SOSim User Manual v. 1.0cr1 A Guide to Operate the Sunken Oil
Contents
1. 3 3 Modeling Area and Grid The modeling area is the geographical area over which the user wishes to predict oil locations in time Due to limitations in terms of computational time and the corresponding resolution of the statistical computations and output assignment of modeling areas larger than 50 km wide by 50 km long approximately 0 50 longitude by 0 5 latitude degrees is not recommended However assignment of larger regions is allowed by SOSim for example to allow the user to first check the overall movement or direction of a spill before modeling smaller areas In that case a warning message will pop up requesting permission from the user to assign such a large area and warning of the disadvantage in doing so An alternate approach for viewing results over a larger area is conduct a run for the principal area first then conduct partial re runs of the program for contiguous regions using the post processing Pan posteriori and Recalculate set of tools located to the right of the screen once their use become available after processing the current run However in this case the results for each area modeled will have a new color scale not continuous with the first because the highest relative concentration predicted for each run will be colored dark red for example though these concentrations will not likely be equal making interpretation more difficult The grid is a set of orthogonal points in the west east and north2. Figure 3 1 32 3 3 3 4 LIST OF FIGURES AND TABLES Main screen of the computer application before starting a new project Spill Information input prompts in the GUI 0 ccc cece cece ees Marked spill site after selection of longitude latitude values and CITE CHONG S E ER E E E E E E s Creation of a sampling campaign file Recording spatial coordinates and observed relative concentration values scale 0 to 100 for each sampling ISIS Gt S TUNG sedis seen orate dae dedua eatiie anaes waa wend iam ea eed ee ene Sampling Campaign input prompts and buttons in the GUI Example uploaded sampling campaign data file cc ee ee eees Modeling Area and Grid input prompts and buttons in the GUI Selecting the modeling area in the GUI 0c cece eee e eee ees Land Boundaries input buttons and spin box in the GUL 4 Prediction Times input buttons in the GUI ccc cece cce eee e eee ees Modify default input settings dialog DOX ccc ccc eee e cece e cece cece eee eeees PROCESSING DUMONS ereenn Sure EE EOAR Pan View DUON S L H ua a R EE H 15 16 19 1 Introduction SOSim is a modeling tool developed to help locate sunken oil in relatively flat bays based on limited available field data collected shortly after a spill when oil has begun appearing on the bottom This User Manual will guide the operator through model installation to model operat
3. Transport Model and Simulator of Sunken Oil Mass Doctoral Dissertation University of Miami July 2010 Englehardt James D Echavarria Gregory M Angelica Avellaneda Pedro 2010 Development of a Predictive Bayesian Data Derived Multi Modal Gaussian Maximum Likelihood Model of Sunken Oil Mass Draft Final Report June 2010 Research Planning Inc 1994 Natural Resource Damage Assessment Emergency Guidance Manual Version 1 1 Columbia SC 1994
4. s geographic parallel closest to the latitude of the spill site and will mark the spill site with an X Figure 0 3 If you wish to modify the spill name at this point you would have to proceed with the change and then repeat the last step Latitude 19 10 xl File Map I BOolleksae SPILL INFORMATION Spill Name my Spill Oil Type la Spill Time 8 09 am 01 01 2000 lt Longitude 83 835 Latitude 27 8 MODEL SET UP FOR SPILL MY SPILL LEGEND SAMPLING CAMPAIGNS Sampling Campaign Number Lo a Started at 12 00 AM oa 01 01 2000 lt Upload Sampling Campaign Data File Remove Use Existing Calibration File MODELING AREA AND GRID Select Area from Map Auto Select Area Number of grid nodes Default Define GLOBAL AREA COORDINATES 0 0 0 0 LAND BOUNDARIES Select From Map fe No Nearby Boundaries CALCULATION TIMES Prediction number fi At 12 00 am 4 1 1 2000 ox Calibrate Calibrate Run Recalculate Calculating 0 CANCEL Figure 0 3 Marked spill site after selection of longitude latitude values and directions 3 2 Sampling Campaign s Oil movement on the bottom is inferred by the model based principally on the relationship between the location and time of the spill and the locations and relative concentrations at subsequent sampling times For purposes of S
5. time of predicted landfall The user should be aware that such predictions in the nearshore environment may potentially occur even in some cases if no a priori warning is issued In general predictions of substantial sunken oil occurrence at the shoreline should be interpreted with care Input is entered as follows e Prediction number spin box Assign a whole number index to each desired time of prediction If the index number is not changed in the spin box when adding a new date and hour a warning message will appear e At hour spin box and time line prompt Complete these fields with respectively a the hour including minutes if desired and b the date from the smart 32 calendar pop up at which predictions are desired The user must confirm final entry by clicking the OK button e OK button Add the current hour and time to a list of prediction times and assign it to the index position indicated by the Prediction number spin box The GUI will then estimate whether the time entered is likely to be within the time required for the oil mass to reach the shoreline if the oil is apparently moving towards the shore If the time entered is determined to be potentially after the time of landfall a warning message will pop up Any prediction time prior to the most current sampling campaign cannot be modeled as the likelihood is affected by all sampling campaigns A message will appear requesting that only sampling campaigns p
