Jump to main content.

Further Technical Details about HAPEM4

Information provided for informational purposes onlyNote: EPA no longer updates this information, but it may be useful as a reference or resource.
Overview of HAPEM4
An air quality dispersion model, such as the ASPEN model, predicts an ambient concentration for a given time period at a given location. If a person could remain at this location for the specified time period (in this case, one year), the concentration predicted by the air quality dispersion model would equal his "apparent" exposure, or the concentration available for the subject to breathe. In the real world, people's day-to-day activities generally move them from one location to another, (e.g., from home to work, or home to school). Also, most people do not spend their entire day outdoors; a majority of our time is spent in indoor locations (e.g., the home, workplace, school, or vehicle). Studies have shown that air quality concentrations in indoor environments can be quite different than those in the outdoor environment. Because of these factors, an exposure model is generally employed to consider these factors and predict the "apparent" inhalation exposure.

Over the past 15 years, EPA's Office of Air and Radiation, in conjunction with the EPA's Office of Research and Development, developed the Hazardous Air Pollutant Exposure Model for Mobile Sources (HAPEM-MS), to assess inhalation exposure to air toxics from highway mobile sources. The initial version of HAPEM used carbon monoxide as a tracer for highway mobile source air toxic emissions. EPA has updated and improved HAPEM to predict inhalation exposure concentrations of air toxics from all outdoor sources, based on ambient concentrations from model or monitor data for specific air toxics, and no longer uses CO as a tracer. It predicts nationwide census tract level annual average human exposure levels and can thus be used in a screening level inhalation risk assessment.

The HAPEM4 model has been designed to predict the "apparent" inhalation exposure for selected population groups and air toxics. Through a series of calculation routines, the model makes use of ambient air concentration data, indoor/outdoor concentration relationship data, population data, and population activity pattern data to estimate an expected range of "apparent" inhalation exposure concentrations for groups of individuals. Details on the HAPEM4 model and its application to the initial national-scale assessment are given below.

Ambient Air Concentration Data
The HAPEM4 requires annual-averaged, diurnally distributed air quality data. In addition, HAPEM4 can also evaluate the contributions of sub-sets of the air quality data (e.g., air concentration values for specific source sectors such as point source, area source, mobile source). While the air concentration data for HAPEM4 must be in a specific format (e.g., annual average and diurnally distributed), the source of the data could be either from an air dispersion model or an ambient monitor. The most common form of ambient air concentration data for HAPEM4 is output data in the ISCLT format.

For the national-scale assessment, annual average ambient concentrations for each US census tract were estimated with the ASPEN model. In order to preserve any characteristic diurnal patterns in ambient concentrations that might be important in the estimation of population exposure, ASPEN annual average concentration estimates are stratified by time of day, with an annual average for each of the (8) 3-hour time blocks (e.g., midnight to 3am, 3am to 6am). ASPEN air quality files were also provided by each of the 4 major source sectors (i.e., major, area, mobile onroad, mobile nonroad). Thus, the results of HAPEM4 can be summarized for each of the 4 groups or a combination of them.

Microenvironments (MEs) are generally described as a small space in which human contact with a pollutant takes place, and which can be treated as a well-characterized, relatively homogenous location with respect to pollutant concentrations for a specified time period. MEs include indoors at home, school, work, inside an automobile or bus, outdoors, etc.

In order to calculate a person's exposure concentration an estimate is required of the concentration in each of the various MEs. In HAPEM4 the concentration in each ME is derived from the ambient concentration estimate for the census tract (obtained from ASPEN) and a set of 3 ME factors: PEN, PROX, and ADD.

C(i,k,t) = [ASPEN]i x PEN x PROX + [ADD]


C(i,k,t) = concentration predicted within census tract i and microenvironment k in time step t
[ASPEN]i,t = ambient concentration estimated from ASPEN in census tract i for time step t
PEN = penetration factor
PROX = proximity factor
ADD = additive factor accounting for sources within the microenvironment

The penetration factor, PEN, is an estimate of the ratio of ME concentration to the concurrent outdoor concentration in the immediate vicinity of the ME. These pollutant-specific estimates are derived from reported measurement studies. The proximity factor, PROX, is an estimate of the ratio of the outdoor concentration in the immediate vicinity of the ME to the outdoor concentration represented by the air concentration data. (ASPEN is designed to predict census tract centroid impacts, not that which would be found adjacent to a point source or on the side of a roadway). ADD is an additive factor that accounts for emission sources within or near a particular microenvironment, i.e., indoor emission sources. For the national scale assessment the ADD term has been set to zero; as additional indoor source data becomes available this factor will be included in future assessments.

