of analyses and presentation of findings for experts from a variety of backgrounds, including environmental science, ecology, toxicology, and engineering, and for non-scientists. She has provided litigation support in the form of depositions, trial testimony, and expert reports. Her work includes investigations of multichemical and multimedia interactions, environmental forensics, spatial and temporal trend analysis, modeling exposure and effect relationships, probabilistic modeling, and study design considerations. With more than 20 years of experience, Ms. Edwards has provided statistical support for projects involving PCBs, dioxins/furans, PAHs, and metals in soil, sediment, dust, groundwater, surface water, tissue, and air.

Ms. Edwards’ understanding of data workflow includes data management practices, data review and cleansing, statistical analysis methods, and clear presentation and communication of findings. Equally important, she understands the weaknesses and limitations of improper data preparation and statistical techniques to avoid over-interpretation of results. She is familiar with traditional methods such as correlation, regression, analysis of variance, non-linear models, factor analysis, principal component analysis, and non-parametric methods, as well as modern analytics techniques as they apply to environmental data.

Read More    Read Less   

Forensics

U.S. Environmental Protection Agency, Ottawa River, Ohio Performed statistical analyses of elevated PCB congener and Aroclor concentrations after a known source area had been capped. Results identified a common composition downstream but variability in the cap area. Individual congeners elevated in some samples suggest seepage from a historical landfill in the area. Data analysis identified laboratory errors with the Aroclor results for a subset of samples.
U.S. Environmental Protection Agency, River Raisin, Michigan Analyzed PCB congener results from an initial sampling event to understand compositional variability relative to potential sources. The identified compositions reflect standard Aroclor compositions with elevated low-chlorinated congeners. Results provided insights to direct a more comprehensive sampling event. Forensic analyses of the congener and Aroclor results from the greater spatial coverage and depth sampling are underway.
DDT Manufacturing Facility, Portland, Oregon Utilized likelihood ratios to combine nondistinct chemical signatures for dioxins/furans and pesticides, which provided evidence to limit the area of influence. Individual forensic analyses of PCBs, dioxins/furans, and pesticides were not sufficient to uniquely distinguish contributions from individual sources.
Massachusetts Department of Environmental Protection, Hudson, Massachusetts Reviewed statistical analyses of per- and polyfluoroalkyl substances (PFAS) results for a variety of sample matrices to distinguish between potential sources and impacts to the surrounding area related to a coating application services plant. Results included visual assessment of bar charts, principal component analysis, and hierarchical cluster analysis of groundwater, surface water, sediment, residuals and sludge, soil, and air emissions and filtrate. Technical comments covered choice of statistical methods, presentation of results, and validity of conclusions.

Site Investigation

Fox River Superfund Site, Wisconsin Evaluated and interpreted spatial and temporal trends in sediment PCB composition and concentration throughout a 25-mile stretch of river, in support of a natural resource damage assessment. Additionally developed a predictive relationship for fish PCB concentrations based on sediment.
Chemical Manufacturing Plant, West Virginia Performed statistical analyses of dioxin and furan concentrations in soil near a former chemical manufacturing plant, which yielded evidence of localized source influences but no broad impacts to the surrounding residential area. Chemical forensic analyses of dioxin and furan levels in blood samples from nearby residents showed that former plant workers were distinct from other residents.
Red Dog Mine, Northern Alaska Assisted with developing a sampling program for metals in multiple tundra environments. Conducted statistical analyses to evaluate the contribution of metals to changes observed in plant communities near road and active mining areas. Analyses encompassed various methods to quantify spatial trends, compare concentrations between potentially impacted and reference locations, evaluate habitat differences in the vegetation community, assess similarities in the benthic invertebrate community, and develop inputs for ecological and human health risk assessment.
Deepwater Horizon Oil Spill, Gulf of Mexico Developed a data assessment and quantification analyses to relate field-collected fluorescence and dissolved oxygen measurements in the water column with laboratory-quantified benzene, ethylbenzene, toluene, xylene (BTEX) and PAH concentrations. Results demonstrated significant biodegradation within the water column. Designed the database and provided oversight for data compilation from multiple diverse sources encompassing three million sample records each with multiple measurements.

