Computer Software Reviews. RISKPRO: Environmental Pollution


Computer Software Reviews. RISKPRO: Environmental Pollution...

0 downloads 112 Views 405KB Size

J. Chem. In$ Cornput. Sci., Vol. 31, NO.3, 1991 421

COMPUTER SOFTWAREREVIEWS

(10) (1 1)

(1 2) ( 1 3) (14) (1 5)

(16) (17)

The other uses will probably not cause severe confusion. The prtscnt paper deals with isospectral points and pairs (including edges) within a single graph, which following RandiE could be termed endospectral points and pairs. We, however, do not use this term here since it does not seem to be widely adopted. (b) RandiE, M. SIAM J . Alg. Disc. Merh. 1985, 6, 145. Lowe, J. P.; Davis, M. W. J . Marh. Chem. 1990, 5, 275; see also RandiE, M.; Baker, 9. J . Marh. Chem. 1988, 2, 249. RandiE, M. J . Compur. Chem. 1980, I , 386. Knop, J. V.; Miiller, W. R.; Szymanski, K.; Randie, M.; TrinajstiE, N. Croor. Chem. Acra 1983, 56, 405. These proofs use Heilbronner’s factorization theorem; see, Heilbronner, E. Hela Chim. Acra 1953, 36, 170 and ref Sf, p 98; see also ref 16. Riicker, G.;Riicker, Ch. Understanding the Properties of Isospectral Points and Pairs in Graphs: The Concept of Orthogonal Relation, in preparation. Lowe, J. P.;Soto, M. R. MATCH 1986, 20, 21. In the graph of n-propylcyclopropane, the pairs made of the primary and the tertiary vertex on the one hand, of a secondary cyclopropane site and the middle vertex of the side chain of the other hand were prematurely concluded to be isospectral,since bridging one or the other of these pairs by a one-vertex bridge results in two isospectral graphs’‘ (those of cyclopropanespirocyclopentane and of 3-methylbicyclo[3.1 .0]hexaneI6). These pairs are neither isocodal, nor do they meet the criterion of equal products of eigenvector coefficients. DAmato, S. S.; Gimarc, B. M.; TrinajstiE, N . Croat. Chem. Acra 1981, 54, I . Since it is impossible to do every conceivable change experimentally,

the above coefficient criterion for the isospectrality of pairs as a sufficient condition is highly welcome. In a forthcoming paper1’we shall disclose in detail and with proofs the exact mathematical relations between isocodality, ‘isamfficiency”, and isospectrelity for pain as well as for vertices. (1 8) The eigenvalues and eigenvectors of a graph were calculated by using the program EIGEN,the core of which is the N A G routine FOZABF. This program is by far faster than, e.g., Balasubramanian’s (Balasubramanian, K.;Liu, X . J . Compur. Chem. 1988, 9, 406). For comparison of the eigenvaluesof graphs resulting from systematic changes of all the vertices or of all the pairs in a graph the programs ISOSPEC and ISOPAIR were developed, the core of which is the N A G routine F02AAF. (19) 3 and 4 are therefore counterexamples to RandiE‘s statement:

(20) (21) (22) (23) (24)

‘Endospectral graphs have different counts of random walks for the isospectral points.”9b Balaban, A. T.; Mekenyan, 0.;Bonchev, D. J . Compur. Chem. 1985, 6,538. Mekenyan, 0.;Bonchev, D.; Balaban, A. T. J . Compur. Chem. 1985, 6, 552. Stankevitch, M. I.; Tratch, S. S.; Zefirov, N . S. J . Compur. Chem. 1988, 9, 303. Mathon, R. Proceedings of rhe 9rh Sourhearrern Conference on Combinatorics, Graph Theory, and Computing, (Congr. Numer.) Utilitas Math., Winnipeg, 1978, p 499. Weisfeiler, B. On Consrrucrion and Idenrificarion of Graphs; Lecture Notes in Mathematics No. 558; Springer: Berlin, 1976. Knop, J. V.; Miiller, W.R.; Szymanski, K.; TrinajstiE, N.; Kleiner, A. F.; RandiE, M. J. Marh. Phys. 1986, 27, 2601.