6. Full set of possible Minimum Maximum Advective velocity in the x direction krnfd 3 00 Adyvective velocity in the y direction Krmrd1 3 00 Coefficient of Diffusion in the x direction km2fd Coefficient of Diffusion in the direction ikmzid Coefficient of Correlation between x and y directions Weighting Parameters Fixed o Restore Defaults Cancel Apply Figure 0 11 Modify default input settings dialog box 34 4 Processing This section describes operation of the model from the moment that the input information is complete until production of the default output requested by the user Output can be 1 only a calibration file to be saved for further processing 2 a calibration file plus map predictions for requested scenarios based on the last or 3 recalculations based on previous predictions and a saved or recently created calibration file Figure 4 1 shows the processing buttons of the GUI Calibrate Calibrate Run Recalculate Figure 0 12 Processing buttons 4 1 The Calibrate Calibrate Run and Recalculate Buttons Calibrate button this button is used 1f it is intended only to calibrate the model by using only the sampling campaign s uploaded files the spill information and the characteristics of the geographic area No prediction times need be selected to run a calibration When finished the calibration files can be saved by the user for future use Calibrate Run button this b
7. GB between the 2 3 GB limit and the kernel memory taken up by idle processes Indirect warning messages provided by the GUI will guide the user in setting the best possible resolution to achieve optimal performance in terms of memory 2 2 Software Installation Python and all prerequisites of SOSim are included in the OSGeo4W console a compilation of open source packages developed by the Quantum GIS project The OSGeo4W console is distributable and therefore is included with the SOSim package distribution Steps required for the installation of SOSim are as follows Note that additional advanced or updated installation procedures may be needed by a developer for future continued development of the Simulator 1 Open the distribution SOSim v 1 0rcl you will see two icons one for SOSim itself and one for OSGeo4W Copy and paste the SOSim icon folder that comes with the distribution directly to the C Program Files directory of your machine 2 The OSGeo4W package is required and should be installed when installing SOSim using the executable file osgeo4w setup exe provided with the distribution This executable has an online installer which means that your machine needs an internet connection at the time of the installation Double click on the icon and choose Desktop express install when prompted for installation preferences and then keep the following default options checked Once installed 13 you will find several shortcut
8. area is the continuation of a coastal region that extends to the north select the button N under the Pan View section The canvas will move towards an area of the same map scale and size in the selected direction e A pop up message will request the user to complete the boundary conditions for the new area along with the desired resolution number of nodes in the east west and south north directions for this contiguous case Of course any change in resolution may result in changes to the grid of the output map If the new area is being modeled with the intention of creating a mosaic of multiple runs a constant resolution 1s recommended e Prediction times can be changed but be aware that in order to see the continuation of the original output the same times of prediction are required and e Press the Recalculate button in the Pan View panel Pressing this button instead of the Run button will instruct SOSim to use previous input information regarding the sampling campaigns and the spill and will require much shorter 39 runtime than would be required to rerun the case from scratch using the Run button To model zoomed in areas of a current output map or larger zoomed out areas e From the panel Modeling Area press the Select Modeling Area button Use the tool as before making sure to select the resolution number of nodes in each direction and new boundary conditions desired
9. before starting a new project Required input is prompted in the area located to the left of the canvas see Figure 3 1 The results of the prediction depend on the input In particular movement of the oil will be predicted based on relative oil concentrations at the field sampling points at and around a spill site the time and location of the spill and of the samples collected and shoreline boundaries Other essential model input determining output characteristics 16 including run time and resolution includes basic geographic information about the spill extent of the modeling area desired spatial resolution of the graphical output geographic conditions within the modeling area such as proximity to the shoreline and the dates at which predictions are desired The GUI s input area is organized according to functionality as described in the following sections 3 1 Spill Information Information on the spill name oil type location and time are entered as shown in Figure 3 2 SPILL INFORMATION Spill Name Oil Type ls Spill Time jizioo am 22 onono Longitude Latitude l Figure 0 2 Spill Information input prompts in the GUI This spill information is used to set the geographic system of reference for the run Required information includes the spill time and the coordinates at which the accident most likely occurred A spill name is also prompted for use in saving and distinguishing output figures and f
10. before starting the run The Zoom in tool under the Map toolbar also works to select a modeling area if the user subsequently confirms the nodes and boundaries e Change times of prediction if desired and e Press the Recalculate button under the Pan View section Pressing this button instead of the Run button will instruct SOSim to use previous input information regarding the sampling campaigns and the spill and will require much shorter runtime than would be required to rerun the case from scratch using the Run button To update the current output map based on new or revised sampling campaign data e If a sampling campaign is no longer needed or requires revision remove it using the Remove Sampling Campaign button making sure to select the correct campaign number from the drop down menu upload the new sampling campaign data as usual and 40 If the modeling area boundaries and prediction times are not to be changed then press the Run button The Recalculate button is not useful in this case because the model run must be computed from scratch using the new field data conserving only the desired prediction date s areas boundaries and initial spill information 41 6 Post Processing Post processing activities include uploading previous raster output for viewing in a GIS geographic information system saving output saving calibration files for further use and pri
