Projects

Click here for past projects.
Note that you may need to right-click on some of the files to download them.

Regression Model Development for Forecasting PM2.5

Clients

Mr. David Krask

Mr. Matthew Seybold and Duc Nguyen

Maryland Department of the Environment

Brief Description
Fine particulate matter is a significant pollutant that endangers human health. Small particles, 2.5 micrometers in diameter or less, penetrate further into the lungs of humans than larger particulates leading to increased cases of respiratory diseases and eventual death. Both annual mean and 24 hour National Ambient Air Quality Standards have been set for fine particulate matter (PM2.5). PM2.5 is one of five pollutants reported in the USEPA’s Air Quality Index. It is critically important that the current PM2.5 value can be accurately forecasted so it can be reported to the public with an appropriate health advisory. Our objective is to develop reliable forecasting regression models to serve as tools for predicting PM2.5. The regression models will take into account various meteorological parameters such as temperature, wind speed, wind direction, and yesterdays PM2.5 measurements. Our client, the Maryland Department of Environment, provided all meteorological and fine particulate matter data. Analyses of selected particulate matter monitoring stations and meteorological sites in the state of Maryland have lead to discoveries of certain PM2.5 patterns. Trends show PM2.5 variations between winter and summer seasons as well as weekday and weekend periods. Various patterns, interaction terms, nonlinear curvature, and other possible confounders will be taken into account. The focus of this effort is to In addition to finding a model to predict the data one day in advance, it is important to develop a model which predicts the midnight to midnight PM2.5 value by 10:00 a.m. in the morning. The prediction is made14 hours before the midnight to midnight measurement is available. This approach is significantly more accurate than the prediction prepared the day before. It is necessary for this model to make use of the newest conditions in order to give the public the most reliable and current prediction possible.

Students

Wilma C. Jackson

Data...

Can Meteorologically Adjusted Ozone Trends Estimate the Impact of the NOx SIP Call?

Clients

Mr. George Bridgers

NC Department of the Environment and Natural Resources

Dr. Ryan Boyles, Director, State Climate Office

Brief Description
Breathing in ground-level or tropospheric ozone can trigger a variety of health problems including chest pain, coughing, throat irritation, and congestion. It increases problems with bronchitis, emphysema, and asthma. The ability to determine the impact of ozone precursor emission controls on ground-level ozone trends is complicated by the impact of meteorology, which can be either conducive to ozone formation or not. How do you know if emission controls are really working? The major precursors to ground-level ozone formation are volatile organic compounds (VOCs) and nitrous-oxides (NOx). The EPA Nitrogen Oxides State Implementation Plan Call (NOx SIP Call) began in 2001 in an effort to mitigate the formation of ground-level ozone. Since ozone is strongly affected by the influence of meteorological variables, many different approaches have been taken to determine the trend in ozone by removing the effects of varying meteorology. The purpose of this project was to build a time series model that removes the effects of meteorology, autocorrelation, and seasonal trends based on ozone and meteorological data from the Maryland and New Jersey Departments of the Environment. This data spans April through October of 1997-2006 for Maryland and Washington , DC and 1997-2005 for New Jersey . As the result of our analysis, a series of models were combined with a filtered time series model and back trajectory modeling to estimate the reduction in ground-level ozone over this ten-year period.

Students

Adrienne M. Wootten

Timothy Brown

Kristan Gore

Jie Zheng

Data

An Exploratory Analysis of Public Databases to Identify the Next Major Lead Source Targeted for Reduction to Improve Blood Lead Levels in Children

Clients

Dr. Barry Nussbaum, Chief Statistician, Office of Environmental Information, U. S. Environmental Protection Agency

Brief Description
For over a year now this research has explored multiple data bases such as the Toxic Release Inventory, air and water quality data, and census data to gain a better understanding of the lead problem affecting the children of our nation. Using these exploratory methods of public data bases the CDC’s Childhood Blood Lead Survey became the primary focus of the research at hand. These data, upon initial viewing, provided cluster locations of elevated blood lead levels among children across the U.S. and a special interest on the cluster in California . Upon further investigation and phone interviews with individuals within these high regions and with the CDC; the findings of the research have been confirmed that these concentrated areas are in fact real and the data are correct. The purpose of this research is to expose the underlying cause of elevated blood lead levels within these clusters and to gain a firm understanding of whether these regions are a result of environmental, industrial, or cultural factors.

Students

Joshua M. Drukenbrod

Data

NHANES Data

CBLS Data for 41 States

SAS Data Sets

SAS Lead Programs

KML Files from California

TRI Data

***California BLL Data- 2006 ***

Crustal Matter/EPA Project

Clients