COMPUTER SOFTWARE REVIEWS RISKPRO: Environmental Pollution Modeling System? WALTER L. MCLEOD* and ROBERT G. BUTZ Jellinek, Schwartz, Connolly & Freshman, Inc., 1015 15th Street, N.W., Suite 500, Washington, D.C. 20005 Received March 28. 1991 RISKPRO is a powerful and user-friendly exposure assessment system designed to evaluate the behavior of toxic chemicals when released into the environment. Two chemical property estimation programs interact with seven environmental modeling programs to simulate the passage and transformation of a chemical through air, soil, aquifers, rivers, lakes, and streams. RISKPRO incorporates most of the same sophisticated models used by the EPA in performing exposure assessments. RISKPRO also allows the user to install other modeling programs into the system. To keep pace with advances in modeling capabilities, periodic upgrades are planned for RISKPRO. RISKPRO was developed by General Sciences Corporation (GSC) of Laurel, MD. The chemical property estimation programs contained in RISKPRO are ATOMLOGP and AUTOEST. These programs are used to generate the necessary chemical properties required to run the environmental models with no input other than the chemical structure (using the SMILES format) or CAS number of the compound. The CAS number/SMILES dataset of RISKPRO contains 20000 chemical entries. It is recommended, however, that if chemical properties data are available, they be entered manually. The environmental modeling programs used in RISKPRO , ENPART, ~ L U and , ISCLT. These are SESOIL, A T ~ D EXAMS-11,



RISKPRO is available from General Sciences Corp., 6100 Chevy Chase Dr., Laurel, MD 20707. Telephone: (301) 953-2700, The list pricc for RISKPRO is 55660.00.

modeling programs are the heart of the RISKPRO software package. The RISKTAB option allows the user to create tables of dose, risk, and hazard index estimates by using data generated from the models. All programs include a series of menus and extensive help functions to aid users with various degrees of experience. Features and applications of each program include the following: ATOMLOGP estimates the octanol/water partition coefficient (log Kow) for a chemical on an individual atom basis. The atoms in the molecule are categorized into various atom types based on the valence geometry of the atom, formal charge density of the atom, and solvent approachability toward the atom. Each atom type is assigned a hydrophobic contribution value obtained from a database of 4500 chemicals. Manual input of known log Kowvalues (if available) is recommended, due to ATOMLOGP’S high deviation in calculated log K, values when compared to measured log KO,values. The data from this program are subsequently processed in the AUTOEST program. AUTOEST estimates the following seven physicochemical properties of a chemical: melting point, boiling point, water solubility, vapor pressure, Henry’s Law Constant, bioconcentration factor, and soil adsorption coefficient normalized for organic carbon content (K,). Physicochemical property values are either calculated (using several different approaches), drawn from the AUTOEST Measured Chemicals data set (containing measured values for 825 chemicals), or manually

428 J . Chem. If. Comput. Sci., Vol. 31, NO.3, 1991

input. After the program has finished running, a summary sheet listing the property values for the chemical and the source of the values (calculation technique or data set) is provided. These physicochemical property estimates may then be used as input values for the environmental models. SESOIL is a seasonal soil column model that simulates long-term vertical chemical transport and transformation in the unsaturated soil zone. This EPA model describes the following information for a soil column extending from the ground surface to the groundwater table: hydrologic cycle of the unsaturated soil zone, pollutant concentrations and masses (in water, soil, and air phases), pollutant migration to groundwater, pollutant volatilization at the ground surface, and pollutant transport in washload caused by surface runoff and erosion at the ground surface. These data may be estimated on a monthly basis for up to 99 years. The soil column may be composed of up to four layers, each of which may be subdivided into a maximum of 10 sublayers. The SESOIL model contains climatological data from 262 weather stations throughout the continental United States. Soil data are available for 14 common soil types ranging from sand to fine clay. A limitation of this model is its inability to estimate transport on a daily basis. The AUTOEST program data may be used to estimate the chemical data needed for SESOIL input. ~ ~ 1 is2an3analytical ~ groundwater transport model that predicts the transient spread of a contaminant through a groundwater aquifer and estimates the chemical concentration at specific times in three dimensions. Chemical release to groundwater may occur from a single point, area, line, or volume source and may be instantaneous, continuous, or of finite duration. Groundwater flow is assumed to be steady and uniform. Isopleth graphs of the contaminant plume may D The user may input data be generated from any A T ~ run. D from a previously created SESOIL run to generate A T ~ data. ENPART is a screening level model that uses the fugacity approach to calculate concentration or mass ratios in the three major environmental compartments: air, soil, and water. Fugacity models are relatively simple and have been used for nearly a decade, being applied to many different classes of chemicals and environmental conditions. Four compartment fugacity models have recently become available (bottom sediment being the fourth compartment) and could be used by the modeler who wants to carry this approach further. ENPART calculates compartment ratios by using the Henry’s Law Constant, octanol/water partition coefficient, and the soil adsorption coefficient of the chemical of interest. It is not appropriate for inorganic or high molecular weight compounds. As a screening model, ENPART is useful for determining the relative importance of the three environmental compartments. Absolute concentrations or masses cannot be calculated, only the ratios between compartments. Also, aqueous and atmospheric photolysis rates are the only environmental fate rates that can be estimated by the model. ENPART is an excellent starting point for the modeler who has limited information on the physicochemical properties of his compound or is just beginning to model a “new” compound. It is user-friendly and has extensive help screens. Sensitivity studies can be conducted efficiently. EXAMS-11 is an extremely complex surface water model created by the EPA to estimate both time-varying and steady-state concentrations of chemicals in almost any type of aquatic ecosystem. Owing to the complexity of the scenarios that can be modeled with EXAMS-11, nearly 100 input parameters can be used to describe the chemical, aquatic environment, and the chemical loads to the environment. This can be a formidable task for even the most experienced modeler. Likewise, the output from EXAMS-11consists of a series of tables with detailed data on the concentration of the chemical in each

