MISKAM Manual

MISKAMManual for Version 6On behalf ofgiese-eichhornenvironmental meteorological softwareAm Spielplatz 255263 WackernheimTel 06132-62947c© 2011 Dr. J. EichhornAll trademarks used in this manual belong to the registered owners, respectively.ContentsPreface 6Software license 8Warranty terms 101 Short information 112 Introduction 142.1 About MISKAM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142.2 Why prognostic modeling? . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152.3 Scope and limitation of application . . . . . . . . . . . . . . . . . . . . . . . 152.3.1 Scope of application . . . . . . . . . . . . . . . . . . . . . . . . . . . 162.3.2 Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183 Theory 203.1 Preliminary note . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203.2 The momentum equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203.3 The turbulence model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213.3.1 Calculation of the diffusion coefficients . . . . . . . . . . . . . . . . . 223.3.2 The prognostic equations for E and e . . . . . . . . . . . . . . . . . . 233.4 The splitting method according to Patrinos . . . . . . . . . . . . . . . . . . . 243.5 The dispersion model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253.5.1 The prognostic equation . . . . . . . . . . . . . . . . . . . . . . . . . 2512 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Contents3.5.2 Sedimentation and deposition . . . . . . . . . . . . . . . . . . . . . . 253.5.3 Momentum sources . . . . . . . . . . . . . . . . . . . . . . . . . . . 253.6 Initial and boundary conditions . . . . . . . . . . . . . . . . . . . . . . . . . 253.6.1 Initializing the flow model . . . . . . . . . . . . . . . . . . . . . . . . 253.6.2 Boundary conditions for the flow model . . . . . . . . . . . . . . . . . 263.6.3 Initialization of the dispersion model . . . . . . . . . . . . . . . . . . 263.6.4 Boundary conditions for the dispersion model . . . . . . . . . . . . . . 274 Numerical methods 284.1 Discretization and grid configuration . . . . . . . . . . . . . . . . . . . . . . 284.2 Treatment of the advection terms . . . . . . . . . . . . . . . . . . . . . . . . 294.2.1 Momentum equations . . . . . . . . . . . . . . . . . . . . . . . . . . 294.2.2 Advection of scalar quantities . . . . . . . . . . . . . . . . . . . . . . 294.3 Treatment of the diffusion term . . . . . . . . . . . . . . . . . . . . . . . . . 304.4 Solution of the Poisson-equation . . . . . . . . . . . . . . . . . . . . . . . . 305 Operating instructions 325.1 Conventions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 325.2 Hardware- and software requirements . . . . . . . . . . . . . . . . . . . . . . 325.3 The MISKAM-CD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 325.4 Installation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 335.4.1 Extracting the program files . . . . . . . . . . . . . . . . . . . . . . . 335.4.2 Installed files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 335.5 The configuration files *.INP . . . . . . . . . . . . . . . . . . . . . . . . . . 355.5.1 Structure of the configuration files . . . . . . . . . . . . . . . . . . . 355.5.2 Precision requirements . . . . . . . . . . . . . . . . . . . . . . . . . . 385.5.3 To save the minimal distance to the ground . . . . . . . . . . . . . . . 385.6 Additional configuration files . . . . . . . . . . . . . . . . . . . . . . . . . . . 395.6.1 Flow-through . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 395.6.2 Vegetation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 395.6.3 Momentum-containing sources . . . . . . . . . . . . . . . . . . . . . 40Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35.7 Programming and meteorological control parameters . . . . . . . . . . . . . . 415.7.1 The initialization file MISKAM.INI . . . . . . . . . . . . . . . . . . . 415.7.2 The control file MISKAM.BND . . . . . . . . . . . . . . . . . . . . . 455.8 Utilization steps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 465.8.1 Initialization and program start . . . . . . . . . . . . . . . . . . . . . 475.8.2 Terminating the program . . . . . . . . . . . . . . . . . . . . . . . . 475.8.3 Result output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 475.8.4 Control output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 485.9 Help program . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 495.9.1 KONFIG: Interactive setting of configuration files . . . . . . . . . . . . 495.9.2 DATXYZ: Interactive evaluation of result files . . . . . . . . . . . . . 495.9.3 MISVIS: Visualization of MISKAM results . . . . . . . . . . . . . . . . 506 Literature 54List of Figures4.1 Grid configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285.1 Example for MISKAM-configuration . . . . . . . . . . . . . . . . . . . . . . . 365.2 Definition of a flow-through area . . . . . . . . . . . . . . . . . . . . . . . . 395.3 Definition of cell areas with vegetation . . . . . . . . . . . . . . . . . . . . . 405.4 Definition of momentum containing sources . . . . . . . . . . . . . . . . . . . 415.5 Example of a MISKAM initialization file . . . . . . . . . . . . . . . . . . . . . 425.6 File MISKAM.BND to set the lateral boundaries. . . . . . . . . . . . . . . . 465.7 Control setting of time series . . . . . . . . . . . . . . . . . . . . . . . . . . 495.8 MISVIS-example 1: building configuration . . . . . . . . . . . . . . . . . . . . 515.9 MISVIS-example 2: Wind fields . . . . . . . . . . . . . . . . . . . . . . . . . 525.10 MISVIS-example 3: Mass concentrations . . . . . . . . . . . . . . . . . . . . 534List of Tables5.1 Content of the self-extracting archive M?SETUP.EXE . . . . . . . . . . . . 345.2 Kind of ground and roughness lengths in MISKAM . . . . . . . . . . . . . . . 375PrefaceThe flow and dispersion model MISKAM has recently become a highly respected expert toolfor evaluations in sectors as e.g. road planning, environmental impact studies, and air hygiene.This manual describes the MISKAM model version 5 (latest published version 6.1), which wasimproved and extended in comparison to the previous releases 3.x, 4.x and 5.x:• Modified stationarity criteria for flow and dispersion calculations, resulting in a moreconsistent convergence.• Time-step-splitting for calculating turbulence variables. All known problems with respectto convergence in complex obstacle configurations could be eliminated by processing eachtime-step in two halves.• Two-dimensional calculations with only one grid cell set in the y-direction, resulting inan greatly reduced calculation time, for instance for street canyons.• Change to programming without fixed array sizes in all program sections. This eliminatesthe necessity to distribute different program versions depending on RAM availability.• More consistent one-dimensional initialization, with more realistic z0-dependencies ofthe wind and turbulence profiles.• Automatic internal generation of lateral up- and downstream flow zones. An equidistantor a spread grid can be used. Further, it can be distinguished between boundary zoneswithout obstacles or a conversion to the boundaries of the last obstacle belonging to theinner model area.• Optional setting of a vertical flow velocity for point sources. Stack emissions can thusbe very realistically simulated.• The influences of vegetation (flow deceleration and additional turbulence production)can optionally be considered.6Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7• Controlled abortion of the simulation with a dump of partial results.• Various possibilities of additional control dumps.• Increased maximum number of grid points for vertical direction.• Improved adcetion calculation by optional use of McCormack scheme for velocity com-ponents and MPDATA algorithm for all scalar quantities.The model was vastly verified and validated while revising the code. The test calculationsmainly oriented at the guidelines of the VDI-Code 3783/9 ”Environmental meteorology -Prognostic microscale wind field models - Evaluation for flow around buildings and obstacles.”All accuracy requirements of the guideline are fulfilled by MISKAM’s versions 5 and 6, sum-maries of the results have been published (Balczo´ and Eichhorn, 2009; Eichhorn and Kniffka,2010).More tests were run recalculating previous research by other authors, for instance within thePEF-projects.MISKAM was mainly programmed in Fortran90. The compiler Lahey/Fujitsu Fortran95 Ver-sion 7.2 was used to generate the executables. A 32Bit-operating system is compulsory, werecommend using Windows XP Professional or Windows 7. A usage under Linux or variousUnix-systems is possible, some adjustments in the source code are needed, however.This MISKAM-manual is structured as follows:• In the first part, short information is given with respect to the MISKAM model, accordingto the requirements of the VDI-Code 3783/9.• The ensuing chapter describes model concepts and usage limitations.• The theory, upon which the model is based, as well as the numeric transformation aresubject of chapters 3 and 4.• Chapter 5 describes the installation as well as the usage of the model and of the helpprograms.Calculations of verifications and validations are documented in scientific publications and re-sults are openly published on the homepage of Lohmeyer GmbH & Co KG.The data and results of all examples given in the manual are on the MISKAM-CD.Wackernheim, autumn 2011Dr. Joachim EichhornSoftware licenseThis software product is protected by copyright and international copyright contracts as wellas other laws and contracts on intellectual property. This software product is not sold, it islicensed.You are authorized to install or use one copy of the software product or any prior version whichis suitable for the same operating system on one single computer. The original user of thecomputer on which the software is installed is allowed to exclusively use a second copy on hishome computer or his notebook.You are also authorized to save or install a copy of this software product on a storage unit, e.g.a network server, however only, if it is solely used to install or execute this software product viaan internal network to your computers. However, you have to purchase one license for eachcomputer on which the software product is installed or executed from the storage unit. Onelicense is not allowed to be used by various users or installed on different computers.You are not authorized to reconvert (reverse engineering), decompile or disassemble the soft-ware product. However, this is only valid as long as the actual law does not explicitly allowsuch a possibility.The software product is licensed as a single product. You are not authorized to separate thecomponents in order to use them on various computers.You are not authorized to rent or to lease the software product.You are authorized to permanently transfer all rights of the license contract, provided that youdo not keep any copies and that you transfer the complete software (including all components,the media and the printed materials, all updates of this license contract as well as the certificate,insofar as it is used). The receiver has to agree to the conditions of the license contract. Ifthe software product is an update, all previous versions must be transferred too.Despite other rights, giese-eichhorn is entitled to cancel the license contract, if you violatethese rules and conditions. In this case, you have to destroy all copies of the software productand its components.8Software license . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9If the software product is an update of another product of giese-eichhorn or any otherdistributor, you are only entitled to transfer the software product in combination with theupdated product except if you destroy the updated product. If the software product is anupdate of a giese-eichhorn product, you are entitled to use the updated product only inagreement with this license contract. If the software product is a component-update of asoftware product which you obtained licensed as a single product, the software product canbe used and transferred only as a part of the single product package. It is not allowed to beseparated to be used on several computers.The property and the copyright of this software product, the printed materials and all copiesof the software product are with giese-eichhorn or its distributors. The software product isprotected by the copyright and regulations of international contracts. The software product isthus to be treated as any other material protected by copyright with the exception that thissoftware product (a) is copied once for safety and archiving purposes or (b) is installed on onesingle computer while the original was stored for safety and archiving purposes. You are notentitled to copy printed materials belonging to the software product.It can be that you obtain the software on several data carriers. You are only entitled to useone medium fitting your computer independently of the type or size of the obtained media.You are not entitled to lend, rent, lease or transfer any other medium to another user exceptas a part of a permanent transfer of the software product (as described above).Warranty termsgiese-eichhorn guarantees within 90 days after delivery that the medium/media (a) is free offailures and (b) functions within normal usage as described in the manual. The guarantee isgiven by giese-eichhorn as producer of the software. Any legal guarantee or liability claimsagainst distributors of the software are limited or impaired by these terms.In case of a warranty claim, you are entitled to receive a replacement or a supplementationof the missing parts of the software, which do not satisfy the limited warranty term of giese-eichhorn and which are returned to giese-eichhorn together with a copy of the bill. Thislimited warranty is not valid if the disfunction of the software is due to accident, misuse orfaulty usage. giese-eichhorn gives a warranty of 30 days to replace the software or of theremaining warranty time, depending which one is longer. giese-eichhorn excludes any furtherliability with regard to the software and the additional literature (printed or in electronic form).Neither giese-eichhorn nor the distributors of giese-eichhorn products are liable to any kindof damage (damages by lost earnings, business interruptions, loss of business information ordata or any other financial loss are included without any restrictions), which occur by usingthis software or by the user’s mishandling. This is also valid, if giese-eichhorn was informedabout the possibility of such a damage. In any case, the liability of giese-eichhorn is limitedto the amount of the actual sales price. This exclusion is not valid for damages which werecaused by intention or gross negligence by giese-eichhorn. Claims due to legal non-waivableregulations based on the product liability also remain unaffected.If individual clauses are invalid due to legal regulations, the further terms are not affected bythese regulations.101 Short informationModel nameMISKAMVersion6.1 (as of February 2011)AuthorDr. Joachim EichhornInstitut fu¨r Physik der Atmospha¨reJohannes Gutenberg Universita¨t55099 MainzTel.: 06131-3922866Fax: 06131-3925567Email: eichhorn@uni-mainz.deModel typeNon-hydrostatic, three-dimensional, obstacle-dissolving flow model;three-dimensional Eulerian dispersion model.Scope of application, size of areas and resolutionFlow and dispersion calculations in build-up areas,area sizes up to about 1000 × 1000 × 300 m,grid resolution of approximately 1 to 10 m.Limitations of applicationNot usable for steep topography, unstable thermal stratification andoversaturation (condensation).Solution algorithmsModel equations:• Three-dimensional momentum equations, non-elastic boussinesq-approximated1112 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1. Short information• E-e-turbulence model• Patrinos-splitting• Advection equation of diffusion for mass concentrationsNumeric solutions using the following schemes:• Forward differences in time• Upstream-advectionoptional McCormack scheme (momentum) and Smolarkiewicz scheme (scalars)• ADI-process for diffusion equations• SOR-process for Poisson-equationInput sizesModel geometry: Cartesian coordinates, orientation of the model areas, position andheight of buildings, roughness length of the terrain (or: type of surface per grid) and ofbuilding areas, optional flow-through building areas (e.g. passageways or arcades)Meteorology: wind speed, wind direction, stratificationControl parameter to run the program: number of time-steps to be calculated, abortioncriteria, control of the advection calculationOutput sizesFlow model: Three-dimensional arrays of the Cartesian wind components, of the dynamicpressure disturbance, of the turbulent kinetic energy, of the energy dissipation and ofthe diffusion coefficientsDispersion model: mass concentrations, if needed the dry deposition ratePrevious EvaluationsThe following verifications of the model were performed:• Evaluation of the flow model according to the draft of the VDI-guideline 3783 sheet9 (prognostic microscale wind field model), internal verification of consistency aswell as comparison with wind tunnel data.• Verification the dispersion model with wind tunnel data and natural data (Go¨ttingerStraße, Hannover).Hard- and Software requirementsStandard PC (starting with Intel Pentium-Prozessor or AMD-Athlon),64 MB RAM, approx. 10 MB hard disk space for the standard installation, some hundredMB’s for the model results,32 Bit-operating system (WindowsXP or newer, Linux-versions upon request). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13AvailabilityThe following versions can be obtained:• Standard version (executable, manual, help programs), operating under WindowsXPor later. Obtainable through giese-eichhorn, Wackernheim.• WinMISKAM (Standard version + Windows-surface), obtainable through LohmeyerGmbH & Co. KG, Karlsruhe.• Soundplan-Modul, obtainable through Braunstein + Berndt, GmbH, Back-nang.Literaturesee Chapter ??.2 Introduction2.1 About MISKAMThe prediction of expected, traffic-related immissions will gain greater importance in the con-text of current discussions regarding new legal guidelines (for instance ”Bundes-Immissions-schutzverordnung”, VDI-guidelines).No planning tasks, be it in the urban sector or in road design, are practically imaginablewithout considering the immission load. Measuring the most important air contaminantsis very important to evaluate the existing pollution. However, measurements, due to theirrelatively high costs, cover only a small area, and, in addition, can only be performed duringa short time-span.Since several years, numeric methods have been developed as a supplement and an upgradeto the existing measurements. The fast technical development of the hardware sector allowsus to run elaborate numeric models on a standard PC nowadays. High-end computers wererequired for these jobs until recently. It is even more surprising that current guidelines, as theTA-Air, only demand Gaussian models even though their weaknesses and limitations are wellknown. Just lately, the official interest has shifted to more complex models, which mainlyconsider the dispersion conditions of developed terrain.The MISKAM model (microscale climatic and dispersion model) is one of the more sophisti-cated models with respect to its physical content within a wide range of available models in themeanwhile. It has been developed at the Institute for Atmospheric Physics at the Universityof Mainz. This institute had been working on the development of regional and local climateand dispersion models since several years.MISKAM is suited for dealing with small-scale dispersion processes (typical model dimensionsof several 100 m). Therefore, MISKAM is especially useful for tasks previously mentioned(road and urban planning), because it considers mainly the physical processes influencing thetransport of pollutants within the direct environment of buildings.142.2. Why prognostic modeling? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15MISKAM is a three-dimensional non-hydrostatc numeric flow and dispersion model for smallest-scale forecasts of wind behavior and immission concentrations, e.g. in roads, as well as up tourban quarters. It was originally planned to treat microclimatic problems (Eichhorn, 1989). Itwas only after users expressed the desire for a PC solution to predict the immisssion in roads,that the current model version was further developed. MISKAM allows for the explicit treat-ment of buildings in the form of rectangular form structures, so that the particular flow aroundbuildings can be realistically modeled. It was further desired to develop a model of a highphysical standard avoiding as much as possible the usage of empiric-diagnostic relationships.The physical basis of MISKAM are the three-dimensional equations of momentum (so-calledprimitive equations) to simulate the flow conditions as well as the advection-diffusion-equationbased on the dispersion calculation of density-neutral substances.2.2 Why prognostic modeling?Even when using the most modern hardware resources, the use of prognostic flow and dispersionmodels is still combined with considerable computer time in comparison to the diagnosticmodels. Diagnostic flow models first estimate the wind field considering empirical assumptionsregarding the flow pattern (Lee-circulations, flow separation at edges, etc.). This estimatedfield is then freed of flow divergencies by iterations. Mass-conserving wind fields are thusobtained also for large areas with many flow obstacles after a relatively short computationaltime. To prognostically calculate a wind field, as done by MISKAM, requires 10 to 20 timesmore computational time. However, various arguments support prognostic modeling:• Prognostic models generate information on the wind and turbulence field, while diagnos-tic models do not allow to consistently calculate the turbulence. Especially the reciprocalinfluence of wind and turbulence is not considered with diagnostic models.• Complex structures of obstacles are combined to single obstacles in the diagnostic model.However, the interaction of flow effects of single obstacles do not correspond with theirreal flow conditions.• Diagnostic models do not consider the stratification effects, which are important for airhygienic investigations, especially the effects of a stable stratified atmosphere.2.3 Scope and limitation of applicationThe possibility to gain secure conclusions about the expected atmospheric load via a validatednumeric model seems to be promising. Cost-intensive measuring campaigns can be reduced.16 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2. IntroductionFurther, numeric models also offer an area-wide, respectively space-wide construct of informa-tion being much more detailed regarding single points of immission then legally required.As long as the numeric model can be used reasonably comfortably according to our presentstandards, the perspective is tempting at the first glance so as to obtain those results withina few steps on a standard PC which previously only could be obtained via complex interactionof measurements, calculations and meteorological knowledge.This way of thinking can be dangerous however, as shown. If a model – for instance themodel software – takes over too much of the user’s work, it can happen that the user overlyemphasizes the model results and does not critically analyze them any longer. It is thereforeimportant to point out the limitations of a numeric model very clearly. This is the purpose ofthis chapter.2.3.1 Scope of applicationMISKAM can be used for the following tasks:• Calculation of quasi-stationary wind fields in the environment of isolated buildings orwithin a complex structured urban development. It has to be kept in mind that structureddevelopments can only be considered with a limited accuracy depending on the selectionof the grid of discretization.• Simulation of the dispersion of density-neutral, non-reactive substances with a randomlyassumed source distribution in previously calculated wind fields.