11. electronic source package uses variables and attributes passed by the OPI module to compute the predictive Bayesian relative concentrations saves output files and passes results back to the OPI module for display by the GUI or for further use The GUI has the capability to communicate with the core module through the OPI module during a model run after a predictive result has been presented to allow the user to request the modeling of contiguous or other areas as needed The GUI contains three basic layouts other than the main toolbars and menus 1 the pre run layout or input section where the user is prompted for input information concerning the spill sampling campaign s and time s land boundary desired modeling area and prediction times 2 the canvas layout where results are displayed in interactive georeferenced maps and 3 the post run layout or output section which contains tools that allow the user to display the results at different times of prediction run the model for contiguous areas save results print images and perform other formatting tasks 1 2 Why Python The Python programming language was chosen for development of the SOSim model based on the following criteria e The programming software is non proprietary or non commercial and open source e The programming environment supports relatively rapid prototyping and testing e The language is robust in terms of functioning in various operating
12. icons on your desktop you only need the desktop shortcut to OSGeo4W console and you can delete all others 3 Once installed go to C OSGeo4W apps qgis bin and find the qgis_core dll and qgis_gui dll libraries Copy and paste them to the _ directory C OSGeo4W apps qgis python qgis 4 Add to your system or edit the environment variable below PYTHONPATH C OSGeo4 W apps qgis python Environment variables must be changed from an account with administrator access privileges and for all users Now the SOSim software is callable from the OSGeo4W console using the following steps 1 2 3 1 Double click on the OSGeo4W shortcut logo on your desktop 2 Type in the console cd C Program Files SOSim and 3 Type SOSimOPI pyw The SOSim GUI will pop up and will be ready to use To change an environment variable in a Windows 32 bit machine right click on your My Computer icon from the Start menu and select Properties In the System Properties box under Advanced select Environment Variables at the bottom System variables are in the lower part of the box and you only have access to then if you have administrator privileges over the machine Screen the list to see of the environment variable that you need already exists if it does click on it and on Edit then type a semicolon after the existing text and type the given path following without spaces If the environment variable does not exist click o
13. in sunken oil movement due to e g tidal action storm events and sediment entrainment Upload Sampling Campaign Data File button this button opens an explorer dialog box that allows you to browse for your txt tab delimited Excel file containing the sampling campaign information which you want to upload It is recommended that every sampling campaign file have a different name and be identified with a number If the sampling campaign entered happens to be during the assessed sinking retardation time the sampling campaign will be invalidated and a warning issued Remove Sampling Campaign button this button automatically removes from the record the data of the campaign that corresponds to the current number in the Sampling Campaign Number spin box A confirmation message pops up to confirm the deletion along with the number of the sampling campaign that was removed However deletion of the file does not change the numbers of any other uploaded campaigns so that there will now be a campaign number with no assigned data Use Existing Calibration File button this button allows the modeler to use an existing calibration file which must have been named and saved by the user after a previous calibration By clicking on this button a search dialog box will open 24 and will prompt for the directory and file location An informational message will confirm that the calibration file has been loaded when found An example uploaded samp
14. spills within a coastal area The progress bar indicates the total number of operations completed 36 5 Output Output is produced as a map representation of predicted relative oil concentrations within the modeling area There are two classes of output the default and the optional output 5 1 Default Output The map at the first requested time of prediction is automatically displayed in the GUI canvas immediately after the processing stage is completed In addition a portable raster map PNG format of this result 1s saved to the desktop along with a corresponding world file required for automatic georeferencing in any geographic information system If more than one time of prediction was initially requested map results corresponding to all prediction times will be available for the user to display on the canvas and save to disk as the user requests with the next and previous buttons located to the right of the display Output maps are color coded such that dark red areas are predicted to have high relative sunken oil concentrations near 100 and dark blue areas are predicted to have low relative concentrations near zero Predicted relative concentrations are effectively averaged across smaller scale patchiness of the sunken oil and across cyclic movement due to tidal excursions Also total sunken oil masses are not known as a function of time In fact although relative concentrations may decrease with time due