COMPUTER SOFTWARE REVIEWS

compartment and phase (dissolved, sediment-sorbed, and biosorbed) as well as tables showing the relative importance of each fate process (Le., photolysis vs biodegradation) and persistence in the modeled ecosystem. The advantages of using EXAMS-11 in the RISKPRO environment are many. Most of the physicochemical properties needed as inputs for EXAMS-11 may be estimated and reviewed in the AUTOEST program before they are used as EXAMS-11 inputs. More importantly, however, the modeler may select the environmental fate processes he/she believes are critical to modeling his/her chemical and then concentrate his/her efforts on providing the best possible values for these inputs while allowing input values for less critical processes to be estimated by the program. Again, as with all of the RISKPRO programs, the help files are excellent, and in the case of EXAMS-11, default values are often explained. Finally, RISKPRO provides the EXAMS-11 modeler with several graphics options essential for visualization of output. With these enhancements, RISKPRO makes using EXAMS-11 a much more manageable task. PTPLU is a Gaussian plume dispersion air model used primarily as an intermediate step for detailed screening. PTPLU, an EPA model, is useful in determining the maximum 1-h concentration from a single point source for 49 internally generated combinations of wind speed and stability. Model results are most accurate within 10 km of the source. PTPLU is based on the following assumptions: (1) wind speed existing at stack top applies to both plume rise and dilution, (2) pollutant release is continuous at a user-specified rate, and (3) calculations are made assuming steady-state conditions. Although PTPLU is excellent as an initial screening model, it is not suitable for estimations involving complex terrain or multiple point sources. For these scenarios, ISCLT is more appropriate. ISCLT is a long-term dispersion air model that assesses the air quality impact of emissions from a variety of multiple point sources associated with an industrial source complex. ISCLT is a sector-averaged model that uses statistical wind summaries to calculate annual concentration or deposition values. Model results are most accurate within a 50-km range of the source. Multiple point sources may be chosen to do modeling. Data may be displayed by using either a polar or a Cartesian receptor grid. One especially attractive feature in ISCLT is the terrain utility, which accounts for variations in terrain height over the receptor grid. Three different source types are available in ISCLT: stack, area, or volume. The steady-state Gaussian plume equation for continuous source emissions is employed to calculate concentrations for stack and volume sources. The area source estimation uses a virtual point approximation. Because industrial sites are located in both cities and rural areas, ISCLT has an urban and rural setting option to enhance simulations. Another special feature of ISCLT is the regulatory default option. This option allows the user to set various model options and parameters to EPA-recommended values (ISCLT was also developed by the EPA). Once the ISCLT modeling results have been generated, these data may be used to set up and run exposure and risk estimations (see RISKTAB below). Annual population exposure, annual population dosage, and lifetime excess cases of cancer may be estimated for the specified population. The user should keep in mind, however, that ISCLT is not very effective for estimating short duration releases into the air, such as accidental releases. The graphics capabilities of ISCLT allow the user to create isopleths of estimated annual atmospheric concentration and to overlay the isopleths on Census BG/ED (Block Groups/ Enumeration Districts) generated maps that display the population density and distribution in the modeled locale. The