• Comparison of the calculated arithmetic mean and percentile of concentrations with theguideline and threshold values.The dynamic effects of various ground conditions can be considered via the roughness lengthfor the calculation of wind fields. The degree of roughness of the various grids is determinedby setting a constant value for the whole model area or by entering a two-dimensional field ofparameters assigning a certain type of surface (low or high vegetation, asphalt, non-explicitlyassigned development, etc.) to each grid. In the last case, the classification of the lengths ofroughness to the various surfaces is model internally determined.The thermal stratification can be considered as a further critical value for the flow behavior. Itwill be considered as being constant in the model area and is specified as input parameter bythe vertical gradient of the potential temperature. The influence of the thermal stratificationconsists of a reduction of the turbulent exchange for stable conditions as well as an increasefor unstable conditions.2.3. Scope and limitation of application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17The simulation of the pollutant dispersion also considers the effect of the sedimentation besidesthe source dispersion, which also allows – at least approximately – the interpretation of nondensity-neutral substances as well as deposits. Both processes are recorded by setting constantcharacteristic velocities.The sedimentation velocity is added to the vertical component of the wind field for the cal-culation of advection. The disposition velocity determines this portion of the transportedsubstance which settles on the ground or on the surface of buildings out of the atmosphere.Both velocities have to be considered as material constants and have to be set by the user.Due to the possible usages mentioned above, the model MISKAM becomes a multi-functionaltool for urban planning and road design.However, the model should not be interpreted as a black box, its usage demands an intensiveamount of work and thinking. This is intended to avoid any uncritical usage of the model.The following steps for numeric model simulations are thus not automatically taken care offby MISKAM:• Generating an optimal grid of discretization specific for the area of investigation.• Positioning of sources and determination of the source type.• Setting emission rates.• Construction of statistics based on several MISKAM runs.• Conversion of calculated fields of immission into planning-relevant parameters (yearlyaverage mean, 98-percentile, etc.).The program KONFIG can be used to convert the selected grids, the containing buildings, aswell as the sources of pollutants to a suited input file. It asks for all required values, performs athorough verification for the data of its plausibility, and finally generates the MISKAM-readableinput file.Sophisticated instruments are available for some of the above mentioned tasks, as for instancethe manual of emission factors which was developed on behalf of the Federal EnvironmentalAgency. In addition, various implementations of the model also exist in the form of compre-hensive program packages to work more comfortably with MISKAM.Lohmeyer GmbH & Co. KG (Karlsruhe) offers the program WINMISKAM, adding a comfort-able Windows interface to the MISKAM kernel. This can be used to generate the neededinput files, to produce the diagrams of the results, to evaluate the statistics, and to producethe most important parameters.18 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2. IntroductionMISKAM also was integrated in the widely-used program package SoundPLAN (Braunstein +Berndt GmbH, Backnang), allowing the SoundPLAN-users to work with their familiar surfacefor data entry and evaluation.More information about both implementations can be directly obtained by the distributors orby giese-eichhorn.2.3.2 LimitationsBesides the requirements mentioned in the previous chapter, several further, principal limita-tions exist, because certain actions cannot be simulated in MISKAM. Moreover, some limita-tions result from the adopted modeling assumptions as well as their numerical implementation.• Thermodynamic processes (energy transformation at the surface of the terrain, alongwalls, and roofs of buildings, as well as thermal dispersion, buoyancy, water balance) arenot considered, because this would cause a dramatic increase of computational time anddisk space which cannot any longer be handled by a standard PC.• MISKAM does not consider any chemical processes. The reactions of NOx to NO2are particularly interesting for road traffic, for which various empirical relationships areavailable. This was not considered in the model equations, because the general validityof this relationship is questionable and a later consideration is possible by the appropriateconversion of the concentrations of immission.• The approximation of slanted roofs by step shaped structures requires a critical evaluationof the results for the near field of the roofs (see Eichhorn, 2003).Considering flow over buildings it is noticed that the height of recirculation zones isgenerally underestimated by MISKAM. This is a result of the E-e turbulence closurewhich doesn’t permit a realistic simulation of the flow separation at windward edges.This flaw of E-e models is well-known and accepted, as the model otherwise representsa reasonable compromise of accuracy and applicability.For dispersal simulations, e.g. for traffic induced emissions this property of the turbu-lence model is of no relevance. For emissions at roof level, however, it is not advisableto use MISKAM as long as the emission sources are not definitely located outside therecirculation zone. In particular, the model is not suitable to specify the optimal mini-mum height of an emission source. For this purpose, alternative data, e.g. from windtunnel measurements or results from more elaborate models (LES) have to be adopted.The usage of the model is restricted by neglecting the thermodynamics because certain condi-tions exist where thermodynamic influences on the flow field cannot be omitted (for instance2.3. Scope and limitation of application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19street canyons with intensive asymmetric solar radiation). However, it can be assumed thatthese effects can hardly be identified in the yearly mean. Thus, one of the most importantareas of usage is not effected by this restriction.The thermic effects can generally be excluded for calculating episodes, since these are mainlycalculations of air-hygenic critical cases (so-called worst case studies). The most importantthermal factors for near-ground immissions, stratification, is considered while calculating theturbulent diffusion coefficient.This listing could lead to the assumption that the usage of MISKAM is very limited. However,its spectrum covers practically all air-hygenic aspects needed by planners as annual averagevalue, percentile, and the maximum load. However, MISKAM expects a competent and diligentuser producing finally reliable results for a particular scenario.3 Theory3.1 Preliminary noteEvery numerical atmospheric prognostic model basically consists of a linked system of dif-ferential equations to predict the conditional variables (wind, temperature, air composition).These differential equations follow from known physical conservation equations of momentum(? momentum equations), mass (? continuity equations) and energy (? energy equation).Various simplifications and omissions are possible, depending on the complexity of the taskand the size of the model area.MISKAM is used to calculate wind fields and immission dispersions. Therefore, MISKAM doesnot consider the thermal exchange and the hydrologic process. This reduces all the prognosticvariables to components of wind vectors and mass concentrations of the substances underinvestigation.The elimination of sound waves, and the density as independent prognostic value are performedwith the help of the Boussinesq-approximation. In addition, the equations are averaged,resulting in variables which have to be considered as microturbulent means. As a consequenceof the averaging, the equations contain subscale processes for the turbulent transportationof momentum, heat and mass. The used closure of turbulences additionally requires theprognostic computation of turbulent kinetic energy as well as energy dissipation.A detailed derivation of the system of equations is not given here. Details of the Boussinesq-approximation as well as of the averaging procedure are given by Eichhorn (1989).3.2 The momentum equationThe prognostic system of the flow part of MISKAM consists of the Cartesian components ofthe Boussinesq-approximated momentum equations. Due to the small size of the model area,Coriolis force is neglected. In addition, buoyancy is also not considered.203.3. The turbulence model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21The turbulent transportation of momentum is calculated by a first-order closure without anydifferences between horizontal and vertical exchange coefficients.Using these assumptions, the momentum equation is the following:?ui?t+?ukui?xk= - 1?0?p'?xi+??xk[Km(?ui?xk+?uk?xi)]- cdnLvui (3.1)wherexi = x, y, z Cartesian coordinates [m]ui = u, v, w Cartesian components of the wind vector [m/s]t time [s]?0 constant reference density at near-ground atmosphere [kg/m3]p' dynamic pressure disturbance [Pa]Km exchange coefficient of momentum [m2/s]cd resistance coefficient [-]n ”degree of vegetation coverage” [-]L one-sided leaf area density (LAD) [m2/m3]v wind velocity [m/s]The last term on the right side describes the deceleration of the flow due to vegetation, i.e.the friction against leaf areas within a grid cells.The normal convention of sums was applied in the momentum equation, which means thatequal indices have to be summed up from 1 to 3.This system of equations is enhanced by the requirement of a non-divergent wind field, whichreplaces the continuity equation for the total mass:?u?x+?v?y+?w?z= 0 (3.2)To assure non-divergence at each time-step, an elliptical differential equation has to be solvedfor the dynamic pressure disturbance (condition of compatibility) in addition to solving theprognostic equations. It is only due to this additional computational load that economic time-steps can be used, since the requirement of a non-divergent wind field enables the eliminationof sound waves from the system.3.3 The turbulence modelThe K-model, which was still provided in previous versions, and which is based on the classicalboundary layer theory, calculated the required diffusion coefficients diagnostically with the help22 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3. Theoryof the three-dimensional wind field, and the specified thermal stratification, as well as the so-called mixing length.In the context of a PEF1-program promoted project (Ro¨ckle and Richter, 1995), validationruns were performed for several models available at that time, amongst them was a previousversion of MISKAM. The comparison was based on wind tunnel data on the one hand flowingaround a U-shaped building and on the other hand through a complex arrangement of buildingsrepresenting a section of the BASF factory site in Ludwigshafen. The results of MISKAMwere partially unsatisfactory concerning the flow around the single building due to the abovementioned diagnostic treatment of turbulences. Therefore, a more sophisticated method ofclosures, the so-called E-e-closure, was integrated for further developing the model. Theobsolete diagnostic closure is no longer used by MISKAM since version 4.Due to the E-e-closure, the model results are based on a much more sound physical foundation.The exchange and diffusion coefficients are now obtained from the local values of the turbulentkinetic energy and the turbulent energy dissipation. These fields have to be calculated bysolving two additional prognostic equations.The mixing problem existing in the K-model is bypassed with this closure. The distance tothe ground of the single grids is only needed to initialize the turbulent energy dissipation. Inaddition, several empirical constants have to be specified as external values. Broadly acceptedvalues can be obtained from literature.Instationary phenomena, as for instance the instationary separation of lee turbulence and theseparation of so-called wakes, are not considered by this method. In principle, there aremodels available for this simulation, however, the time of computation needed does not makethem feasible for a standard PC. The E-e-closure seems to be appropriate to simulate quasi-stationary conditions corresponding to a temporal mean of the flow regimes (see, for instance,Paterson and Appelt, 1986, 1989). However the computational complexity is significantlyhigher than with the K-model.3.3.1 Calculation