15. z SAMPLING CAMPAIGNS Sampling Campaign Number p a Started at 5 00 4M 4 01 20 2000 lt Upload Sampling Campaign Data File Remove Use Existing Calibration File MODELING AREA AND GRID Select Area from Map Auto Select Ares Number of grid nodes Default Define LAND BOUNDARIES Select From Map e 5 No Nearby Boundaries CALCULATION TIMES Prediction number 1 4 At 12 00 AM 1 1 2000 gt Ss PAN VIEW CN Cw CE Calibrate Calibrate Run Recalculate Calculating 0 cancer rS Figure 0 8 Selecting the modeling area in the GUI e Auto Select Area button this button automatically zooms in and selects a region of 0 25 degrees longitude by 0 25 degrees latitude around the spill site approx 25 by 25 km or approx 15 6 by 15 6 miles This feature can be used along with the post processing Pan posteriori and Recalculate set of tools when the user plans to consecutively run SOSim for contiguous geographical regions of the same size so as to obtain a mosaic of a large area and 28 e Number of Grid Nodes as mentioned previously resolution of the output is defined by the number of nodes desired in the west east and north south directions within the modeling area Two options are available to the user The first is the Default radio button which assigns 25 nodes in each direction Using this option if the m
16. OSim input a sampling campaign is defined as a set of qualitative measurements of relative oil concentration on the bottom taken at approximately the same time Spill and sample coordinates are 20 entered in WGS World Geodesic System units decimal degrees of longitude and latitude Time differences from sampling point to point of up to a few hours may not need to be considered as the model was not designed to track possible cyclic excursions of the oil due to the tidal cycle SOSim version 1 0rcl accepts and has capability for processing up to 10 different sampling campaigns Each sampling campaign file must be created in a separate Microsoft Excel spreadsheet closely following the instructions below e The spreadsheet must use only the first 4 columns A B C and D e A title field name for the column must be included in the first line of each column e In Column A enter index numbers for the data points beginning with the numeral 1 to be entered on line 2 e In Column B enter the longitude coordinate in decimal degrees at which the sample was collected starting on line 2 e In Column C enter the latitude coordinate in decimal degrees at which the sample was collected starting on line 2 e In Column D enter a positive or zero value for the measured relative oil concentration on a relative scale of range scale of O 100 that is enter a percentage of oil without the percent sign starting on line 2 If q
17. SOSim User Manual v 1 0cr1 A Guide to Operate the Sunken Oil Mass Simulator By M Angelica Echavarria Gregory In Conjunction with the Dissertation Development of a Predictive Bayesian Data Derived Multi Modal Maximum Likelihood Gaussian Model for Simulation of Sunken Oil Mass in Time TABLE OF CONTENTS LISTOF FIGURES AND TABLES rrace E N Section l INTRODUC TION crara EE E EEEE E ES E WAGE 1S 5 OS Ty C E I2 WAY Ts 1 a 157 e E 1 3 Objectives of this User Manual A Guide to the SOSim GUI 1 4 Scope of Model Applicability ccc cece ccc cce eee e cece eeeeaees 2 INSTALLA TION icense sies anea a a e A eas eae ZA Hardware Requirements aeni ne A E R 22 SONWANE a AE nears E E binds 3 IPU T or en eee E ne Terr eee er nore ne SL SPUL Oma OM eda era R hoe acne 32 Sampling Campalsn S T 3 3 Modeling Area and Grid TTT 3 4 Ele BOUNGAMES 056 csicdeescatacswawhieeedadens vide datasenewhiawedadees E SoD Prediction K Te STR 5 21T Mapu CTT 4 PROCESSIN E beeen oiseda a E ea eaae tia 4 1 The Calibrate Calibrate Run and Recalculate Buttons 4 2 R n ime and Progress Bar aaa aa a Re RR R RRR KRA ire 5 CORT N R cee secencsecs wee casa tous ne icnscaueeua tease E A E IL Delat QUM erei a E E E AE eden D2 Opona OUM eee NAE E AEE NE 6 POSTPROCESSING cerner nE E 7 PORTABILITY OF RESUL S raria a E 8 SOFTWARE PORTABILMT Y 2 cccssccuvarevacatenswennsnwaneweteacsawniiwedass REFERENCES
18. didate version 1 0rc1 to a 1 0 2 version as soon as the complete data sets are gathered after a spill occurence SOSim has been programmed in Python in its entirety The SOSim model although using and relying for much of its functionality on several existing Python packages and modules consists of three principal Python modules developed by the author of this User Manual the graphical user interface GUI module the operating and processing interface OPI module and the core module The GUI module ui_SOSim py in the electronic source package automatically layouts and keeps characteristics of widgets labels canvases and templates in the graphic user interface holds raw user s input and imports Windows palettes and display The OPI module SOSimOPLpy in the electronic source package 19 the executable file It imports and links all other modules together it captures the input information that is entered by the user in the GUI module and operates interrelated buttons and activities of the GUI it filters organizes and processes the input passes ready to use variables and attributes to the core code accepts modeling results back from the core module processes the results and sends display signals to the GUI module s canvas layout for it to depict relative sunken oil concentrations on a map for further user interaction It also controls modal behavior of the main windows pop outs menus and toolbars The core module SOSimCore py in the
19. e Remove Use Existing Calibration File Figure 0 5 Sampling Campaign input prompts and buttons in the GUI The following descriptions and conditions apply Sampling Campaign Number adjust the number in the spin box to that of the sampling campaign that you want to upload and process A campaign number greater than 10 19 not accepted Start at Time Date set the spin box to the time representing the midpoint of the sampling campaign Note that you can edit the hour and minutes by using either the mouse cursor or the arrows of the spin box or both Set the day month and year of the sampling campaign using the smart calendar that pops up when you click in the drop down menu located to the right Notice that the date line edit can also be changed manually with the mouse cursor It is not necessary that the dates of previously uploaded campaigns have earlier sampling dates than the campaign currently being uploaded but the dates of all sampling campaigns must be subsequent to the spill date entered and subsequent to the assessed retardation gap during which oil may be still sinking depending on its type A warning 23 message will guide you in case of error Sampling dates of different campaigns can even be the same different response teams can take samples in different areas at the same time Nevertheless it is recommended that sampling be conducted at different times in order for the model to better account for changes
20. ed includes only open sea 3 5 Prediction Times CALCULATION TIMES Prediction number AL 12 00 am H 1 1 2000 Delete Figure 0 10 Prediction Times input buttons in the GUI The selection of prediction times is the last segment of required input In order to enter these times all previous input sections must be completed An error message pops up if input information is missing Projection times are the dates and hours the user wishes to view predicted sunken oil relative concentrations The user can request up to five times of prediction per model run However computational run time will be longer for each projection time requested and it is recommended that one time at a time is 31 modeled with the aim to plan and make timely decisions about subsequent run needs Times are input to the panel shown in Figure 0 10 SOSim v 1 0rc1 calculations are based on the assumption that sunken oil impacting the shoreline will be reflected back into the water This assumption allows modeling of accumulation at the shoreline However the assumption may not be realistic at times after which the majority of the oil mass is predicted to reach the nearshore environment At these later times predictions may indicate a bouncing of the patch off of the shoreline To attempt to avoid this potentially unrealistic result a warning message will pop up 1f the requested prediction time is estimated by the model a priori to be after the
21. een the user and the processing modules which perform all required processing of raw input from users and provide the core module with the appropriate information Among the most important tasks of the GUI aided by the OPI module are Collection of basic information on the spill in terms of time of occurrence and coordinates longitude and latitude Display of the location of the spill on the map canvas within a 2 degree radius around the spill to allow selection of the desired modeling area Storage and processing of input data from single or multiple sampling campaigns Assignment of the resolution and scale rectangular of the results based on default values or a user defined number of modeling nodes in each direction north south and west east within the desired or default modeling area Setting of customized dates and times at which sunken oil mass prediction is desired Processing storage and passage to the core module of user input Acceptance of output from the core module and geo referencing of it on the local map presentation of relative probabilities of finding sunken oil O 1 on a relative color coded scale with hotspots in red 1 fading to blue 0 Although probabilities of sunken oil decrease in time the hue is re rendered in every time calculation such that the areas with the highest probability of finding sunken oil are shown in red at each modeled time independently of relative probabilities shown for
22. environments with minimal damage alteration or loss of functionality and e The source code is viewable at all times Python is a general purpose high level programming language first released by Guido Van Rossum in 1991 in The Netherlands Python is flexible applicable to many programming domains and available on many platforms The language has an open community based development model managed by the non profit Python Software Foundation Python s design philosophy emphasizes programmer productivity and code readability core syntax and semantics are minimalistic employing white spaces and easy commands Python supports multiple programming paradigms primarily object oriented and features a fully dynamic type system and automatic memory management similar to Perl Ruby Scheme and Tcl The Python programming language has an open general public license GPL that allows source code modification addition of original code as imported packages and publication of derived work as GPL Packages that are created around the world are available to the Python community Python has also been used as an extension language for many existing systems including GIS Web programming numerical tools and the R statistical package 1 3 Objectives of this User Manual A Guide to the SOSim GUI The objective of this user manual is to guide the operator through installation and operation of SOSim The GUI is the point of connection betw
23. he use of a polyline If the number is not changed before clicking on the Select from Map button the number of vertices defaults to 6 Select from Map button clicking on this button enables two functionalities First is a message box that instructs the user to Please select X points on the land border shown on the canvas where X is the number in the spin box 6 by default Straight shorelines can be approximated by two points slightly curved shorelines may be approximated with three points and so on offering the potential to greatly reduce run time Second the map tool with which the user defines these points is enabled To select each point click the left mouse button at a point on the shoreline depicted in the GUI canvas release it in the same position and then proceed to the next point A thin yellow line will follow and connect these points to create a 2 to 10 vertex polyline approximating the actual boundary Note that if a very large modeling area is selected that includes land but all shorelines are far 30 from the spill and dispersing oil the predicted results will be negligibly different from the results that could be obtained much more quickly by selecting the No Nearby Boundaries option and e No Nearby Boundaries button this button communicates to the model the lack of need for computing reflection of sunken oil from shoreline boundaries within the modeling area for example when the geographic area to be model