In this national-scale assessment, HAPEM4 considers cohort activities in 37 ME locations. ME factors for HAPEM4 have been compiled for the 33 air toxics being assessed and are summarized in a detailed report on ME factors. It is important to note that data to support quantitative ME factors are not well developed for many of the air toxics compounds and for most of the 37 microenvironments. As a result, representative values were used in many cases based on measured values of similar compounds in similar situations. Further, for many of the compounds/microenvironments with measured data supporting ME factor estimation, few measurements were available and the representativeness of the available measurements is unknown. These limitations in the database available to support ME factor estimation introduce uncertainty into the analysis of exposures.

Population Activity Data
One of the keys to determining someone's long-term inhalation exposure is to develop a composite of that person's breathing level concentration over that period of time. This can be done by following a person around in time and space and tracking the person physical activities. HAPEM4 uses two types of population activity data: activity patterns data and commuting patterns data.

An activity pattern is a series of discrete events of varying time intervals describing an individual's lifestyle and routine as she moves through different microenvironments. The HAPEM4 model needs estimates of the amount of time spent in each of a set of microenvironments (e.g., at home, at work, in an automobile, etc.), and a description of what the individual was doing in each microenvironment (e.g., sleeping, eating, exercising, etc.). The HAPEM4 model utilizes activity pattern data from the EPA's Consolidated Human Activity Database (CHAD). The CHAD, contains more than 22,000 person-days of activity pattern records from 12 studies (for more information about this database see the CHAD website). However, it is important to note some of the limitations associated with the CHAD database. First, much of the activity pattern data in the CHAD database are for individuals over a one or two day period. The lack of activity pattern data that extend over periods longer of times presents a challenge for HAPEM4 to predict the long-term (yearly) activity patterns that are required to determine chronic exposures. Second, the CHAD database may not be representative of all demographic groups, particularly ethnic minorities (e.g., black males) or low-income populations. Therefore, HAPEM4 may not accurately characterize air pollutant exposures of these demographic groups. Human activity varies with cohort demographic characteristics as well as other factors such as climate (i.e., temperature) and day of the week (i.e., workday vs. non-workday). HAPEM4 considers these factors when estimating long-term inhalation exposure levels. An approach of selection of a series of single day's patterns (from CHAD) to represent an individual's activity pattern for a year was developed. If too few days are selected in building this pattern it is likely to overestimate the individual-to-individual variability, if too many days are selected it is likely to underestimate the individual-to-individual variability. Nevertheless, given the available data from CHAD, for the national scale assessment, an approach for estimating inter-individual variability in activity patterns with HAPEM4 was developed by sampling a series of activity patterns for each of 3 specified day types: weekends, summer weekdays, non-summer weekdays. To implement this approach, first all the activity pattern data are grouped according to cohort group (see below) and day type combinations. For each census tract/cohort group combination an activity pattern is selected for each day type from the appropriate activity pattern group. The corresponding exposure concentration is calculated for each of the three activity patterns. Then a weighted average of the three exposure concentrations is calculated to represent the annual average concentration, where the weightings represent the number of days per year for each day type (i.e., 104 for weekends, 65 for Summer weekdays, 196 for non-Summer weekdays)

An additional capability of HAPEM4 that is included in this assessment is the cohort commuting feature. The commuting pattern data contained in HAPEM4 were derived from a special U.S. census database that specifies for each census tract the number of residents that work in each tract, i.e., the population associated with each home tract/work tract pair. This model feature, which permits specific cohorts to physically move between "home" and "work" census tracts, allows air quality estimates utilized in ME concentration calculations to vary with the cohort's physical location. In areas with significant ambient concentration changes over short distances (i.e., cities versus suburbs) it is possible for a person to live in an area with relatively low ambient concentration and work in an area with relatively higher concentration. The net effect of the HAPEM4 commuting feature (rather than have the person stay located in his home tract) would be to increase this person's overall exposure.

While HAPEM4 currently includes commuting features for the adult population, it does not include the feature for children who commute to school outside their home census tract. A future version of the model is expected to add this important feature.

Population Demographic Data
For an exposure assessment, the population is usually divided into a set of cohorts such that (1) each person is assigned to one and only one cohort, and (2) all the cohorts combined encompass the entire population. A cohort is generally defined as a group of people whose characteristics may result in different exposure than other cohorts. The use of cohorts is a simplifying assumption for modeling exposures of a large population. Because adequate data on the exposure of each individual within a population do not exist, information about people who are expected to have similar exposures (i.e., people with similar demographic characteristics such as age and gender that are expected to influence exposure to air toxics) is aggregated together to make exposure estimates for these groups. This approach is necessary in order to make better use of the limited data that are available.