of the diffusion coefficientsThe diffusion coefficients for the transportation of momentum, Km, are calculated as followsfor the E-e-closure:Km = cµE2e(3.3)whereE turbulent kinetic energy [m2/s2]e turbulent energy dissipation [m2/s3]cµ = 0.09 empirical constant [-]1PEF = Projekt Europa¨isches Forschungszentrum3.3. The turbulence model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23For simplification, the relationship valid for neutral stratification is used for the diffusion coef-ficient for heat KhKh = 1.35Km (3.4)3.3.2 The prognostic equations for E and eThe following prognostic equations have to be solved to determine the kinetic energy ofturbulence:?E?t+?ukE?xk=??xk(Km?E?xk)+Km(?ui?xk+?uk?xi)?ui?xk-Kh gT0?T?xkd3k - e+ cdnLv3 - 4cdnLvE (3.5)?e?t+?uke?xk=??xk(Kms?e?xk)+ c1eE[Km(?ui?xk+?uk?xi)?ui?xk-Kh gT0?T?xkd3k]- c2 e2E+32eEcdnLv3 - 6cdnLve (3.6)wherec1 = 1.44 empirical constant [-]c2 = 1.92 empirical constant [-]s = 1.3 empirical constant [-]The numerical values of the empirical constants correspond to the ”classical” E - e-model(for instance Rodi, 1980). The validity of atmospheric calculations of turbulence was verifiedfor instance by Ramanathan (1995).The last two terms contain the parameterization of the influence of vegetation. The positiveterms describe the increased mechanical production of turbulence energy and dissipation be-cause of leaves. The last term was derived from a suggestion by Greene (1992), leading toa significantly better correlation between simulated and observed wind fields of trees (see forinstance Lauerbach and Eichhorn, 2004).It should be noted that this closure with the required stationarity can only lead to plausibleresults for neutral and stabile thermal stratification. With unstable stratification, the thermalproduction leads to a continuing increase of the turbulence energy and of the diffusion coeffi-cient. But this contradicts the requirement of quasi-stationary conditions, like in reality, wherea unstable stratification cannot be kept upright for a prolonged time due to the increasedexchange.E and e are diagnostically calculated viaE1 =u2*vcµ(3.7)24 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3. Theorye1 =u3*?z1(3.8)in the lowest grids with the help of the friction velocity u*. z1 is the height of the first gridabove ground or building. Equal relationships are formally applied for vertical walls, however,z1 is replaced by the roughness length of the building wall.These boundary conditions are strictly valid only for neutral stratification. The influence ofthermal stratification on the turbulence along the horizontal boundary areas is taken intoaccount via the calculation of u*. This is calculated byu* =| v||(?) |Cm(?+z0z0, ?+z0?)(3.9)Cm is the Clarke-function for momentum, whose values are taken from existing tables (Panhansund Schrodin, 1980). ? is the minimal distance of the considered grid point to the fixedboundaries of the model, ? is the stability length. Details of the Clarke-function are givenby Eichhorn (1989). To calculate u* along building walls, a neutral stratification is assumedreferring to a logarithmic wind profile vertically along the wall.3.4 The splitting method according to PatrinosThe splitting method of Patrinos und Kistler (1977) is used to solve the prognostic system.Here, the momentum equations are first numerically solved by neglecting the pressure distur-bances, resulting in temporary wind components u˜, v˜, w˜:?u˜i?t=?ui?t+1?0?p'?xi(3.10)The following equation is obtained by applying the divergence operator and temporal forwarddiscretization?u˜n+1k?xk- ?u˜nk?xk=?t?0?2p'?x2k(3.11)where ?t is the time-step, n and n+1 denote time t and t+?t. The compatibility conditionshave to be fulfilled before starting with a time-step. Therefore, the second term on the leftside is zero and the following Poisson-equation remains for the pressure disturbance?2p'?x2k=?0?t?u˜k?xk(3.12)After solving and inserting intoui = u˜i?t?0?p'?xi, (3.13)the desired non-divergent velocity field is obtained.3.5. The dispersion model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253.5 The dispersion model3.5.1 The prognostic equationThe dispersion model mainly consists of the prognostic equation for a density-neutral airconstituent with the mass concentration m:?m?t+?ukm?xk=??xk(Kh?m?xk)+Q (3.14)Q refers to the sum of the sources and sinks of the considered air constituents.As normally, the heat exchange coefficient, Kh, is also used for the mass transportation.3.5.2 Sedimentation and depositionA constant sedimentation velocity can be set to consider the sedimentation of substances witha greater density than air. It is added to the vertical wind for the advection calculation.Dry deposition on horizontal planes can be considered by setting a deposition velocity. Theamount which is deposited on the ground per time unit is proportionally added to the depositionvelocity and the mass concentration in the grid cell located above.Both velocities are set as material constants of the considered air additives, values can beobtained from literature.3.5.3 Momentum sourcesA fixed vertical velocity (velocity of air emission, for instance out of a stack) can be set forpoint sources. This is already incorporated in the flow calculation, so that wind and turbulencefields close to a stack outlet can react to the additional momentum. This results in muchmore realistic pollutant plumes than without considering the emission velocity or applicationof the effective height of the source.3.6 Initial and boundary conditions3.6.1 Initializing the flow modelThe three-dimensional model calculations are preceded by a one-dimensional initialization. Forthis purpose, wind and turbulence profiles are calculated up to a height of 2000 m.26 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3. TheoryTo implement the MISKAM model thereafter, the three-dimensional profiles are converted insuch a way that the simulated wind velocities at the height of the used anemometer correspondto the set value at all times.Before starting the 3D-simulation, the profiles of the wind velocity and the turbulence en-ergy are homogeneously transferred to the three-dimensional model area. To incorporate theelevated energy dissipation around buildings already in the initial distribution, an increase,depending on the distance to the model’s boundaries, is applied while implementing the dissi-pation:e3D(x, y, z) = e1D(z)z?(x, y, z)(3.15)e3D is the energy dissipation used in the initial distribution of the 3D-simulation, e1D is theone-dimensional profile which was simulated in advance. ? is the minimal distance of eachmodel point to the solid model boundary.3.6.2 Boundary conditions for the flow modelThe profiles resulting from the 1D-initialization of the wind components and the turbulencevariables are temporally kept constant on the inflow areas and the upper boundary.So-called no-flux-boundary conditions are valid on the lateral outflow boundaries which meansthat the disappearance of the normal gradients on these boundaries is required.The outflow boundary values are corrected for the solution of the Poisson-equation of thedynamic pressure disturbance in such a way that the conservation of the total mass is assuredfor the entire model area. An extensive discussion of the boundary conditions is given byEichhorn (1989) as well as by Eichhorn et al. (1997).All wind components disappear on the lower boundary as well as along the building walls.Consequently, the disappearance of the pressure gradient perpendicular to the relevant planesis required there as well as at the upper boundary.The values of the lower boundary of the diffusion coefficients are - as already mentioned -calculated by applying the Clarke-function. This also determines the corresponding frictionvelocity u*, which again can be used for the calculation of the boundary values of turbulenceenergy and dissipation.3.6.3 Initialization of the dispersion modelA wind field, which was calculated before, is read to initialize the dispersion calculations,including its exchange coefficients. The mass concentrations are set to zero at the start of thesimulation.3.6. Initial and boundary conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27An optionally added background level is not considered during the simulation, it is only addedto the results.3.6.4 Boundary conditions for the dispersion modelInflow and outflow grid cells are treated differently in the lateral boundary conditions of massconcentrations. Clear air from outside of the model area is assumed at the inflow boundarieswhich means that the concentration value is kept at its initial value of 0.The boundary value of the inner field is extrapolated for outflow grid cells with the assumptionthat the concentration gradient along the area normal remains constant.In the case of a non-disappearing disposition velocity, the mass concentration at the lowermodel boundary is diagnostically calculated for each time-step, otherwise m is set to zero atthe surface and walls, respectively.4 Numerical methods4.1 Discretization and grid configurationA staggered discretization grid of the type Arakawa C is used in MISKAM. The componentsof the wind vectors are defined on the corresponding grid surfaces, while the components uiare set on the center points of the cell planes, being perpendicularly oriented to the directioni. All scalar prognostic variables (p', E, e) as well as the exchange coefficients are defined atthe cell center points. Figure 4.1 illustrates the grid configuration.111122223333 zz yy x 1 u2 v3 w• p', K, E, eFigure 4.1: Grid configurationThe selected grid structure allows a simple usage of flow obstacles. Assuming that the grid boxis located either completely free or completely within an obstacle, it is sufficient to define fieldsof multipliers which define whether a grid wall belongs to an obstacle (value of the multiplieris 0) or not (1). These multipliers are used for the Cartesian wind components as well as forthe pressure gradient, and therefore assure that wind components will always disappear onbuilding areas.284.2. Treatment of the advection terms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29A modeling of overhanging obstacles (bridges, driveways, or similar) is also possible in MISKAMvia this procedure, because the multipliers contain the entire information about the obstaclesin the model area. It only has to be assured that the configuration will be clearly interpretedby the model and the calculation program respectively. This has to be done via input fileswhich have to be set according to well-defined rules. More details can be found in chapter 5.4.2 Treatment of the advection terms4.2.1 Momentum equationsUp to MISKAM version 5 advection terms in the equations of motion were discretised by simpleupstream differencing. The upstream scheme is known to produce considerable numericaldiffusion, but is nevertheless suitable for the purposes of MISKAM, i.e. simulations of quasi-stationary wind fields. Due to its moderate computational requirements, it is preferable incomparison to the much more time-consuming schemes of higher accuary, on particular onless powerful PC systems.A significant accuracy improvement of the momentum advection calculation kann be achievedusing the predictor-corrector method of McCormack (1969). This scheme yields the solutionof the advection equation as the arithmetic mean of the solutions of an upstream step and aconsecutive downstream step.4.2.2 Advection of scalar quantitiesTo calculate advection of positive definite scalar variables (turbulent kinetic energy, dissipation,mass concentrations) instead of the upstream scheme the MPDATA algorithm (Smolarkiewiczand Grabowski, 1989) can be used. In MPDATA, optionally, one or more correction stepsfollow the upstream step, thus partially reversing the numeric diffusion. For the turbulencemodel, only one corrective step is performed to keep computation time in reasonable limits.For dispersal simulations and in particular for the treatment of point sources, use of theMPDATA scheme is advisable, for line sources (e.g. streets) normally the upstream schemeshould be sufficient.For flow and dispersal simulations in complex configurations (Balczo´ und Eichhorn, 2009)the overall best performance of MISKAM was achieved using the combined McCormack- andMPDATA scheme for the flow model, while only minor differences between upstream andMPDATA scheme occured in the dispersal results.30 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4. Numerical methods4.3 Treatment of the diffusion termThe diffusion term is numerically evaluated by the ADI (Alternating Direction Implicit)-methodof Douglas and Rachford (1956). The values of a variable ? at time n+ 1 are calculated viatwo interim solutions ?