Statistical Modeling

Hudson River, New York Modeled relationships between mink mortality and reproduction (decline in fecundity) and PCB concentrations using a generalized linear model. The analyses provided cumulative probability curves and estimates of exceedance probabilities for specified response limits to inform an assessment of potential ecological impacts.
Pesticide Registration Modeled the complex time-dependent dose-response relationships of cholinesterase inhibition across multiple rat exposure studies to estimate the benchmark dose with uncertainty safety limits.  Model results supported sensitivity assessment between adults and pups and across gender.
Berry’s Creek Superfund Site, New Jersey Advised academic scientific researcher on appropriate statistical analyses to assess differences between treatment methods for reducing mercury methylation in sediment.  Results supported modification of the study design and sampling strategy to reduce the confounding effects of tidal changes.
Global Waste Stream Modeling, Petroleum Company, Worldwide Locations Developed a probabilistic model to manage global waste stream volumes for a large petroleum company, including liquid and solid waste from sites ranging in size from local gas stations to major port loading facilities.  Results from the model provided a baseline volume estimate prior to implementation of waste reduction practices.

Data Sufficiency

Department of Natural Resources, Michigan Recommended changes to the fish monitoring program sample design to achieve equivalent monitoring information at lower costs using simulation methods to quantify relative variability in fish tissue concentrations of PCBs and mercury.
Reid Gardner Power Plant, Nevada Demonstrated the limit of available data to explain variability in the spatial distribution of arsenic concentrations that justified measurement of additional parameters in groundwater near the waste ponds.  Predicted most likely source contributions based on groundwater compositions and defended results to responsible parties because conclusions contradicted prior understanding.  Statistical analysis confirmed hydrologic assessment to identify gaining and losing portions of surface water onsite.
Dermal Absorption of Arsenic, New York Determined dermal absorption of arsenic from soil as not different from background levels using analysis of variance (ANOVA) with nested factors.  Evaluated statistical power of the study design and quantified minimum detectable absorption as a recommendation to EPA for updating the human health risk assessment guidance for arsenic.

Aerial Deposition

Wood Treatment Facility, Grenada, Mississippi Demonstrated that historical releases from a wood treatment facility were not a prominent source of dioxin and furan concentrations detected in a nearby residential neighborhood. The chemical fingerprints of residential dust samples did not reflect a consistent gradient away from the claimed source.
Aerial Deposition Analysis, Multiple Locations Quantified the lack of a consistent spatial pattern expected from deposition of aerial emissions of lead, arsenic, and dioxin and furan congeners. Analyses included distance from the claimed source and suggested alternative sources.

Litigation Support

Litigation Support, Nitro, West Virginia Generated upper limit blood concentrations for background levels of exposure to dioxin and furan congeners and toxicity equivalents based on the National Health and Nutrition Examination Survey (NHANES) complex survey study design. Comparisons provided evidence to conclude that there was limited to no excess exposure at residential areas near a chemical manufacturing site.
Litigation Support, Seattle, Washington Estimated the reduction of daylight exposure that would not be sufficient to incur health effects.  Used the American Time Use Survey with location-specific astronomical data to quantify average daylight exposure times.  Compared survey sample demographics with census statistics for the urban area confirmed representativeness of the estimates.

R&D

Dermal Absorption of Arsenic, New York Determined dermal absorption of arsenic from soil as not different from background levels using analysis of variance (ANOVA) with nested factors.  Evaluated statistical power of the study design and quantified minimum detectable absorption as a recommendation to EPA for updating the human health risk assessment guidance for arsenic.

Expert Testimony

New Jersey Department of Environmental Protection v. The Sherwin-Williams Company Docket No. CAM-L-005032-19 in the New Jersey Superior Court, Camden County Provided expert opinions about the plaintiffs’ derivation of background concentrations of lead and arsenic in soils. Expert report submitted June 2024.
Zina G. Bibb, et al. v. Monsanto Company, et al. Civil Action No. 04-C-465 in the Circuit Court of Putnam County, West Virginia Provided demonstration to court of statistical analysis of NHANES data. Responded to questions from the judge and Plaintiffs’ attorney. Affidavits submitted August and September 2011.
RETURN TO OUR TEAM PAGE