24. iles The oil type refers to the classification of oils into six generally 17 accepted classes by weight and other properties affecting their behavior in the environment Research Planning Inc 1994 as explained below under Oil Type The following conditions apply to the Spill Information input section Only a point source spill occurring at a pair of coordinates in degrees of longitude and latitude can be modeled Spill Name prompt allows the user to set the title of the spill e g DBL 152 Characters lt gt and are not accepted but blanks between words are recognized If a change in typing is required an informational message will pop up after all inputs to the Spill Information panel have been entered Oil Type set the spin box from 1 to 6 according to the type of spilled oil as follows Research Planning Inc 1994 Type 1 oils are very light perhaps 31 API gravity including gasoline and very volatile hydrocarbons Type 2 are moderately volatile and soluble including jet fuels diesel fuel number 2 fuel oil and light crude oils Type 3 includes most crude oils characterized by their persistence and diminished propensity to evaporate about one third of the total mass evaporates within 24 hours Type 4 oils may have 10 API gravity little propensity to evaporate or dissolve and high likelihood of sinking Type 5 oils have essentially no evaporation potential weather very slowly and sink
25. immediately including heavy industrial fuel oils Type 6 oils include heavy animal or plant oils 18 Spill Time set the spin box to the time at which the most significant oil loss occurred Notice that you can edit the hour and minutes using either the mouse cursor or the arrows of the spin box or both Then set the day month and year of the spill using the smart calendar that pops up when you click in the drop down menu located to the right The date line edit can also be changed manually using the mouse cursor and the keyboard Longitude this line prompts for a decimal quantity in degrees WGS corresponding to the longitude coordinate at which the spill occurred All quantities must be greater than zero and must have decimal figures that can be zero In front of the prompt the user shall select the radio button E longitude east or S longitude south which will assign a direction to the quantity you typed and will guide the canvas to the world s geographic meridian closest to the longitude of the spill site Latitude this line prompts for a decimal quantity in degrees WGS corresponding to the latitude coordinate at which the spill occurred All quantities must be greater than zero and must have decimal figures that can be zero In front of the prompt select the radio button N latitude north or S latitude south which will assign a direction to the quantity you typed will guide the canvas to the world
26. ion and results management to obtain maps of relative probabilities of finding sunken oil at user specified times of prediction that are not conditional upon the values of uncertain parameters of the model 1 1 What is SOSim SOSim is a predictive Bayesian multi modal Gaussian model of relative probabilities of finding sunken oil at points on a bay bottom and in time designed to accept primary information in the form of limited field data at one or more sampling times The predictive relative probabilities produced are not conditional on the values of uncertain model parameters such as the water velocity and coefficients of dispersion on the bay bottom These probabilities can be interpreted as relative oil concentrations depicted to occur on the bottom in somewhat more disperse patches than are actually occurring due to uncertainty in the advective and dispersive forces acting on the oil at depth Due to the lack of information on the total oil sinking as a function of time the model cannot assess absolute concentrations but rather relative concentrations showing oil hotspots and areas where oil may not be collecting All the functionalities designed for the model have been tested and verified but a formal verification of the results has not been possible yet as it depends on deployable data distributed in two sets in time one to be used as input and one more to compare results with It is expected to convert the current release can
27. ling campaign data file is shown in Figure 0 6 SyntheticCampaign3_Char txt Notepad File Edit Format View Help Point Lon Lat on 1 79 37600 32 70000 0 12151 2 09 373 0U 32 70000 0 87663 3 79 37000 32 70000 0 87663 d 79 37900 32 70300 O 49654 4 79 37000 a2 70300 3 99130 5 79 37300 a2 70300 4 423594 7 79 37000 a2 70300 O 8 663 z 79 38500 32 70600 0 01982 9 79 38200 32 70600 1 28053 10 79 37900 32 70600 12 44053 11 79 37000 32 70600 18 17246 le 79 37300 32 70600 3 99130 1s 79 37000 a2 70600 0 13181 14 9 38800 a2 70900 O 02895 15 79 38500 a2 70900 2 26075 16 79 38200 a2 70900 26 544534 1 79 37900 32 70900 46 86520 14 79 37600 32 70900 12 44053 193 79 37300 32 70900 0 49654 ZU 00 321 0U 32 71200 0 02895 el 7 9 38800 32 7 1200 2Z 72322 7 22 79 38500 32 elle 00 38 7 601 23 9 38200 2 7 1200 B2 73948 24 f9 4 900 See L200 26 54434 25 f9 3 7600 32 7 1200 1 28053 26 9 39400 32 71500 0 01982 27 f9 391700 32 71500 2220075 26 9 38800 32 71500 38 7 601 eg 79 38500 32 71500 100 00000 30 79 38200 32 1500 38 7 601 ERI 79 37900 32 1500 2 20075 a2 79 37000 32 1500 0 01982 33 79 39400 a2 1800 1 28053 a4 f9 391700 32 7 1800 26 54534 35 9 38800 32 7 1800 Be 72948 36 9 38500 32 7 1800 38 7 601 af 9 38200 32 7 1800 2 73237 an F9_ 4 7900 aA 77800 O 04894 Figure 0 6 Example uploaded sampling campaign data file 25
28. n New and type the name in capital letters then the given path 14 If your Windows version cannot open the file and requests to select a program from a list click cancel and close the OSGeo4W console and then call SOSim using the following alternative steps 1 2 and 3 1 Double click on the OSGeo4W shortcut logo on your desktop 2 Type in the console python 3 Type execfile C Program Files SOSim SOSimOPL pyw The SOSim GUI will pop up and will be ready to use Repeat the last 3 steps of your convenience always when need to execute SOSim 15 3 Input 01x File Map I BO lekzRBae SPILL INFORMATION Spill Name SEE Oil Type l Spill Time fiziooam 2 foijoijzo00 lt Longitude SAMPLING CAMPAIGN S Sampling Campaign Number i a Started at 12 00 AM 01 01 2000 lt Upload Sampling Campaign Data File Remove Use Existing Calibration File MODELING AREA AND GRID Select Area from Map Auto Select Area Number of grid nodes Default C Define LEGEND GLOBAL AREA COORDINATES 0 0 0 0 TIME VIEW LAND BOUNDARIES Select From Map B No Nearby Boundaries CALCULATION TIMES Prediction number fi At 12 00 am 4 1 1 2000 r spain x Delete Calibrate Calibrate Run Recalculate Calculating 0 cancel Figure 0 1 Main screen of the computer application