* and ?**. Implicit calculations are performed in one space directionwhich means that the derivatives of the variables ? with respect to the space coordinate haveto be applied at the new. The method can be described in the following form:?* -?n?t= ?z(?*) + ?x(?n) + ?y(?n) +R??** -?*?t= ?x(?**)-?x(?n) (4.1)?n+1 -?**?t= ?y(?n+1)-?y(?n)The abbreviations ?x, ?y, ?z stand for differential operators which are applied to the par-ticular quantity ?. The remaining parts of the prognostic equation are summarized in theremaining term R?.In principle, the working order is extraneous in the direction of the space. The order speci-fied here is used because other boundary conditions are formally used in the vertical (closedboundaries, i.e. prognostic variables are prescribed) versus along the lateral boundaries (openboundaries).The inversion of a tridiagonal matrix is only respectively needed to calculate the various interimsolutions, for which a simple standard-algorithm is used.4.4 Solution of the Poisson-equationA so-called Red-Black-SOR-method is used to solve the Poisson-equation. The pressure valuesof the six adjacent cells are needed, in addition to the pressure value of the considered grid cellin order to discretize the Laplace-operators in the Arakawa-C-grid. If the discrete equation issolved for the pressure value of the central grid, the following is obtained:p'i,j,k =Axp'i-1,j,k +Bxp'i+1,j,k + Ayp'i,j-1,k +Byp'i,j+1,k + Azp'i,j,k-1 +Bzp'i,j,k+1 +DAx +Bx + Ay +By + Az +Bz(4.2)whereD =?0?t?u˜k?xkThe coefficients Ax, Bx, Ay, . . . are obtained by discretizing the Laplace-Operators, theycontain the multipliers described in chapter 4.1, to assure the disappearance of the normalgradient of the pressure disturbance on and in obstacles.4.4. Solution of the Poisson-equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31At first, in order to solve this, a first approximation of the pressure disturbance is used inthe right side of the equation (4.2). The pressure field is set to zero at the beginning of thesimulation. For future time-steps, the pressure field of the preceding time-step is used as aninitial approximation.The solution itself is obtained in two half-steps for each iteration step ?. In the first step, newpressure values are calculated for each second point on the grid. Thereby, those points on thegrid obtain new pressure values which take over the adjacent values on the right side of theequation (4.2) in the second half-step. The name of this treatment is based on the checkeredarrangement of the grids containing the new values of the half-steps, however why red is usedand not white is unknown to the author.The two half-steps result in a interim solution p*, which is combined with the pressure field ofthe previous iteration step ? - 1 according top? = ?p*(?) + (1- ?)p(?-1). (4.3)? is a relaxation parameter without dimension, whose optimal value depends on the gridgeometry. The iteration is stopped when the maximal divergence of a wind field (corrected forthe calculated pressure values) falls below a previously set value. This threshold is successivelydecreased during the simulation to increase the precision of the wind field calculation.5 Operating instructions5.1 ConventionsSpecial fonts are used for the following descriptions / actions in the following chapters:command line cd miskamfile, directory name NAME.INIMISKAM refers to the model itself as well as to the total program package. The variousexecutable program parts are MISKAM?.EXE1, KONFIG.EXE and DATXYZ.EXE. ADemo version of the visualization program MISVIS.EXE is also part of the package.Attention: The hyphen in a file name at the end of a line has to be ignored.5.2 Hardware- and software requirementsMISKAM runs under all Windows operating systems starting with Windows 95, recommendedis Windows XP. A computer with a Pentium II processor or higher is recommended as well as64 MB RAM at least. The available memory limits the number of processed grid points. Therecommended RAM size, for instance, allows for a grid of about 100 × 60 × 50 points.5.3 The MISKAM-CDThe MISKAM-CD contains all files needed to run the model MISKAM in a self-extractingZIP-archiv M?SETUP.EXE.The input files as well as the results of the sample calculations are in the directoriesBEISPIEL\\\\EIN and, respectively, BEISPIEL\\\\AUS on the CD. The files GOETTING.*1The ? in the program name stands for its actual version number at the time of printing.325.4. Installation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .33contain the input data as well as the results of a simulation for the Go¨ttinger Straße (Hannover).The input file is a revised version of the afore mentioned PEF-project.The results saved under the name QUERU.* contain the flow and the concentration fieldaround a single building with a flow through (for instance passageway). The file QUER.INPdescribes the basic configuration, the flow-through areas are defined in QUER.001.All input files as well as the most important results of all validation calculations are finally puttogether in the directory VALIDIERUNG with a separate subdirectory for each calculatedcase. Descriptions and results of the various calculations are given in chapter ??.5.4 Installation5.4.1 Extracting the program filesThe following steps are required for a new installation under Windows:1. Starting of M?SETUP.EXE in the root directory of the MISKAM-CD.2. Setting of the target directory to unzip the program files.Further subdirectories below the target directory are created while unzipping. The directoryMISKAM contains the executable program as well as the needed control files. The subdirec-tory MISKAM\\\\EIN is created for input files and MISKAM\\\\AUS for model results. A copyof this manual is stored in the directory MISKAM\\\\DOC as as PDF file.5.4.2 Installed filesAfter executing M?SETUP.EXE according to the instructions above given, the files listed intable 1 are stored in the MISKAM-directories, assuming an installation on drive C: as well asin the directory MISKAM.The batch file MSTART.BAT is used to start MISKAM with the help of a pre-generated INIfile, more details can be found in chapter 5.8.The files STROEM1.INI and STROEM0.INI contain the necessary startup information forthe flow calculations (STROEM1: new run, STROEM0: serial run with using the previousresults), AUSBR1.INI and AUSBR0.INI are the corresponding startup files for the dispersioncalculations.The file README.TXT contains the most important information on how to install anduse MISKAM. Program changes which are implemented after the printing of this manual aredescribed in the text file WHATSNEW.TXT.34 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5. Operating instructionsFile name ExplanationC:\\\\MISKAM\\\\MISKAM?.EXE MISKAM - main programC:\\\\MISKAM\\\\MSTART.BAT batch program to start MISKAMC:\\\\MISKAM\\\\MSTOP.BAT batch program to interrupt MISKAM runsC:\\\\MISKAM\\\\STOPTEXT help file for MSTOP.BATC:\\\\MISKAM\\\\KONFIG.EXE producing INP-filesC:\\\\MISKAM\\\\DATXYZ.EXE producing result tablesC:\\\\MISKAM\\\\MISVIS-D.EXE plot routines to evaluate the results(demo version)C:\\\\MISKAM\\\\README.TXT general information regarding MISKAMC:\\\\MISKAM\\\\WHATSNEW.TXT additional informationC:\\\\MISKAM\\\\STROEM1.INI startup file for the flow calculation,new runC:\\\\MISKAM\\\\STROEM0.INI startup file for the flow calculation,serial runC:\\\\MISKAM\\\\AUSBR1.INI startup file for the dispersion calculation,new runC:\\\\MISKAM\\\\AUSBR0.INI startup file for the dispersion calculation,serial runC:\\\\MISKAM\\\\EIN\\\\CLARK.TAB table of the Clarke-functionsC:\\\\MISKAM\\\\EIN\\\\KONFIG.INP Example of the configuration fileC:\\\\MISKAM\\\\AUS\\\\STROEM.PRS results of the sample files,log file of the flow calculationsC:\\\\MISKAM\\\\AUS\\\\STROEM.UVW results of the sample files,table of wind components, pressure irritationsC:\\\\MISKAM\\\\AUS\\\\STROEM.TUR results of the sample files,table of turbulence variablesC:\\\\MISKAM\\\\AUS\\\\STROEM.ZWU results of the sample files,binary file of wind fieldC:\\\\MISKAM\\\\AUS\\\\STROEM.ZWT results of the sample files,binary file of turbulence fieldC:\\\\MISKAM\\\\AUS\\\\AUSBR.PRA results of the sample files,log file of dispersion calculationC:\\\\MISKAM\\\\AUS\\\\AUSBR.KON results of the sample files,table of mass concentrationsC:\\\\MISKAM\\\\AUS\\\\AUSBR.ZWK results of the sample files,binary file of concentration fieldC:\\\\MISKAM\\\\DOC\\\\HANDBUCH.PDF MISKAM manualTable 5.1: Content of the self-extracting archive M?SETUP.EXE5.5. The configuration files *.INP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .355.5 The configuration files *.INPMISKAM needs information about the grid structure and the layout of the model area regardingobstacles (buildings). This information has to be delivered via an ASCII file in a fixed structure(NAME.INP, NAME variable, DOS compatibility is required). Form and content of theconfiguration files are described in this chapter.5.5.1 Structure of the configuration filesAs an example, the file KONFIG.INP of the M?SETUP.ZIP archive is illustrated in figure5.1. Further examples are given in the configuration files on the CD directory BEISPIEL\\\\EINas already mentioned, as well as in various subdirectories of the directory VALIDIERUNG.The different parts of the configuration file contain the following information:Cartesian grid: The first 4 rows of the configuration file define the number of used gridboxes as well as the grid resolution.The number of grid boxes in each of the three dimensions (sequence x, y, z) is set inthe first row, the fourth number in this row refers to the angle between the x axis of themodel area and the northern direction. The three following rows contain the Cartesiancoordinates of the cell walls in the three dimensions. Hence, one more value has to beset in each row than grid boxes were defined in each dimension.Arrangement of buildings: The size of the following data blocks depends on the numberof grid boxes in the x- and y- directions which was set before. The following rowsrepresent a top view of the model area, where the number of cells, being filled inthe vertical direction with flow obstacles (buildings), are given for each horizontal gridsurface. 0 thus means no obstacle.Roughness length: Either an unique value for the roughness length z0 of the ground or atwo-dimensional field for the surfaces with internally determined roughness lengths canbe set in MISKAM. In addition, roughness lengths for building walls and roofs have tobe set in any case, which are identical for all buildings recorded in the model area.To set a unique roughness length, ”j” has to be inserted in the row after the buildingconfiguration. The next row thus contains the roughness lengths of the ground as wellas of the building areas. All values have to be set in cm.To set the inhomogeneous roughness distribution, ”n” has to be set in the row followingthe building configuration. The roughness lengths of building areas are then only to beset in the following row.36 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5. Operating instructionsModellgebiet aus .INP-Datei Gebäudegrundriß Linienquelle Von MISKAM generierter Seitenrand 30 22 18 90. 0.00 4.00 8.00 12.00 ... 112.00 116.00 120.00 0.00 4.00 8.00 12.00 ... 80.00 84.00 88.00 0.00 1.00 2.00 4.00 ... 75.00 85.00 95.00 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 7 7 7 7 7 7 7 7 7 0 0 9 9 9 0 0 0 0 9 9 9 0 0 0 0 9 9 9 0 0 7 7 7 7 7 7 7 7 7 0 0 9 9 9 0 0 0 0 9 9 9 0 0 0 0 9 9 9 0 0 0 0 0 0 0 0 0 0 0 0 0 9 9 9 0 0 0 0 9 9 9 0 0 0 0 9 9 9 0 0 0 0 0 0 0 0 0 0 0 0 0 9 9 9 0 0 0 0 9 9 9 0 0 0 0 9 9 9 0 0 0 0 0 0 0 0 0 0 0 0 0 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 0 0 0 0 0 0 0 0 0 0 0 0 0 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 0 0 0 0 0 0 0 0 0 0 0 0 0 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 ... 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 5 5 5 5 5 0 0 0 0 5 5 5 5 5 8 8 8 0 0 0 0 0 0 0 0 0 0 0 0 0 5 5 5 5 5 0 0 0 0 5 5 5 5 5 8 8 8 0 0 0 0 0 0 0 0 5 5 5 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 8 8 8 0 0 0 0 0 0 0 0 5 5 5 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 8 8 8 0 0 0 0 0 0 0 0 5 5 5 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 8 8 8 0 0 0 0 0 0 0 0 5 5 5 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 j 10.0 1.0 NOX 0.00 0.00 0.00 lx 1 9 1 0.1 ... lx 30 9 1 0.1 lx 1 11 1 0.1 ... lx 30 11 1 0.1Figure 5.1: Example for MISKAM configuration as top view (top) and in form of the INP file(bottom, not completed)5.5. The configuration files *.INP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .37value kind of land allocation z0 in cm0 roof of a building z0,w1 asphalt or similar without obstacles 12 meadow 53 meadow with single trees, shrubs 104 more dense low vegetation 255 low buildings, not explicitly determined 506 higher buildings, not explicitly determined 100Table 5.2: Kind of ground and roughness lengths in MISKAM. The set value of walls androofs is used for z0,w.In previous versions (up to 5.x) no distinction between building walls and roofs hadbeen made. This inconsitency was eliminated in MISKAM 6, but to ensure backwardcompatibility, the model also handles configuration files which only contain one valuefor all building surfaces.A two-dimensional data field filed follows, in which each grid area receives a valuebetween 0 and 6. The values of the corresponding ground types as well as its roughnesslengths are given in table 2.Distribution of sources: The INP file lists the considered pollutant sources after the geom-etry of obstacles as well as the specifications of the roughness of the ground.The name of the considered substance, its background values (in mg/m3), as well asan eventual sedimentation velocity and deposition velocity (for each in m/s, a positivesedimentation velocity refers to a sinking of the considered substance) have to be sethere.Then, a table follows describing the source type, the grid positions, and the strengthof the source. Roads are normally presented as line sources. However, attention has tobe paid, if the considered road is not parallel to the horizontal axis of the coordinates.In these cases, appropriately corrected emissions have to be set, because a descriptionas a source parallel to one axis would result in a too large total extend of the source.Alternatively, the intensity of a source can also be converted into a point source and setin the configuration file.It has to be mentioned that in grid cells more than one source can be defined. Thestrength of the sources are converted into volume sources within the model, multiplesources will be added to each other.38 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5. Operating instructions5.5.2 Precision requirementsThe following settings are absolutely required during configuration of the buildings:• The height of the highest building is not allowed to exceed 30 % of the model height.• The degree of construction (the proportion of the area of obstacles to the total inflowarea) is not allowed to exceed 10 %.• Investigation points (points for which results should be deducted according to 23. BIm-SchV) should have a minimal distance of 10 grid cells to the respective boundaries.• Investigation points are not allowed to be within grid cells containing sources as well asin their neighboring cells.• Investigation points are not allowed to be in direct neighboring cells of buildings andthey must have a minimal distance of two grid cells to the ground.• Relevant street canyons have to be at least divided into 6, when possible into 8 gridboxes perpendicular to their longitudinal axis.• A precise grating of the buildings is needed in the area of investigation, some uncertaintiescan be allowed in the boundary zones. The geometry of the buildings should be recordedwith a precision of = 1 m in the area of investigation, in order to decrease errors asmuch as possible which can occur in the Cartesian grid.Within the limits of the possible grid resolution, the shape of the roofs can be consideredwhile grating the buildings.• The geometric dimension of buildings in the upstream area has to be determined witha precision of = 20 %. It is not required - and also impossible for the somewhat lessprecise grid resolution in the upstream area - to consider the shape of the roofs, thebuildings can be considered as cuboids with an averaged eaves height.Regarding the position of investigation points, the mentioned restrictions imply, that, near theground, vertical mesh sizes of e.g. 0.5 m - 0.6 m - 0.8 m - 1.1 m can be used accordingly to23. BImSchV. This assures that the center of a cell is located at the height of 1.5 m and thatthere are two further grid cells in between the point and the ground.5.5.3 To save the minimal distance to the groundThe minimal distance of each grid point to the fixed boundaries of the model are calculatedafter launching a INP file. This value is needed to initialize the turbulence model. This calcu-lation can be relatively time consuming for complex configurations. Therefore, the distances5.6. Additional configuration files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39to the ground are written into a separate file called NAME.XXL after the calculation. Forfurther simulations (for instance for other wind directions), only this file is then read. Thisshortens the startup phase of the flow model considerably. Attention: If the configuration fileis changed (coordinates of buildings), the additional file has to be deleted before a new run5.6 Additional configuration filesMISKAM recognizes up to 999 additional configuration files, whose name consist of the prefixof the INP file and the ending nnn (001 < nnn < 999). Additional information about theconfiguration is given to the model via these files. Flow-throughs and obstacles with overhangs,vegetation, momentum-containing sources can thus be defined in MISKAM.5.6.1 Flow-throughThe used configuration file has to contain the string ”Durchstroemung” (please note: smalland capital letters) in the first row. Any number of flowable cell areas can be assigned in thefollowing rows. In the layer perpendicular to the flow direction, the extension of the flowablearea has to be at least two grid cells wide in each direction.The following example defines a flow in the x-direction, which should be spread over the gridcells 6 to 8 in the x-, 4 to 7 in the y- and 1 to 4 in the z-direction. The corresponding inputfile would then look as follows:Durchstroemungx 6 8 4 7 1 4Figure 5.2: Configuration file to define the flow-through areaSetting the direction of the flow-through is important for the correct internal processing,because it determines which cell walls have to be freed again.The results, under the consideration of the flow-through defined in the fileBEISPIEL\\\\EIN\\\\QUER.001, are given in the files BEISPIEL\\\\AUS\\\\QUER-1.*. The filesBEISPIEL\\\\AUS\\\\QUER-2.* contain the appropriate model results of a non-flowable build-ing.5.6.2 VegetationThe influence of vegetation could only be considered in the model via the roughness lengthz0 up to now. This, however, does not lead to realistic results, because the part which is40 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5. Operating instructionsinfluenced by the vegetation is explicitly dissolved by MISKAM, but z0 only includes thosesurface roughnesses that cannot be directly described as flow obstacles. The implementedconcept simulates the influence of vegetation via an additional restraining force in the momen-tum equations, as well as via modified production rates of turbulent kinetic energy and energydissipation. This leads to a realistic simulation of of the flow retardation and the turbulenceincrease through trees and hedges.The user has to provide information in an additional input file (naming convention as withflow-through, which means name.00x with x in the range 1...5) concerning the position of theleaf-containing grids, the degree of vegetation coverage (top view), as well as the leaf areadensity in (m2 of leaf area)/(m3 of air). Values for leaf area density of various forest speciesare given in the literature, for instance Groß (1993).Sample data of an explicitly dissolved solitary tree (BAUM1.INP, BAUM1.001 in\\\\BEISPIEL\\\\EIN) are given. The following table shows a part of the file BAUM1.001. Thevalue ”Vegetation” (please note: small and capital letters) in the first row is mandatory. Thecolumns of the further rows contain the following information:1st column type of vegetation, ”L” or ”l” are mandatory.2nd + 3rd column index-area in x-direction4th + 5th column index-area in y-direction6th + 7th column index-area in z-direction8th column leaf area density in (m2 of leaf area)/(m3 of air) of the cell areadefined in columns 2–79th column degree of vegetation coverage of the cell area.VegetationL 22 22 19 19 6 6 8.400 1.000L 23 23 25 25 7 7 0.781 1.000... ... ... ... ... ... ... ... ...Figure 5.3: Configuration file to define the cell areas with vegetation5.6.3 Momentum-containing sourcesA fixed vertical velocity (outlet air velocity, for instance out of a stack) can be set for pointsources. This is already integrated in the flow calculation, so that wind and turbulence fieldclose to stack exits can respond to the additional momentum.Much more realistic pollutant plumes are obtained with this method than without consideringthe escaping velocity or with using the effective source height.Please note:5.7. Programming and meteorological control parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41• The results are not any longer independent of the flow velocity, because the wind fieldwhich is close to the source directly depends on the exit velocity or, respectively, on itsratio to the undisturbed flow velocity.• The time needed to calculate a stationary wind field can increase considerably.A further input file is needed to set the exit velocity, the files KONFIG.INP and KONFIG.001contain an example. The additional file has to contain ”Quellen” (please note: small andcapital letters) as a first value, thereafter exit velocities (column 4, in m/s) can be allocated toseveral grids (column 1 to 3 contain grid indices in x-, y-, and z-direction). The source positionhas to be consistent with the one of the .INP file (see source position in KONFIG.INP).Quellen14 15 11 5.0Figure 5.4: Configuration file to define momentum containing sources5.7 Programming and meteorological control parametersMISKAM needs a number of control parameters to steer the program, as well as for themeteorological initialization. These parameters will be explained in the following chapter5.7.1, ”confidence intervals” are additionally given for the various values.Lateral up- and downstream zones are program internally defined by default. These defaultvalues can be modified via an additional control file being described in chapter 5.7.2.5.7.1 The initialization file MISKAM.INIThe parameter values to be used are transferred to MISKAM via the file MISKAM.INI. Thefile STROEM1.INI of the installation archive is shown in Figure 5.5. The results obtainedby the program settings given in the file STROEM1.INI correspond to the installed filesEIN\\\\STROEM.*.The meaning of the various parameters, including those of other possible INI files is listedbelow. An S preceding the text refers to parameters needed exlusively for flow calculations,an A to those needed only for dispersion calculations.Control parameter (s: flow model, a: dispersal model):Self explanatory, specification wether flow or dispersal simulation shall be started.42 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5. Operating instructionsSettings for flow calculations N E W S T A R T------------------------------ ===============Controling parameter (s: flow model, a: dispersion model) ............ sconfiguration file (prefix) .......................................... konfigistart (1: new Start, 0: consecutive run) ............................ 1isteps (number of time-steps until intermediate storage) ............. 1000z0ein (roughness length for 1d-initialization in cm) ................. 10.0zanem (Anemometer height in m) ....................................... 10.0uv0 (wind velocity in m/s in Anemometer height) ...................... 5.0winku (wind direction in degree against N in Anemometer height) ...... 225.dtdz (stratification in K/100m, 0: neutral, >0: stabile, <0: unstable) 0.eeps (closure, e: E-eps-model, k: K-model) ........................... eabbr (termination criterion, s for stationarity or number of seconds) soutput file (prefix) ................................................. stroemAttention has to be paid when changing the default values,that the various values start in the 70th column ^Figure 5.5: Example of a MISKAM initialization file for the flow calculation.To control the use of different advection schemes for flow simulations, instead of ”s”also ”s1” or ”s2” can be specified:”s”: use of upstream scheme for momentum and turbulence advection”s1”: use of McCormack scheme for momentum advection and of upstream scheme forturbulence advection”s2”: use of McCormack scheme for momentum advection and of MPDATA scheme forturbulence advectionConfiguration file (prefix):Specification of INP file to be used, the file PREFIX.INP must be located in the folderEIN.New start (1) or consecutive run (0):This parameter tells the program, whether a new calculation of all fields should be started(one-dimensional initialization for the flow model, start with a ”clean” atmosphere forthe dispersion model), or whether a still to be specified batch of existing result files ofa previous run should be used.Values other than 0 and 1 are not allowed.Number of time-steps needed up to the intermediate storage:5.7. Programming and meteorological control parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43The binary files are created at latest after the number of time-steps which was set here,and the calculation is terminated. The calculation is terminated earlier, if the definedstationarity criterion is reached before.Turbulence closure (e or k):S This parameter is obsolete due to the fact that the E-e-model is used exclusively.However, it is kept to ensure the compatibility of older INI files.Roughness length of 1d-initialization:S This is used to calculate the one-dimensional wind profile, which is transferred tothe three-dimensional model area after the required scaling.As already mentioned, five additional grid cells are internally added to the lateral modelboundaries. The outer three of these cells obtain the z0-value of the 1d-initialization, aninterpolation to the first value of the actual model area is performed for the two innergrid rows.The roughness length has to be given in cm.Flow direction:S The angle of the given flow is set according to the common notation (N = 0?, O =90? and so on) in meteorology.Flow velocity:S Wind velocity in anemometer height, setting in m/s. It is principally sufficient in mostcases to calculate one single flow velocity and to determine the distribution of immissionfor other velocities by an appropriate scaling, because the concentration values calculatedby MISKAM are inversely proportional to the wind velocity.Anemometer height:S Optionally available wind measurements (for instance roof-level measurements) canbe considered in the model calculation by setting the anemometer height (setting in m)explicitly. The pre-calculated wind profile is scaled in such a way that the settings ofthe wind velocity and the anemometer height are kept.Thermal stratification:S The stratification is set as vertical gradient (K/100m) of the potential temperature.The value 0 refers to neutral conditions, positive values characterize stable stratification.Special attention is needed to set the non-neutral thermal stratification to producephysically reasonable combinations of wind velocity and stratification. Difficulties canarise when thermal stable stratification is combined with relatively high near-ground windvelocities (for instance > 1 m/s in 2 m height), because this produces unrealisticallyhigh wind velocities in high altitudes. It makes more sense to set low near-ground wind44 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5. Operating instructionsvelocities in the case of stable stratification, or a large anemometer height, for instanceat the upper boundary of the model. The gradient of the potential temperature shouldnot be more than 1 K/100m.Negative values of the temperature gradient (unstable stratification) are internally setback to 0 (neutral stratification), because the combination of unstable stratification ?stationarity is physically not reasonable as mentioned above.Steps to correct the advection calculation:A To calculate the mass advection in the dispersion model, either the simple upstream-scheme or the scheme of Smolarkiewicz and Grabowski (1989) is used, where numericdiffusion effects are partially reversed by one or more correction steps.The upstream-scheme is used if 0 is set, a maximum of two correction steps can becalculated. Values > 2 are set to 2.It is generally justified to use the upstream-scheme for traffic emission (line sources).It has to be noted that each Smolarkiewicz correction-step increases the computationaleffort for the dispersion model by about 80 %.Termination criterion (”s” or number of seconds):Optionally, different termination criteria can be used for the simulations of flow anddispersion.The fixed time termination criterion is identically handled in both parts of the model,the user has to set the number of seconds. This number is directly written in the INIfiles as an termination criterion (see example on the distribution disk).It is especially advisable to set a fixed time for dispersion calculations, where variousbuilding variants are compared, because it assures that the same total mass is emittedfor each model calculation.The stationarity criterion is differently interpreted depending on the kind of simulation:For flow calculations, the simulation is terminated, when the following unique criteriaare fulfilled:• The maximum of the relative changes (change per time-step / inflow velocity in10 m height) of the three wind components,• as well as the maximum change of the diffusion coefficient, also related to an inflowvalue in 10 m heightBoth have to drop below 0.1 %. According to this stationarity criterion, stationarityshould normally be reached after approximately 1000 to 2000 time-steps depending onthe complexity of the model construction.5.7. Programming and meteorological control parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45The dispersion calculations offer a further criterion. The setting of ”s” implies a ter-mination of the calculation as soon as the concentration changes do not exceed 0.01% of the maximal concentration value in any grid. This needs approximately the samenumber of time-steps as the one for flow calculations. This criterion assures that noconcentration changes occur close to sources any longer. It is therefore usable if, forinstance, immission concentrations have to be analyzed close to roads.This criterion is not sufficient to investigate the pollutant concentration in a largerdistance of sources, the stationarity criterion ”s2” can be set in this case.The terminationoccurs in the latter case when the concentration value of any cell is not changed by morethan 0.1 % comparing with the cell‘s own value. Approximately 2 to 4 times more time-steps are needed for this criterion depending on the configuration of the buildings.5.7.2 The control file MISKAM.BNDAn additional lateral boundary zone consisting of five respective grid rows is used by defaultfor MISKAM simulations. The grid resolution is extrapolated from the model area, the threeoutermost grid cells are equidistant. All boundary cells are free of obstacles. Other settings ofthe lateral boundaries can possibly make sense in different scenarios. For instance, additionalboundary cells are not needed, if the actual model area already contains sufficient up- anddownstream zones. It also can be suitable to extend the flow obstacles beyond the lateralboundaries to regard them as being quasi infinite (for instance situations in street canyons,which should not be treated as being two-dimensional due to existing road crossings, approxi-mative consideration of the topography). The file MISKAM.BND in the MISKAM directorycan be used to overwrite the default values in these cases. The structure of the file is givenas follows in Figure 5.6.In the given example, the first three additional grid cells obtain a higher grid spacing by thefactor of 1.2, respectively. The fourth and fifth cell receive the same resolution as the thirdone. The boundary cells would be free of flow obstacles in this case.The variable igeb is responsible not only for the obstacle height but also for the roughnessof the ground and the sources of the pollutants at the lateral boundaries, these values alsobecome ”infinite” if igeb=1.The boundary cells remain free of sources by default. The roughness of the ground is set for thethree outer cells to this value, which is also used for one-dimensional initialization of the flowcalculations. These cells are supposed to be representative of an undisturbed environment.For the two remaining rows of cells, the roughness length is interpolated between the value ofinitialization and the value of the boundary cell of the inner model area.For two-dimensional calculations, the variable igeb is utilized only for the x-direction, homo-46 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5. Operating instructions# Usage of the lateral boundaries# ==============================## Extrapolation of the grid spacing# =================================# xyfact = 0 <==> 1 additional boundary cells as in MISKAM3.6# xyfact = 1 <==> equidistant continuation of the grid over 5 cells# xyfact > 1 <==> spread continuation of the grid over 5 cells# xyfact is program internally limited to 2.## Extrapolation of the obstacle structure# ===================================# igeb = 0 <==> no obstacles on boundaries# igeb = 1 <==> last inner obstacle height is transfered to boundaries.# igeb should ONLY be set to 1 to approximate the topography.# Exceptions are two-dimensional calculations, igeb=1 is automatically set#xyfact 1.2igeb 0#EndeFigure 5.6: File MISKAM.BND to set the lateral boundaries.geneous conditions are assumed for the y-direction.To return to the default setting, the file MISKAM.BND has to be arbitrarily renamed ormoved to a different directory.5.8 Utilization stepsThe model calculation is started with MISKAM via the DOS-window (Win 95/98) or the com-mand prompt (Windows NT, Windows 2000). The program obtains the needed informationabout the program control and the meteorology from the file MISKAM.INI, as mentionedabove. Results are stored in binary files and ASCII-tables. Because no interactive processeshave to be worked through to start, MISKAM also can be used for batch processes, for instanceto produce statistics.5.8. Utilization steps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 475.8.1 Initialization and program startBefore MISKAM can be started, the appropriate INI file for the planned run has to be produced.The files STROEM1.INI, STROEM0.INI, AUSBR1.INI, and AUSBR0.INI are copied as”templates” into the program directory during the installation. The purpose of an initializationfile is stated in its first line.MISKAM searches for a file MISKAM.INI in the installation directory to start. The nameis obtained via the provided batch file MSTART.BAT. After saving the start conditions forinstance by editing the file STROEM1.INI, under an arbitrary file name, ending with ”.INI”(for instance C:\\\\MISKAM\\\\FLOWTEST.INI), the command line to start with looks like:mstart flowtestSuccessive runs or dispersion calculations, respectively, have to be prepared in the same way(editing of the appropriate INI file) and started via MSTART.BAT.Depending on the task, different parameters are needed (see 4.7), which also changes thestructure of the INI file. Examples are copied into the installation directory while unzipping thearchive MISKAM-INSTALL.ZIP. Names and content of the sample data are self explanatory,therefore a detailed description is omitted at this place. The meaning of the various parameterswas already explained in chapter 5.7.1.5.8.2 Terminating the programRunning MISKAM calculations can be terminated in a controlled manner with the help of aprovided text file STOPTEXT as well as the batch file MSTOP.BAT, both being copiedinto the MISKAM directory during installation. The execution of the batch file causes thetermination after the next