J . Chem. If. Comput. Sci., Vol. 31, No. 3, 1991 429

COMPUTER SOFTWARE REVIEWS user may select exact geographic locations by inputing the longitude and latitude coordinates of the desired site. The resulting output is displayed in a three-dimensional format, allowing visual inspection of the dispersion rate and concentrations of the chemical over the geographic region and/or population of interest. RISKTAB enables the user to generate tables of risk, dose, and hazard index estimates for single or multiple exposure pathways using the data from the model results. These estimates can therefore be made for very specific populations, generally the population within the modeled locale. For example, the predicted concentration of a chemical in a river (output from EXAMS-11) could be used to estimate the hazards and risks associated with drinking surface water from the river of interest. RISKTAB generates data using methods from the EPA document Risk Assessment Guidance for Superfund, Volume 1: Human Health Evaluation Manual (Part A). Three types of tables may be created by the RISKTAB option: carcinogenic risk estimations, chronic hazard index estimates, and subchronic hazard index estimates. All data produced

in these tables are averaged over the entire period of model simulation. As with the chemical estimation programs, users are given the option of entering their own data (either partially or completely) or not accessing any model data. SUMMARY RISKPRO is a collection of software used by exposure assessment professionals that allows the user to estimate physicochemical properties of a chemical, model the fate of that chemical in air, soil, and water, and then visualize these results without leaving the user-friendly environment of RISKPRO. It therefore offers modelers an environment where exposure assessments can be conducted with very little file handling and minimal knowledge of computers. We would especially recommended RISKPRO to the novice modeler who has some knowledge of environmental chemistry. For the experienced modeler, RISKPRO offers the convenience of reduced file handling, time savings, and untroubled visualization of model output.

Printer Fonts: Bitstream Facelift and HP Type Director 2.0 STEPHEN R. HELLER USDA, ARS, Building 005, Beltsville, Maryland 20705-2350 Received June 4, 1991

While laser printers are becoming quite common as the printer of choice for most offices and laboratories, the version that is usually purchased is generally a rather stripped down machine compared to its actual potential. Stripped down basic printers often have only 512K of memory, which is not sufficient to take advantage of their ability to print out complicated graphics like a chemical structure. Printing out a full page of graphics often requires 2-3 Mbytes of memory, which can easily add $500 to the cost of a laser printer. In addition, most of the basic models of laser printers lack the built-in capability to print out character sets in different sizes and styles. This review discusses two of the more popular and well-known software packages that provide for the ability to create more professional-looking documents and thus overcome the second problem. Both software packages can run on 5 12K of (printer) memory, but this may prove to be too little in some instances. The packages are the Bitstream Facelift and Hewlett-Packard (HP) Type Director, Version 2.0. Both packages perform the same basic tasks. They both allow one to create different styles of type (typefaces) in different fonts (typesizes). The main difference between the two is that the H P Type Director typefaces are loaded and saved on your hard disk, while the Bitstream Facelife typefaces are, for the most part, created on the fly as you go to print

your document. This means in general that, for the same typefaces, the Bitstream system takes up less space on your hard disk, but takes a little longer to actually print out. Bitstream also makes a Fontware software package which is similar to the HP Type Director system in that it does not use the Facelift font scaling technology. lnstallation of the packages was both simple and straightforward. I installed the software on Compaq, IBM, Epson, and TI computers with H P and Epson laser printers. The software 1 used for testing both software packages was Wordperfect, Version 5.1. Both instruction manuals are clear and easy to understand. Both companies provide hotlines for help. The HP line is open 11 hours a day, Monday through Friday. The Bitstream hotline is open 10 hours a day, Monday through Friday, and offers a FAX hotline (617-497-7514), primarily for non-USA users. I tried both and was pleased with the speed 1 was able to get help, and more importantly, the quality of the help. The H P hot line (208-323-2551) is not toll free, while the Bitstream hot line is (800-522-FONT). Type Director with Wordperfect supports font sizes (the physical size of the character being printed) from 4 to 200 points. The Bitstream Facelift package defaults to supporting font sizes from 1 to 32, but can be changed to range up to a point size of 999.