The U.S. Census Bureau is the primary source of most population demographic data. The Census collects information on where people live, their demographic makeup (e.g., age, gender, ethnic group), and employment. The HAPEM4 model uses 1990 U.S. Census data reported at the spatial resolution of census tracts, which are small, relatively permanent statistical subdivisions of a county. Census tracts usually contain between 2,500 and 8,000 residents

For the national-scale assessment application, HAPEM4 divided the population into 10 cohorts, based on combinations of age (5 categories), and gender (2 categories). Predicted inhalation exposure levels for cohorts were then aggregated across cohorts to estimate general population exposure levels.

The use of distributions for many of the exposure parameters included in HAPEM4 is prohibitive because of the large number of calculations required to apply the model on a national scale. It is expected that future versions of the model will make better use of statistical sampling techniques (i.e., Monte Carlo) to better characterize the range of predicted exposure concentrations. The model as currently configured and applied in the national scale assessment, utilizes averages, medians, and "best-estimates" for many of the exposures parameters discussed above. Thus, the predictions as presented are best viewed as population estimates rather than individual exposure estimates.

HAPEM4, as applied in the national scale assessment, does incorporate two random variables. The first is the selection of activity patterns to represent each cohort group. As noted above, HAPEM4 builds long-term activity patterns by randomly selecting daily (with replacement) activity patterns for 3 data types (i.e., weekends, summer weekdays, and nonsummer weekdays) and combining them to find the average fraction of time in each of the 37 microenvironments for each of the (8) 3-hour time blocks. EPA constructed 100 such annual activity patterns for each cohort group. Then for each census tract, 30 of the 100 annual patterns were randomly selected (with replacement) (30 patterns were selected to predict a range of activity patterns and yet minimize model run time). The result is a set of 30 annual exposure concentration estimates for each cohort group in each census tract.

In addition, for activity patterns that indicated time spent at work, EPA randomly selected a work district from the set of work districts associated with each home district, using the proportion of workers commuting to each work tract for its selection probability. For more information on variability in the national-scale assessment follow this link.

Comparison with Personal Monitoring
It is EPA's general practice, when using computer simulation models of pollutant levels in the environment and human exposures to pollutants, to provide an evaluation of the model whenever possible. This evaluation generally consists of a comparison of the model's estimates to corresponding pollutant measurements. As part of NATA, EPA has conducted this kind of model performance evaluation for the ASPEN model, using methods that have been peer reviewed by the Agency's Science Advisory Board. However, a model performance evaluation has not been conducted for the HAPEM4 exposure estimates. This is because much of the currently available personal monitoring data that would be used for such an evaluation was used in the development of the HAPEM4 microenvironment factors. Therefore, we are able to characterize how well the ASPEN estimates of ambient pollutant concentrations compare with "actual" pollutant concentrations in the environment, but we are not yet able to characterize how well the HAPEM4 exposure estimates compare with actual human exposure data. EPA is currently working with the Mickey Leland Center to help identify new and independent sources of personal monitoring data for use in comparison with the national scale model results. When these data become available, EPA will make comparisons to the HAPEM4 exposure estimates and include these results in future technical reports describing the comparison process and model validation findings.

HAPEM4 Peer Review
In the summer of 2000 the EPA had an external peer review by technical experts for both the ME factors utilized in the HAPEM4 model and the overall application of the HAPEM4 model for the national scale assessment. In Appendix A of the National Scale Assessment technical report, several of the issues addressed by these reviewers are discussed. In addition, the EPA's Science Advisory Board (SAB) reviewed the application of the HAPEM4 model as part of the 1996 national-scale assessment. While the peer reviewers identified several limitations inherent in the current methodology, it was still acknowledged as an appropriate tool to help better understand the relation of human exposures to ambient concentration levels. Where practical, many of the suggestions made by the SAB panel were incorporated into the national scale assessment. In addition, the EPA will continue to use these comments and suggestions to improve future versions of EPA exposure models and help guide the agency's exposure research needs.

HAPEM4 Downloads

HAPEM4 User Documentation

Chapter 1- Provides a brief introduction to HAPEM4 modeling fundamentals including a brief history of the development of HAPEM4.

Chapter 2 - Provides an overview of the various components of HAPEM4 and basic information needed to run the model.

Chapter 3 - Provides a description of the format, data, and options for each of HAPEM4 input files.

Chapter 4 - Provides a description of the format and data associated with each of HAPEM4 output files.

Chapter 5 - Provides a description of the purpose, inputs, and outputs, including a brief description of the computer code, for each of HAPEM4 computer programs.

Chapter 6 - References.

Appendix - Application of HAPEM4 for National Air Toxics Assessment - this section descibes how the HAPEM4 model was applied on the National Scale Assessment as well as addresses many of the models peer review issues.

HAPEM4 ME Factor Full Report
HAPEM4 ME Factor Table Only
HAPEM4 Model Installation Package

Local Navigation

Jump to main content.