soft copy on the screen (maximum after 10 further time-steps forflow calculations, maximum after 100 time-steps for the dispersion calculations). The resultsare written in the normal output file.5.8.3 Result outputAll data which is needed for further runs (for instance the continuation of the simulation, forwhich the stationarity criterion was not reached yet) are saved in an unformatted binary fileat the end of a model calculation. These files are produced after 50 calculated time-steps forflow calculations. The binary files of the previous dispersion calculation are used as input filesfor dispersion calculations.The binary files contain all the coordinates and building configurations, the distribution ofroughness, and the positions and intensities of the pollutant sources, respectively. The com-48 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5. Operating instructionsplete wind field and the field of the pressure disturbance are additionally saved in the files*.ZWU, the diffusion coefficients, the kinetic turbulence energy, and the turbulent energydissipation in the files *.ZWT.A ”readable” result output is additionally given in the form of ASCII tables.The Cartesian wind components, the amount of wind velocity as well as the pressure distur-bance are saved in form of horizontal cross sections in the files *.UVW. The files *.TURcontain the variables of the turbulence model (diffusion coefficient turbulence energy, andenergy dissipation). The calculated mass concentrations are finally saved in the files *.KON.A scaling is performed for all output values except the one of the wind component. The scalingresults in the largest value of each variable, which is reached in the field, being dumped as afour digit number. The wind components as well as the value of the wind velocity are usuallysaved in mm/s.To control the program, log files *.PRS (flow model) or *.PRA (dispersion model) areproduced, where all information is saved about the simulation run (input and output files,parameter values, information about the calculated time-steps).5.8.4 Control outputThe logging of flow calculations was considerably improved in MISKAM 5. The logs are savedin the file MISKAM.PRO, which has to be located in the MISKAM directory.If MISKAM.PRO does not exist, the log file NAME.PRS is normally generated. If MIS-KAM.PRO exists regardless of its content, a detailed protocol is generated with the nameNAME.PRS. Besides the maximal remaining divergence, the individual changes of the windcomponents, as well as the turbulent diffusion coefficient together with their respective gridpositions are also given in the table. To add this detailed logging is useful, if convergenceproblems arise (which still cannot be excluded for complex, extensive building configurations).Time series of model variables at individual grid points can be used as a further online-control.The file MISKAM.CTR in the MISKAM directory is needed for this purpose. The grid pointindices of up to nine grid points can be listed in this file, the time series of u, v, w, p', E,e, and Km are then written in an own file CONTROL-n.OUT. The files are numbered, then-th file contains the results of the n-th grid point of the input file MISKAM.CTR.5.9. Help program . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49# MISKAM-Control# ----------------------## Up to 9 tripels (i,j,k) can be set,# each number HAS to have four digits and HAS to be right aligned.#25 7 1025 8 1225 9 11# ENDFigure 5.7: File MISKAM.CTR to control time series, as an example of three grid points.5.9 Help program5.9.1 KONFIG: Interactive setting of configuration filesThe program KONFIG belonging to MISKAM allows the generation of the INP files withoutthe detour via an ASCII-editor. However, it is currently only possible to set the roughnesslength in this way, the setting of inhomogeneous surfaces has to be done by hand in the INPfile afterwards.KONFIG is a dialog program which successively prompts the necessary information to con-struct a MISKAM-suited input file. An extensive plausibility verification of the required datais performed. The following settings are rejected or questioned:• Not strictly monotonic increasing coordinates of cell walls (new setting is required)• Change of the grid resolution from one to the next cell by more than a factor of 2(program asks, whether the user is sure about the setting)• Only two free grid boxes above an obstacle (new setting is required)• Less than 6 free grid boxes above an obstacle (program asks, whether the user is sureabout the setting)5.9.2 DATXYZ: Interactive evaluation of result filesFurther processing of the calculated data with graphic or spread sheet software is often desired.To make it easier to import MISKAM results in such programs, the program DATXYZ isincluded.50 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5. Operating instructionsDATXYZ transfers the grid positions as well as the values of the variables into tables withthree columns for individual sections (horizontal, x- z- or y- z-cross section). The followingvariables can be evaluated:• wind vectors• wind components• pressure disturbance• diffusion coefficients• kinetic turbulence energy• turbulent energy dissipation• mass concentration (with or without background value)• deposited mass / deposition rateThe first two table columns contain the coordinates of the grid points (in m), the third thecorresponding value of the variable.For wind vectors, the results are listed in 4 columns, the third contains the value of the windvelocity, the forth the angle of the wind vector to the axis turned towards the right of theselected cross section plane.All necessary settings (name of the result files, variables which have to be analyzed, cuttingcoordinates) are prompted in a dialog.5.9.3 MISVIS: Visualization of MISKAM resultsThe previous versions of MISKAM used rudimentary graphic functions provided by MS-DOSto visualize model results and input files. Since the needed program libraries are not availableany longer, a new visualization program became necessary with the introduction of pure 32-BitWindows environments. The result of this development by giese-eichhorn is MISVIS. Thegraphic library DISLIN2 was used.MISVIS offers graphic displays of all prognostic variables calculated in MISKAM as well as thevisualization of the respective building configurations.2www.dislin.de5.9. Help program . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51Figure 5.8: Example of a MISVIS graphic: Presentation of a building configuration of thefile KONFIG.INP52 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5. Operating instructions(a)0.0 20.0 40.0 60.0 80.0 100.0 120.0x (m)0.014.729.344.058.773.388.0y (m)(b)0.0 20.0 40.0 60.0 80.0 100.0 120.0x (m)0.014.729.344.058.773.388.0y (m)0.09.1*10-11.8*1002.7*1003.6*1004.6*100Figure 5.9: Example of a MISVIS graphic: Presentation of a near-ground wind field of thefile STROEM.ZWU. (a) vector presentation (b) wind velocity (m/s) as color raster graphic.5.9. Help program . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53(a)0.0 20.0 40.0 60.0 80.0 100.0 120.0x (m)0.014.729.344.058.773.388.0y (m)(b)0.0 20.0 40.0 60.0 80.0 100.0 120.0x (m)0.014.729.344.058.773.388.0y (m)0.02.1*10-24.2*10-26.3*10-28.4*10-21.1*10-1Figure 5.10: Example of a MISVIS graphic: Presentation of a near-ground field of massconcentrations (mg/m3) of the file AUSBR.ZWK. (a) Isolines presentation (b) color rastergraphic.6 LiteratureBalczo´, M. and J. Eichhorn, 2009: Refined MISKAM simulations of the Mock Urban Set-ting Test. Proceedings of the XXIII. MicroCAD International Scientific Conference,Miskolc / Hungary, pp. 7-12; ISBN 978-963-661-866-7.Douglas, J. and H. Rachford, 1956: On the numerical solution of heat conduction prob-lems in two and three space variables. Trans. Amer. Math. Soc., 82, 421–439.Eichhorn, J., 1989: Entwicklung und Anwendung eines dreidimensionalen mikroskaligenStadtklima-Modells. Dissertation, Universita¨t Mainz.Eichhorn, J., 1996: Validation of a microscale pollution dispersal model. In: Air PollutionModeling and its Application IX, Plenum PRess, New York.Eichhorn, J., K. Cui, M. Flender, T. Kandlbinder, W.-G. Panhans, R. Ries, J.Siebert, T. Trautmann, N. Wedi, and W.G. Zdunkowski, 1997: A three-dimensionalviscous topography mesoscale model. Beitr. Phys. Atmosph., 70/4, 301-317.Eichhorn, J., 2003: Numerical Simulation of Effects of Roof Shape on Microscale Flow:Extension of the Numerical Model MISCAM. 4th International Conference on Urban AirQuality, Prague/Czech Republic).Eichhorn, J. und A. Kniffka, 2010: The Numerical Flow Model MISKAM: State of Devel-opment and Evaluation of the Basic Version. Meteorol. Zeitschrift, 19/1, 81–90.Green, S.R., 1992: Modeling turbulent air flow in a stand of widely-spaced trees. ThePHOENICS Journal of Computational Fluid Dynamics and its Application 5, 294–312,Wimbledon.Groß, G., 1993: Numerical Simulation of Canopy Flows (Springer Series in Physical Envi-ronment), Springer-Verlag, Heidelberg.Lauerbach, H. and J. Eichhorn, 2004: Flow Through Decidious Tree Crowns - Compari-son of Measurements and High Resolution Numerical Modelling. NATO Advanced Study54Literature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55Institute ”‘Flow and Transport Processes in Complex Obstructed Geometries”’, 4-15 Mai2004, Kiew, Ukraine.Panhans, W. G. and R. Schrodin, 1980: A one-dimensional circulation and climate modeland its application to the lower atmosphere. Beitr. Phys. Atmosph., 53, 264–294.Paterson, D. A. and C. J. Apelt, 1986: Computation of wind flows over three-dimensio”-nal buildings. J. of Wind Engineering and Industrial Aerodynamics, 24, 193–213.Paterson, D. A. and C. J. Apelt, 1986: Simulation of wind flow around three-dimensio”-nal buildings. Building and Environment, 24, 39–50.Patrinos, A. N. A. and A. L. Kistler, 1977: A numerical study of the Chicago lake breeze.Boundary Layer Meteor., 12, 93–123.Ramanathan, N., K. Srinivasan and B. V. Seshasayee, 1995: Numerical simulation ofboundary layer variables using e-e closure scheme. J. Appl. Meteor., 34, 542–548.Ro¨ckle, R. and C. J. Richter, 1995: Ermittlung des Stro¨mungs- und Konzentrationsfeldesim Nahfeld typischer Geba¨udekonfigurationen - Modellrechnungen. AbschlussberichtPEF 92/007/02, Forschungszentrum Karlsruhe.Rodi, W., 1980: Turbulence Models and Their Application in Hydraulics - A State of theArt Review. Iahr Monograph Series, A. A. Balkema, Delft.Smolarkiewicz, P. K. and W. W. Grabowski, 1989: The multidimensional positive def-inite advection transport algorithm: Nonoscillatory option. J. Compu. Physics, 86,355–375.VDI, 2005: Environmental meteorology - Prognostic microscale windfield models - Evalua-tion for flow around buildings and obstacles. – VDI guideline 3783, Part 9; Beuth-Verlag,Berlin, 53 pp. Preface Software license Warranty terms Short information Introduction About MISKAM Why prognostic modeling? Scope and limitation of application Scope of application Limitations Theory Preliminary note The momentum equation The turbulence model Calculation of the diffusion coefficients The prognostic equations for E and The splitting method according to Patrinos The dispersion model The prognostic equation Sedimentation and deposition Momentum sources Initial and boundary conditions Initializing the flow model Boundary conditions for the flow model Initialization of the dispersion model Boundary conditions for the dispersion model Numerical methods Discretization and grid configuration Treatment of the advection terms Momentum equations Advection of scalar quantities Treatment of the diffusion term Solution of the Poisson-equation Operating instructions Conventions Hardware- and software requirements The MISKAM-CD Installation Extracting the program files Installed files The configuration files *.INP Structure of the configuration files Precision requirements To save the minimal distance to the ground Additional configuration files Flow-through Vegetation Momentum-containing sources Programming and meteorological control parameters The initialization file MISKAM.INI The control file MISKAM.BND Utilization steps Initialization and program start Terminating the program Result output Control output Help program KONFIG: Interactive setting of configuration files DATXYZ: Interactive evaluation of result files MISVIS: Visualization of MISKAM results Literature