Development and Custodial Organization:
Allwine, K.J., and X. Bian.  1995. PGEMS 2.0 – An Atmospheric Dispersion Model for Routine Air Quality Assessments and Emergency Response Applications. PNL-12088. Battelle Northwest Laboratories, Richland, Washington.

 Glantz, C.S., K.J. Allwine, M.A. Pelton, and K.J. Pattison. 2002. User’s Guide for the Air Pollutant Graphical Environmental Modeling System (APGEMS). PNNL-14043. Pacific Northwest National Laboratory, Richland, Washington.

APGEMS Documentation


The APGEMS SQA gap analysis review (Sept 2012) is available for viewing. 



Another limitation of APGEMS, especially in areas of complex terrain and complicated meteorology, is that empirical diffusion coefficients based on historical plume studies, typically conducted over relatively flat terrain, describe the turbulent diffusion of the puffs. This limitation in representing diffusion is partially offset in APGEMS by adding a component of variance to the horizontal and vertical diffusion coefficients that represents enhanced diffusion caused by vertical wind shears. Note that the need to constrain computational time for emergency response modeling, rather than limitations in the current state of air quality modeling, dictates the limitations described above. Several prognostic full-physics models have been applied to air quality modeling, and in some cases, to emergency response modeling. A significant limitation on the use of these full-physics models to emergency response situations is their intensive computational requirements and therefore lengthier simulation times in comparison to models like APGEMS. When new meteorological data are available every 5 to 15 minutes, and updated information on the emergency response event is arriving every few minutes (resulting in the need to modify release parameters and source term information), a model that takes only a seconds to adjust input and run a new simulations is often preferred over a prognostic, full-physics model that may take from many minutes to hours to complete a simulation.

The principal limitation of APGEMS is that the flow model is diagnostic, and consequently, forecasts of winds cannot be predicted directly from the basic equations of the flow model.  Forecasts of meteorological fields are computed by APGEMS from forecasts of meteorological "observations" at the input meteorological stations.  Simple persistence is routinely assumed to produce forecasts. Additionally, thermally developed Mesoscale flows (e.g., mountain-valley winds, slope flows, and land-sea breezes) are not simulated directly byAPGEMS. The model represents these flows to the extent that the observations are sufficient to resolve the flow feature. However, as mentioned previously, flow blocking and channeling by major terrain features and slope flows are parameterized in the flow model. 





TDM uses the output files from MET and TER and produces concentration and deposition fields on a user-specified receptor network for each time step of a simulation. The model computes source term characteristics, plume rise, contaminant transport, diffusion, puff depletion, radiological decay, and output parameters on the ground-level receptor grid.  Detailed information is provided in Allwine and Bian (1995). All output fields are written to a binary file.  These file are post-processed to generate graphical contour plots using ESRI’s MapObjects software.  The user can select any of the model’s output products for display, look at results at specified output times, or animate the output display. All output products can be printed or transferred for display on other devices (including display using other GIS software products).

MET produces wind fields, gridded mixing heights, stabilities, and precipitation categories, using upper-air and surface weather observations. The meteorological data required by MET can be specified at variable time increments over variable periods for a variable number of stations. That is, MET can accommodate meteorological data with missing records and different averaging times. During the routine assessment mode, MET will interpolate or average data according to user-selected modeling time step (5, 10, 15, 30 or 60 minutes). Surface data will be interpolated for up to 8-hour separations in observations, and upper-air data will be interpolated for up to 12-hour separations. Surface stations with data gaps larger than 8 hours (or 12 hour for an upper-air station) will not be used during the period of the gap. The three-dimensional wind field produced by MET is described analytically, for each model time step, on up to nine terrain-following surfaces that are between the ground and a user-specified upper boundary. The two-dimensional boundary layer parameter fields (horizontal atmospheric stability categories, vertical atmospheric stability categories, mixing heights and precipitation categories), produced by MET for each time step, are derived by interpolating the surface-station data to a 10 x 10 grid using an inverse-distance-squared weighting approach. The three-dimensional wind field is initially determined at each model time step by applying a mass-conserving interpolation technique to upper-air and surface-wind observations. The wind field is then subject to modification. For example, the model sets the surface winds to the local downslope direction to approximate drainage flows during stable atmospheric conditions (experienced during day- or nighttime).  A dividing streamline height is also used during stable atmospheric conditions when there is flow channeling and blocking from terrain features. Further adjustments to the wind field are described in Allwine and Bian (1995). 

The APGEMS modeling engine consists of three main computer programs named TER, MET and TDM. TER uses a base terrain file (typically a USGS digital elevation model in latitude-longitude coordinates) for the region of interest and produces disk files of gridded spatially-averaged terrain statistics (e.g., heights, drainage direction) in Universal Transversal Mercator coordinates (the primary coordinate system of APGEMS) for use by the MET and TDM programs.