29. nting images The post processing tools are located in the Map toolbar of the GUI SOSim Version 1 0rc1 provides only for saving the output of a current project but not for saving the contents of the prompts to be opened as a project later 42 7 Portability of Results To allow the posting of SOSim output data to the World Wide Web for rapid long distance widely accessible transfer of information output image data is produced and saved in Portable Network Graphics PNG extension format PNG is an extensible file format for the lossless portable well compressed storage of raster images As a complement of any output map created by SOSim a corresponding world or georeferencing file with PGW extension is created and saved The PGW file must accompany its parent PNG raster file in order for any output map to be ported to any GIS geographic information system that supports common PNG raster images including but not limited to ArcGIS QGIS GRASS and SOSim itself PNG figures can be imported inserted printed and used like any other common image file 43 8 Software Portability SOSim version 1 0rc1 comes in an executable package for the Microsoft Windows 32 bit platform allowing easy portability among computers having this common operating system Portability to other OS platforms may be considered for future versions 44 REFERENCES Echavarria Gregory M Angelica 2010 Predictive Data Derived Bayesian Statistic
30. odeling area is been defined using the Auto Select Area button then a node is placed every 0 01 degrees approximately every kilometer The second is the Define radio button which makes available one prompt line in each direction for the user to assign equal or different numbers of east west and north south nodes 3 4 Land Boundaries Shoreline boundaries are accounted for by SOSim if the user indicates that land is present in the area to be modeled Two options are available to the user as indicated in Figure 0 9 LAND BOUNDARIES Select From Map le No Nearby Boundaries Figure 0 9 Land Boundaries input buttons and spin box in the GUI Limitations on the boundary conditions accepted by SOSim v 1 0rc1 include a only one mass of land can be defined in each modeling area e g a single polyline within a 29 single modeling area b irregularly shaped geographic features close to shore such as islets and mangroves should be considered part of a single land mass and c the maximum number of vertices 1s set at 10 This latter restriction is not inherent to internal model computations but was fixed so as to limit run times on desktop computers to the order of one day Input is entered as follows Boundaries spin box by default this box is set to 6 but can spin from 2 to 10 The number set in this box is the number of vertices that the user wishes to select on the map so as to mimic the boundary conditions by t
31. orms including the Mac OS could be considered for future versions The software was proved to work in Microsoft Windows 64 bit machines although the development was not intended for this platform To achieve reasonable performance in terms of computational speed hours a 3 0 GHz processor or better is required In cases in which the machine will be expected to perform tasks in addition to SOSim an active duo or second processor is required SOSim can run on a computer with a page file virtual memory of minimum 2 3 GB Nevertheless it is recommended that the memory card is of a minimum of 3 0 GB Memory requirements of SOSim are determined by the fact that Python can allocate memory only up to a total of 2 3 GB including memory required for all machine functions prior to running the model when implemented on the Windows 32 bit platform this limitation is not expected if the model is developed in the future for the Windows 64 bit OS The total memory used by all processes before starting to run SOSim is typically about 512 MB on machines not having many applications installed and many idle processes to run by default except for Windows 7 and some editions of Windows Vista which may consume up to GB when idle Therefore for the majority of spill cases to be solved with optimal resolution and including recalculations it is estimated that a computer would require an available memory of about 1 7 GB that is a difference of 12 about 1 7
32. other times of prediction and e Processing of optional post run operations include display of results at different prediction times and performance of partial recalculations to view results for contiguous modeling areas Useful documentation functions including saving the results as images and printing are also included in the GUI 1 4 Scope of Model Applicability Version 1 0rc1 of SOSim is designed for application to e Sunken oil e Relatively flat bay bottoms dredged bays reef flats and lagoons or pools protected by offshore rocks bays with steeply sloped bottoms would require capability for the use of bathymetric data as prior information a possible future enhancement e Resolution down to the scale of the tidal excursion oil locations effectively averaged across this excursion e Prediction up to the time that the majority of the oil mass is predicted to reach the shoreline e Discrete accidental oil releases as opposed to natural progressive oil seepage and 10 Relatively uncomplicated concave and convex shoreline geometries modeling in straits inland water bodies harbors islet areas and like geographies are not addressed due to computational limitations and the sometimes transient nature of small scale features 11 2 Installation 2 1 Hardware Requirements SOSim Version 1 0rcl has been developed for use within the Microsoft Windows 32 bit operating system environment though porting to other platf
33. rior to the requested prediction time be on file e Delete button If desired you can use this button to cancel the addition of the prediction time that corresponds to the current value in the Prediction number spin box If a prediction time is deleted by mistake the hour and date can be added using a different new index or Prediction number 3 6 Default Input Default input corresponds to internal parameter ranges that may be modified by an expert in oceanography or Bayesian modeling with the aim to optimize the precision of the predictive capabilities of the simulator based on the changing conditions of the modeling scenario or the specifics of a known situation given the occurrence of an oil 33 spill Default input may be modified every time that SOSim is launched if the default ranges need to be changed To modify the default input go to the Options menu of SOSim and click on Modify Default Settings A tabbed widget will guide you through the 3 settings that can be modified including the default initial parameters range the default number of grid nodes and the default resolution Change the settings as required observing that some values are restricted by the software given mathematic and physical constraints Figure 3 11 shows the preview of the Parameters Range tab Modify Default Settings Preview Parameters Range Grid nodes Resolution 505im uses the Following default parameters ranges as the
34. south directions defined on a Cartesian plane representing the area to be modeled Results of the prediction are calculated and plotted at every node in the grid The more nodes requested in each direction the better the spatial resolution of the mapped output and the longer the 26 run time These characteristics are entered in the panel Modeling Area and Grid shown in Figure 0 7 MODELING AREA BND GRID Select Area From Map Auto Select Dres Number of grid nodes C Default T Define Figure 0 7 Modeling Area and Grid input prompts and buttons in the GUI Information is entered as follows e Select Area from Map button this button is a map tool that enables a selection cursor To select the modeling area left click on the map point that corresponds to the north east corner of the desired modeling area Then drag the pointer without releasing the left button to the south east corner of the area to be mapped Release the mouse left button to set the area Figure 0 8 It is not necessary to include the spill site marked with an X within the modeling area though it may be helpful to model the vicinity around the spill site first E ISE Ie BDI Fao SPILL INFORMATION Spill Name My Spil oil Type j H Spill Time lans an 2 ono200 lt Longitude Iaa ase Latitude 27 8 MODEL SET UP FOR SPILL MY SPILL ELEGEM i GLOBAL AREA COORDINATES 0 0 0 0 TIME VIEW
35. to oil dispersion total sunken oil mass may simultaneously be changing due to sinking and re suspension 37 Therefore colors are not related from one time of prediction to another that is the same colors in different maps do not indicate the same relative concentration values Rather a scale of dark red to dark blue is presented at each time indicating relative spatial probabilities of finding sunken oil at each time independently The same interpretation applies whether individual prediction times are requested as part of the same run or in separate runs It should be noted that the internal calculations of SOSim assume a constant total sunken oil mass in time so as to maintain internal consistency and quality control checks on the integration computations However this assumption does not affect model output 5 2 Optional Output Optional output includes 1 map results for areas that are either contiguous to the current modeling area zoomed in within the current modeling area or zoomed out to include a larger area and 2 maps updated with new or revised sampling campaign data Optional output is obtained as follows Figure 0 13 Pan View button set 38 To model areas contiguous to a current output map e Under the section Pan View Figure 5 1 select the direction in which the desired contiguous region is located with respect to the current output map in the display For example if the desired new modeling
36. uantitative measurements are available measurements can be entered with accuracy of up to several decimal places 21 When complete the Excel file must be saved as a txt tab delimited file in the location of your preference Figure 0 4 shows an example of the Excel file EJ Microsoft Excel SyntheticCampaign3_Cbar txt File Edit View Insert Format Tools Data Window Hell NaBH eBSsiagkivals4aa s j Sie 10 BZ 015 SS i deskPOF Z oe Goto Office Live Open Save A fx Point Con 79 376 32 7 0 13151 79 370 32 7 0 67663 79 37 32 7 0 67663 2 E d 79 3709 FU 0 49654 B 9376 32 03 3 9913 b 79 373 34 03 402394 i Os 32 03 U 6 bp2 Q 79 05 32 06 0 01962 2 79 302 34 06 1 26053 WO 79 379 34 06 12 44053 110 2 79 376 34 706 15 17246 q2 79 373 32 706 3 9915 13 9 37 32 706 0 13161 14 79 300 34 709 0 02095 15 79 305 32 709 2 26075 Ib 79 302 32 709 26 54534 VF 09 379 32 709 46 0652 41a THA TC Ja T 1439 a ao Figure 0 4 Creation of a sampling campaign file Recording spatial coordinates and observed relative concentration values scale 0 to 100 for each sampling point in an Excel file To upload each sampling campaign file make use of the prompts and buttons in the Sampling Campaign s panel of the user interface shown in Figure 0 5 22 SAMPLING CAMPAIGNS Sampling Campaign Number 1 Started at J12 00 AM Jo1f01 2000 Upload Sampling Campaign Data Fil
37. utton is used to instruct SOSim to first calibrate the model using the sampling campaign files and then immediately begin computations for prediction This button is intended to obtain the results of a complete scenario by 39 investing time in only one initial configuration Overnight calibration and runs sessions are possible using this button No changes to the input are allowed at this point until the core module has completed the Bayesian processing stage Recalculate button this button has several uses First the button may be used to generate results for the same spill scenario at a different time prediction without data entry and recalibration Second the button may be used to obtain a new prediction following a change in resolution boundaries or geographical area post processing and optional output again without additional data entry or recalibration Third the button can be used to generate results using an existing calibration file that has been imported using the Use Existing Campaign button in the Sampling Campaign s section 4 2 Run Time and Progress Bar Run time increases as function of a the number of output nodes or geographical resolution selected by the user b the existence of boundary conditions within the selected modeling area and c the number of vertices in the polyline representing the shoreline if present A longer time of calculation about 5 times longer should be allowed when modeling
Download Pdf Manuals
Related Search