To run the model in the first responder mode, the user simply needs to select a release location from a pull down menu or identify the release location using the program’s map display. The default simulation duration (4 hours), start time for the release (the current time), and release duration (1 hour) can be quickly adjusted if required. As a default, APGEMS will run with a ground-level, unit release (1unit/sec). When additional information is available, the user can open a menu to reset release parameters and another menu to select a source term scenario or define a new source term.


 couples an atmospheric transport and diffusion model for complex terrain applications (developed by Jerry Allwine) with a user friendly graphical interface (developed by Mitch Pelton, Nino Zuljevic, and Cliff Glantz). The user interface supports the rapid and intuitive entry of contaminant release and simulation parameters and the smooth display of model output graphics. 



The current version of APGEMS is hardwired for operation at the DOE’s Hanford Site. Other versions of APGEMS are hardwired to their locations and meteorological monitoring systems. APGEMS can be modified to run at other sites, but it require a substantial level of effort to do this, particularly to perform the required verification and validation testing.

In the initial stage of an emergency event, APGEMS is designed to support the needs of first responders. APGEMS rapidly generates graphical plots of the cumulatively affected area and the current area being affected by the effluent cloud. Initial model runs use only the limited information generally available near the start of an event (i.e., location of the release and the start time of the release). Running on a “standard” desktop computer or laptop, a four–hour first responder simulation only takes a few seconds.  As additional information becomes available (e.g., the type and quantity of materials released to the atmosphere, the duration of the release, the temperature of the effluent, the actual release height of the effluent) the user can quickly re-run the model and generate output products. In the Detailed mode, when running with 20 or more radionuclides, it takes less than half a minute to complete a four-hour simulation and generate all output products for display. 

For emergency response operations APGEMS can be run in three different modes: the Simple, Detailed, and Exercise modes. The Simple mode may be used during an actual emergency or drills when real-time meteorological data are being used and simulations from field monitoring values are not needed. The Detailed mode is used by exercise developers to prepare exercise data packages and during an actual emergency when projections of readings by radiological monitoring instruments are of interest. The Exercise mode is used during drills by exercise participants and controllers. In the Exercise mode, APGEMS ignores the actual real-time meteorological data and instead accesses "canned" meteorological data that is typically based on archived data from a day that matches the desired environmental conditions for the exercise. 


generates cumulative exposure, time-averaged concentration, total effective dose equivalent, and maximum organ dose values on a dense receptor grid.  Deposition fields are also calculated and these and other output products can be used to generate plots of simulated radiological readings from 11 different monitoring instruments (e.g., activity grab air alpha & beta samples, Cutie Pie [CP] measurements in the air and on the ground, micro rem meter readings). The time resolution for graphical output can be set to 15, 30, 60, or 120 minutes. Output products can be viewed at a specified time or by using an animated display that rapidly cycles through the output from the entire simulation.  The user can readily zoom in and out of the graphical display to focus on areas of interest, turn on or off contour labeling, pan across the modeling domain, and turn on or off various GIS layer information (e.g., water bodies, major and minor roadways, trails, building profiles, fencelines, railroads).  Supplemental data for up to ten monitoring locations that are typically close to the release location (e.g., within a few hundred meters) are provided using a simple plume model.  All model graphics, animations, and numerical displays are designed to efficiently convey information to hazard analysts and decision-makers.  In addition to being viewed on the screen, model products can be printed or shared via web connections.  This allows information to be rapidly shared with personnel within an emergency operations center and at other sites. 


is a PC-based modeling system with an easy-to-use graphical user interface. The model system balances the competing needs for technical precision, speedy performance, and simplicity of use. APGEMS can be used for areas with relatively uniform terrain or complex terrain environments. Source-to-receptor transport distances can range from as little as one hundred meters to a few hundred kilometers. APGEMS uses a three-dimensional diagnostic wind model to compute the vertical and horizontal spatial variation in winds at each time step in the simulation. The wind field is determined by applying a mass-conserving interpolation technique to the surface- and upper-air observations supplied to the model. The model accounts for flow channeling, blocking by major terrain features, and drainage flows. The model runs in a nested mode allowing for finer resolution near the source while covering a broader region at an appropriate resolution.  APGEMS employs a Gaussian puff formulation to mathematically describe the concentration distribution of the released materials as they move in the mean wind field. Mathematical reflections of the concentration distribution from the ground and the top of a mixed layer modify the initial Gaussian distribution. Primarily designed to model the dispersion and dose impacts of radionuclides, it can also be used to estimate the dispersion of non-radiological contaminants. The model treats wet and dry deposition, radioactive decay, and first-order chemical transformations of the released material.

The Air Pollutant Graphical Environmental Monitoring System (APGEMS) is an important atmospheric dispersion and dose assessment modeling tool used for emergency planning, preparedness, and response applications at the US Department of Energy's Hanford Site. Versions of the model are also used by the State of Idaho and by nuclear power plants to assist their emergency management programs. 





Nino Zuljevic, APGEMS Code Custodian, (nino.zuljevic@pnnl.gov; 509.372-4434)

Cliff Glantz, Senior Staff Scientist/Project manager,  (cliff.glantz@pnnl.gov; 509.375.2166)

Key Contacts at PNNL: