Home

Public Hurricane Risk and Loss Model Primary Document Binder

1. Wind Speed Bands mph Figure 3 4 12 Histogram for PVxz 3 D 16 Wind Speed Probability greater than mid Max Wind Speeds for Zip 33156 0 35 0 3 0 25 0 2 0 15 0 1 point of the band Pv gt y 0 05 P amp amp PA 75 SLE LS qegqPq gqegg Wind Speed mph Figure 3 4 13 Histogram for PV gt y 3 D 17 Cumulative Wind Speed Probability Cumulative Frequency of Max wind speeds for Zip 33156 0 4 0 35 0 3 0 25 M 0 15 M 0 05 0 AA BELG NE x Peg AN Qe ERICA DS ISS NS Wind Speed mph Figure 3 4 14 Cumulative frequency diagram for maximum wind speed probabilities for zip code 33156 3 D 18 Section 4 Insurance Loss Model ILM Use Case VII Section 4 1 General Description of ILM Insurance Loss Model ILM calculates the expected losses during storms There are two variations of ILM Scenario ILM and Probabilistic ILM Scenario ILM takes actual observed wind speed or modeled wind speed per each zip code and calculates the expected losses using Vulnerability Matrices provided by the engineering team per loss type given the input exposure insurance policy data Probabilistic ILM on the other hand uses all possible wind speeds from 50 to 250 mph together with
2. 2 2 3 4 Class Diagram and Description A Class Diagram Client submit loginCheckBea n SD Passwd BSisetlD BSisetPasswd BSigetlD BSigetStatus BSilogin SGTIndex BSisubmit y SGTSimul ation BSisubmit BSIIMSL Table Setup SIMSL_Seed BSIIMSL Random General Figure 2 2 7 Class Diagram for SGT Database ESstrQuery gt BMregisterDrive BSigetConnection BSicreateStatement BSexecuteQuery A getSGTDataBean SGTDataEntry EDatasetName EjulianDate ESstrQuery ESgenesis Time BSisetDatasetName BsisetEntry BSisetQuery BSisetJulianDate BSigetDatasetName BsisetGenesisTime 7 igetQuery BSigetJulianDate BSigetSize BSigetGenesisTime BSigetDataTable BSigetData MathModel BSKNFunc S E Es gtDataArray gt BsMcDF SHXXIKNFunc BSigenerateSGT _ BBDWFunc BSigetSGTArmay BSSGTSimulation va BSigensGTValues IMSL Library 2 B 13 B Classes Descriptions Here we would like to give a brief introduction of the functions of the classes we use Generally our design follows the flow chart we developed gt Client A virtual class It means the user who uses this system Or we could say it is the web browser of user ma
3. sees 1 2 Computer Model and Implementation eee 1 3 System Architecture DOSIBI iore ceo ver yo ONG EE EX EO ER E es 2 Storm Forecast Module 2 1 Use Case I Annual Hurricane Occurrence ccc cece cece eee cena eee 2 1 1 General Description Of AHO 0 e ence eee e tee neeneennenaes 2 1 2 AHO Design Requirement cece cee I me mee 2 1 3 AHO Interface Design Requirements sese 2 1 4 Computer Model Design csse 2 1 5 Implementation of AHO csssssseesee II ee 2 2 Use Case II Storm Genesis LIO Severo PE PIRE OS Fe RR ode tates 2 2 1 General Description Of SGT eee cece cece mee 2 2 2 SGT General Requirements csseseeseeee mee 2 2 3 Computer Model Design csssese eceeneee ee eeeeneeneeas 2 2 4 Implementation of SGT 0 cece cence ne I 3 Wind Field Module 3 1 Use Case INE Storm Track Model re ere heh nonne enr 3 1 1 General Description of Storm Track Model sees 3 1 2 Technical Description of Storm Track Model esessesse 3 1 3 Computer Model Design amp Implementation eese 3 2 Use Case IV Wind Field Models 3 2 1 General Description of Wind Field Model eeeeeesss 3 2 2 General Requirements of Wind Field Model essssss 3 2 3 Technical Description of Wind Field Model
4. sseueesse 3 2 4 Computer Model Design cece ee cee cee ne eee e 3 2 5 Implementation of Wind Field Model cece cee ee sence eee enee 3 3 Use Case V Wind Speed Correction sisi en EY hn X He Xn 3 3 1 General Description Of WSC 0 0 0 0c cece ence tence I eneenaees 3 3 2 WSC General Requirements sese 3 3 3 WSC Interface Design Requirement eesess 3 3 4 Computer Model Design c sss 3 3 5 Implementation of WSC cssssssssssssse IH e 3 4 Use Case VI Wind Speed Probability eese 3 4 1 General Description Of WSP ssssessesseee ene eneeneennees 3 4 2 WSP General Requirements csse 3 4 3 Computer Model Design csse 3 4 4 Implementation of WSP 2 0 0 0 cece ce cceee eee e eee e neces ene eneeneenaenaes 3 D 12 4 Insurance Loss Model ILM 4 1 General Description of ILM 0 c cece cee ence ence eee eeeeeeneeenaeeeas 4 A 1 4 1 1 Design Requirements iere eU bc 4 A 2 4 2 Detailed Design and Implementation of Insurance Loss Model 4 A 5 4 2 1 ILM Implementation Steps cssssse cece ence nee ne enone eeneenes 4 A 7 4 3 Computer Model Design se dsenseuenine Eon HA iene aa ques 4 A 15 4 3 1 Use Case View of Insurance Loss Model ILM usesss 4 A 16 43 2 System Design ced EE EH Pe 4 A 16 4 3 3 Class Diagram and D
5. 2 edus8650 FDOI fdologin jsp EDE Free Hotmail 2 Windows 2 Windows Meda E Dictionary al FIUAHRC Public Hurricane Risk and Loss Model Relogin Page UserID PassWD LOGN Wrong user name password y Ee ol a Figure 2 2 13 Snapshot of the Login error page for SGT 2 2 4 2 SGT page If the login is successful the user can go to access Use Case One Annual Hurricane Occurrence via selecting the option Online Demo of Use Case 2 in the Service Selection Page as illustrated below Use Case Two is used to estimate the storm genesis time and generate corresponding time for simulated hurricanes which are obtained from Use Case One Annual Hurricane Occurrence Several steps are conducted to achieve that task 2 B 20 aio File Edit View Favorites Tools Help A e i E m Qe O 9 2 5 95 Dra D2 S Address la http irene cs Fiu edu 8888 FDOI fdoiLogin jsp So Links Service Selection Page Please choose an online service Online demo ofusecase 1 z View documentations FDOI related Publications Online demo of usecase 1 Online demo of usecase 2 Roughness Model Testing Online demo of Engineering Module IDL RPC Testing Prepare standard document Q amp A Figure 2 2 14 Snapshot of the Service Selection Page Step 1 To accomplish the above task first the users need to select a year range The Dataset Selection Page is designed
6. i damageRatio i i i f E companyProcess I 1 1 I 1 gt 1 l company i N l l policy I 1 matrices i i i 1 1 I 1 populate i i i i 1 i i 4 i 4 3 5 Sequence Diagram for Scenario ILM ILM Matrices Damage Ratio WindSpeeds Windborne Company Policy I I li I I I LE MEM M cl A windborne y L L 1 L I i windProbability i i gt i damageRatio i I 1 1 gt i i companyProcess i i i 1 i gt i i i i I company l 1 I I i I Policy i i i matrices i l i i 1 i populate I I li li I References 1 2004 National Renovation amp Insurance Repair Estimator J Russell Craftsman Book Company Carlsbad CA 2 CEIA Cost 2002 R Langedyk V Ticola Construction Estimating Institute Sarasota FL 4 A 34 Section 5 Database Document 5 1 Specification for the Project The North Atlantic best track is maintained by the forecasters and researchers at the National Hurricane Center in Miami Florid and the National Climatic Data Center in Asheville North Carolina Currently the Database extends from 1851 to 2001 Based on the provided data we are going to develop a Database system using Oracle software so tha
7. FLORIDA DEPARTMENT OF FINANCIAL SERVICES PUBLIC HURRICANE RISK AND LOSS MODEL PHRLM PRIMARY DOCUMENT BINDER The Document This binder contains a complete set of documents specifying the model structure detailed software description and functionality Project Supervisors Dr Shu Ching Chen Dr Mei Ling Shyu Associate Professor Associate Professor School of Computer Science Electrical and Computer Engineering Florida International University University of Miami Development Team Members Min Chen Kasun Wickramaratna Team Leader MS Candidate Ph D Candidate School of Computer Science School of Computer Science Florida International University Florida International University Na Zhao Xiaosi Zhou Ph D Candidate Ph D Candidate School of Computer Science School of Computer Science cee Florida International University Florida International University Testing Team Members Indika Priyantha Shaminda Subasingha MS Candidate MS Candidate Electrical and Computer Engineering Electrical and Computer Engineering University of Miami University of Miami Khalid Saleem Kasturi Chatterjee Ph D Candidate Ph D Candidate School of Computer Science School of Computer Science Florida International University Florida International University Table of Contents 1 The Public Hurricane Risk and Loss Model cssceeeecseseeseees 1 1 General Description of PHRLM Model
8. 8 1 System Architecture System is implemented in three tier architecture Following diagram gives a big picture of the system software arrangement HTTP Request HTTP Response i i l SSL Request Beans Oracle9i Database SSL Response HTTP Request ETTP Response IMSL Statistical and Mathematic Library Figure 8 1 System Architecture 8 2 Software List Java 1 5 JDK 1 3 1 IMSL library CNL 5 0 OC4J v1 0 2 2 1 Oracle 91 AS 9 2 JNI 1 3 1 IDL Version 6 0 MapInfo Data 2004 Dynamic Census Block Group amp Zip code Boundaries Math works Real Time Workshop 6 2 Geronesoft s Code Counter Pro Software 1 23 Matlab 7 0 8 A 2 8 3 Hardware Configuration PHRLM is a large scale system which is supposed to store retrieve and process huge amount of hurricane historical data and the simulated data And also intensive computations are required for hurricane analysis and projection Correspondingly high speed CPU and large RAM are necessary The hurricane data may be regularly updated and the related mathematical models for the hurricane data model and the projection results are also potentially changeable The system aims to support both professional and general users in a very convenient way Therefore a distributed environment and high bandwidth network are needed to handle the simultaneous requests Considering all these facts following hardware co
9. Database BStrQuery BregisterDriver gyetComection SC reateStatement BSexecuteQuery Bibackward gt DSSelecti getDBean submit en trQuery Bisubmit ResultSet DataTable submit pem _ BsetDSName N SimuSele MBsetQuery ction BSgetDSName BgetSize Es ub mit WSgetDataTable T N etData S fitDistriBean WBsisetParam BSitDistribution doS imulation BSCompare MathModel EU inomial BSisimulation o NumericSet zi Eset myButton B3ength EbuttonName ESmaxlength Baction Bgetlength V BSigetmax BSsetfromstring BSvalue Button from awt BSButton BSB utton addActionListener addNotify BSigetActionCommand BSgetLabel aramString rocessActionE vent rocessEvent emoveActionListene r SsetActionCommand setLabel dataEntry Year ESstormNo BSisetEntry BSisetYear BSisetStormNo BSgetYear BSigetStormNo CalMVSBean a nean B3standDevia variance BSgetMean BSgetVariance etStandDeviation WSprocess IMSL Library BSIMSL Poisson BSIMSL Binomial random poisson random binomial Figure 2 1 8 Class Diagram for AHO 2 A 14 B Classes Descriptions This section add
10. Updates to ZIP codes historical meteorological events and the related characteristics will also trigger the updates to the model results Whenever the new data or new modeling methodologies become available which results in a non trivial improvement in the modeling results a new model version number is assigned The PHRLM project development team will maintain archive and document the features of each model version The first version is released with version number 1 PHRLM Version 1 Version number will be incremented by one at each yearly update to the system Each time a new version is released mid of the year the version number is incremented by one decimal fraction When a new model version is released a release document with detailed documentation for users and the programs and data that are used in this release will be packaged and tested by crosschecking Standard test cases are also packaged with the release to allow later verification This assures the correctness and consistency of each release PHRLM s software development team employs source revision and control software for all software development In particular PHRLM employs Concurrent Versioning System CVS an accepted and effective system for managing simultaneous development of files It is in common use in large programming projects to track modifications of all source code CVS maintains a record of the changes to each file allowing the user to revert to a pr
11. JO search Le Favorites Que O C2 2 Sa Wos 3 Address El http firene cs fiu edu 8888 FDOI SGT SGTsimulation jsp SGT simulated value table 2631 3 3722 1195 g 4858 2952 11 772 2329 15 2485 2981 19 2875 4092 23 2722 1220 27 3565 4066 31 3134 3932 35 3104 2330 39 3173 3568 43 3144 3304 47 3609 1177 51 2216 2349 55 3757 4425 59 1914 4057 63 4003 4050 67 2996 3964 71 3830 2287 75 869 2501 79 2649 2545 83 4756 3237 87 1600 3102 91 3234 ls cac I ea Figure 2 2 17 Snapshot of the result table in the result page for Use Case Two 2 B 23 Section 3 Wind Field Module Module II Section 3 1 Storm Track Model Use Case III 3 A 1 3 1 1 General Description of Storm Track Model Strom track model is aimed at generating the storm tracks for simulated storms based on data obtained from Use Case II and stochastic algorithms The storm track model consists of two main components the empirical probability distribution generator GENPDF and the storm track generator STORMGEN Descriptions of these components are given below 3 A 2 3 1 2 Technical Description of the Storm Track Model 3 1 2 1 The empirical probabilit
12. Hurricanes Xo Xx Observed Frequency Expected Frequency Oo s Ox Eo Ex 0 Oo Eo 1 O E 2 O E2 Xk Ox Ex Output from Step 3 Table 2 1 7 Probability distribution Year Range 1851 2003 Type of fit Mean Variance p value Goodness of fit Poisson Negative Bin 2 A 8 4 The system presents the following question to the user 4 1 How many years would you like for your simulation 4 1 1 The user input the number of years for simulation For example 100 000 years 5 The system uses IMSL routines to generate a random sample from the chosen distribution obtained in Step 3 and generates a number of simulated years SYo SYn and their associated numbers of hurricanes SH SH Output from Step 5 Table 2 1 8 Simulated years and their associated numbers of hurricanes Year SYo SY Hurricanes SH SH SYo SHo SY SH SY SH SY SH 6 The system stores simulated years and number of hurricanes see Table 2 1 8 Note Steps 2 4 are repeated for each year range that the user requests 2 A 9 2 1 3 AHO Interface Design Requirements This part designs the GUI Graphic User Interface for the Annual Hurricane Occurrence AHO The user interface design aims at providing a friendly and easy to use environment for the users to log in to the PHRLM system and access the AHO us
13. SUBMIT QUIT Figure 2 1 4 Simulation selection interface D The fourth step simulation results display Figure 2 1 5 portrays the Simulation Results Display Interface The system displays the year range selected at step 2 the probability model used the number of simulated years user designated at step 3 and the simulation results Year Range 1851 2003 Probability Model Negative Binomial Number of Simulated Year 10000 Number of storms Simulated Years NEW SIMULATION QUIT Figure 2 1 5 Simulation results display interface 2 A 11 2 1 4 Computer Model Design 2 1 4 1 Use Case View of AHO A Actors There is one actor scientists in AHO Scientists use this use case to find a statistic modal with satisfying goodness of fit conduct the simulation and observe the simulation results B Use Case Use case AnnualHurricaneOccurrence is used to estimate the probability distribution for annual hurricane occurrence and to generate a series of simulated years along with their associated numbers of hurricanes occurrences with respect to the probability distribution that has the best goodness of fit C Use CaseDiagram Figure 2 1 6 shows the use case diagram for AHO AS een N AN Scientist AnnualHurricaneOccurrence Figure 2 1 6 Use Case Diagram for AHO 2 1 4 2 System Design This part describes the system design Appropriate diagrams are provided to describe t
14. To input more than one set of input data the user can change the number of input data sets by using the number of sets change option provided at the top of the dataset input page The user inputs the desired number of data sets in the blank after How many sets and then clicks Set button Figure 3 3 14 shows an example dataset input web page after the user requests three sets of input data 3 C 19 F Roughness Calculation Microsoft Internet Explorer Ele gdt vew Favorites Tools Help O x aA f Os yere US E I Ga 173 ess El http firene cs fiu edui8868 FDO1 WSCindex jsp Wind Field Roughness Calculation How many sets ies Input From Wind Model Zip Code wind Speed m s Wind Direction Roughness length m 33133 90 0 01 33133 120 0 02 33133 360 0 03 Zip Code Zip Wind Speed Vo Surface wind speed for open terrain produced by the wind model m s Wind Direction Wd Surface wind direction Deg from North Roughness Length Zoo Roughness length m for open terrain 0 03 m Site directory Documentations FDOI Publications Demo Usecasel Demo Usecase2 Engineering Module Figure 3 3 14 Snapshot of the web page after user specify the sets number to 3 Step 2 Secondly the system constructs a query using the user s data to obtain desired data from the underlying Oracle database This desired data along with the user data is used to carry out the correct
15. Total WSCCalVamphBean java 269 202 3 80 554 WSCSpeedCheckBean java 78 99 5 29 211 WindSpeed jsp 75 580 1 10 666 WSCindex jsp 210 0 0 31 241 Use Case VI Wind Speed Probability WSP Filename Source Comment Both Blank Total WPHeader h 7 10 0 3 20 WPStruct h 103 14 6 31 154 WProbability cpp 520 198 59 351 1128 WPStruct cpp 7 10 0 10 27 WPUtils cpp 6 11 0 2 19 WSPCalc jsp 123 98 2 11 234 WSPtask jsp 154 0 0 24 178 General Insurance Loss Module ILM Filename Source Comment Both Blank Total IL MInputs h 437 196 28 65 726 InsuranceLossModel cpp 5 4 0 1 10 InsuranceLossModel h 1082 258 42 138 1520 Scenario Insurance Loss Module ILM Filename Source Comment Both Blank Total ILMInputs h 431 176 26 53 686 InsuranceLossModel cpp 11 18 0 6 35 InsuranceLossModel h 1133 212 45 148 1538 Use Case IV Wind Field Model Filename Source Comment Both Blank Total dus pro 12 3 1 3 19 fixshots15 pro 27 4 1 9 41 gemf m 5 2 0 0 7 gemf pro 10 5 1 5 21 gemfplex pro 17 4 1 4 26 genstrex m 11 7 0 3 21 IItoxy pro 14 5 1 1 21 mnrdsg pro 6 2 1 2 11 mnrdu pro 6 2 1 2 11 obc m 7 3 0 0 10 onefix m 27 12 0 7 46 pkwinds pro 75 28 1 20 124 reach pro 7 0 1 0 8 rsdsg pro 22 5 1 6 34 rsdu pro 21 5 1 6 33 selset m 3 3 0 0 6 sgdv
16. We examined the trace file carefully and chose the statement that consumes less resource and has the better overall performance 5 A 27 Section 6 PHRLM Quality Assurance Section 6 1 Coding Guide Lines 6 1 1 About the Coding Guidelines This document is prepared as a part of the PHRLM project All the developers involved in the system development are asked to read and follow the instructions given in here In general this document may be read as a guide to writing robust and readable codes Examples given in here are mainly focused on programs written in C flavors but the content is generally applicable for programs written in any other programming language 6 1 2 File Organization 6 1 2 1 Source files e Keep your classes files short don t exceed 2000 lines of code e Divide your code up make structures clearer e Put every class in a separate file and name the file like the class name This convention makes things much easier 6 1 2 2 Directory Layout Developer s own structure e Create a directory for every use case and keep all the related codes in that e For each major revision create a subfolder with the revision number e Do the CVS before any change and keep your own backup always System Directory Structure e All the codes C and JAVA codes Create a subdirectory under the use case name inside home irene 1b oracle j2ee home default web app WEB
17. v sd Vs Normalized storm relative tangential wind component u s Normalized storm relative radial wind component Ve IMS uta C a Friction coefficient h mean boundary layer height 3 B 9 Ca Drag Coefficient c a normalized translation speed g max g s 2vo s s f A1 d s 39 vs f A2 where a dot represents a derivative with respect to s 8 5 and ds depend only on Vo and f O s 9 v s v5 Normalized departure from gradient balance Scaling of the governing equations prior to implementation Substituting the terms from the above definitions and changing the radial coordinate from r tos the steady state form of the governing equations 1 and 2 become uduts v 0 9 u o g s o Q u csind w c 0 A3 udo s v 0 0 0 u d s o av 0 ccos w c 0 A4 w 4 u csin 9 v 0 ccos Hp A5 where w is the total normalized earth relative wind In the event that c vanishes so that the cyclone is stationary these equations reduce to the ordinary differential equations uu o g s o auw 0 A6 u s o d av o w 0 A7 v V 0 w 0 8Vu v w w 1 fw f 0 132653 A8 ee 99 for the radial profiles u s and o s Here indicates differentiation with respect to S Equations A3 and A4 supplemented by A5 constitute two coupled time independent partial differential equations for the storm relative rad
18. 8 wide nodes with dual 375MHz Power3 II Winterhawk II processors 27 thin nodes with single 375MHz Power3 II Winterhawk II processors e Personal Computers 10 machines 2 Machines with following Configuration Dell Dimension 4550 21 inches Monitor Windows XP operating System Pentium 4 3 06GHz Processor 1GB RAM 230GB Disk Space 16X DVD ROM 3 5 1 44 MB floppy drive 3 Machines with following Configuration Dell 1400 GX 400 Minitower 21 inches Monitor Windows 2000 Operating System Pentium 4 1 4 GHz Processor 1 GB RAM 256K Cache Two 40 GB EIDE 7200 rpm ATA 100 Hard Drive 16X DVD ROM Harman Kardon 19 5 Speakers 10 Machines with following Configuration Dimension 4100 Series 19 inches Monitor Windows 2000 Operating System Pentium III 1GHz Processor 256MB SDRAM 40GB Ultra ATA 7200 rpm hard drive CD ROM 3 5 floppy drive 8 A 4 e Laptops 4 machines Two Laptops with following Configuration Latitude C600 Windows 2000 Operating System Pentium II 850 MHz 256MB RAM 14 1 TFT 20GB Hard Drive 8X DVD with software Nylon Carrying Case Two Laptops with following Configuration Dell Latitude D800 Pentium 256 M Processor 755 2 0GHz w 15 4 WSXGA Display 1024MB DDR SDRAM 2 DIMMS NVIDIA 256 128MB DDR Video Memory 128MB 60GB Hard Drive 9 5MM 5400RPM DELL LOGITECH USB OPTICAL MOUSE Internal 56K Modem 8 24 24 24X SWDVD CDRW Combo Drive Intel 256 PRO Wireless 2200 WLAN 802 11b g 54Mbps miniPCI Card
19. Microsoft Internet Explorer Ele Edt Yew Favontes Toos Help Qu O MAD seach Prats Qs O E 1 138 3 Address dE hetp firene cs fiw edu 8888 FDOLWSC WindSpeedCalc jsp Wind Field Roughness Calculation Results Please input a valid Wd Wind Direction Integer 0 WD lt 360 in the dataset 1 Please input a valid Zoo Roughness Length double in the dataset 2 Please input a valid zip code in the dataset 3 Please input a valid Wd Wind Direction Integer 0 lt WD lt 360 in the dataset 3 1 Input from wind model for each zip code centroid Zip Zip Code Vo Surface Wind Speed for open terrain produced by the wind model m s WD Surface Wind Direction Deg from North Zoo Roughness Length m for open terrain 0 03 m 2 Input from roughness table for a given wind direction Zoa Actual roughness length based on FEMA HAZUS conversion table relating land use land cover LULC to aerodynamic roughness m 3 Output U o Open terrain friction velocity m s U a Actual terrain friction velocity m s Va Surface wind speed for actual terrain m s Vamph above with english units of statute miles per hour Site directory Documentations FDOI Publications Demo Usecasel Demo Usecase2 Engineering Module Q amp A Software Documentations Related Model Papers DataSet Selectionl Dataset Selection Main Page Document Preparation A Internet Figure 3 3 18 The result webpage catches al
20. Validation http www epa gov quality vandv html 6 B 4 Section 6 3 Model Maintenance and Revision 6 C 1 6 3 1 Model Maintenance and Revision PHRLM has developed a clearly documented policy for model revision with respect to methodology and data Any enhancement to the model that results in a change in any Florida residential hurricane loss costs also results in a new model version number PHRLM uses version control and tracking software to identify all errors as well as modifications to code data and documentation 1 PHRLM employs consistent methods for data and documentation control for all software development including both server and client programs written in C and Java The installation date program specification personnel involved current version number and date of most recent changes are documented for the individual components in the system The data and model is maintained and updated each year At each year the ZIP Code information is updated to reflect the most recent changes within the past 12 months In particular the ZIP Code boundaries and the centroids are updated and using this updated information the ZIP Code related features are updated including distance from the coastline population centroid elevation and surface roughness etc The historical hurricane data for Atlantic Basin is periodically updated to take into account new hurricane events
21. displayed in the web browser 2 A 18 C Fit Distribution Process sim ulation fitD istributiion m athModel IMSL IMSL SimuS election fitD istriBean MathModel Library setP aram itD istribution i i g poisson zy IMSL Poissonf binom iat MSL Binomial lt doSimutation sim ulation random poisson lt Figure 2 1 11 Sequence diagram for fit distribution process Stepl The simulation object passes the calculated mean variance standard deviation values and other parameters to the fitDistribution object Step 2 The simulation object then calls the fitDistribution object to fit the distribution using Poisson and negative binomial model and then identifies the better one Step 3 The fitDistribution object achieves the distribution fitting task by calling the math models written in C Step 4 The math model calls the IMSL libraries to get the results Step 5 After identifying the better model the simulation object calls the fitDistribution object to do the actual simulation Step 6 The math model does the actual simulation work and return the result set back to the original caller 2 A 19 D Plot Process client Client plotObject plot
22. the number of damage ratio intervals is N Portfolio is replacement cost Get damage ratio vectors i e the middle point values for N intervals Xs Xc Xap XaLE TE For a wind speed W get the vectors of the probability of damage Pp Pp LMc LM c LMap Ppap Pare provided DMs Z Pp Xs DMc EPy Xe DM ap PPop Xar es Lap LMap SumDM DMs DMc DMap Ds pMg D SumDM De DMc D SumDM L 1 25 LMc Lap 1 25 LMap Dap DM ap D SumDM Ii I I I I Mobile S N Use weighted Use weighted 1 mobile home Misc matrices l matrices I I I eee m 1 N Vi 1 25 LMs II L 0 5 LMs Lap 0 1 LMs LMc LMap LMale provided Le 1 25 0 5 LMs Lap 1 25 0 1 LMs Late 1 25 0 2 LMs provided provided 4 A 20 Ve 2 Xc AP Vap Xap LMate XALE 2 ic T A ALE II 4 A 2 INPUT Probability of wind speed Pw for given wind speed in the given zip code Mobile home SumLs SumL Ls Pp Output SumL s SumL SumLc Lc Pp SumL c SumL app SumL app SumL app Lap z Dio SumLA E SumLA E t Late Puis Y Output SumEL SumLarr N Finish portfolio Y Output SumAEL Q N Finish company oop 3 Y 4 A 22 REMARKS I Map the portfoli
23. 1013 W 10 627 7 Else if latitude lt 25 N P 1013 W 12 016 7 Else if latitude 35N P 21013 W 14 172 3 A 5 Else P 1013 W 16 086 Where P is central pressure in mb and W is wind speed in kt 3 1 2 4 Appendix B Relative Intensity Calculation The relative intensity calculation is based on Darling 1991 The calculation is as follows ry 2 461 rh 0 80 e ts to ts es 6 112 exp 17 67 ts 273 ts 29 5 Pda 1013 rh es Lv 2 5 10 2320 rs 273 a ex Ly es l e rv ts Pda b rh 1 es log rh Pda a Then solve for x in x exp a I x b and then finally the relative intensity is given by RI 1013 Pmsl 1 rh es 1 x 1013 rh es 3 1 2 5 Data Sources This calculation requires as input the mean sea level pressure Pmsl which in our case is the storm central pressure the outflow to and sea surface temperatures ts The outflow temperature is taken to be the monthly mean 100 millibar temperature derived by the Climate Diagnostics Center CDC using National Center for Environmental Prediction Center NCEP Reanalysis II data This data is available online at http www cdc noaa gov ncep_reanalysis The sea surface temperature data is monthly mean Reynolds Optimal Interpolation Version 2 Olv2 data Reynolds et al 2002 3 A 6 3 1 3 Computer Model Design amp Implementati
24. 113 117 August 10 13 2003 Newark New Jersey USA 10 A 1
25. 13 Student must adhere to the FIU Code of Computing Practice a policy produced by University Technology Services 14 If a computer security violation is suspected the Systems and Networking Group Manager and or the Associate Director for Computing have the authority to investigate the suspected violation by reviewing and modifying system and user files in an effort to ascertain the extent of the violation and restore system security If a violation of the computer security policy has occurred the Assoc Dir for Computing notifies the SCS Director Assoc Director s and UTS Security Officer of the security violation Once the appropriate steps are taken to restore security the SCS Director is notified and a public statement is made to the SCS user community of the incident as deemed appropriate by the SCS Director 15 SCS systems which require presentation of credentials should use the appropriate encrypted channels SSL SSH etc SCS will be discontinuing application service support of unencrypted logins as we are able to migrate legacy applications services 7 A 6 7 3 FIU SCS Hurricane Preparation Procedures During Hurricane season June November the Director Associate Director or designee may issue an alert to the staff to prepare for an impending storm The Lab Manager may use his master key or one will be made available to him her to enter Faculty offices to begin preparations to safeguard computin
26. 2 0 3 17 rooflayout6044 m 106 37 13 28 184 wall_loading m 100 45 13 16 174 window_pressure_check m 174 20 101 11 306 Use Case Ill Storm Track Module Filename Source Comment Both Blank Total stormgen f 823 274 275 203 1575 genpaf f 1142 403 244 280 2069 Igenpdf h 13 34 5 1 53 6 E 4 Section 7 Security 7 1 Security Procedures PHRLM has implemented security procedures for access to code data and documentation that are in accordance with standard industry practices PHRLM employs a number of physical and electronic security measures to protect all code data and documentation against both internal and external potential sources of damage Summary 1 The application server IRENE and the database server ANDREW as shown in Table 7 1 are considered mission critical servers see its definition in the Security Procedures Manual Section II and are kept and maintained in a secure server room which limits non authorized access Access to the server room is granted by electronic key card and is limited to essential personnel only All servers and desktops are protected with Norton Antivirus software Table 7 1 PHRLM servers HOSTNAME Operating System Purpose andrew cs fiu edu Solaris 8 DB Server and File Storage irene cs fiu edu Red Hat Linux Application Server and File Storage 2 As outlined in the Security Pr
27. A 3 OLA Comments A debel EE Phe mace eens 6 A 4 6 1 5 Variable Declarations ii YN RO RUE eR cre 6 A 5 T6 dd A SS ES 6 A 5 6 1 White Space init Ri de ita 6 A 7 6 1 8 Naming Conventions esses I e ee eem emen 6 A 8 GO EO MRETETEN CE o dodo etes eec eee dea Sire se refe PRINS a tee da 6 A 9 6 2 Data Validation and Verification 0 cece cece cence cece eeeneaeeeeees 6 B 1 6 2 1 About the Document eocen en n pe eee eee 6 B 2 6 2 2 Introduction cc ccc ccc ccc cece ccc eee ence cece ee eese esee eee eee eee 6 B 2 6 2 3 NA oe ENS 6 B 2 6 2 4 Data Security and Integrity cece ee eee scene eH 6 B 3 62 Referentes nara e er op LS Eee nds 6 B 3 6 3 Model Maintenance and Revision and eese 6 C 1 6 3 1 Model Maintenance and Revision eeeeeeeeeeeeee cece cece eens eee e es 6 C 2 6 4 PHRLM Testing Procedures 0 cece esc e ence ence eee eenee teas eeneeeenes 6 D 1 6 4 1 Software Testing Procedures cece cece cence nee ne e 6 D 2 6 3 Code Couiit Tables erro nM 6 E 1 Security T l Security Procedutes uota A A 7 A 2 7 2 FIU SCS Computer and Networking Security Procedures Manual 7 A 4 7 3 FIU SCS Hurricane Preparation Procedures ooooccocccconcccncccnncconcconos 7 A 7 7 4 Non Disclosure Agreement cce cece eeee ence ene eteneeeneeeeseeeneeeees 7 A 8 System Hardware and Software Configurations 5 1 System ATCHItE
28. Code to be mixed together with static HTML or XML templates The Java logic handles the dynamic content generating while the markup language controls structuring and presentation of data Client Side Application Logic Database Server pee ceco tose Sete eee oe oe eee ot eee Li i Li 1 r HTTP SS i Web i4 i Browser i Li Li Li Li OC4J Web Container Server JavaBeans ORACLE DB JNI Interface Library L L IMSL Math Model i L t L Figure 1 4 Detailed system architecture OC4J is short for Oracle9iAS Containers for J2EE It is a complete J2EE 1 2 container that includes a JSP Translator a Java servlet engine and an Enterprise JavaBeans EJB container OC4J also supports the Java Messaging Service and several other Java specifications Advanced techniques such as JavaBeans and JNI are employed in the second layer JavaBean is a Java class that defines properties and that communicates with other Beans via events Properties can be defined within the JavaBean class definition or they can be inherited from other classes JNI stands for Java Native Interface it is part of the Java Developer Kit The actual mathematical and statistical computations are implemented in C C language for the sake of speed JNI then serves as a bridge between java side and native side of an applica
29. Execution Plan 0 SELEC STATEMENT GOAL HINT FIRST_ROWS 47 SORT GROUP BY 466 NESTED LOOPS 1274 TABLE ACCESS FULL OF ATMOSEVENT LIST 466 INDEX RANGE SCAN OF ENSO STORM IDX NON UNIQUE ACKCKkCk kk ck ck ck ck ck ck ck k ck ck ck ck ck k ck ck kCk ck AAA ck ck ck ck ck k ck ck k ck ck k ck ck k ck ck k ck ck k ck ck k ck ck k ck ck k ck k k ck k k kk kkk k Structure 2 select first rows Year count Cyclones from select to char s when t yyyy Year from atmosevent list s where exists select os year from oscillation constant list where os year to number to char when t yyyy and s basin 1 and s type gt 4 group by Year call count cpu elapsed disk query current rows Parse 1 0 00 0 088 w 0 Q0 Q Execute 1 0 00 0 00 0 0 0 0 Fetch 5 0 06 0 07 0 1286 0 47 Edel m wae An a 0289 e 7 C d Misses in library cache during parse 1 Optimizer goal FIRST ROWS 5 A 26 Parsing user id 29 CZHANGO2 Rows Row Source Operation 47 SORT GROUP BY 466 FILTER 1274 TABLE ACCESS FULL ATMOSEVENT LIST 464 FILTER 464 INDEX RANGE SCAN object id 28035 Rows Execution Plan 0 SELEC STATEMENT GOAL HINT FIRST ROWS 47 SORT GROUP BY 466 FILTER 1274 TABLE ACCESS FULL OF ATMOSEVENT LIST 464 FILTER 464 INDEX RANGE SCAN OF ENSO STORM IDX NON UNIQUE
30. Fix Data Storm Storm Genesis SGT Genesis Lat Lon Max Pre ID Name Date Time W S 4 Not Named 20 Aug 1851 2706 180000 21 9 80 4 70 0 4 The system stores in the database the calculated hours between the genesis of hurricanes and 0 00 hours May 01 5 The system uses the data from the output generated in step 3 to estimate the probability distribution of SGT In the following we denote the random variable of SGT by T 5 1 9 2 5 3 5 4 5 5 5 6 The system calculates the number of years of the year range the user entered at step 1 Let it denoted by M The system calculates the number of hurricanes in each year in the year range Let n denote the number of hurricanes in yeari The total number of hurricanes we have is N y n The system sorts all the hurricanes in ascent order according to their SGT Assume now that these hurricanes occurred at t mesOsT ST SXT ST whereW lt N The system also calculates the number of hurricanes that occurred at timeT Let it denoted by f The system calculates the empirical CDF Cumulative Distribution Function for T an estimate of the true CDF F t P T lt t using the following equation 0 ift T t hastet N x if T lt t lt T i 1 2 W 1 l if t gt Ty The system calculates the smooth estimator of F t For a suitable kernel function K and a positive bandwidth sequence Ay t Note
31. INF classes FDOlclasses e Save the latest copy of the codes there e All the JSP files has to be saved in home irenelb oracle j2ee home default web app FDOl useCaseName e DoCVS Example Ex Ccode ojx File Edit View Favorites Tools Help EJ Back gt gt Gsearch Pyrolders lt 4 31808 x uw EE Address Ica Z home default web app WEB INF classes FDOIclasses usecaseone Ccode e Go Folders Size Type Modified home OKB C Source 4 4 2005 1 30 default web app WEB INF classes C FDOIcasses 0 usecaseone Ccod idis Sud gt Type C Source Size O bytes 0 bytes BR Local intranet 6 A 2 6 1 3 Code Indentation 6 1 3 1 Wrapping Lines When an expression will not fit on a single line break it up according to these general principles e Break after a comma e Break after an operator e Prefer higher level breaks to lower level breaks e Align the new line with the beginning of the expression at the same level on the previous line Example Breaking up method calls longMethodCall expr1 expr2 expr3 expr4 expr5 Example Breaking an arithmetic expression PREFER var a b c g f 4 z BAD STYLE AVOID var a b c g f 4 z The first is preferred since the break occurs outside the parenthesized expression higher level rule Note that you indent with tabs to the indentation level and then with spaces to the breaking position in our example this would be
32. Lj If Lsijkn is 2 Linijks then Lsijkn Linijk If Lsijkn is 0 then let Lsijkn 0 Repeat step 15 for C AP and ALE Here these variables are means conditional on the wind speed Generate Le Lap and Late Next to get the expected insured loss for the observed wind speed w multiply each element Lii of the vector Lijk by its corresponding probability Pii to compute Lijknw and then sum over the N intervals Steps 15 17 can be represented by Expected Structure Loss E Ls Y DM Ds Ps Y LMsPs Ds Expected Content Loss E Lc 2 c De Pe x LMP Expected Appurtenant Loss E Lar gt AP Dap Par b LMarPar Expected ALE Loss E Late E ALE Dare Pare x LMarePaLE where Lijkwn LMij if DMijn Dsijk gt Lwijk and if DMijn Dsijx lt 0 then let DMi Dsix 0 i e replace negative values of net of deductible loss with zero The same applies to C AP and ALE 4 A 12 18 19 20 21 22 23 24 25 26 27 ExpectedLoss E L jk E Ls E Lc E Lar E Lave for property k Steps 7 through 18 are repeated for all dwellings of type i in zip code j to generate E Lijx for all properties k 1 K The expected losses are then summed to get the Expected Aggregate Loss for property type i in zip code j K Expected Aggregate Loss E L Y E L Variance will now need to be computed empirically si
33. NYLON DELUXE CASE 3 Year Limited Warranty plus 3 Year NBD On Site Service e Printers Two Printers With following Configuration LaserJet Printer 4100N 2 Extra Memory for two printers e Other Accessories 2 DVD RW Drives 8 4 Safety and Backups Nightly backups of all UNIX data disks and selected Windows data disks at user request are performed over the network onto Exabyte Mammoth M2 tapes Full dumps are taken periodically it works out to every 2 3 weeks and incrementals are taken daily between them A separate full dump of all department UNIX data disks and selected Windows data disks at user request is taken once per semester including IRENE and ANDREW and kept at an off site location usually a different building on campus 8 A 5 Section 9 Training Plan 9 1 Introduction Training of PHRLM must be made available to the development team members computer group who will implement the system and to those who will use it to successfully complete their tasks This document describes the plan that will be used to train the technical staff on PHRLM system It also describes the plan being used to train the other professional group members engineering group meteorology group statistics group and finance group 9 2 Technical Training Plan The Technical training plan is intended for the development team members who will assist with the system development installation conf
34. Of WSP WSP short for Wind Speed Probability is the sixth use case of the FIU IHRC Public Hurricane Risk and Loss model It aims at calculating the probabilities of the 3s gust wind speeds affecting each of the zip codes in the threat area 3 D 2 3 4 2 WSP General Requirements Name Wind Speed Probability Description The user provides the system with the surface corrected wind speed time series for each of the storm The wind speeds are in units of miles per hour mph The system computes the following 1 For each zip code the annual probability of the maximum wind speeds being within the 5 mph interval for all storms starting at 22 5 mph and ending at 302 5 mph 2 For each zip code the annual probability of the wind speeds exceeding the mid point m for each of the 5 mph interval for all storms starting at 22 5 mph and ending at 302 5 mph 1 The end user enters the following input data Y Number of years in simulation T Wind speed time series for each of the storms one file for each storm Z All the zip code in threat area Each of the T consists of following pertinent information Storm_name Storm name Storm_date_time Storm landfall date and time Zip Zip Code V3 Surface corrected 3Sec gust wind speed in mph obtained from the WSC use case A sample input is as follows storm0000002 8 19 1 15 00 Num Zip Vo Wd Zoo Zoa U o U a Va Vimph V3mph 1 32008 21 6816 150 0 03 0 664 1 492 1 857 12 596 38 2
35. Ratios Loss Estimation Module Calculates financial loss by multiplying the damage ratios by values Figure 1 1 Model Flowchart User Input Central Pressure Storm Track Rmax 3 1 A 2 1 2 Computer Model and Implementation 1 2 1 Use Case View of the System Use case diagram is one diagram in UML for modeling the dynamic aspects of a system Use case diagrams are central to modeling the behavior of a system a subsystem or a class Figure 1 2 presents the use case diagram of our computer model for the PHRLM and it shows a set of use cases and actors and their relationships A Actors There are two actors in this system the scientists and the statisticians The scientists can access all use cases related to the Storm Forecast Module and Wind field Module while the statisticians can interact with the Damage and Loss Estimation Modules B Use Cases Use Case I Annual Hurricane Occurrence Use Case I is used to estimate the probability distribution for annual hurricane occurrence and to generate a series of simulated years along with their associated number of storms according to the selected the probability distribution Use Case II Storm Genesis Time Use Case II is used to generate the probability distribution of the origin dates for the historical storms and simulated storms produced by Use Case I Use Case III Storm Track Generation Use Case III is used to generate the
36. Section 3 3 Wind Speed Correction WSC Use Case V 3 C 1 3 3 1 General Description Of WSC WSC short for Wind Speed Correction is the fifth use case of the FIU IHRC Public Hurricane Risk and Loss model It aims at refining open terrain wind speed produced by the hurricane wind model with respect to the actual terrain based on land use land cover 3 C 2 3 3 2 WSC General Requirements Name Wind Speed Correction Description The inputs are zip code surface wind speed for open terrain produced by the wind model surface wind direction and roughness length for open terrain The system generates the following 1 Surface wind speed for actual terrain m s 2 One hour sustained wind speed for actual terrain mph 3 3 Second gust wind speed for actual terrain mph 1 Following are the input data Zip Zip Code Vo Surface Wind Speed for open terrain produced by the wind model m s WD Surface Wind Direction Deg from North Zoo Roughness Length m for open terrain 0 03m Zoa Roughness length based on upstream terrain Lat Latitude 2 Based on the input data from step 1 the system queries the database and returns Zoa parameter which corresponds to the actual roughness length based on FEMA HAZUS conversion table relating land use land cover LULC to aerodynamic roughness m Roughness represents a weighted average of all roughness pixels within a 45 degree sector
37. Selection Page is designed for that purpose which is the first page for AHO Figure 2 1 16 illustrates the snapshot of the Dataset Selection Page 3 Dataset Selection JSP TAR File Edit View Favorites Tools Help Q e amp x le f Search Sf Favorites A meda E Address amp l http irene cs Fiu edu 8888 FDOlI task jsp Annual Hurricane Occurrence Atlantic Basin Oniy DataSet Selection Page DataSet Selection 1851 2003 v 1851 2003 1944 2003 ENSO Multi Decadal Figure 2 1 16 Snapshot of the first web page for AHO A dropdown list containing all valid year ranges is provided to make the selection simpler to the users and to avoid any illegal year range specified by the user There are five possible year ranges 1851 2003 1900 2003 1944 2003 ENSO and Multi Decadal For detailed explanation of these terms see to AHO design requirement part The user selects a year range from the dropdown list and submits his her selection to the system 2 A 23 Step 2 Upon the user s year range selection the system constructs a query and questions the underlying Oracle database to get the data set pertaining to the user s year range selection A Statistical computation is carried out upon the retrieved data set to analyze its numerical characteristics several statistical values of that data set such as mean value variance and standard deviation are derived through this process and a
38. This information is passed to the loginCheckBean class which really communicates with the database If the password or username is not matched with the information stored in the database the user will get an error message and be asked to login again If the username and password are matched user can continue to access all use cases of the system 2 B 16 B Simulation Process client Client dataSelection simulation genSGTData getSGTData database SGTIndex SGTSimulation SGTBean getSGTDataBean Database ml bil e E igetConnection lt rexecuteQuery Cy lt lt generateSGT A Figure 2 2 10 Sequence diagram for simulation process Step1 The user selects the data set e g 1851 2001 and then clicks the Submit button Step 2 The dataSelection object verifies the data set the user selected and then calls the simulation object to get the related data from database Step 3 The simulation object will call the getSGTData object which will connect with the database create the query and then get the desired data from database Step 4 The simulation object will call genSGTData object to generate the SGTs 2 B 17 C Generate Genesis Time genSG
39. all the instances But in general most of these procedures are applicable and should be followed by the developers In case all these procedures are not applicable it is advised to develop your own methods and properly document the procedure followed e Format check Check if the data is in the right format This can be done manually or using any commercially available data manipulating tools such as Excel or Access Mainly data is received in text file format If the input data set is too large do the format test on randomly selected files e Length check 6 B 2 Check the data isn t too short or too long For this check the whole file and then check the expected length of the each field This is applicable to text fields only e Range check Checks a number isn t too big or too small For an example a zip code has to be greater than 0 and less than 40000 e Presence check Checks that a field has been entered Once the above checks are completed and successful it is ready to be imported to the system It is recommended to use data manipulating software or a simple program written by the developer for the data importing process rather than manual entry If it is unavoidable you may use manual entry In either case it is recommended to double check the imported data The above mentioned steps can be repeated and in addition following tests are recommended e Double entry Type the data in twice and compare the tw
40. and lying over the mid point the interval C Use Case Diagram 2 WindSpeedP robability Usecase Scientist Figure 3 4 1 Use Case Diagram for WSP 3 D 6 3 4 3 2 This section includes the appropriate diagrams to describe the system classes System Design components activities and the overall flow chart of WSP General Flow Chart of WSP The flow chart of WSP is depicted in Figure 3 4 2 Co User Specifies the Input Dataset T Y Z For each zip code system calculates the wind speed probabilities lying in the 5 mph intervals l For each zip code the system calculates the wind speed probabilities above the mid point of each 5 mph interval System writes them to a file End Figure 3 4 2 Flow chart of WSC 3 4 3 3 Calculation of WSP The implementation for Use Case six WSP has been completed and meets the requirements specification The back end for the implementation of WSP use case has been coded in C Input 1 Zip codes in threat area 2 For each storm passing within 3 Rmax of the threat area e Peak max sustained wind speed and associated peak gust wind speed already corrected for the upstream terrain of the zip code e Storm name date and time for each peak wind Implementation 1 For each zip code 2 For each storm select the maximum 3s Gust wind speed V3 Note These
41. and password again 2 A 21 PES a ELE 3 FIU IHRC Public Hurricane Risk and Loss Model Relogin Page UserID PassWD LOGIN Wrong user name password E BE PT eme Figure 2 1 14 Snapshot of the Login Error Message 2 1 5 2 AHO page If the login is successful the user can go to access Use Case One Annual Hurricane Occurrence via selecting the option Online Demo of Use Case 1 in the Service Selection Page as illustrated below lol Ele Edit View Favorites Tools Help ay Om O AO la Rem O12 E Address fe http irene cs Fiu edu 8888 FDOI fdoiLogin jsp M Go Links Service Selection Page Please choose an online service Online demo of usecase 1 R4 View documentations FDOI related Publications Online demo of usecase 1 Online demo of usecase 2 Roughness Model Testing Online demo of Engineering Module IDL RPC Testing Prepare standard document Q amp A Figure 2 1 15 Snapshot of the service selection page 2 A 22 AHO is used to estimate the probability distribution for number of hurricanes per year and to generate a number of simulated years with its associated number of hurricanes based on the estimated probability distribution Several steps are conducted to achieve that task Step 1 To accomplish the above task first the users need to select a year range The Dataset
42. clearly as recursive SQL statements in the output file You can suppress the listing of recursive calls in the output file by setting the SYS statement line parameter to NO The statistics for a recursive SQL statement are included in the listing for that statement not in the listing for the SQL statement that caused the recursive call So when you are calculating the total resources required to process a SQL statement you should consider the statistics for that statement as well as those for recursive calls caused by that statement The following examples are two formatted SQL statements with TRPROF Example Structure 1 select first rows Year count Cyclones from to response with the first row quickly select to char s when t yyyy Year from atmosevent_list s oscillation_constant_list o where os_year to_number to_char when_t yyyy and s basin 1 and s type gt 4 group by Year 5 A 25 call count Parse 1 Execute 1 Fetch 5 total 7 elapsed disk query current rows 0 18 0 0 0 0 0 01 0 0 0 0 0 04 0 17 0 47 0 24 0 1 7 0 47 isses in library cache during parse 1 Optimizer goal FIRST_ROWS Parsing user id 29 CZHANGO2 Rows Row Source Operation 47 SORT GROUP BY 466 NESTED LOOPS 1274 TABLE ACCESS FULL ATMOSEVENT_LIST 466 INDEX RANGE SCAN object id 28035 Rows
43. consist of 9 columns The names of these columns are as following Storm ID Storm Name Genesis Date Julian Date Genesis Fix Time Lat Lon Max Wind Speed and Pressure For example in year 1851 there was one hurricane in the threat area Table 2 2 3 illustrates the content of the returned data Table 2 2 3 Record of First Fix Data Storm Storm Genesis Julian Genes Lat Lon Max Pre Id Name Date Date is W S Time 4 Not Named 20 Aug 1851 2397355 180000 21 9 80 4 70 0 3 The system uses data from the output of step 2 to calculate the hours between the genesis of each hurricane in 6 hour resolution and 0 00 hours May 01 The system generates a new matrix consisting of data from the output of step 2 and the calculated hours of each hurricane The matrix also contains 9 columns The column names are the same as those in step 2 See Table 2 2 4 Each day storm data is collected in the one of the following intervals I1 0 00 6AM I22 6AM 12 Noon I32 12Noon 6PM I4 6PM midnight For the sake of simplicity each interval is associated with its starting point So for example since the hurricane with Storm ID 4 happened in the interval I3 on August 20 1851 the 2 B 4 number of hours recorded for this hurricane will be 24 2397355 Julian date of 20 Aug 1851 2397243 Julian date of 1 May 1851 18 2706 Table 2 2 4 New Record of First
44. damage matrices rather than the traditional continuous vulnerability functions Damage ratios are grouped and intervals or classes of various length are used No statistical distributions are fitted or tested The engineering component produces non parametric estimates of damage probabilities for various intervals or classes of damage ratios for structures and contents They do not fit any statistical distributions to the damage ratios Thus for the insured loss model a choice must be made to either fit a parametric statistical distribution for the damage intervals using some of the same standard techniques mentioned above but applied to grouped data or to use a non parametric technique presented as a broad algorithm The advantages and disadvantages of using parametric vs non parametric techniques are well known For our purpose the main advantage of the parametric technique would be computational efficiency Once a statistical distribution has been fitted and its parameters estimated it is relatively easy to estimate expected losses and formulas are available for estimating the mean and dispersion etc for many distributions in the presence of deductibles or limits truncated data Predictions can be made relatively easily if the distributions are stable Computationally there are fewer steps involved It is also easier to test hypothesis or to investigate the effect of e g changes in deductibles and limits on expected losses The major disadva
45. for that purpose which is the first page for Use Case Two Figure 2 2 15 illustrates the snapshot of the Dataset Selection Page 2 B 21 4 Storm Genesis Time Simulation Microsoft Internet Explorer File Edit view Favorites Tools Help Qs J x a f Search Sg Favorites QU necia J a Address 2 http j irene cs Fiu edu 8888 FDOIJtask jsp Storm Genesis Time Atlantic Basin Only Dataset Selection Page Description This program is used to analyze the genesis time of storms in Atlantic Basin please select a dataset from the following list as your simulation basis DataSet Selection 1851 2003 v 1900 2003 1944 2003 ENSO Multi Decadal 4D Internet Figure 2 2 15 Snapshot of the first web page for SGT A dropdown list that is consisted of some possible year ranges is offered to make the selection simpler to the users and on the other side to avoid the user type in any wrong year range There are five possible choices 1851 2003 1900 2003 1944 2003 ENSO and Multi Decadal Please see to user requirement documentation for detailed explanation of these terms The user then selects a year range from the dropdown list and submits his her selection to the system Step 2 Upon the user s year range selection the system constructs a query and questions the underlying Oracle database to get the data set pertaining to the user s year range selection which contains the hurricanes and
46. get the property value Vijx its policy limits LMijk and its deductible Dijk The limit LM is the default value of the property k default is V LM if value is not available Value is contingent on the type of policy specified and is either replacement cost or actual cash value replacement cost minus depreciation 10 Select a wind speed from the distribution Apply the damage ratio vector Xj to the property k of type i in zip code j For each damage interval n calculate the damage DMijww Vik X Xi Thus a Nx1 damage vector DMi is generated for property k This vector is associated with the chosen wind speed 11 For the above selected wind speed estimate the row vector of wind conditional mean content damages where each element is the mean content damage for the given wind speed meanCijkw LMcx meanCinw ratio 12 For the selected wind speed estimate the row vector of wind conditional mean AP damages where each element is the mean AP damage for the given wind speed meanAPijkw LMar X meanAP in ratio 13 For the selected wind speed estimate the row vector of wind conditional mean ALE damages where each element is the mean ALE damage for the given wind speed meanALEiw LMar x meanALEi ratio 14 Using the wind conditional mean structural damage DMijx and combining it with the wind conditional mean C mean APjjxy and mean ALE xy calculate the deductibles Ds Dc Dap on a pro rata basis to the
47. idl uvstr holds the wind profiles calculated for each fix en VGHGEN PRO This module calculates the gradient wind profile and its second derivative USG PRO This module computes the radial and tangential wind profiles for a stationary storm with surface friction for exactly one fix Wind profiles are calculated from two directions inward and outward from center Then the results are combined to get the complete profile 1 Form the inward boundary value at s 20 using obc m 2 Numerically integrate momentum equations for stationary storm profiles LSODE 3 Match solutions across the shock and obtain uz and sgz For 0 lt s lt 1 uz uout Sgz sgout for 1 lt s lt 20 uz uinw Sgz sginw 4 Sub grid smoothing process simulates turbulent diffusion 5 Sub sample to recover original resolution USNOADV M Ignore radial advection and algebraically solve for u and sigma equation A6 A7 starting the numerical integration outwards from the origin USADV M This module iteratively improves the results of USNOADV M by including radial advection terms first derivatives evaluated from the previous iteration OBC M This module computes the outer boundary values for u and sg to start the inward numerical integration of u and sg using LSODE Refer to IDL manual for LSODE Procedure DUS is used to calculate the derivatives of u and O from A6 A7 A8 3 B 17 1X DUS PRO Calculates radial derivat
48. in the original storm track file TYPE A 92620 08 16 1992 M 13 2 SNBR 899 ANDREW XING 1 SSS 4 1 92620 Card 2 08 16 1992 MM DD Year Days 3 M 13 S 4 2 Total 5 ANDREW Name 6 XING 1 US Hit 7 SSS 4 Hi US category Card Sequential card number starting at 00010 in 1851 MM DD Year Month Day and Year of storm Days Number of days in which positions are available note that this also means number of lines to follow of type B and then one line of type C SH Storm number for that particular year including subtropical storms Totals Storm number since the beginning of the record since 1886 Name Storms only given official names since 1950 5 A 8 US Hit 1 Made landfall over the United States as tropical storm or hurricane 0 did not make U S landfall Hi US category 9 Used before 1899 to indicate U S landfall as a hurricane of unspecified Saffir Simpson category 0 Used to indicate U S landfall as tropical storm but this has not been utilized in recent years 1 to 5 Highest category on the Saffir Simpson scale that the storm made landfall along the U S 1 is a minimal hurricane 5 is a catastrophic hurricane TYPE B 92580 04 2282450610 30 100352490615 45 100252520620 45 100252550624 45 1003 1 92580 2 04 22 3 8 Card MM DD Storm category 4 2450610 30 1003 5 2490615 45 1002 6 2520620 45 1002 7 2550624 45 1003 LatL
49. point uxstre holds the zonal component of the wind at each grid point zxystre Complex array containing the zonal and meridional wind components at each grid point xiii GENSTREX M This module places the profile functions and the auxiliary functions g and d which we calculated earlier on the polar grid yielding fields with no azimuthal dependence for exactly one fix sstre array containing radial distance to each of the grid point from the centre xiv SHIFT M This module simply shifts the polar coordinate system so that the origin of the coordinate system lies on the center of the storm Center of the storm is the point where wind speed is zero 3 B 19 xv PKWINDS PRO This module produces an output file which lists the peak marine and open terrain wind components experienced at each zip code for the current storm If the storm makes landfall then the peak marine and open terrain winds are listed at the tie and the site of the landfall If the storm only bypasses the state then the peak marine and open terrain winds are listed for the fix exhibiting the lowest central pressure 1 2 3 restore zipcodes idl This contains the longitude and latitude of all zip codes restore fixshots idl We generated this in previous step Initialization of other variables elonk east longitude of the track every minute nlatk north latitude of the track every minute kmax maximum time life time of the storm in Min
50. pressure B is the Holland pressure profile shape parameter R is the radius of maximum wind speed in nautical miles and Ap is the pressure deficit defined earlier The central pressure is modeled according to the intensity modeling in concert with the storm track The resulting expression for B is B 1 38 0 00184Ap 0 00309R max 5 Where is a random term from a zero mean normal distribution with a standard deviation of 0 05 3 2 3 1 4 Land See Flag Isfg Gives the position of the wind at the storm fix 0 Over Ocean 1 Land Fall 2 Sea Fall 3 Over Land 4 Closest Approach of bypassing Storm 3 B 8 3 2 3 2 Definitions and Equations of the wind model R Radius of maximum surface wind speed specified ct storm translation speed specified cdir storm translation direction compass heading specified AP Central pressure deficit specified R max m E p r pot Ape sea level pressure B 1 38 0 00184Ap 0 00309R _ Holland profile parameter Azimuthal coordinate measured counterclockwise from east s R normalized radial coordinate 2 v s Gradient wind Ys Rfv 1 op s p ds f 2Qsin Coriolis parameter 9 latitude of storm center v Gs vo 5 normalized gradient wind symmetric ao where V max is the maximum g max gradient wind in the radial profile Rf f Normalized Coriolis parameter g max
51. program there is also a training feedback given to each student upon the completion of the training session 9 A 3 Section 10 PHRLM Related Publications 10 PHRLM Related Publications 1 Shu Ching Chen Sneh Gulati Shahid Hamid Xin Huang Lin Luo Nirva Morisseau Leroy Mark D Powell Chengjun Zhan Chengcui Zhang A Web based Distributed System for Hurricane Occurrence Projection Software Practice and Experience Volume 34 Issue 6 pp 549 571 May 2004 2 Shu Ching Chen Shahid Hamid Sneh Gulati Na Zhao Min Chen Chengcui Zhang and Paresh Gupta A Reliable Web based System for Hurricane Analysis and Simulation IEEE International Conference on Systems Man and Cybernetics 2004 pp 5215 5220 October 10 13 2004 Hague The Netherlands 3 Mei Ling Shyu Shu Ching Chen Min Chen Chengcui Zhang and Chi Min Shu MMM A Stochastic Mechanism for Image Database Queries Proceedings of the IEEE Fifth International Symposium on Multimedia Software Engineering MSE2003 pp 188 195 December 10 12 2003 Taichung Taiwan ROC 4 Shu Ching Chen Sneh Gulati Shahid Hamid Xin Huang Lin Luo Nirva Morisseau Leroy Mark Powell Chengjun Zhan and Chengcui Zhang A Three Tier System Architecture Design and Development for Hurricane Occurrence Simulation Proceedings of the IEEE International Conference on Information Technology Research and Education ITRE 2003 pp
52. storm center and the zip code centroid Then using interpolation calculates the marine wind speed at the zip code centroid Use ZMAR2ZOT to convert above calculated marine windspeeds into Open Terrain windspeeds After the construction of the time series record maximum total OT windspeed at each zip code Obtain marine and OT peak winds at landfall or lowest pressure for bypassing storms At the same time record the time and location of landfall or lowest pressure fix Write the output file if at least one zip code is affected by the storm 3 B 20 xvi REACH M This function determines the influence radius Influence radius 12 3246 0 162 rmw If the calculated value is less than 4 then set it to 4 xvii LLTOXY PRO This module converts east longitude and north latitude into zonal distance xmerc and meridional distance ymerc in meters from the cyclone center elo gglo ymerc mercator y coordinates from latitudes xmerc mercator x coordinates from longitudes xviii ZMAR2ZOT PRO This module converts marine wind speeds m s into Open Terrain windspeeds m s xix GEMF M This module is used to set the time step and call the executable 1 set the time step for storm series calculations and load gemplex exe kinc 15 time step is set to 15 minutes flcnt 0 start the output file numbering from 1 outputl dat output2 dat References 1 Vickery P J and L A Twisdale 1995 Wind field and fil
53. that h is a function of the point and the sample size N 2 B 5 D 10 11 5 7 This estimator denoted by is defined as By t K t n OF wx S Fy T Fy T j 23 W and S F T Also amp u is the integral of K x that is k u KkG ax 5 8 In the above function one has a wide variety of choices available for the kernel function and the corresponding bandwidth We will try the following kernel function and bandwidth a The kernel function is the Epanechnikov kernel K that K x 3 45 1 x 5 V5 lt u lt V5 1 3 b The LOCAL bandwidth Ay t 2 where S is the standard deviation of the calculated SGTs of all the hurricanes The system presents a list of simulated events sets The user selects a set of simulated events and submits it to system The system checks the selected simulated events If they already exist in database the system query the database to get the data of selected simulated years or the system triggers Use Case One AHO to generate a set of simulated events The system uses IMSL routines to generate the SGT for each hurricane of the selected simulated events The selected simulated events give the number of hurricanes in a given year Assume there are M hurricanes in yeari The system will sample M hurricanes from CDF t to get the genesis time for those M hurricanes in yeari The system stores the generated S
54. the passed array e queryZoa Uses the connection to the database to send a query and retrieve the value for Zoa based on the zip and the returned string value from a call to the findCol method e calVamph Method is used to calculate the Vamph Va Uo and Ua using the following input values from the user Vo Zoo and Zoa e calcGust Method is used to calculate the gust factors G1n 60 and Gyn 3 using the following input values Lat WD Zoa and Va e GetRoughness def Method uses the established connection to the database to find and return the column names and starting and ending degrees corresponding to each column of the roughness def table e findCol int wd Method takes a wind direction WD as a parameter It uses the established connection to the database to find and return the correct string using the WD which represents the column in the lookup table for Zoa e get Zoa Return the value of Zoa e getVamph Return the value of Vamph e getVimph Return the value of Vimph e getV3mph Return the value of V3mph e getUa Return the value of Ua e getUo Return the value of Uo e getVa Return the value of Va e disconnect Method is used to disconnect from the database 3 C 13 3 3 4 4 State Chart Diagram Figure 3 3 8 depicts the state chart diagram for Use Case Five This diagram illustrates states that the use case goes through from beginning to end Page requested m
55. utility No root console logins on goedel are authorized except for scheduled installations and emergency work which is disclosed to the A D 5 All security incidents will be log in the FIU SCS Computer and Security Activities Log Depending on severity security incidents will be reported to the SCS Director and or other FIU management 6 Computer or networking devices whose security has been compromised may be disconnected from the FIU SCS network until the system security is restored 7 Computer accounts whose security has been compromised may be disabled until the appropriate credentials are properly reassigned 7 A 5 8 All computer system operating systems will be maintained with critical security patches as indicated by OS provider and or security community 9 Mission critical servers will be maintained in a physical location which limits non authorized access Access to the server room is granted by electronic key card and is limited to essential personnel only 10 If a mission critical server goes down the server room security system will immediately page the system administrators to report the incident 11 The School maintains anti virus software on all networked computers and regularly updates the anti virus software 12 The School s security policies shall be consistent with those security policies which govern the State of Florida and Florida International University Academic Affairs
56. with origin at the population weighted centroid of the zip code and extending outward to 20 km from the centroid The weighting function for averaging the roughness values is a Gaussian filter with a half power point at 3 km The format of the lookup table is as following Zip Lon Lat Zol Zo2 Zo3 Zo4 Zo5 Zo6 Zo7 Zo Where Zol Actual Roughness for wind directions inclusive of 46 90 Zo2 Actual Roughness for wind directions inclusive of 1 45 Zo3 Actual Roughness for wind directions inclusive of 316 0 360 Z04 Actual Roughness for wind directions inclusive of 271 315 Z05 Actual Roughness for wind directions inclusive of 226 270 Zo6 Actual Roughness for wind directions inclusive of 181 225 Zo7 Actual Roughness for wind directions inclusive of 136 180 Zo8 Actual Roughness for wind directions inclusive of 91 135 3 C 3 Table 3 3 1 shows a sample record from the lookup table Zip 33172 Lon 80 24401855 Lat 25 73268509 Zol 0 2399817854 Zo2 0 3124250770 Zo3 0 3429141343 Zo4 0 3098731637 Zo5 0 3196663558 Zo6 0 2674820721 Zo7 0 5406716093E 01 Zo8 0 7273393869E 01 Table 3 3 1 A sample record for the lookup table 3 Given the wind direction for each zip code centroid the appropriate value for actual terrain roughness is extracted from the lookup table The system then computes the output values as b
57. zip code use x axis with wind speeds at 5 mph resolution starting at 22 5 mph and ending at 302 5 mph A sample calculation result is as follows PVxz annual probability of a wind Gust speed within the band indicated at left greater than the first number and less than or equal to the second number PV gt y annual probability of the wind Gust speed exceeding the mid point of the band ZipCode 41 Gust Band mph Mid Point PVxz Pv gt lt 22295 24 5 25 0 0 017 0 262 2YqT b 325 30 0 0 022 0 247 32 531 25 3520 0 022 0 229 3725 425 40 0 0 023 0 211 42 5 47 5 45 0 0 022 0 192 4145 5215 50 0 0 019 0 176 52245 57425 55x0 0 017 0 160 Seb 62525 60 0 0 017 0 146 62 5 67 5 65 0 0 017 0 131 3 D 4 67 72 TT 82 87 92 97 102 107 12 117 122 127 1324 I3 7 142 147 152 LST 162 167 1712 TU 182 187 192 197 202 207 212 217 222 227 232 237x 242 247 252 257 262 267 272 277 282 287 292 Z9 Ts 72 Ths 82 87 92 97 1 02 55 aaa a a O1 TIZ 113 122 127 132 137 142 147 152 l5 1 62 167 1525 177 182 187 192 197 202 207 212 217 222 227 232 237 242 247 252 297 262 267 272 277 282 287 292 297 302 O10101010101010 010101010101010101010101010101010101 aa aaa aaa a 01 O1 O1 O1 a wo 70 75 80 85 90 95 100 105 1
58. 000007 11 20 3 05 00 For 33133 and for this simulation of years with events IN event 2 of years with no events No event 99998 Here is the standard error of the command if any Site directory AD Internet 8 Done l Figure 3 4 11 WSP results page 3 D 14 3 4 4 5 Results of WSP Results for zip code 41 ZipCode 41 Gust Band mph Se 22 27 32 3T 42 4T 525 Sis 62 67 TDs 77 82 87 92 974 102 117 172 212 2g 222 221 2924 237 242 247 252 257 262 sor d eT VOL 107 112 T22 127 1324 LITA 142 147 152 157 162 167 LTT 182 187 192 T1975 202 207 lt De v2 5 324 beg Se 2 59592 ow e 529924 5e DTZ 9c ds De OB 95792 3 97 527 102 5 woo LO 52 1124 j llY som 12253 5 X274 DS 2 4 5 L3 5 142 ao 147 5 1524 Solo 5 162 5 167 DANZA sone RET s 5 1825 5 187 5 192 5 WO 5 202 5 203 52 2124 RL D Z222 5 224 sor L232 5 237 2 242 5 241 9 4923 52 294 522062 5 267 O10101010101010101010101 a ow al O101010101010101010 010101010101010101010101010101010101010101 01 a al Mid Point 25 0 30 35 40 45 50 D 60 65 20 Tos 80 85 90 95 100 105 110 115 120 12 55 130 135 140 145 150 1555 160 165 170 MT Des 180 185 1905 195 200 205 210 215 220 225 230
59. 03 55 113 2 32009 14 6592 163 0 03 0 818 1 009 1 274 7 978 24 786 36 494 3 32011 14 2175 162 0 03 0 699 0 978 1 222 8 132 24 949 36 345 2 For each of the zip code a From each storm wind speed time series we select the maximum 3S gust wind speed V3 b Wind speeds V3 affecting each of the zip code for N number of storms in the simulation affecting each zip code are then sorted in an ascending order c The sorted vector of wind speeds is then distributed into bands of 5 mph starting at 22 5 mph and ending at 302 5 mph Wind speeds over the midpoint of each band are assigned to the corresponding bin 3 D 3 Count the number of instances of maximum wind speeds within each of the wind speed bands and below the midpoint of each band Compute the annual probability of the maximum wind speed being within each band defined by wind speed x on low side and z on high side as PVxz Number of Instances of V3 gt x and lt z N Compute the probability of the wind speed exceeding the mid point v of the band as _ Number of Instances of V3 lt v N N number of storms affecting that zip code Estimate A total of storms affecting that zip code simulation Years Pv v Estimate the probabilities P V v 21 e d P Also P V lt v 1 e e P oo n i xn P z lt V lt x Eramos j we n j l n 3 Plot histogram and cumulative frequency diagram of maximum wind speeds for each
60. 10 LES 120 125 130 1395 140 145 150 155 160 165 170 175 180 185 190 195 200 205 210 215 220 225 230 2399543 240 245 250 255 260 265 270 275 280 285 290 295 300 C C9 OO CO QOO somo OC OC 010000 0 OO CO O C OC OQ 070 O C 00000 OG GG Q OO cC O O C O S 2 0 0 OS 10 0 00 00 00 00 00 00 OO CO CO CO C9 0 0 OrO 0 00 OOO C369 CO CY Oe C9 00 0 CO O 0 010 0 0 70 00 000 0 0 OO 0 0 0 O 00 00 CO CO 700 00 100 OOO CO CO CO CO LC CO C oO0oO0O0O00O0O0O0O0O0O0O0O0O000000000O0000000o00000o00o0o00o00o000000o00000 0 00 0 0 O O OUO 0 00 C 00 0 C0 CO 0 10 0 00 0 0 O 0702 12 NGC ds E OIL Oy 00 OPRRRRPRR c O PB Ba JJ C3 OO C CO C OO OO OOO CO CO CO CO C23 0 00 CO E CX CO CO C3 CO 0 CO CO COS COS CO COD CO C3 CO CO O OOS OOO 010 O 090 0 0 0 00 7000 0000 Or O Or Oro CFO C9 CO CS CX 0 00 0 00 0 00 CX CO OC S Ce O1 Oo J CO A A 1 00 A O1 N N W Ww our Oper Own 0 000 0000000000 OOO0 00000 0 cO0O0 00 0 O0oO0 0o 00 CO 00 00 CO CO CO CO CO CO CO CO CO CO CO CS 0050 C FD OG OT OF 116 102 3 D 5 3 4 3 Computer Model Design 3 4 3 1 Use Case View of WSP A Actors There is one actor scientist in WSP B Use Case WSP is used to calculate the annual probabilities of the wind speeds lying in a wind speed interval
61. 235 240 245 250 255 260 265 C3 C CO C CX CO CS O SO CO CO O QOO OO 00 CO QC C 010 0 C O OO O0 OQ OSC OC 00 OO C C O 00 PVxz CO C2 C2 2 CO COO C 000 Co COG GOGO GOG CC c OO C3 CO CO CO C2 C2 Or 0 0 CO CO 00 0 O c COO OO cO cC cO OO 0 0125 O OV OOO CO CO C CO 00 000 00 00 01 070 0 0 00 900 0 00 CO C C CO CO CO CO CO ae 0 CO 0 0 0 00 Ooo CO 0 0 C0 O 0 O Q C C C CO CO 0 OO COO 0 0 0 CO CO CO OQ 0 CO CO 0 0 0 CO 17 22 22 23 22 No oooocnooocncoooocuoceoeoooceoeooorrRR NMP NN WHKRUDOAWOOF FA A ar a YY I tg lt Ovi GOO C C O Or OOS 000 CO C OO OO 0 00 OOO 0 0 CO CO CO CO CO CO CO C OOO OOO OC 0 0 OO OOO OO O GOOO Or C CO 070 0 00 10 CO 0070 00 OO 0 00 00 CO 0 00 CY OC 1 A 01 J 0 O00 IN oor O1 CON o Ol OY A A ad O1 N N W CO OU Pm J ore vouw 00 oO CO CO O0 0 OO O O O O C O O C C OO OO CX Oo CC QC Oc O0 0 OO 0 COO OO OC SO OQ HR INN wina 262 247 229 211 Ww E 3 D 15 267 5 272 5 270 0 0 000 0 000 272 5 277 5 275 0 0 000 0 000 277 5 282 5 280 0 0 000 0 000 282 5 287 5 285 0 0 000 0 000 287 5 292 5 290 0 0 000 0 000 292 5 297 5 295 0 0 000 0 000 297 5 302 5 300 0 0 000 0 000 Max Wind Speed Probabilities for 33156 0 07 Wind Speed Probability Pvxz 0 06 0 05 0 04 0 03 0 02 0 01
62. 86762 oz Where 60 represents 1 min or 60 seconds and the integral scale time parameter J is hh 3 132 IL 4 96 In which Z 10 meters is used I 0 68 0 z 3 042 1 0 199 0 1 02 3 0 8682564210 z Where 3 represents 3 seconds 3 C 5 4 5 Compute the wind fluctuation cycling rates 0 654 o007 02142 60 I C 60 C 60 0 00982 I 0 654 omn coa I CA 3 C 3 0 061 4 6 Compute the Peak factors for the max 1 min 60 sec and max 3 second winds PIGS OO e Oe 2Ln 600C a z ON 216000 9097 jz 2Ln 600C z 4 7 Compute the longitudinal turbulent intensity _ 942 Un Til 4 8 Compute the gust factors Gio min oo 1 TuP 60 Gio min3 1 TiPs 3 3 C 6 A sample calculation is as follows Input Lookup value Output Zip 33133 Zoa 0 219 m U o 3 44 m s U a Vo 50 m s 3 96 m s Va 37 845 WD 60 m s V1 101 504 Zoo 0 03 mph V3 134 605 mph Lat 25 73 3 C 7 3 3 3 WSC Interface Design Requirements This section presents the Graphic User Interface design for the Wind Speed Correction WSC 1 The first step the user logs in the system Figure 3 3 1 shows the Login Interface User needs to enter the user id and password to enter the system The system verifies the user s information with the login data extracted from the database
63. 992 1992 1992 1992 1992 1992 1992 BA HAHAHAHAHAHA HASHA 3 1 4 0824 0824 0824 0824 0824 0824 0824 0824 0824 0824 0824 0824 05 00 06 00 07 00 08 00 09 00 09 05 10 00 11 00 12 00 13 00 14 00 15 00 25 4 25 4 25 4 25 4 25 5 25 5 25 5 25 6 25 6 25 6 25 6 25 7 References 79 3 79 3 79 6 80 0 80 4 80 3 80 8 81 2 81 2 81 5 81 9 82 3 937 937 939 942 945 922 948 951 951 950 949 948 19 19 18 18 18 19 18 17 17 18 19 20 1 4772400 1 4772400 1 4727061 1 4681721 1 4636379 1 5048399 1 4591039 1 4545699 1 4545699 1 4541880 1 4538059 1 4534241 SOON WWrROOCC SO Reynolds R W N A Rayner T M Smith D C Stokes and W Wang 2002 An Improved In Situ and Satellite SST Analysis for Climate J Climate 15 1609 1625 3 A 9 3 A 10 Section 3 2 Wind Field Model Use Case IV 3 B 1 32 1 General Description Of Wind Field Model Wind Field model aims at estimating the terrain wind speed with respect to the actual terrain based on land use land cover To be precise it calculates wind speed time series for each of the zip code affected by the storm The time series includes the date landfall time of the storm It also includes the zonal wind speed m s surface wind speed m s and the wind direction in degrees at regular time intervals 3 B 2 3 22 Genera
64. CtUre SS ic 8 A 1 8 2 Software List ecrire a a ad SEIU aden 8 A 2 8 3 Hardware Conti Ural OM dE 8 A 3 o A Con suc S UE Pea avg E tan oes EPOR UE Pam 8 A 5 y CPN PRAM ARE AAA A bua was 9 A 1 9 1 Intro UCA E 9 A 2 902 Techical Tranne Pl O E 9 A 2 93 dnd User Tranine At SN CE cam ol 9 A 3 10 PHRLM Related Publications eee 10 A 1 Section 1 The Public Hurricane Risk and Loss Model PHRLM 1 1 General Description of PHRLM Model The PHRLM model is a probabilistic model designed to estimate the damage and insured losses due to the occurrence of hurricanes in Atlantic Basin The PHRLM estimates the full probabilistic distribution of damage and loss for any significant storm event The modeling methodology of it can be partitioned into four major components Storm Forecast Module Wind field Module Damage Estimation Module Loss Estimation Module The high level flow chart is shown in Figure 1 1 Historical Storm Database L HURDAT PONES Storm Forecast Module DEEP Determines the storm Stochastic Storm properties to be used in Simulated Storms Database the analysis Storm Properties Wind Field Module Generates the wind field based on geo coded location Information obtained from geo database Ground Elevation Exposure Classification Wind Speed A S Damage Estimation Vulnerability Module Statistics Calculates Damage
65. DLLLLLLLLLLLLLLLLLLLEL Rows Statistics about the processed rows appearing in the ROWS column ROWS Total number of rows processed by the SQL statement This total does not include the number of rows processed by subqueries of the SQL statement For SELECT statements the number of rows returned appears for the fetch step For UPDATE DELETE and INSERT statements the number of rows processed appears for the execute step 5 A 24 Resolution of Statistics Timing statistics have a resolution of one hundredth of a second therefore any operation on a cursor that takes a hundredth of a second or less may not be timed accurately Keep this in mind when interpreting statistics In particular one should be careful when interpreting the results from simple queries that execute very quickly Recursive Calls Sometimes in order to execute a SQL statement issued by a user Oracle must issue additional statements Such statements are called recursive calls or recursive SQL statements For example if you insert a row into a table that does not have enough space to hold that row Oracle makes recursive calls to allocate the space dynamically Recursive calls are also generated when data dictionary information is not available in the data dictionary cache and must be retrieved from disk If recursive calls occur while the SQL trace facility is enabled TKPROF produces statistics for the recursive SQL statements and marks them
66. Display Interface in case of failure 2 B 10 2 2 3 Computer Model Design 2 2 3 1 Use Case View of SGT A Actors There is one actor scientists in SGT They will use this use case to estimate the probability distribution model and to generate the storm genesis time of the simulated hurricanes generated in Use Case One AHO B Use Case SGT is used to estimate the probability distribution model for HBG Hours between Genesis and generate the genesis time of a series of simulated hurricanes generated in Use Case One C Use Case Diagram N Fi gt A gt bh p Scientist StormGenesisTime Figure 2 2 5 Use Case Diagram for SGT 2 2 3 2 System Design This part includes the appropriate diagrams to describe the system classes components activities and the overall flow chart of SGT 2 B 11 2 2 3 3 Program Flow Chart of SGT Here we give out the flow chart of SGT we could see clearly from the chart the relations of different parts User selects the dataset IMSL library Begin System gives out selection form of dataset System gets data from database Y System estimates the CDF of HBG System generates the SGT Save the SGT into database Display result graph to user Figure 2 2 6 Flow chart of SGT Oracle DB 2 B 12
67. F lists the statistics for a SQL statement returned by the SQL trace facility in rows and columns Each row corresponds to one of the three steps of SQL statement processing The step for which each row contains statistics is identified by the value of the CALL column 5 A 22 PARSE This step translates the SQL statement into an execution plan This step includes checks for proper security authorization and checks for the existence of tables columns and other referenced objects EXECUTE This step is the actual execution of the statement by Oracle For INSERT UPDATE and DELETE statements this step modifies the data For SELECT statements the step identifies the selected rows FETCH This step retrieves rows returned by a query Fetches are only performed for SELECT statements The other columns of the SQL trace facility output are combined statistics for all parses all executes and all fetches of a statement These values are zero 0 if TIMED_STATISTICS is not turned on The sum of query and current is the total number of buffers accessed COUNT Number of times a statement was parsed executed or fetched CPU Total CPU time in seconds for all parses executes or fetch calls for the statement ELAPSED Total elapsed time in seconds for all parses executes or fetch calls for the statement DISK Total number of data blocks physically read from the data files on disk for all pars
68. GT into database The system displays the generated SGT and the corresponding simulated events on screen 2 B 6 Note Steps 2 11 are repeated for each year range that the user requests Steps 7 11 are repeated each set of simulated events the user selects at step 6 2 B 7 2 2 2 1 SGT Interface Design Requirements This part designs the GUI Graphic User Interface for the Storm Genesis Time SGT It describes the process by which scientists or statisticians log in to the PHLRM system to view the genesis time of events generated in Use Case One AHO if they exist or to trigger the events if they do not exist A The First step the user logs in the system Figure 2 2 1 is the Login Interface The user enters the User ID enters a password and submits to the system The system checks the user name and password and let the user log into the system if the User ID and Password are correct Or the system gives the invalid user password error to the user and asks the user to reenter the User ID and password User ID FDOIUSER Password Login Figure 2 2 1 Login Interface B The second Step the user selects a year range Figure 2 2 2 is the Year Selection Interface The system presents a list of year ranges to the user The correct year ranges are shown at Table 2 2 5 The user selects a year range and submits to the system The system also shoul
69. Holland G J 1980 An analytic model of the wind and pressure profiles in hurricanes Mon Wea Rev 108 1212 1218 11 Dunion J P C W Landsea and S H Houston 2003 A re analysis of the surface winds for Hurricane Donna of 1960 Mon Wea Rev 131 1992 2011 12 Willoughby H E and E Rahn 2002 A new parametric model of hurricane wind profiles 25th AMS Conference on Hurricanes and Tropical Meteorology San Diego 29 April 3 May 2002 13 Powell M D P J Vickery and T Reinhold 2003 Reduced drag coefficient for high wind speeds in tropical cyclones Nature 422 279 283 14 Large W G and S Pond 1981 Open ocean momentum flux measurements in moderate to strong winds J Phys Oceanography 11 324 336 15 Moss M S and S L Rosenthal 1975 On the estimation of planetary boundary layer variables in mature hurricanes Mon Wea Rev 106 841 849 16 Powell M D 1980 Evaluations of diagnostic marine boundary layer models applied to hurricanes Mon Wea Rev 108 757 766 17 ASTM 1996 Standard practice for characterizing surface wind using a wind vane and rotating anemometer D 5741 96 Annual Book of ASTM Standards Vol 11 03 18 Anctil F and M Donelan 1996 Air water momentum flux observations over shoaling waves J Phys Oceanogr 26 1344 1353 19 Reinhold T and K Gurley 2003 Florida Coastal Monitoring Program http www ce ufl edu fcmp 3 B 22
70. Ide ho d mouseClicked User data sent Processing 1 Results Displayed Figure 3 3 8 State Chart for WSC 3 C 14 3 3 4 5 Sequence Diagram A Sequence Diagram for the Vamph Calculation Process Scientist 7 d s d N WindSpeedCal jsp WSCCalVamph Page Requested Bean java setParameter string lt Database setZip int setVo double setWD double setZoo double connect getConnection string queryZoa executeQuery String Result Set findCol int lt executeQuery string Result Set calVamph getZoa double getUo double getUa double getVa double getVamph double disconnect close Figure 3 3 9 Sequence diagram for Vamph calculation process 3 C 15 Sequence Diagram Steps for the Vamph Calculation Process Step 1 The user requests the html page Step 2 The user enters the number of data sets to be calculated Step 3 The user s input data for zip code Vo WD and Zoo are passed to WSCCalVamphBean object Step 4 WindSpeedCalc jsp requests WSCCalVamphBean object to establish a connection to the database Step 5 WSCCalVamphBean establi
71. If there is a match the user logs into the system successfully Otherwise system displays the wrong user name password error and requests the user to login again UserID FDOIUSER PassWD 3K k k kK k k k k LOGIN Figure 3 3 1 Login Interface 2 The second step select the use case from the service selection page Figure 3 3 2 is the service selection page interface System presents a list of available use cases to the user User selects Roughness Model use case and clicks Go to submit Please choose an online service _ Go Figure 3 3 2 Service Selection Interface 3 The third step The user provides the input from the wind model In this step system provides the interface for the user to input data generated by the wind model The following inputs are required and are illustrated by Figure 3 3 3 e Zip code 3 C 8 e Wind Speed Surface wind speed m s for open terrain produced by the wind model e Wind Direction Surface wind direction Degree s from the North e Roughness Length Roughness length m for open terrain 0 03 m Wind Speed Wind Roughness length m s Direction m 2 _ Jr Jh__ Figure 3 3 3 Input from Wind Model Interface Num Zip Code 4 The forth step The system displays the result in the interface In this step system calculates the result and displays both the input and the output to the user as shown in Figure 3 3 4 The output inc
72. It also provides buttons to allow the end user to change the graph type move back forward in the figure NumericSet This class is used to store the simulation result data it is used by the class myPlot MyButton This class is for new button customization which is used by the class myPlot to let the end user to change the graph type move back forward in the result figure PlotApplet This class provides some basic functions in plotting and is the base class for class myPlot Note This class is implemented by MIT CS department Applet Button A base class provided by Java API 2 A 16 2 1 4 4 Sequence Diagram Sequence diagrams are helpful in understanding the relations among the classes This section shows sequence diagrams that describe four major activities in use case AHO which are login fit distribution simulation and plot respectively A Login Process client Client checkerBean database loginCheckBean Database login EtegisterDriver gt lt getConnection gt lt executeQuery gt lt getStatus lt gt Figure 2 1 9 Sequence diagram for login process e Step 1 The user enters the user ID and password in the web browser and click login button e Step 2 The login information is passed to the loginCheckBean class which communicates with the
73. Plot JSP Ele Edit View Favorites Tools Help he 3 8 JO search Shp Favorites Qe O 2 6 Ba Woes 3 firene cs fiu edu 8669 FDOLJAHO plotSimulation jsp Eo Simulation Result Plot Page You want to conduct your simulation in the following way Distribution model Poisson The number of years to simulate 1000 Dataset based on 1900 2003 k value 19 9 The simulation has completely successfully The simulation result is plotted as follows Annual Hurricane Occurrence Result Ai perte 10 AIDA een e d ot Tm mA 7 6 H u 21 1 r i c a4 n e 3 N u m b2 e r 1 0 5 10 15 20 2 30 35 40 45 50 55 60 65 70 75 80 85 90 95 Simulation Number Goto SGT Y Figure 2 1 19 Example of line plot of the simulation result 2 A 26 Section 2 2 Storm Genesis Time SGT Use Case II 2 2 1 General Description Of SGT SGT short for Storm Genesis Time is the second use case of the FIU IHRC Public Hurricane Risk and Loss model It aims at estimating the probability distribution for storm genesis time and generating the genesis time of a series of simulated hurricanes generated in Use Case One In this use case only the historical hurricanes falling in threat area are considered For the detailed documentation on threat area please check use case 1 AHO documentation 2 B 2 2 2 2 SGT General Requirements Name Storm Gene
74. Statistical models model implementation query design query optimization web user interface design web graphical demo applet design e Engineering Module Wind field Module and Insurance Module model implementation performance optimization data characteristics system integration e The policies for using CVS and documenting program data files e The security policies Each student will receive a copy of all the seminar presentations and a copy of the most recent technical report maintained by the designated administrator coordinator The 9 A 2 training will be conducted in ECS building at Florida International University A training feedback form will be given to each student upon the completion of the training session The form will be designed to gather feedback on the seminar content and the instructors The feedback will be used to update and or improve future training sessions 9 3 End User Training Plan The end user training plan is intended for PHRLM end users and managers The intended audience will consist of the professional group members engineering group meteorology group statistics group and finance group They may also expand to Federal customers as needed The objective of this training plan is to train the end user with PHRLM so they can use it for their tasks immediately The training will entail one to one instruction between attendees and technical instructors using computer based tr
75. TData mathmodel IMSL IMSL SGTBean MathModel Library genSGTValues IMSL_Table_Setu _ Table Setup gt IMSL_Seed lt IMSL_Random_General E gt Figure 2 2 11 Sequence diagram for simulation process e Stepl The genSGTData object calls the mathmodel object to generate the SGTs e Step2 The mathmodel object will do all the pre process work and then pass the related parameters to the IMSL library functions e Step3 The IMSL library functions will create the related data and then pass the data back to the mathmodel object Then an array will be returned to the genSGTData object that contains the generated SGTs 2 B 18 2 2 4 Implementation of SGT Currently the implementation for Use Case two SGT has been finished The demo is online at http www cs fiu edu PHRLM 2 2 4 1 Login page The users need a username and a password to access the FIU IHRC Public Hurricane Risk and Loss Model Following is the snapshot of the web page for login purpose PAE a ELE FIUAHRC Public Hurricane Risk and Loss Model User Login Page UserID PassWD LOGIN Figure 2 2 12 Snapshot of Login page for SGT If the username password is wrong error message will show and the user is required to input the username and password again 2 B 19
76. Type Determine from portfolio file info eliminate matrices which do not apply if unknown use other matrix Select Applicable Weighted Vulnerability Matrices 4 A 24 Region Type South amp Keys Select Applicable Weighted Vulnerability Matrices 4 A 25 Select a company C Select a portfolio P Get Observe wind speed Property value Vi Wind deductible get information of Construction type Zip code County Region D Loop 1 limits for structure LMs content LMc App LMap ALE LMarg If Hurri_Deductible 0 Use Other Deductible Else Hurri_Deductible Hurri_Deductible ISO classification available Use weighted matrices concrete matrices Vulnerability Matrices for Structure S Content C Appurtenant AP and ALE for a given construction type region based on a given mix of construction features Assume the number of damage ratio intervals is N Get damage ratio vectors i e the middle point values for N intervals Xs Xc Xap Xap T Based on Wo get the vectors of the probability of damage Poy Poe Pop Pp i5 whose corresponding wind speed interval Wi 2 5 mph includes Wo in case of a tie break the tie by picking the larger one DMs L Pp Xs Dmc EP X DM ap Y Pp p Xap SumDM DMs DMc DMap Ds DMg D SumDM D DMc D SumDM Dap DM ap D SumDM Use wei
77. _deg longitude_deg max_windspeed_mps min_pressure_mb height_m stage rmax crossing TYPE NEWFIX latitude_deg NUMBER 10 4 longitude_deg NUMBER 10 4 max_windspeed_mps NUMBER 10 4 min_pressure_mb NUMBER 6 height_m NUMBER 8 3 stage NUMBER 2 rmax NUMBER 4 crossing VARCHAR2 10 when_t fix_id fixobj at_time for_event event_id produced_id produced_by TYPE STORMFIX fix_id NUMBER when_t DATE at_time CHAR 6 event_id NUMBER 6 for_event REF ATMOSEVENT produced_id NUMBER 4 produced_by REF PLATFORM_TYPE fixobj NEWFIX 5 A 3 key_id stm_nbr when_t name type basin ATCF_name TYPE ATMOSEVENT key_id stm_nbr when_t name type basin ATCF_name key_id type description TYPE PLATFORM_TYPE key_id type description category_no state_code TYPE LANDFALL_TYPE category_no category_no state_code state_code TYPE LANDFALL_TYPE_ARR NUMBER 6 NUMBER 6 DATE VARCHAR2 30 NUMBER 2 NUMBER 2 VARCHAR2 20 NUMBER 4 VARCHAR2 50 VARCHAR2 50 NUMBER 2 RCHAR2 4 5 A 4 category_no state_code NUMBER 2 RCHAR2 4 This object is an array of the LANDFALL_TYPE object There are 5 tables in HURDAT Schema Table 1 type description TABLE 1 PLATFORM_TYPE_LIST Key_id PRIMARY KEY CONSTRAINT This table is based on Object Platform_type Table 2 PK when_t PK at_time PK FK2 event_id fix
78. _id for_event produced_id TABLE 2 STORMFIX_LIST when_t constraint constraint constraint constraint NOT NULL FIX_ID_UN UNIQUE FIX_ID EVENT_ID_FK FOREIGN KEY EVENT_ID PRODUCED_ID_FK FOREIGN KEY PRODUCED_ID FIX_ID_PK PRIMARY KEY EVENT_ID WHEN_T AT_TIME 5 A 5 Table 3 stm_nbr when_t name basin Un ATCF name TABLE 3 ATMOSEVENT LIST constraint AL KEY ID PK PRIMARY KEY KEY ID constraint AL stm nbr UN UNIQUE STM NBR constraint AL ATCF name UN UNIQUE ATCF NAME Table 4 TABLE 4 STORM CATEGORY category no NUMBER 2 constraint S CATEGORY NO UN UNIQUE description VARCHAR2 30 constraint S_DESCRIPTION_PK PRIMARY KEY Table 5 FK storm id landfall obj TABLE 5 LANDFALL storm id NUMBER 6 landfall_obj LANDFALL_TYPE_ARR 5 A 6 constraint LD_EVENT_ID_FK FOREIGN KEY STORM_ID REFERENCES ATMOSEVENT_LIST KEY_ID ON DELETE CASCADE NESTED TABLE landfall_obj STORE AS landfall_obj_list Table 6 fee TABLE 6 LANDFALL_STATE state_code VARCHAR2 4 constraint STATE_CODE_UN UNIQUE name VARCHAR2 30 constraint LD_NAME_PK PRIMARY KEY 5 4 Data Processing 5 4 1 Original Data Processing Following is the data format of the original text file recording the storm tracks of Atlantic basin In order to populate the data into the database schema we have to process the data and convert them into some suitable format according
79. ach record in rmax dat except the records marked as new Update table stormfix list VII Append the records marked as new into table stormfix list VIII Update table atmosevent list based on the updated table stormfix list 5 A 13 5 6 Export and Import the Data The next step is to migrate the whole database from fdoi georges cs fiu edu to hldp andrew cs fiu edu We make use of the Oracle export and import utility to complete the task Before we begin using the Export utility the following steps are necessary Export the Schema Step 1 Run catexp sql This job is done by Lin Luo the DBA of HLDP database The script performs the following tasks to prepare the database for export e Creates the necessary export views in the data dictionary e Creates the EXP FULL DATEBASE role e Assigns all necessary privileges to the EXP FULL DATEBASE e Assigns EXP FULL DATEBASE to the DBA roll e Records the version of catexp sql that has been installed Step 2 Ensure that there is enough disk space to write the export file Since our database is not very big in size there is no problem about the storage Step 3 Verify that we have the required access privileges To use Export you must have the CREATE SESSION privilege on an Oracle database To export tables owned by another user the EXP FULL DATEBASE role has to be granted to the user who will perform the export Step 4 Prepare the parameter file We specif
80. aining Before the actual training session a computer usage survey will be passed to each attendee to collect the information such as their computer skills This is very useful for the instructor to assess the student s needs and to better meet his her training expectations A user s manual is under development and will be used in the training The following topics will be covered in detail Logging into the system Performing simulation and specifying the model parameters Displaying the simulation results through web interface Viewing the documentations and PHRLM related publications Using the on line Question and Answer facility to submit questions answers and browse other people s questions answers e The security policies Each student will receive a user s manual at the training session The materials used in technical training program will be tailored and used as the auxiliary guide in the end user training sessions All the training materials can be accessed via PHRLM s web documentation system We also implemented an on line Q amp A facility for end users to submit their questions and get answers by using the same interface In addition an end user request response policy is set up for user questions submitted by email The designated technical staff will be responsible for answering users PHRLM related question during regular working hours and try to response to end users as soon as possible Similar to technical training
81. ale HD D Nature_Coverage County region Vi NZ ICompany Company_Name T IPolicies ICompany in Company Name string Damage Ratio Damage Ratio Windborne windborneZips windborne toString Figure 4 1 3 Class Diagram for ILM Scenario Based 4 A 32 Class Description This section addresses the major classes used and their functionalities e ILM This is the main class of the application which instantiates all of the other classes e Matrices This class forms the vulnerability matrices for Content Appurtenant Ale and Structure in the required format e Damage Ratio This class reads and stores the Damage Ratio required in calculating the expected loss of properties e Windspeed This class stores the wind speed corresponding to each zip code e Policy This class reads the input file and categorizes data e Company This class gets the input data and formulates it for each company in the proper format e Windborne This class holds the windborne derby region information 4 A 33 4 3 4 Sequence Diagram for ILM ILM Matrices Damage_Ratio Wind_Probability Company Policy I I 1 I I I I woo li I I I I Uger WindProbabiliy
82. ameters and their values in a parameter file Storing the parameters in a file allows us to be easily modified or reused and is the recommend method for invoking Import We create the parameter file using the DOS text editor as follow FILE dba dmp the name of the export dump file OWNER czhang02 we import the schema from czhang02 s account IGNORE n display object creation errors SHOW y list the contents of the export file which are not imported GRANTS y imports objects grants ROWS y rows of table data are imported LOG dbaemp save the import report and error information to file dbaemp Step 3 Invoking the Import Utility In our case we use the User mode to import the exported dump file dba dmp to HLDP Execute the following command in DOS gt imp username password PARFILE paramsi dat 5 A 15 5 7 Data Checking Since the import was terminated with warnings we have to check that the entire schema in the old account is moved to the new account After importing Chengcui a team member made a first pass check to make sure that the schema in the new account is the same as the one in the old account Although there are warnings with the import but actually all the data and tables as well as database objects are all successfully imported It is very important to ensure that the imported data is consistent with the original data file We randomly retrieved some records from the table in
83. ass performs all the calculations needed for WSP use case It includes following methods Data count zip Read the zipcodes txt file to count and return the number of lines in the file e build zip array Read the zipcodes txt file and store the zip codes in to an array wp max windspeeds Search through one storm to find the maximum 3s Gust wind speeds at each zip code and record it as intermediate output Iterate this process through all storms e Order Max WindSpeeds Search through all the storms and record in an array the maximum wind speed for specified zip for each storm Sort Max WindSpeeds Sort the wind speed array in descending order bin decider maps the specified wind speed to a bin e Distr Bands Estimates the wind speed probabilities of the wind speeds lying in the band PVxz and above the midpoint of the band Pv gt y e CalcPVxz Calculate the wind probabilities Pv and PVxz and produces the CalcP__ txt output file 3 D 9 3 4 3 6 State Chart Diagram Figure 3 4 4 depicts the state chart diagram for WSP use case This diagram illustrates states that the use case goes through from beginning to end Page requested 5 mouseClicked User data sent Processing Results Displayed e Figure 3 4 4 State Chart for WSP 3 4 3 7 Sequence Diagram M W spCalc jsp WP Data Scientist page_r
84. ations on a given platform Native programs writing in languages other than java such as C C can be integrated into Java applications and it is ensure these programs are completely portable across all platforms By programming through the JNI you can use native methods to create inspect and update Java objects including arrays and strings to all Java methods to perform runtime type checking 1 A 8 Section 2 Storm Forecast Module Module I Section 2 1 Annual Hurricane Occurrence AHO Use Case I 2 1 1 General Description Of AHO AHO short for Annual Hurricane Occurrence is the first use case in the FIU IHRC Public Hurricane Risk and Loss model It aims at estimating the probability distribution for annual hurricane occurrence and generating a series of simulated years along with their associated number of hurricanes according to the probability distribution that has the best goodness of fit 2 A 2 Threat Area In our latest PHRLM model only the hurricanes fall in the threat area are considered Here the threat area is defined as a radius of 900 km centered at 29 0 North Latitude and 83 0 East Longitude which is actually the region of the interest The threat area surrounding Florida is shown in s1 In other words a hurricane will be considered if it ever passed the threat area and it ever had the w
85. atus of users In addition for system security and reliability purpose we also deploy a development environment besides the production environment Modifications to the code and data are done in the development environment and tested by in house developers The final production code and data can only be checked into the production environment by authorized personnel Baseline tests are always run to ensure the model is functioning properly and reproducing known results The models resulted from PHRLM project can only be used by authorized users Authorized user accounts are created by the project manager The models are accessible to authorized users via web applications using JSP The source code is stored in server side and cannot be tampered with by unauthorized users The output of the models is always coupled with the analysis parameters and other information needed to reproduce the analysis results which are documented in each technical report to maintain the information integrity Passwords will be kept private in a not shared disk Passwords will consist of a minimum of 6 alphanumeric characters no common names or phrases Passwords will be changed every 120 days this will be enforced by an automatic expiration procedure to prevent repeated or reused passwords User accounts will be frozen after 3 failed logon attempts All erroneous password entries will be recorded in an audit log for later inspection and action as necessary Session
86. chine No need to implement it LoginCheckBean This class is in charge of user login It gets the username ID and password the user enters and then checks with the information stored in Oracle Database We will explain the login process later with a sequence diagram SGTIndex This class is used to get user selections e g year range It will then pass control to classes that will get data from database SGTSimulation This class will call other classes to get the needed data from database and then call the related classes to generate the SGT and then display the results using a table on end user s web browser getSGTDataBean The true class to get related hurricane information from database SGTDataEntry It is a class served for get the Julian Date and Genesis Time Database It is an abstract concept It includes all the system provided database operations SGTBean The class is used to interface with the true math model MathModel The C class using IMSL library functions to fit distribution and generating the simulation It will communicate with the Java main application using JNI interface IMSL Library Library functions provided by IMSL 2 B 14 2 2 3 5 Activity Diagram Figure 2 2 8 depicts the activity diagrams consisting of the major activities in Use Case Two This activity diagram offers a clear and direct understanding of the business logic of Use Case Two System gi
87. d provide the user the option to go back to the first step or to quit the system Year Range Selection 1851 2003 1900 2003 1944 2003 ENSO Multi Decadal SUBMIT GO BACK QUIT Figure 2 2 2 Dataset Selection Interface 2 B 8 Table 2 2 5 Valid year ranges Valid Year Range 1851 2003 1900 2003 1944 2003 ENSO Multi Decadal C The third step storm genesis time results display After the user submits the selected year range the system generates the storm genesis time for each simulated hurricane store the results into database and display the success message to the user if succeeded or display error message when failed The system also should provide the user the option to go back to the third step to start a new operation or to quit the system Figure 2 2 3 is the Simulation Results Display Interface in the case of success Figure 2 2 4 is the Simulation Results Display Interface in the case of success Year Range 1851 2003 System has successfully generated 10 000 simulated storms and stored the results into database GO BACK NEW OPERATION QUIT Figure 2 2 3 Storm Genesis Time Results Display Interface in case of success 2 B 9 Year Range 1851 2003 System failed to generate 10 000 simulated storms and no results were stored in database GO BACK NEW OPERATION QUIT Figure 2 2 4 Storm Genesis Time Results
88. database If the password or username is not matched with the information stored in the database the user gets an error message and is asked to login again If the username and password are matched the user can continue to access the system 2 A 17 B Simulation Process client Client dataSelect simulation calculateParam getData database DSSelection SimuSelection CalMVS Bean getDBean Database submit submit getData j igetConnection lt rexecuteQuery gt lt E lt process KL acd ioe SS Figure 2 1 10 Sequence diagram for simulation process Step 1 The user selects a data set e g a year range and then clicks the Submit button Step 2 The dataSelect object captures the data set selected by the user and then calls the simulation object to get the data from database and processes the retrieved data Step 3 The simulation object connects with the database creates the query and then gets the desired data from database Step 4 The simulation object calls the calculateParam object to calculate some statistic values of the data set such as mean variance and standard deviation values Step 5 The simulation object returns the mean variance and standard deviation values back to the user
89. dcalcjsp _ Wind Field Roughness Calculation Results The Zoa can not be found in database for the zip code in dataset 2 a1 Sets Number 3 an zip zoo E Uto m s Uta m s Va n 2 Yank spl 1 33133 o 01 0 239981785 3 4428486324656387H 308821623501995 40 1773632127498489 99729359655964 3 33133 l0 0310 34291413400000004 5 026559003399833 P 96994200884561850 33945090290669 112 760370022511 1 Input from wind model for each zip code centroid Zip Zip Code Vo Surface Wind Speed for open terrain produced by the wind model m s WD Surface Wind Direction Deg from North Zoo Roughness Length m for open terrain 0 03 m 2 Input from roughness table for a given wind direction Zoa Actual roughness length based on FEMA HAZUS conversion table relating land use land cover LULC to aerodynamic roughness m 3 Output U o Open terrain friction velocity m s Uta Actual terrain friction velocity m s Va Surface wind speed for actual terrain m s Vamph above with english units of statute miles per hour Site directory 4 internet Figure 3 3 20 The result webpage which caught the exception when the Zoa value can not be fetched for some specific Zip code 3 C 23 3 C 24 Section 3 4 Wind Speed Probability WSP Use Case VI 3 4 1 General Description
90. deg ter day hour min zonal meridional total m s dir deg OT al 8 0 8 13654 10 2191 13 0626 321 zipcode 32 longitude 80 2700 deg latitude 25 3400 deg ter day hour min zonal meridional total m s dir deg OT T 8 0 8 26174 4 41157 9 36580 298 zipcode 33 longitude 80 1310 deg latitude 25 7100 deg ter day hour min zonal meridional total m s dir deg OT l 8 15 7 02892 10 4825 12 6210 326 zipcode 34 longitude 80 3200 deg latitude 25 6100 deg ter day hour min zonal meridional total m s dir deg OT 1 8 15 6 55432 6 86560 9 49187 316 3 B 4 3 2 5 Technical Description of Wind Field Model Input storm track Centre pressure R max Longitude Latitude Isflg Holland B date amp time for each of hourly fixes of the storm Wind field Model Output Land fall or by passing location longitude latitude of storm Maximum OT wind speed time direction any where in the storm Maximum Marine Exposure any where in the storm Maximum wind speed time direction at each zip code affected by the storm Figure 3 2 1 Input Output of Wind Field Model Once a simulated hurricane moves to within a distance threshold of Florida communities the wind field model is turned on Gradient balance represents a circular flow caused by the balance of forces on the flow whereby the inward directed pressure gradient force is balanced by outward Coriolis and centripetal accelerations The coordinate system translates with the
91. e e Username under which each parse occurred e Each commit and rollback We can enable the SQL trace facility for a session or for an instance When the SQL trace facility is enabled performance statistics for all SQL statements executed in a user session or in an instance are placed into a trace file The additional overhead of running the SQL trace facility against an application with performance problems is normally insignificant compared with the inherent overhead caused by the application s inefficiency TKPROF Facility After executing the SQL trace we need to run the TKPROF facility to format the contents of the trace file and to place the output into a readable output file Optionally TKPROF can also e Determine the execution plans for SQL statements e Create a SQL script that stores the statistics in the database TKPROF reports each statement executed with the resources it has consumed the number of times it was called and the number of rows it processed This information lets us easily locate those statements that are using the most resource 5 A 20 The steps to use the SQL trace and TKPROF facilities 1 Set initialization parameters for trace file management 2 Enable the SQL trace facility for the desired session and run your application This step produces a trace file containing statistics for the SQL statements issued by the application 3 Run TKPROF to translate the trace file creat
92. e type of the tropical cyclones Instead of using string we can use number 4 to represent a tropical storm and number 5 11 to represent hurricane level 1 5 5 A 16 According to the new schema we change all s fixobj stage like H or s fixobj stage Tropical Storm statements in the old schema to s fixobj stage gt 4 In the new schema s fixobj stage gt 4 functions the same way as the old one using string matching Shown below is the revised query which works correctly in the new schema and is more effective compared with the original queries select Year count Cyclones from select to char when t yyyy Year from atmosevent list s where s basin 1 and s when t between 01 JAN 1851 and 31 DEC 2001 and s type gt 4 group by Year We made the same changes for all the queries of Use Case One and Use Case Two The execution speed of queries is nearly three times faster than before 5 9 Database Tuning 5 9 1 Tuning SQL Statements Although the execution speeds of the SQL statements have been greatly improved by revising the schema additional SQL tuning efforts are necessary to improve the performance of the statements The Goals of SQL Tuning Oracle SQL tuning is a phenomenally complex subject and we will begin with a high level description of the goals of SQL tuning and get into details later on There are some general guidelines that all Oracle SQL developers must follow in order to im
93. e case A The first step the user logs in the system Figure 2 1 2 depicts the Login Interface The user enters a User ID and a corresponding password and then submits the login request to the system The system verifies the user ID and password If the User ID and Password are correct the access right is granted to the user otherwise the system gives the invalid user password error message and asks the user to reenter the User ID and password User ID FDOIUSER Password SIS Login Figure 2 1 2 Login interface B The second step the user selects a year range Figure 2 1 3 illustrates the Year Range Selection Interface The system presents a list of year ranges to the user The valid year ranges are shown at Table 2 1 9 The user selects a year range and submits to the system Year Range Selection 1851 2003 1900 2003 1944 2003 ENSO Multi Decadal SUBMIT QUIT Figure 2 1 3 Year range selection interface 2 A 10 Table 2 1 9 list of the valid year ranges Valid Year Range 1851 2003 1900 2003 1944 2003 ENSO Multi Decadal C The third step the user specifies the number of years for simulation Figure 2 1 4 shows the Simulation Selection Interface The system displays the year range selected at step 2 The user then specifies the number of years and submits to the system Year Range Selected 1851 2003 Number of Years for 10000 Simulation
94. e corresponding damage probabilities VII SumEL is across all wind speeds which is obtained by aggregating all the expected losses at different wind speed with respect to the corresponding probability for the wind speed VIII SumAEL aggregates all expected losses for one company NOTE Save information zip code county region construction type 4 types of coverages property value company for SumLs SumLc SumL pp SumL x and SumEL For SumLs SumLc SumLapp and SumLa y save wind speed too and for SumEL save V sum of V where sum of V is for each construction type Masonry Timber Mobile home and is calculated offline Variance of SumA EL can be calculated for a company for a Zipcode or for a construction type 4 A 23 gt The non parametric approach for generating Scenario Based Expected Loss Costs Portfolio File Determine Construction Type Manufactured Wood Masonry Others Zone Type Zip Code Determine region amp eliminate all matrices which do not apply Select Applicable Weighted Vulnerability Matrices Sub Region Determine based on zip code amp eliminate all matrices which do not apply Neither Windborne Debris Region High Velocity Hurricane Zone least stringent replacement more stringent replacement more stringent replacement requirements requirements apply requirements apply to windows apply to windows and roof Structure
95. e list of simulations Then the user has the option of entering a zip code in order to view a summery of results for that zip code In order to do this user has to click on the Enter zip code that you want to process radio button and then type the required zip code in the input box provided below If the user doesn t want to enter any specific zip code just press Submit button If the user selects the Simu_other option from the select base data set drop down list then he she has to enter the start file number end File number and the number of years in the simulation This option is provided for advanced users who may require to do test runs on arbitrary data sets 3 4 4 4 WSP results page Note The results listed below are just placeholders Actual values will be different 4 Wind Speed Probability Microsoft Internet Explorer File Edit View Favorites Tools Help Bak gt gt 5 A Asah Garavortes Breda BE 8 BOSD Address amp http sirene cs fiu edu 8888 FDOL WSP WSPCalc jsp Q Go Links y Y oy Search web E NEW Toolbar Update Uf Bookmarks gt gt Wind Speed Probability Results Number of Simulation years 100000 Total number of input files 9 Here is the standard output of the command 3s Gust WindSpeeds for Zipcode 33133 WS 0 38 677314 Storm Name date time storm0000005 10 14 1 01 00 WS 1 108 718022 Storm Name date time storm0
96. e not hurricanes In the latest version of model only the hurricanes in the threat area are considered 3 The system uses the data retrieved in Step 2 the years and their associated numbers of hurricanes and the Statisticians equation to generate the parameters of probability distribution 2 A 5 Detailed steps are as follows 3 1 The system fits the distribution for the historical data from Table 2 1 4 To do so system uses historical data from Table 2 1 4 to calculate the mean and the standard deviation 3 2 The system stores the output mean amp standard deviation from 3 1 in the database 3 3 The system determines the distribution fits for each range using the Poisson model 3 3 1 For Poisson model the system calculates the Mean u 3 3 2 The system determines the goodness of fit for the Poisson model 3 3 2 1 The system calculates k the maximum number of hurricanes in the data set 3 3 2 2 The system calculates the number of hurricanes Xo Xx and the observed frequencies Oo Ox where O represents the number of years in which there were 1 hurricane occurring 3 3 2 3 The system calculates n the sum of the observed frequencies generated in 3 3 2 2 3 3 2 4 The system uses the Mean u and IMSL library functions to calculate the expected frequencies for the number of hurricanes Ey Ex where E represents the expected number o
97. ecord all the access rights and responsibilities of each CVS user Therefore in PHRLM general users can submit patches to a the maintainer authorized user and the maintainer will commit changes directly to the repository In the future we plan to implement an easy to use CVS commit log search interface for scanning CVS commit logs form any part of the repository over any time period for all users or for a particular user It is required that all CVS users need to use ssh to access a repository on a remote machine This is set in CVS s configuration file In addition the development team members are required to add meaningful change notes for the appropriate files By doing that it is much easier to locate the correct version in roll back operations when needed 6 C 3 6 C 4 Section 6 4 PHRLM Testing Procedures 6 4 1 Software Testing Procedures PHRLM software testing and verification is done in three stages i Code inspection and the verification by the code developer Code developer should carry out sufficient amount of testing on the code and should not deliver the code until and unless he she is convinced of proper functionality and robustness of the code In this level of testing should code level debugging walk through the code to ensure proper flow inspection of internal variables through intermediate output printing and error logging use of exceptio
98. ed for those Database objects For example you can create a standard data type used for all address data 3 Defined Access Paths For each object you can define the procedures and functions that act upon it which means you can unite the data object and the methods that access it Having the access paths defined in this manner allows you to standardize the data access methods and enhances the reusability of the objects 5 A 1 latitude_deg longitude_deg max_windspeed_mps min_pressure_mb height_m stage rmax crossing when_t fix_id key_id fixobj stm_nbr at_time when_t for_event A emn name event_id type produced_id basin produced by ATCF name key id type description PK PK PK FK2 when t at time event id stm nbr when t name type basin ATCF name for event produced id produced by fixobj Px pet type description category_no state_code state_code storm_id landfall_obj category_no state_code EIE L gt PK Primary Key EN Un Unique Constraint Type Table FK Foreign Key es A sl Relationship Relationship between Relationship between Tables Tables and Types between Types 5 A 2 5 3 Description of the Objects and Tables There are 6 object types in HURDAT Schema 1 latitude
99. ed in Step 2 into a readable output file This step can optionally create a SQL script that stores the statistics in the database 4 Interpret the output file created in Step 3 Step 1 Set Initialization Parameters for Trace File Management Before enabling the SQL trace facility one should check the settings of the TIMED STATISTICS USER DUMP DEST and MAX DUMP FILE SIZE parameters TIMED STATISTICS This parameter enables and disables the collection of timed statistics such as CPU elapsed time by the SQL trace facility and the collection of various statistics in the dynamic performance tables The default value of FALSE disables timing The value of TRUE enables timing Enabling timing causes extra timing calls for low level operations This is a session parameter MAX DUMP FILE SIZE When the SQL trace facility is enabled at the instance level every call to the server produces a text line in a file in your operating system s file format The maximum size of these files in operating system blocks is limited by the initialization parameter MAX DUMP FILE SIZE The default is 500 If you find that your trace output is truncated increase the value of this parameter before generating another trace file This is a session parameter USER DUMP DEST This parameter specifies fully the destination for the trace file according to the conventions of your operating system The default value for this parameter is the default destina
100. el Casing This convention capitalizes the first character of each word except the first one E g testCounter e Do use descriptive names which should be enough to determine the variable meaning and it s type But prefer a name that s based on the parameter s meaning e Remember a good variable name describes the semantic not the type e An exception to this rule is GUI code All fields and variable names that contain GUI elements like button should be post fixed with their type name without abbreviations Example System Windows Forms Button cancelButton System Windows Forms TextBox nameTextBox 6 1 8 2 Variable Names e Counting variables are preferably called i j k 1 m n when used in trivial counting loops Note Indexer variables generally should be called i j k etc But in some cases it may make sense to reconsider this rule In general when the same counters or indexers are reused give them meaningful names 6 1 8 3 Method Names e Name methods with verbs or verb phrases 6 A 8 6 1 9 Reference The C Coding Style Guide by Salman Ahmed is used as a template for this guideline development 6 A 9 6 A 10 Section 6 2 Data Validation and Verification 6 2 1 About the Document This document is prepared as a part of the PHRLM project The primary audience for this guidance is practitioners directly
101. elow 3 1 Compute open terrain friction velocity U o Unit m s Uo Vo 0 4 Ln 10 0 0 03 3 2 Compute actual terrain friction velocity U a Unit m s using equation 3 of Powell et al 1996 Ua Uo Zoo Zoa 0 0706 3 3 Compute actual terrain wind speed at 10 m Va Va Ua 0 4 Ln 10 Zoa 3 4 Convert wind speed to the unit of MPH Vamph Va 2 24 Compute gust factors for peak 1 min wind over the hour G1h 60 and peak 3s wind over the hour Glh 3 based on the actual roughness See gust factor calculations below 3 5 Compute max 1 min wind m s occurring within 1 hour period V1 Va Glh 60 3 6 Compute max 1 min sustained wind speed in mph Vimph V1 2 24 3 7 Compute peak 3s gust in mph V3 Va Glh 3 2 24 3 C 4 4 Gust factor calculations 4 1 Compute friction velocity u 0 4Va Ly Zoa u 4 2 Compute Height scaling parameter based on a height of 10 m e u where f 2 7 292 10 sin Lat is the Coriolis parameter p 4 3 Compute the standard deviation of the wind speed 7 5Nu 7 Zoa 12 0 156Ln f Zoa 7 T Pd ose Oz 4 4 Compute the standard deviation of the low pass filtered wind speed considering a filter with a cut off frequency of 1 cycle per 3 seconds for the peak 3s gust and 1 cycle per 60 seconds for the maximum 1 min sustained wind speed calculation I 0 68 02 60 oxz 10 193 4 0 1 o z 60 0 3
102. encies for the number of hurricanes Ey Ex where E represents the expected number of years in which j hurricanes will occur The system calls the IMSL gamma function to calculate the expected values The system repeats Steps 3 3 2 5 3 3 2 6 for the Negative Binomial model The system generates a frequency table for the number of hurricanes The frequency table is a matrix consisting of 3 columns The number of hurricanes Xy Xx the observed frequencies Oo Ox and the expected frequencies Eo Ex See Table 2 1 6 2 A 7 3 4 2 5 The system stores the generated frequency tables Table 2 1 6 back to the database 3 4 2 6 The system stores the chi squared statistics Table 2 1 7 back to the database 3 5 The system selects the distribution that gives the highest p value to be the final selected distribution for the number of hurricanes per year 3 6 The system plots the observed frequencies versus the fitted frequencies in a histogram Table 2 1 5 Frequency table of number of hurricanes yearly frequencies and expected frequencies for the Poisson Model Hurricanes Xo Xx Observed Frequency Expected Frequency Oo s Ox Eo Ex 0 Oo Eo 1 O Ei 2 O E Xk Ok Ek Table 2 1 6 Frequency table of number of hurricanes yearly frequencies and expected frequencies for the Negative Binomial Model
103. eps 1 through 28 Sum across the insurance companies to get the Overall Expected Loss 4 A 14 Section 4 3 Computer Model Design 4 A 15 4 3 1 Use case View of Insurance Loss Model ILM A Actors There is one actor engineers in ILM Engineers use this use case to find the expected losses for particular companies for all wind speeds B Use Case It represents the expected losses for particular companies for given scenario or non scenario wind speeds The total expected loss is actually the summation of expected loss of the property for a given wind speed which is calculated by aggregating the losses at different intervals with respect to the corresponding damage probabilities C Use Case Diagram Figure 4 1 1 shows the use case diagram for ILM Engineer InsuranceLossModel Figure 4 1 1 Use Case Diagram For ILM 4 3 2 System Design This portion describes the system design The overall Flowchart Classes and activities for ILM is provided 4 A 16 A Program Flowchart of ILM gt The Non Parametric Approach for Generating Expected Loss Costs for a Given Exposure Portfolio File Determine Construction Type Manufactured Wood Masonry Others Zone Type Zip Code Determine region amp eliminate all matrices which do not apply Select Applicable Weighted Vulnerability Matrices Sub Re
104. equest gt calc_probability gt count_zip build_zip_array wp_max_windspeeds Order Max WindSpeeds V V V V Sort Max W IndSpeeds E bin decider Distr Bands CalcP V xz Figure 3 4 5 State Chart for Wind Probability Calculation Process 3 D 10 3 4 3 8 Program Flow Chart Input zipcodes txt Contains all the zipcodes in the threat area Input WSCoutput__ txt Set of files with zip codes and wind speeds One file a en wp_max_windspeeds Y Y Count_zip gt build zip array mo Output lr i integer array of zip number of zip codes codes y Loop 1 gt foreach of the zip code zipli Y Output Op tbt One output file for each corresponding input file Each line shows the zip code and maximum wind speed at that zip code Order_Max_WindSpeeds Search through all the Op txt files record the wind speed for zip i Y Output WSArr Windspeeds for zip i Y Sort Max WindSpeeds Sort the windspeeds in WSArr Y Output WSArr Sorted Windspeeds for zip i y gt bin_decider maps the wind speed to a bin wind band Distr_Bands ca
105. erty value company for SumLs SumL app SumL arr Variance of SumAEL can be calculated for a company for a Zipcode or for a construction type 4 A 29 4 3 3 Class Diagram and Description A Class Diagram for ILM Matrices vulnerability_matrix_structure vulnerability_matrix_content vulnerability_matrix_appurtenant vulnerability_matrix_ale Matrices populate matrix content populate matrix appurtenant populate matrix ale populate matrix structure IMatrices matrices 4m Ls v 4m Lc v m Lap v m Lale v Iterator iterators for every vector minBand maxBand pWi data windProbability string fileName toScreen getCorrectPwi double speed double getZip policyProcess Ipolicy pol companyProcess midPoint double start double end isISO string Constr void companyProcess Icompany company IPolicy ld Zipcode ConsType Lms Lmc Lmapp Lmale HD D Nature_Coverage County region Vi r T Policies ICompany in Company Name string Damage Ratio Dam Rat Damage Ratio Figure 4 1 2 Class Diagram for ILM 1 4 A 30 Class Description This section addresses the major classes used and their functionalities e ILM This is the main class of the application which instantiates all the other classes and performs all the
106. es executes or fetch calls QUERY Total number of buffers retrieved in consistent mode for all parses executes or fetch calls Buffers are usually retrieved in consistent mode for queries CURRENT Total number of buffers retrieved in current mode Buffers are retrieved in current mode for statements such as INSERT UPDATE and DELETE 5 A 23 Example select first rows Year count Cyclones from to response with the first row quickly select to char s when t yyyy Year from atmosevent_list s oscillation_constant_list o where os year to number to char when t yyyy and s basin 1 and s type gt 4 group by Year call count cpu elapsed disk query current rows Parse T 0 00 0 18 0 0 0 0 Execute 1 0 00 0 01 0 0 0 0 Fetch 5 0 03 0 04 0 17 0 47 total 7 0 03 0 24 0 17 0 47 isses in library cache during parse 1 Optimizer goal FIRST_ROWS Parsing user id 29 CZHANGO2 Rows Row Source Operation 47 SORT GROUP BY 466 NESTED LOOPS 1274 TABLE ACCESS FULL ATMOSEVENT_LIST 466 INDEX RANGE SCAN object id 28035 Rows Execution Plan O SELECT STATEMENT GOAL HINT FIRST_ROWS 47 SORT GROUP BY 466 NESTED LOOPS 1274 TABLE ACCESS FULL OF ATMOSEVENT LIST 466 INDEX RANGE SCAN OF ENSO STORM IDX NON UNIQUE LELLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLELLL
107. escription 0 ccc cece eee ee eee ne eect ne eee ee en ens 4 A 30 4 3 4 Sequence Diagram for ILM ccc cee ee cece cence HH m 4 A 34 4 3 5 Sequence Diagram for Scenario ILM cee ce sence nee ee nee e eens 4 A 34 5 Database Document 54 Specification for tlie Project eiii 5 A 5 2 Data Modeling eir tur ere th nont a tX 5 A 5 3 Description of the Objects and Tables eese 5 A 4 2 21 Data PROCES SIND oo dace Lat singer cui be Un So cat edie sey UE e ead tees 5 A 7 5 4 1 Original Data Processing sicco ceii eee e cece a 5 A 7 5 4 2 New Data Processing eoe etie TeS genes sceageiaes ca a bei 5 A 11 359 DAA O E AI 5 A 12 5 5 1 Original Data Loading ccc cece cence cece cence eee 5 A 12 5 5 2 New Data Lodi ee eke eet 5 A 13 5 6 Export and Import the Data vicio pain 5 A 14 3 Data DECKING cay cone Oy O 5 A 16 N S QUITE orat a A ode ULT C NESS EERE SBS 5 A 16 5 8 1 Change the Query Based on the New Schema eceeee scene ee es 5 A 16 3 9 Data PMI oraes A aaa 5 A 17 5 9 1 Tuning SQL Statements cece n cence eee mem m m meer 5 A 17 6 PHRLM Quality Assurance 6 L Coding Guide Lines d o ext at dessa vem ER AO RE RETE 6 A 1 6 1 1 About the Coding Guidelines 0 cece cece scene eee e eee ne ene en ens 6 A 2 6 1 2 File Organization oce e tco E ee PR RE einen deans 6 A 2 6 1 3 Code Indentation i ecce enm reb Rat ea dans 6
108. ess Figure 1 2 Use Case Diagram of the System 1 A 4 1 3 System Architecture Design Figure 1 3 gives a high level system architecture abstraction which follows the popular three layer architecture User Interface gt Application Logic gt Figure 1 3 The Three layer System Architecture A User Interface User Interface is the first layer of the system and also the only layer visible to the user Due to the popularity and convenience of the Internet a web interface is preferred so that the users are able to access the system online B Application Logic The second layer is used to glue the user interface and the underlying database OC4J is chosen to serve as the second layer C Database The database layer adopts Oracle9i database due to its advanced features for extensibility availability high performance and management 1 3 1 Detailed System Architecture Design Figure 1 4 is the general system organization There are five major components client OC4J container Java application Oracle database and math model Client Side The users can gain access to the system through any commonly used commercial browser such as Internet Explorer Netscape and etc The user interface should be friendly and can offer the user required functionalities as best as possible JSP Java Server Page technique is used to dynamically generate the content in the web page The basic idea of JSP is to allow Java
109. essure report is not available then an attempt is made to interpolate from reports that are within a 24 hour period including the target report Otherwise pressure is obtained using an empirical wind pressure relation see Appendix A Intensity changes are only computed for similar report types observed pressure or wind derived pressures Mixing observed and wind derived pressures was found to create spurious pressure changes Pressures over land were excluded 3 A 3 Due to sparsity of data in some regions or parameter space the PDFs may be coarsened bins widened so that a sufficient number of observations are available to create a robust PDF This is done in the RESIZE function in GENPDF Pressure changes are converted to relative intensity changes The relative intensity calculation is described in Appendix B PDFs for pressure and relative intensity are created though only one is used in STORMGEN By default the relative intensity PDF is used by STORMGEN Input Data GENPDF requires the following input files The HURDAT database A control file which contains the dates of the historical record to use Land Mask file the land mask is based on USGS land use data Outflow temperature for the relative intensity calculation see Appendix B Sea surface temperatures for the relative intensity calculation see Appendix B Output Data e Initial storm location motion and intensity of all selected
110. esting again on the modified units and also Regression testing should be carried to check if the modification affects any other parts of the code Note Please refer to the testing document for more details 6 D 2 Section 6 5 Code Count Tables Use Case Annual Hurricane Occurrence AHO Filename Source Comment Both Blank Total fitDistriBean h 18 22 0 5 45 AHOmath c 417 149 22 95 683 calcMVSBean java 39 84 6 24 153 dataEntry java 22 97 0 20 139 fitDistriBean java 81 150 1 56 288 getDBean java 239 193 7 66 505 myplot java 87 72 4 32 195 OutputFormatBean java 11 29 0 12 52 plotBean java 7 11 0 2 20 simulationBean java 40 122 11 32 205 storeAHOBean java 72 115 7 38 232 banner jsp 8 0 0 0 8 DSSelection jsp 41 0 0 4 45 plotSimulation jsp 109 18 0 19 146 simuSelection jsp 85 12 0 12 109 Use Case ll Storm Genesis Time SGT Filename Source Comment Both Blank Total SGTBean h 14 13 0 3 30 sgtmath c 302 106 2 222 632 getSGT DataBean java 39 35 2 28 104 SGTBean java 23 97 0 15 135 SGTDataEntry java 140 63 10 44 257 SGTbanner jsp 8 0 0 0 8 SGTindex jsp 45 0 0 3 48 SGTsimulation jsp 91 0 1 10 102 Use Case V Wind Speed Correction WSC Filename Source Comment Both Blank
111. evious version merge 6 C 2 versions and track changes This software is able to record the information for each file the date of each change author of the change file version and the comparison of the file before and after the change 7 The software development process is carefully monitored by designated personnel using CVS tracking tools and procedures For example cvs annotate provides a quick of finding who made what changes and when by displaying the last change information for each line of a file in the repository Such information includes the revision number for the last change of each line the user the date and the contents of the line The CVS history file records commits merges conflicts tagging updates to the working directory additions deletions and modifications of files in the repository The oginfo file controls where log information for cvs commit is sent This allows the project manager to keep track of the changes made by the development team and to maintain a central log of the project progress 8 PHRLM employs an access control mechanism that allows only authorized user accounts to modify parts of the hierarchy in the repository Authorization control is for commits only everyone can check out any part of the repository That is to say for user accounts other than the designated ones they do not have write access to the restricted area An access list is maintained to r
112. f the mid points of the interval of damage ratios for structure type 1 in zip code j It has N elements Now rather than use the MDR Mean Damage Ratio of the whole matrix the mid point of the damage ratio interval n X is used to represent an outcome and the probability of this outcome for a given wind speed is Paw In general for structure i in zip code k the mid point of damage intervals is Xjj and its probability of outcome for a given wind speed is Pijnw 7 Select the damage matrix for contents for structure of type 1 on its construction date d If the construction date d is not available another set of vulnerability matrix is used The matrix is provided by the Engineering team and consists of the simulated probabilities for various content damage ratio intervals and wind speeds The row represents a given interval n of content damage ratios and the column represent a given wind speed w The interpretation of the cells values etc is similar to the description given above for structure damage matrix Although the content damage depend indirectly on structural damage there is no stipulated functional relationship between the two matrices and their damage intervals 8 Select the AP and ALE damage matrices accordingly The Engineering team has generated independent matrices for AP and ALE based on indirect relationships between structural damage and both ALE and AP 4 A 7 9 From the insurance policy file
113. f years in which j hurricanes will occur 3 3 2 5 The system generates a frequency table for the number of hurricanes The frequency table is a matrix consisting of 3 columns The number of hurricanes Xo Xy the observed frequencies Oo Ox and the expected frequencies Ey Ex See Table 2 1 5 3 3 2 6 The system stores the frequency tables Table 2 1 5 generated back to the database 2 A 6 3 4 3 3 2 7 3 3 2 8 3 3 2 9 3 3 2 10 The system reconstruct the frequency table to make sure that no expected value is less than 1 and no more than 20 are less than 5 If either of the two conditions is violated then some categories are combined so that the conditions are always satisfied The system calculates chi squared statistics 1 e the goodness of fit statistics The system calculates the p value by calling the IMSL CHIDF routine The system stores the resulted chi squared statistic and p value Table 2 1 7 back to the database The system determines the distribution fits for each range using the Negative Binomial model 3 4 For Negative Binomial model the system calculates m and k estimates 3 4 2 The system determines the goodness of fit for the Negative Binomial modal 3 4 2 1 312 5 3 4 2 3 3 4 2 4 Repeat Steps 3 3 2 1 Through 3 3 2 3 The system uses Xo Xx Oo Ox k and n to calculate the expected frequ
114. for purposes of prediction hundreds or thousands of sample outcomes of losses can be generated for a given representative insured property Next policy modifications such as deductibles and limits can be applied to each outcome to generate a set of net of deductible losses which are then averaged to generate expected loss The losses can be aggregated Exceedance probabilities curves can be generated to estimate the likelihood the portfolio of policies will suffer losses in excess of a given level Alternatively the Value at Risk PML can be estimated at a given exceedance probability In the above traditional practice only the insurance policy files and claims data are used Typically the losses are modeled directly and are not derivative of other variables In our project however the catastrophe modeling process requires that in addition to the See Hogg and Klugman Loss Distribution 1984 particularly Ch 4 and 5 and the appendix and Klugman Panjer and Willmot Loss Models 1998 4 A 2 insurance data output data from the meteorology and engineering components must be utilized The distributions of losses are driven by both the distribution for damage ratios generated by the engineering component and by the distribution of wind speeds generated by the meteorology component The wind speeds and damage ratios are estimated through extensive simulations The engineering group has produced non parametric
115. ftware Documentations Related Model Papers DataSet Selectionl Dataset Selection2 Main Page Document Preparation LE hetpiffirene cs Fiu edu 6886 FDOI WSCindex jsp 4 Internet Figure 3 3 16 The input webpage shows the exception that the inputted set number is not an integer 3 C 21 User need to follow the instructions fill in all the blanks and input the data sets in the correct format Figure 3 3 17 shows some error inputs For example some blanks are not filled the zip code field is filled by words the wind direction is not in the interval 0 360 etc Figure 3 3 18 displays the webpage which caught these exceptions and displays several error messages 2 Roughness Calculation Microsoft Internet Explorer Ble Edt Mew Favorites Tools Help Q O x i f search frre Queda 2 2 JB Address 48 http firene cs fiu edu 8888 FDOLWSCindex jsp Wind Field Roughness Calculation How many sets Set Input From Wind Model Zip Code Wind Speed m s Wind Direction Roughness length m 33133 50 33133 Zip Code Zip Wind Speed Yo Surface wind speed for open terrain produced by the wind model m s Wind Direction Wd Surface wind direction Deg from North Roughness Length Zoo Roughness length m for open terrain 0 03 m Site directory Documentations FDOI Publications Demo Usecasel Demo Usecase2 Engineering Module Q amp A AB Internet 2 Wind Speed Caculation
116. g equipment The Director designee will contact the Lab Manager with final instructions to begin securing the School s equipment In the case where the Director designee is unable to contact the Lab Manager the Lab Manager shall report to campus when the National Weather Service issues a Hurricane Watch 36 hours before land fall If necessary the Lab Manager will call in additional personnel to assist securing equipment If the Lab Manager is out of town an alternate staff member will be designated to respond If the hurricane passes without major incident to our area and the hurricane warning is lifted the lab manager will return to FIU Additional personnel may be requested to assist in restarting systems Damage assessment will occur during this period If however the area suffers severe damage the ability of lab personnel to return to FIU may be hampered Communication with the above mentioned will be attempted If that fails an attempt to reach the campus within 72 hours after the lifting of the Hurricane Warning will be made If it is safe to enter the building damage assessment will occur and the systems restarted 7 A 7 7 4 Non Disclosure Agreement NON DISCLOSURE AGREEMENT The undersigned hereby agrees and acknowledges 1 That during the course of my employment at Public Hurricane Risk and Loss Model PHRLM there may be disclosed to me certain confidential information consisting but not necessa
117. ge as shown in Figure 3 4 9 To view the WSP use case page from the drop down list select WSP and click Go button 3 task page Microsoft Internet Explorer File Edit View Favorites Tools 1 Help El Back gt A A Gsearch fgravorites meda 4 ES 5 c3 gt Address http firene cs fiu edu 8888 FDOI FdoiLogin jsp Go Links gt x 7 e search web E3 NEW Toolbar Update X a Service Selection Page Please choose an online service Wind Speed Probability y Go d Done TOF d p internet Figure 3 4 9 Service Selection Page 3 4 4 3 WSP input selection 4 Windspeed Probability Calculation Microsoft Internet Explorer f File Edit View Favorites Tools Help EA Back gt amp 2 A Gsearch gravorites meda C EV dE Bea i 14 Address la http firene cs fiu edu 8888 FDOI task jsp Go Links gt gt yr y e Search web E3 NEW Toolbar Update Uh gt Bookmarks gt gt il Input From Wind Speed Correction a Select base y data set Simu_10000_1900 2003 y Start File End File No of Years y PAS use default Zip code 33133 C Enter zip code that you want to process Submit Reset If you enter an invalid zip Code the output will be generated to the default zip code 33133 ES ij Done A zi Figure 3 4 10 Input Selection page 3 D 13 User should select the input from the availabl
118. ghted Use timber matrices mobile Pij Pi jut Mobile NN home weighted home Misc matri matrices Portfolio is replacement cost LMc LMap provided Le 1 25 LMc Lap 1 25 x LMap Lc 0 5 LMs Lap 0 1 LMs LMc LMap LMale provided Le 1 25 0 5 LMs Lap 1 25 0 1 LMs Use weighted ces V 1 25 LMs 4 A 26 LMale provided LALE 0 4 LMc LALE LMale LMale provided Lae 1 25 LMale Late 1 25 0 2 LMs Late 1 25 0 4 LMc 4 A 27 DMs Vi Xs G Ve Xc AP Vap Xap ALE LMaLe XALE 4 A 28 Ls DMs Ds Le C De Lap AP Dap LALE ALE Ls gt LMs Lc LMc Mobile home SumLs SumLs Ls Pp SumLc SumL Lc Pp SumL app SumLapp Lap Ppap SumLare SumLare Lare Pp ALE Output SumL s SumL SumL app SumLarLe Finish portfolio Y SumAEL SumAEL SumLs SumL SumL app SumL ay Output SumA EL wa N Finish company Y REMARKS The steps in scenario based ILM are similar to the general ILM except that the wind speed for a certain portfolio is given SumL is expected loss of the property for a given wind speed SumAEL aggregates all expected losses for one company Save information zip code county region construction type 4 types of coverages prop
119. gion Determine based on zip code amp eliminate all matrices which do not apply Neither Windborne Debris Region High Velocity Hurricane Zone least stringent replacement more stringent replacement more stringent replacement requirements requirements apply requirements apply to windows apply to windows and roof Structure Type Determine from portfolio file info eliminate matrices which do not apply if unknown use other matrix Select Applicable Weighted Vulnerability Matrices Region Type 1994 Present 1994 Present Prior 1969 Y Weak Stand Y ard Weak Medium Select Applicable Weighted Standard Vulnerability Matrices Ca E Prior 1970 1984 1994 1969 1983 1993 prese nt Maid Dd 4 A 18 Select a company C Select a portfolio P j get information of Construction type Zip code County Region Property value Vi Wind deductible D limits for structure LMs content LMc App LMap ALE LMarg If Hurri_Deductible 0 Use Other Deductible Else Hurri_Deductible Hurri_Deductible ISO classification available Use weighted matrices Is it concrete Use timber matrices Use weighted weighted concrete matrices INPUT Vulnerability Matrices for Structure S Content C Appurtenant AP and ALE for a given construction type region based on a given mix of construction features Assume
120. gt var a b c g f iden 4 z Where gt are tab chars and are spaces 6 1 3 2 White Spaces Don t use spaces for indentation use tabs An indentation standard using spaces never was achieved Always use tabs Tab characters have some advantages Everyone can set his or her own preferred indentation level e t is only 1 character and not 2 4 8 therefore it will reduce typing even with smart indenting you have to set the indentation manually sometimes or take it back or whatever If you want to increase the indentation or decrease mark one block and increase the indent level with Tab with Shift Tab you decrease the indentation This is true for almost any text editor Here we define the Tab as the standard indentation character 6 A 3 6 1 4 Comments 6 1 4 1 Block Comments e When you wish to use block comments you should use the following style Line 1 Line 2 Line 3 xh As this will set off the block visually from code for the human reader e Alternatively you might use this old fashioned C style for single line comments even though it is not recommended In case you use this style a line break should follow the comment as it is hard to see code preceded by comments in the same line blah blah blah e Incase this kind of block comment is not applicable it is recommended to follow a similar standard 6 1 4 2 Single Line Comments e You should use the co
121. hart of WSC 3 C 11 3 3 4 3 A Class Diagram Class Diagram and Description WSCCalVamphBean java amp zzip I ati vo amp SW D zoo amp 37oa E8Gust_Factor JSP interface B Classes Descriptions BSiconnect BSdisconnect _ BalcGust BSicalVam ph BSigetRoughnes det BSifindCol ge tZo BSiqueryZoa BSigetZoa BSigetVam phi BSgetV 1m ph BSigetV3m phi BSigetUa BSigetUo BSigetVa BSigetV1 BSigetgst60 BSiseiLat BSisetZip BSisetVo BSisetZoa BSisetWD BSisetZoo Figure 3 3 7 Class Diagram for WSC Here is a brief introduction of the functions in the class we used gt WCSCalVamphBean Database This class performance all the functionalities needed for the wind speed correction calculation It includes the following main methods e connect Method is used to establish a connection to the database 3 C 12 e setLat double 1 Method takes an array of doubles and sets the latitude array to the passed array e setZip int z Method takes an array of integers and sets the zip array to the passed array e setVo double z Method takes an array of doubles and sets the Vo array to the passed array e setWD nt w Method takes an array of integers and sets the WD array to the passed array e setZoo double z Method takes an array of doubles and sets the Zoo array to
122. he system classes activities and the overall flow chart of AHO 2 A 12 2 1 4 2 1 Program Flow Chart of AHO The overall flow chart of AHO is illustrated as follows Begin User selects the dataset System gives out selection form oae a o Oracle DB System gets data from database mao System calculates the mean standard deviation etc IMSL library Fit distribution Poisson is better Simulation Simulation with with poisson binomial Display result graph to user Figure 2 1 7 Flow chart of AHO 2 A 13 Class Diagram Client plotSimula tion sub mit myPlot Binit BSigraphit iisetCoord BStorward Biiswitch plotApplet Applet from applet BSApplet BSdestroy BSigetAppletContext BSigetAppletlnfo EgetAudioClip 8getAudioClip BSigetCodeBase BgetDocumentBase BSgetlmage BSigetlmage BSgetLocale BigetParameter BSgetParameterlnfo Binit is Active BSinewAudioClip lay lay esize Biresize BisetStub BishowStatus BSstart Bstop 2 1 4 3 Class Diagram and Description loginCheckBean SD amp PasswD BSiSetlD BSSetPassWD 7 BBigetlD BSgetStatus Bogin
123. he overall flow chart of Wind Speed Model Implementation 3 2 5 1 Program Flow Chart of Wind Speed Model Wind field model has been implemented using Interactive Data Language IDL To be precise it calculates wind speed time series for each of the zip code affected by the storm The time series includes the date landfall time of the storm It also includes the zonal wind speed m s surface wind speed m s and the wind direction in degrees at regular time intervals General structure of the main IDL modules is given below GEMFPLEX Time series FIXSHOTS15 output PKWINDS Figure 3 2 5 General structure or flow of the IDL GEMFPLEX is analogous to a main or the entrant procedure in C C It reads g trackfile and separates it into individual track files for processing GEMF processes each single track Each of the procedures TRACK SUV FIXSHOTS15 and PKWINDS call other procedures TRACK reads the necessary input parameters from the storm track and thins out the fixes based on the storm category and saves track related quantities for future use SUV generates radial profiles from stationary cyclone equations FIXSHOTSIS generate field snapshots with azimuthal variation for each fix PKWINDS is responsible for picking the maximum wind for each zip code If the storm happens to encompass or run through the entire state of the FL then this step would end up consuming a lot of resources Note AII the e
124. hurricane vortex moving at velocity c The vortex translation is assumed to equal the geostrophic flow associated with the large scale pressure gradient In cylindrical coordinates that translate with the moving vortex equations for a slab hurricane boundary layer under a prescribed pressure gradient are udu v vou op 2 u a E du or r rag or pg op dQ PR ot 1 dv v v dv gt v 20u dv E nt LN y Dc e ee eps Ph 2 ES a fe rog r r a EN ot 2 where u and v are the respective radial and tangential wind components relative to the moving storm p is the sea level pressure which varies with radius r f is the Coriolis parameter which varies with latitude is the azimuthal coordinate K is the eddy diffusion coefficient and F c u F c v are frictional drag terms All terms are assumed to be representative of means through the boundary layer The motion of the vortex is determined by the modeled storm track The hurricane windfield model is based on a fully two dimensional time independent scaled version of the tangential and radial momentum equations 1 and 2 for the mean boundary layer wind components The model makes use of a polar coordinate representation grid Fig 1 centered on the moving cyclone The nested 3 B 5 circles are separated from their inscribed and circumscribed neighbors by a radial separation of 0 1 in units of Rmax Radius of maximum winds the azimuthal interval is 10 degree
125. ial velocity u and the storm relative departure from gradient balance 9 The storm relative tangential wind is then given by v v 0 3 B 10 Unfortunately the direct numerical solution of A3 and A4 is time consuming even though the equations are time independent because the non linear coupling of the terms necessitates an iterative numerical approach However equations A6 and A7 can readily be numerically integrated to furnish a completely symmetric windfield fully described by the radial profiles u s and v s 2 v s os The functions u s and o s so obtained can serve as radial profiles for the construction of basis functions for a more realistic attack on A3 and A4 Namely we put forth the ansatz u s 9 ffu 9 u s A9 o s 9 ffo P o s A10 where the azimuthal dependence is introduced through the form factors ffu a a cos 9 a sin A11 ffo 9 b b cos b sin A12 Now the six coefficients a0 al a2 and b0 bl b2 can be variationally determined by substituting A9 and A10 into the left hand sides of A3 and A4 supplemented by A5 to form the residuals RA3 and RA4 We then form the functional RA3 I JI A3 l X ludus v 0 0u 0 g 5 o a ut csing w c I RA4E Y 1 A4 I 2 RA3 RA4 J a 5 AGRID A13 Where the sum is taken over every spatial point for which the profiles and trigonometric functions are known polar grid and NGRID is the tota
126. iguration and maintenance of PHRLM These staff will include the personnel assigned to tasks such as loss model implementation client side user interface implementation web design report development and maintenance database development and maintenance and mainframe system integration They may also include the personnel with duties relating to the administration and maintenance for PHRLM and its components The objective of the training is for the student to gain enough knowledge and hands on experience to get started on his her tasks relating to system development The training will include several seminars and demos A lecture will be presented in each seminar Following the lecture each student will have an overall idea about the system structure and its individual components Then one to one computer based instruction will be entailed to reveal the appropriate technical details that the student needs to know The following is an overview of the topics that will be presented in each of the training seminars e Introduction to PHRLM system architecture system configuration software components web application interface etc e Database Component HURDAT data engineering data wind field data and insurance data Oracle 9i basics Oracle DBA basics for DBA only e AHO Annual Hurricane Occurrence Module Statistical models model implementation IMSL C library JNI web interface design using JSP e SGT Storm Genesis Time Module
127. ind speed of larger than 74mph at least Category hurricane when it was in the threat area The wind speed ranges for category 1 5 hurricanes is shown in Table 2 1 1 For instance hurricane Andrew 1992 12 as shown in Figure 2 1 1 is one of the qualified hurricanes which are considered in our model Table 2 1 1 The wind speed ranges for category 1 5 hurricanes Category Wind Speed mph 1 74 lt WD lt 95 2 95 lt WD lt 110 3 110 lt WD lt 130 4 130 lt WD lt 155 5 155 lt WD In our database 1851 2003 there are totally 1274 historical tropical cyclones which include all hurricanes and tropical storms that were not hurricanes After the filtering process using the threat area definition only 309 of them are considered as the valid historical hurricane records Figure 2 1 1 Threat area and the storm track of hurricane Andrew 1992 2 A 3 2 1 2 AHO Design Requirements Name Annual Hurricane Occurrence Threat Area Only Description The user enters a choice of year range and the system generates the following 1 A probability distribution for the number of hurricanes per year 2 A simulated number of years with their associated numbers of hurricane occurrences 1 The user enters a year range from the following selections 1851 2003 1900 2003 1944 2003 Multi Decadal ENSO NOTE Neutral Years All non ElNino and non LaNina years are conside
128. involved in implementing or managing data verification or data validation efforts This guidance should provide this audience with a conceptual overview on how to verify and validate the data All the personals involved in the implementing or managing data are asked to read and follow the instructions given in here 6 2 2 Introduction If the information being used is not credible there is no point in using it Decisions based on inaccurate or unreliable data can adversely affect the decision making process Data verification and validation is used to evaluate whether data has been generated according to specifications satisfy acceptance criteria and are appropriate and consistent with their intended use 6 2 2 1 Data Verification Data verification is a systematic process for evaluating performance and compliance of a set of data when compared to a set of standards to ascertain its completeness correctness and consistency using the methods and criteria defined in the project documentation 1 6 2 2 2 Data Validation Data validation follows the data verification process and uses information from the project documentation to ascertain the usability of the data in light of its measurement quality objectives and to ensure that results obtained are scientifically defensible 1 6 2 3 Procedures In the context of PHRLM project data validation and verification is mostly a one time process Under mentioned procedures may not applicable in
129. ion computation Once the correction computation is done the system displays the whole data set the input data set retrieved data and the computation results back to the user Figure 3 3 15 shows a result page as an example 3 C 20 Wind Speed Correction Microsoft Internet Explorer Ele gdt view Favorites Tools Help Back y A A Qserh Gurevortes Leda Y Ey S aeaa Address http irene cs fiu edu 8888 FDOI W5C Windspeed isp T Link Y e Search Web J5 NEW Toolbar Update C jMail E my Yahoo Ej Games Basketbal Personals b Music iM Finance Sign In Wind Speed Correction Calculation Results Total Sets Number 1 Num Zip ML rea U o m s U a m s Va m s Vamph mph Vimph mph V3mph mph 1 33133 50 0 60 0 03 0 2193 4428486324656387 3 96158084353828337 84566162629004484 77428204445771101 50465396025234 134 60524860791972 1 Input from wind model for each zip code centroid Zip Zip Code Vo Surface Wind Speed for open terrain produced by the wind model m s WD Surface Wind Direction Deg from North Zoo Roughness Length m for open terrain 0 03 m 2 Input from roughness table for a given wind direction Zoa Actual roughness length based on FEMA HAZUS conversion table relating land use land cover LULC to aerodynamic roughness m 3 Output U o Open
130. it k from the insurance policy database Determine the zip code j of the policy Extract the distribution of wind speeds for the zip code j from the wind database 4 A 10 5 6 7 8 9 10 11 Next determine the building type i and the building construction date d if available for the selected policy Select the damage matrix for structure of type i based on its construction date d If the construction date d is not available another set of vulnerability matrix is used The matrix is provided by the Engineering team and consists of the simulated probabilities for various damage ratio intervals and wind speeds The row represents a given interval n of damage ratios and the column represent a given wind speed w Note that in the scenario analysis the observed wind speeds are used Thus only the column corresponding to the observed wind speed for the zip code is used Let Xij be the vector of the mid points of the interval of damage ratios for structure type i in zip code j It has N elements Now rather than use the MDR Mean Damage Ratio of the whole matrix the mid point of the damage ratio interval n X is used to represent an outcome and the probability of this outcome for a given observed wind speed is Paw In general for structure i in zip code k the mid point of damage intervals is Xijn and its probability of outcome for a given observed wind speed is Pijnw Select the damage mat
131. ives from momentum equations A6 A7 A8 vZ votsg dz d sg s gz g sg s du first derivative of u dsg first derivative of sigma FIXSHOTS15 PRO This module calculates the field snapshots and their second time derivatives at each retained fix time on a polar grid extending outward from the storm center to 15 RMW in steps of 0 IRMW and 10 angle This would give a matrix of 151 x 36 points But three extra lines are added for the convenience of future calculations making the matrix dimension 151 x 39 S X9 m S K S SE Ge Figure 3 2 6 Polar grid Restore suv idl Restore nrmraysel0_15 idl which contain some trigonometric values corresponding to each of the grid point 3 For each retained fix construct the polar grid of earth relative marine surface winds onefix m 4 onefix m gives the polar grid of earth relative marine surface winds for exactly one fix 5 reform converts this 151x39 matrix in to a raw matrix of 1x 5889 6 zsnapi is a complex matrix which contains the snap shots of the retained fixes retained fixes X 5889 7 usnap contains the earth relative zonal winds and vsnap contains the earth relative meridional winds 8 Compute second time derivative of fields for time interpolation Time interpolation is done in order to find the details of the storm every minute Noo 3 B 18 Xi ONEFIX M This module constructs zonal and meridional windfield components f
132. l Requirements of Wind Field Model Name Wind Speed Model Description The user enters Category Year Date Time Latitude Longitude Centre pressure Rmax Holland B and lsflg for each of hourly fixes of the storm The system generates the following Landfall or bypassing location 1 e longitude latitude of storm Maximum open terrain OT wind speed time direction anywhere in the storm Maximum Marine MA Exposure at landfall or bypassing position Maximum wind speed time direction at each zip code affected by the storm PRI 1 The end user enters the input file as the following format lt number of fixes gt lt storm Number gt lt m d yyyy gt lt hh mm gt storm category gt lt year gt lt mmdd gt lt hh gt lt minute gt lt latitude gt lt longitude gt lt center pressure gt lt R gt lt Holand gt lt lsflg gt Example 13 storm00001 8 24 1992 05 00 4 1992 0824 0500 25 4 79 3 937 19 1 4772400 0 4 1992 0824 0600 25 4 79 3 937 19 1 4772400 0 2 Based on the input data from step 1 the model generates the output as the following Given below a partial output file showing the wind field for some of the zip codes affected by the storm Andrew while the original file contains wind fields for all of the zip codes in the threat area which were affected by this storm A Land falling storms ANDREW 8 24 92 5 00 UTC landfall longitude 80 3000 deg latitude 25 5000 deg ter day hour min zonal meridional total m s di
133. l described Section 5 Otherwise update pressure 6 Check if maximum relative intensity is exceeded cap if necessary If pressure is greater than 1011 mb dissipate storm 7 Calculate new Rmax Beta 8 If storm outside threat area terminate Otherwise go to step 3 9 After storm track is generated it is trimmed based on the distance criteria described in the Use Case for Zip Code Criterion Input Data e Initial storm location motion and intensity if using specified initial conditions e Initial storm location motion and intensity PDFs from GENPDF e Storm motion and intensity change PDFs from GENPDF e Hurricane genesis time output from Use Case for Hurricane Genesis SGT e Zip code locations used for distance criteria described in Use Case for Zip Code Criterion e Land Mask file e Outflow temperature file see Appendix B e Sea surface temperature file see Appendix B Output Data e Track positions of stochastic storms in original HURDAT format Note small changes are needed as the original format is not capable of handling large number of storms e Track positions in special format for use in wind model e Landfall data for diagnostic purposes e Diagnostic output file 3 1 2 3 Appendix A Wind Pressure Relation An empirical wind pressure relation is used to convert HURDAT wind reports to pressure The relation is dependent on region The relation is If longitude is gt 81 5W and latitude gt 20N P
134. l number of such grid points J then depends solely on the unknown coefficients a0 al a2 and b0 b1 b2 These coefficients are chosen to minimize J and so furnish us with an approximate solution for u s d and O s Q from which we form the storm relative radial and tangential wind components ur and vt namely ur s 0 u s 0 and vt s d vg s O s 0 A14 3 B 11 By adding the translational velocity c in polar coordinates to ur and vt we obtain the earth relative components of the windfield uer and ver uer s ur s csing A15 ver s d vt s 0 ccoso A16 where c is the normalized translation speed c Vgmax Finally since A3 A4 and A5 refer to a cyclone moving along the y axis the entire generated windfield grid must be rotated so that the y axis of the calculation coincides with the actual compass direction of motion of the translating cyclone 3 B 12 3 2 4 Computer Model Design 3 2 4 1 Use Case View of Wind Speed Model A Actors There is one actor scientist C Use Case Wind Speed model is used to estimate terrain wind speed C Use Case Diagram X O WindSpeedCalUseCase Scientist Figure 3 2 4 Use Case Diagram 3 B 13 3 2 5 Implementation of Wind Field Model This model is implemented using Interactive Data Language IDL language in Unix console based environment This section includes appropriate diagrams and t
135. l the available exceptions 3 C 22 Another possible exception is caused by the invalid zip code Sometimes the user inputs the zip code within the correct format but this zip code cannot be found in the specific lookup table As shown in the Figure 3 3 19 user inputted a zip code as 12345 which is not in the lookup table In Figure 3 3 20 the webpage shows this message so that the user can identify this error 2 Roughness Calculation Microsoft Internet Explorer Ele Edt yew Favorites Toos Help Q x 2 f Os yore Queda OL 03 LJ B Address dE http firene cs Fiu edu 8888 FDOI WSCindex isp Wind Field Roughness Calculation How many sets Set Input From Wind Model Zip Code Wind Speed m s Wind Direction Roughness length m 33133 0 01 0 02 0 03 Zip Code Zip Wind Speed Vo Surface wind speed for open terrain produced by the wind model m s Wind Direction Wd Surface wind direction Deg from North Roughness Length Zoo Roughness length m for open terrain 0 03 m Site directory Documentations FDOI Publications Demo Usecasel Demo Usecase2 Engineering Module Figure 3 3 19 The input webpage contains a Zip code value which is not included in the lookup table 2 Wind Speed Caculation Microsoft Internet Explorer Ble Edt Mew Favorites Tools Help QOQ O x A fn ser Perote edo G A i wW S Address http irene c fiu edu 8888 FDOL WSC WindSpee
136. lculates the wind speed probabilities of the wind speeds lying in the band PVxz and above the midpoint of the band Pv gt y Y Output Opf txt One file for each zip i CalcPVxz re calculates the wind speed probabilities of the wind speeds lying in the band PVxz and above the midpoint of the band Pv gt y using set of equations Output CalcP__ txt One file for each zipcode Figure 3 4 6 Unit flow diagram 3 D 11 3 4 4 Implementation of WSP WSP is online at http www cs fiu edu PHRLM 3 4 4 1 Login page Users need a username and a password to access the FIU IHRC Public Hurricane Risk and Loss Model Following is a snapshot of the login web page FIU IHRC Public Hurricane Risk and Loss Model User Login Page UserID PaseWD LOGN sibs DO D aane Figure 3 4 7 Login webpage for FIU IHRC PHRLM If the username password is wrong an error message will be displayed User will be required to input the username and password again to enter FIU IHRC Public Hurricane Risk and Loss Model Relogin Page UserID PaseWD LOGIN Wrong user name password BS LOIS meme Figure 3 4 8 Login webpage shows the inputted user ID or password is wrong 3 D 12 3 4 4 2 WSP page If the login is successful the user can see the web page named Service Selection Pa
137. lined in FIU SCS Hurricane Preparation Procedures The computing equipment associated with PHRLM will be secured and safeguarded by designated personnel such as the Lab Manager when the hurricane warning is issued When hurricane warning is lifted the lab manager will return to FIU and take in charge of system recovery 7 A 2 10 Security policies are documented and all PHRLM personnel are trained in security requirements and procedures When personnel Graduate Research Assistants supported by PHRLM Professional Programmers etc leave the PHRLM project they are required to sign non disclosure agreements to not keep nor disclose any confidential information documents at the proprietary level and above For details please check the attached documents Any sensitive or confidential data insurance data for example are kept on a local unshared disk on a system which has user access control and requires a login Screen locks are used whenever the machine is not attended Backups are done for that disk at daily basis In addition sensitive data should never be sent via unencrypted email Access to all PHRLM computers workstations is controlled by passwords A screen keyboard lock or login screen should be active on all machines when they are not in use A designated project manager is responsible for providing initial approval for a PHRLM computer account and for notifying the computing center of a change in st
138. ling models for hurricane wind speed predictions Journal of Structural Engineering 121 1700 1709 2 Ho F P J C Su K L Hanevich R J Smith and F P Richards 1987 Hurricane climatology for the Atlantic and Gulf coasts of the United States NOAA Tech Memo NWS 38 NWS Silver Spring MD 3 Kaplan J and M DeMaria 1995 A simple empirical model for predicting the decay of tropical cyclone winds after landfall J App Meteor 34 4 Ooyama K V 1969 Numerical simulation of the life cycle of tropical cyclones J Atmos Sci 26 3 40 5 Shapiro L 1983 The asymmetric boundary layer flow under a translating hurricane J Atmos Sci 40 1984 1998 6 Thompson E F and V J Cardone 1996 Practical modeling of hurricane surface wind fields Journal of Waterways Port Coastal and Ocean Engineering Division ASCE 122 195 205 7 Vickery P J P F Skerjl A C Steckley and L A Twisdale 2000a A hurricane wind field model for use in simulations Journal of Structural Engineering 126 1203 1222 3 B 21 8 Vickery P J P F Skerjl and L A Twisdale 2000b Simulation of hurricane risk in the United States using an empirical storm track modeling technique Journal of Structural Engineering 126 1222 1237 9 Kurihara Y M M A Bender R E Tuleya and R J Ross 1995 Improvements in the GFDL hurricane prediction system Mon Wea Rev 123 2791 2801 10
139. lones from to response with the first row quickly 2 select to char s when t yyyy Year 3 from oscillation constant list o atmosevent list s 4 where to number to char when t yyyy os year and s basin 1 and s type gt 4 5 6 group by Year 7 Elapsed 00 00 00 02 Structure 2 SQL gt select first rows Year count Cyclones from 2 select to char s when t yyyy Year 3 from atmosevent list s 4 where exists 5 select os year from oscillation constant list 6 where os year to number to char when t yyyy 7 and s basin 1 and s type gt 4 8 9 group by Year 10 Elapsed 00 00 00 05 From the example above it is obvious which statement has a better performance However this environment parameter cannot show us how much time the CPU uses for the issued statement and how much time used on the I O and those detailed information is very important We solve this problem by using SQL Trace and TKPROF facilities 5 A 19 Using SQL Trace and TKPROF The SQL trace and TKPROF facilities enable us to accurately assess the efficiency of the SQL statements SQL Trace Facility The SQL trace facility provides performance information for individual SQL statements It generates the following statistics for each SQL statement e Parse execute and fetch counts e CPU elapsed time e Physical reads and logical reads e Number of rows processed e Misses on the library cach
140. ludes e Zoa Actual roughness length based on FEMA HAZUS conversion table relating land use land cover LULC to aerodynamic roughness m U o Open terrain friction velocity m s U a Actual terrain friction velocity m s Va Surface wind speed for actual terrain m s Vamph Above with English units of statute miles per hour Input Vo U o U a Va Ms Pn ds 33133 acc E 0 219 3 96 Em lm TE Figure 3 3 4 Wind Field Roughness Calculation Result Interface 3 C 9 3 3 4 Computer Model Design 3 3 4 1 Use Case View of WSC A Actors There is one actor scientist in WSC D Use Case WSC is used to determine a more accurate model of terrain winds produced by the hurricane wind model C Use Case Diagram X O WindSpeedCalUseCase Scientist Figure 3 3 5 Use Case Diagram for WSC 3 C 10 3 3 4 2 System Design This section includes the appropriate diagrams to describe the system classes components activities and the overall flow chart of WSC 3 3 4 2 1 Program Flow Chart of WSC The flow chart of WSC is depicted in Figure 3 3 6 User Enters Data Values Vo WD Zoo and zip System displays form to the user for data System connects and gets data from the Oracle database Zoa Database System performs calculations Vamph Vimph V3mph System displays results to the user and or produces output file Figure 3 3 6 Flow c
141. milar to the ones mentioned in the previous subsection 5 A 11 5 5 Data Loading 5 5 1 Original Data Loading The output of data pre processing is the desired data format we need for populating the data into the new schema For testing purpose we first loaded the data into the FDOI at georges cs fiu edu using SQL Loader All the constraints in the database schema have been disabled in order to facilitate the loading process The loading codes for the three major database tables atmosevent list stormfix list landfall are listed as follows 1 Loading data into Table atmosevent_list load data infile atmosevent dat append into table atmosevent list fields terminated by trailing nullcols stm nbr when t date mm dd yyyy name type basin key id atm key seq nextval 2 Loading data into Table landfall LOAD DATA INFILE landfall dat TRUNCATE INTO TABLE landfall trailing nullcols storm_id TERMINATED BY landfall obj nested table TERMINATED BY dummy name COLUMN OBJEC state code TERMINATED BY category no TERMINATED BY I E 3 Loading data into Table stormfix_list load data infile stormfix_list_test dat append into table STORMFIX_LIST_TEST fields terminated by trailing nullcols event_id when_t date mm dd yyyy at_time fixobj column objec
142. mment style to comment out code It may be used for commenting sections of code too e Single line comments must be indented to the indent level when they are used for code documentation e A rule of thumb says that generally the length of a comment should not exceed the length of the code explained as this is an indication of too complicated potentially buggy code 6 1 4 3 In line File Documentation e At the beginning of the each file the purpose of the file should be documented using the following template e For each code revision lt Revision History gt has to be updated Filename Creation Date WProbability cc 05 13 2004 description Calculates wind speed probabilities Input Surface corrected wind speeds 3S gust from the WSC module Output Probabilities of wind speeds from 20 300mph interval is 4 mph Revision History date developer Description 05 19 2004 kwick001 initial code f S S SSS9 S S S S S S S S S 6 A 4 6 1 4 4 Inline Function Documentation e At the beginning of the each file the purpose of the file should be documented using the following template e For each revision lt Revision History gt has to be updated EL lt Function gt lt Creation Date gt count zip 02 04 2005 Parameters none dd Return Number of lines i
143. myPlot plotset buttono bject plotSim ulation NumericSet m yB utton submit init gt pum action l lt graphit gt lt t lt Figure 2 1 12 Sequence diagram for plot process Step 1 The user submits request for result visualization Step 2 The plotObject object initializes an instance of myPlot class to do the plot task Step 3 The initialized plot object creates a plotset object to store the result obtained from the simulation Also it creates several buttons instance of myButton class to give user the choices to move forward backward or change the figure type Step 4 The plotObject object then calls the plot object to plot the resulted simulation data set and displays it to the user In user s web browser 2 A 20 2 15 Implementation of AHO The implementation for use case AHO has already been finished The demo is online at http www cs fiu edu PHRLM 2 1 5 1 Login page The users need a username and a password to access the FIU IHRC Public Hurricane Risk and Loss Model Following is the snapshot of the web page for login FIU IHRC Public Hurricane Risk and Loss Model User Login Page UserID PassWD LOGIN Figure 2 1 13 Snapshot of the Login page If the username password is wrong error message is given and the user is required to input the username
144. n handling mechanisms calculation cross checks and verification of the output against sample calculations provided by the system modeler It is the developer s responsibility to collect at least one sample calculation from the system modeler and to compare the results against the results generated through the code ii Verification of results by the person who developed the system model Once the first level of testing is done the developer should send the sample inputs and the generated results back to the modeler Then the system modeler should double check the results against his her model Code is not put in to the production environment with out the OK from the modeler iii Review and extensive testing of the code by external group of software engineers System is rigorously checked for correctness precision of the output and robustness amp stability of the whole system Calculations are performed outside the system and compared against the system generated results to ensure the system correctness Extreme and unexpected inputs are given to the system to check the robustness Wide series of test cases are developed to check the stability and the consistency of system Unit testing Regression testing and Aggregation testing both white box and black box should be performed and documented Any flaws in the code are reported to the developer and the bug corrected code is again sent to the tester The tester should perform unit t
145. n the zipcodes txt file Description read the zipcodes txt file count the number of lines in the file Revision History 02 04 2005 kwick001 generate initial code 6 1 5 Variable Declarations 6 1 5 1 Number of Declarations per Line e One declaration per line is recommended since it encourages commenting In other words int level indentation level int size size of table e Do not put more than one variable or variables of different types on the same line when declaring them Example int a b What is a What does b stand for e The above example also demonstrates the drawbacks of non obvious variable names Be clear when naming variables 6 1 5 2 Initialization e Try to initialize local variables as soon as they are declared For example int val 10 6 1 6 Statements 6 1 6 1 Simple Statements Each line should contain only one statement 6 1 6 2 Return Statements A return statement should not use outer most parentheses Don t use return n n 1 2 Use 6 A 5 return n n 1 2 6 1 6 3 If if else if else if else Statements if and if else statements should look like this if condition DoSomething if condition DoSomething else DoSomethingOther 6 1 6 4 For Statements A for statement shoud have following form for int i 0 i lt 5 i DoSomething Note Generally use bracke
146. nce all the terms in the calculations for the expected losses are correlated We compute the variance as follows The variance for all dwellings of type i in zip code j will be o tao t Ly 0 4 Eu k 1 k To get the Expected Loss mean loss for structure type i in zip code j the E AL is calculated as the weighted average of the Expected Loss of all properties of type i The weight is the relative value of the structure Vi gt Vi Cor relative exposure of the structure LM j y LM y Expected Loss for property type i in zip code j gt va Vin Li k 1 To estimate the expected loss as a percentage of exposure for structure type i in zip code j use E Li Eto rus k l Repeat steps 5 through 21 for all property types i 1 I to get the Expected Aggregate Loss and Expected loss for all property types in zip code j Sum the E ALjj across all property types i to get the Expected Aggregate Loss for all exposure in zip code j E AL Y ElAL Sum 0 to get o the variance for all exposure in zip code j Pick another zip code and repeat steps 4 through 26 to generate E AL for all zip codes Sum across the zip codes j 1 J to get the Expected Aggregate Loss for insurance company m 4 A 13 E ALm for company m Y 4L j 1 J 28 Sum 0 to get o the variance for insurance company m j OS hi pany 29 Pick another insurance company m and repeat st
147. nd model include the following fields e Zip Zip code e Vo Surface wind speed for open terrain produced by the wind model m s e Wad Surface wind direction Deg from North The data range for this field is from 0 to 360 e Zoo Roughness length m for open terrain 0 03 m Latitude Latitude of the corresponding zip code a value between 20 and 40 3 C 18 A data input page is provided to facilitate the users to input the corresponding data easily Users are allowed to input a variety of collections of input data and submit this data for wind speed correction calculation By default the number of input data sets is one So initially there is only one set of blanks for the user to input his her data Figure 3 3 13 illustrates the snapshot of the dataset input page for wind speed calculation E Roughness Calculation Microsoft Internet Explorer File Edit View Favorites Tools Help Back ERAN A Search EgFavorites media 4 5r E 3 a Se i Go Links gt Address E http irene cs fiu edu 8888 FDOItest FDOL task jsp Y ee i search web E3 NEW Toolbar Update Ai Bookmarks gt Input From Wind Model From a File Select the base data set zl Manual Input C Manual Input How many sets Set Num Zip Code u Wind Direction ees a Latitude deg 1 Subrnit Stma 4 E Done 0 Internet y Figure 3 3 13 Snapshot of the first web page for WSC
148. nfigurations are employed in the system e Oracle application server runs on a Linux Server IRENE Dual CPU P4 Xeon 3 06GHz 2GB RAM 146GB 6 SCSI Disks 100Mbps connection to network Runs Linux Fedora Core 2 e Oracle database runs on a Sun Workstation ANDREW SunFire V250 Dual CPU UltraSparc III Processors 73GB 2 SCSI disks 100Mbps connection to network 2GB RAM Detailed information about the disk partition for the database server is showed in Table 8 1 DISK SIZE CONTROLLER MOUNTED ON TABLE SPACE c0t4d0s6 36GB controller 0 home andrew1 ORACLEO1 04G ORACLE03 16G ORACLEOS 16G c0t5d0s6 36GB controller 0 home andrew2 ORACLEO2 2G ORACLE04 12G ORACLE06 12G ORACLEO07 10G c1t1d0s7 36GB controller 1 home andrew Table 8 1 Detailed disk partition for the Oracle database server 8 A 3 e Other machines CHARLEY Backup Sun Server Sun UltraSparc Blade 1000 Dual CPU UltraSPARC III E 750MHz 1GB RAM 35GB 2 7200RPM Ultra160 SCSI disks 35GB 1 7200RPM FC AL internal disk 100Mbps connection to network Runs Sun Solaris 2 8 Generic_108528 29 ISABEL Backup Application Server Dual CPU Intel Pentium III 1 2GHz 1GB RAM 35GB 4 10K RPM Ultral60 SCSI disks 100Mbps connection to network Runs Redhat Linux SCS 7 3 kernel 2 4 26 IBM Cluster Property of School of Computer Science IBM RS 6000SP running AIX 5 1 PSSP 3 5 with 35 nodes
149. ns iv Category 3 Select fixes in 3 hour separations v Category 4 Select fixes in 2 hour separations vi Category 5 Select fixes in 1 hour separations 4 Merge retained fixes and additional landfall and seafall fixes ii SUV PRO This module computes the radial and tangential wind profiles u and v as well as the gradient wind profile and the functions g and d and their second derivatives 1 Restores variables saved in track idl 2 If Holland b is zero calculate it using B 1 38 0 00184Ap 0 00309R this step is rarely done 3 Calculates radial ur and tangential vt wind profiles for each storm fix Wind profiles are calculated at 201 points starting from O the storm center to 20 in steps of 0 1 in units of RMW 3 B 16 1v vi vil Vill 4 Calculates the gradient wind profile for each fix using VGHGEN PRO Calculate g and d using equation A1 and A2 6 Calculate initial estimate for alpha alfi neglecting first derivatives of u ando 7 Alpha is iteratively estimated until correct alpha is obtained USG PRO is used To check the correctness of the estimate the boundary condition following boundary condition is used Peak wind should be at s 1 i e if iw is 10 answer of alpha is correct 8 Momentum equations are used to furnish tangential and radial profiles USG PRO 9 Collects radial and tangential profiles into a structure 10 Saves the variables for use by the other procedures as suv
150. ntage is that with limited data we may never be sure if the right distribution has been fitted and the errors in the estimated parameters can be significant In non parametric estimation empirical functions are fitted to the data There is no worry whether a correct distribution has been fitted and uncertainties are likely to be lower The disadvantage is that a complex algorithm may be required that involves many steps and long computational time Hypothesis testing is also more complicated and stability may be a concern For various reasons we have decided to pursue the non parametric option initially Given the large computing power available computational time is not a major concern Thus it may be prudent to develop a logical non parametric and deterministic algorithm that should produce low uncertainties The broad algorithm utilized to estimate insured losses is discussed below 4 A 3 4 A 4 Section 4 2 Detailed design and Implementation of Insurance Loss Model Input Both the meteorology and engineering components provide outputs that constitute critical inputs into the insured loss model The meteorology component provides for each zip code the associated probabilities for a common set of wind speeds Thus zip codes are essentially differentiated by their probability distribution of wind speeds The Engineering component produces damage matrices tha
151. ntervals with associated probabilities The third major set data utilized are the insurance policy and claim data provided by several property and casualty insurance companies operating in Florida 4 A 6 4 2 1 ILM Implementation Steps gt The Non Parametric Algorithm for Generating Expected Loss Costs for a Given Exposure Here the exposure data by zip code is the only given observed data The wind probabilities and damage matrices are all modeled In practice to generate expected loss costs the method we adopt involves an algorithm with the following steps 1 Start with a particular insurance company m 2 Next pick a residential policy exposure unit k from the insurance policy database 3 Determine the zip code j of the policy 4 Extract the distribution of wind speeds for the zip code j from the wind database 5 Next determine the building type i and the building construction date d if available for the selected policy 6 Select the damage matrix for structure of type 1 based on its construction date d If the construction date d is not available another set of vulnerability matrix is used The matrix is provided by the Engineering team and consists of the simulated probabilities for various damage ratio intervals and wind speeds The row represents a given interval n of damage ratios and the column represent a given wind speed w Each cell represents the probability Paw Let Xi be the vector o
152. o construction type to the ISO classification and select the corresponding vulnerability matrix II The structure limit LMg is applied based on the portfolio type replacement cost or actual cash value to obtain V Vc and Vap HI The deductibles applied to structure content and appurtenant Ds Dc and Dap are calculated based on the mean damages obtained from the vulnerability matrices IV The damages of structure content appurtenant and ALE DMs C AP and ALE are calculated at different damage ratio intervals Here DMs C and AP are calculated based on V Vc and Vap respectively Then the losses of structure content and appurtenant Ls Lc and Lap are computed by applying the deductibles Two exceptions are handled as follows 1 Tf Ls Lc or Lap 0 set it to 0 2 If Ls Lc or Lap gt LMs LMc or LMap set it to LMs LMc or LMap The loss of ALE Lay is set to ALE which is calculated based on the ALE limit LM 5 V The structure loss Ls is compared to the structure limit LMs Two situations are handled 1 For mobile home if Ls gt 0 5 LMs Ls is set to LMs 2 For other construction type if Ls gt 0 5 LMs Ls is set to LMs NOTE It will be implemented However at this stage it will be commented out for later use VI SumL SumLs SumLc SumL app or SumLA g is expected loss of the property for a given wind speed which is calculated by aggregating the losses at different damage intervals with respect to th
153. o copies This can take much more time and means higher costs e Proofreading data This method involves somebody checking what is in the system is the same as the original input Always make sure to make a copy of the data in the data after importing them to the system and give this copy and the original copy of the data to a person who is not involved in the data manipulating process to compare and certify the correctness 6 2 4 Data Security and Integrity This section describes precautionary measures that must be taken in the event that computer malfunctions natural disasters or human error or actions occur that could affect collected data e Duplicate copies or back up system for data Florida International University School of Computer Science takes regular backups generally every Friday All the databases and data files are included in the backup Developer must make sure that they store all the data in those places that are backed up e Data security protocols are in place and effective Firewalls password protection access levels etc are established Accountability for data integrity clearly rests with the person entering the data and the responsible program specialist and manager Only those who are skilled and trained in proper data handling procedures are allowed the direct access to the database 6 B 3 6 2 5 References 1 EPA Quality System Quality Management Tools Data Verification and
154. ocedures Manual section IV part 6 backups are performed on a daily basis and are kept for six weeks Nightly backups of all UNIX data disks and selected Windows data disks at user request are performed over the network onto Exabyte Mammoth M2 tapes Full dumps are taken periodically it works out to every 2 3 weeks and incrementals are taken daily between them Off site backups are performed at the end of every semester and stored off site in the PC building at the University Park Campus which is the Monroe County hurricane shelter and has emergency power and climate control 3 The tape drives have built in diagnostics and verification to ensure that the data is written correctly to tape This ensures that if the tape is written successfully it will be readable provided no physical damage occurs to the tape The off site backup procedure performs a level O full dump of every disk in the department This means that each disk in the department will be backed up to tape in its entirety The dumps can be restored from tape preserving the original file structure and all permissions All read errors during the backup process are reported so if a file system fails to dump correctly the dump can be re done In the past we have successfully restored data from both our offsite and daily backups for many times and no problems have been occurred 4 In case of disasters we have implemented a set of preparation procedures and recovery plans as out
155. on 3 1 3 1 Use Case View of Storm Track Model A Actors There is one actor scientist B Use Case Strom track model is aimed at generating the storm tracks for simulated storms based on data obtained from Use Case II and stochastic algorithms C Use Case Diagram X D StormTrackUseCase Scientist Figure 3 1 1 Use Case Diagram 3 A 7 3 1 3 2 Storm Track Model Implementation This model is implemented using FORTRAN language in Unix console based environment This section includes the overall flow chart of Storm Track Model Implementation 3 1 3 3 Program Flow Chart of Storm Track Read control file Initilize bins Read i 2 HURDAT Desired year Yes Interpolate pressures if possible or i a No use wind pressure relation Pressure Report Yes f threat area and hurricane Yes Y Compute motion Bin initial motion and Compute translation 8 lt a speed and heading change and bin intensity angle Pressure report Yes Compute change in Compute relative compatible intensity intensity and bin Resize bins if Compute PDFs lt _ needed Y Output for Stormgen 3 A 8 3 1 3 4 Storm Track Output 12 storm00004 8 24 1992 01 00 1992 1992 1992 1992 1992 1
156. ongWindPress LatLongWindPress LatLongWindPress LatLongWindPress Card Sequential card number starting at 00010 in 1851 MM DD Month Day and Year of storm Storm category S Subtropical stage tropical cyclone stage E extra tropical stage W wave stage rarely used LatLong Position of storm 24 5N 61 0W Wind Maximum sustained 1 minute surface 10m windspeed in knots in general these are to the nearest 5 knots Press Central surface pressure of storm in mb if available Since 1979 central pressures are given every time even if a satellite estimation is 5 A 9 needed Position and Positions and intensities are at 00Z 06Z 12Z 18Z intensity TYPE C 92760 HRCFLABFL3 LA3 1 92760 Card 2 HR Tp 3 CFL BFL LA Hit 4 4 3 Storm Category Card Sequential card number starting at 00010 in 1851 Tp Maximum intensity of storm HR hurricane TS tropical storm SS subtropical storm Hit U S landfallings as hurricane LA Louisiana etc and Saffir Simpson category at landfall 1 minimal hurricane 5 super hurricane Note that Florida and Texas are split into smaller regions AFL Northwest Florida BFL Southwest Florida CFL Southeast Florida DFL Northeast Florida ATX South Texas BTX Central Texas CTX North Texas The first step is to extract
157. operations specified by the flowchart e Matrices This class forms the vulnerability matrices for Content Appurtenant Ale and Structure e Damage_Ratio This class reads and stores the Damage Ratio required in calculating the expected losses e Wind_Probability This class reads and stores the wind probability e Policy This class reads the input file and categorizes data e Company This class gets the input data and formulates it for each company in the proper format 4 A 31 B Class Diagram for Generating Expected Loss Costs for a Scenario ILM Matrices vulnerability_matrix_structure vulnerability_matrix_content vulnerability_matrix_appurtenant vulnerability_matrix_ale Matrices populate_matrix_content populate_matrix_appurtenant populate_matrix_ale populate_matrix_structure allWindProbability allWindProbability getPwi in zip double in speed double double minBand maxBand pWi data windProbability string fileName toScreen getCorrectPwi double speed double getZip ILM IMatrices matrices m_Ls v m Lc v m_Lap_v m_Lale_v Iterator iterators for every vector m_Vc policyProcess Ipolicy pol companyProcess midPoint double start double end isISO string Constr void companyProcess Icompany company IPolicy ld Zipcode ConsType Lms Lmc Lmapp Lm
158. or exactly one fix 1 Load single fix profiles and corresponding data 2 Calculate purely radial no azimuthal dependence functions on a polar grid GENSTREX M 3 Then introduce azimuthal dependencies and calculate storm relative wind field Equation A11 through A13 4 Calculate the form factors ao ai a2 bo bi b2 of equation A11 amp A12 cfu coefficients of u ag a1 a2 cfsg Coefficients of Sigma bo bi b2 initial estimate 1 0 0 initial estimate 1 0 0 5 Keep changing the estimate to minimize J equation A13 using AMOEBA MNRDU and MNRDSG AMOEBA is a built in function in IDL MNRDU Calculate a s keeping b s fixed MNRDSG Calculate b s keeping a s fixed 6 Form the earth relative wind field assuming that the storm moves northwards 7 Calculate u o uer and ver using equations A14 through A16 8 Storm rotates counter clockwise Once the northward storm translation speed is induced storm center tend to move towards west Shift m takes this into account and shifts the polar origin to the storm center 9 Advance phase 10 Orient the wind field to track direction Initially we assumed that the storm is moving northwards In this step some interpolation is required since the actual direction of the storm unlikely to lie exactly on a radial of the grid 11 Convert the radial and tangential wind fields to zonal and meridional components vystre holds the meridional component of the wind at each grid
159. prove the performance of their systems The goals of SQL tuning are as follows e Remove Unnecessary Large table Full table Scans Unnecessary full table scans cause a huge amount of I O and can drag down an entire database We first evaluate the SQL query statements in terms of the number of rows returned by the query If the query returns less that 40 percent of the table rows on an ordered table or 7 percent of the rows in an un ordered table the query can be tuned to use an index in lieu of the full table scan The most common tuning remedy for unnecessary full table scan is adding indexes Standard B tree indexes bitmapped indexes and function based indexes can all be added into the tables in order to eliminate full table scans In some cases an unnecessary full table scan can be converted to an index scan by adding an indexes hint to the SQL statement e Share SQL Statements 5 A 17 ORACLE holds SQL statements in memory after it has parsed them so the parsing and analysis do not have to be repeated if the same statement is issued again The single shared context area in the shared buffer pool of the System Global Area SGA is shared by all the users We have to set the appropriate INIT ORA parameters for the context areas The larger the area the more statements can be retained there and the more likely statements are to be shared e Use Hints In general hints serve a dual purpose They can be used to alte
160. quations referenced in the following are from Wind field Model Technical description Please see the document for the detailed information 3 B 14 Wind model code IDL code flowchart Note The files names without any extension in them are idl procedure files pro files Coded by Dr George A Soukup Frozen Model TRACK I1 THINNER track idl VGHGEN v USG DUS suv idl nrmrayse10 15 idl FIXSHOTS15 onefix m fixshots idl Y zipcodes idl PKWINDS LLTOXY output dat REACH ZUV2ZOT i TRACK PRO l Reads in the trackfile to arrays ctg storm category zhour fix hour zmin fix min nlat latitude elon longitude cpr centre pressure rmx Rmax hdb Holland B Isflg land sea flag Mark the fix of lowest central pressure unless it coincides with landfall sflg is set to 4 Thins out the storm fixes based on the adjusted fix frequency THINNER PRO is used to accomplish this task Calculate the time in minutes for each fix from the start of the storm rack ktime 3 B 15 5 Samples the data at regular 1 hrs intervals prior to the smoothing using cubic spline interpolation Calculate fbarx Rmax f where f 0 14544 sin nlat 7 Sub samples the smoothed input data to recover the original resolution unequal in
161. r deg MA 1 9 5 52 1610 16 7574 54 7866 72 OT 1 9 5 41 1068 26 5020 48 9094 57 zipcode 3l longitude 80 1000 deg latitude 25 5900 deg ter day hour min zonal meridional total m s dir deg OT 1 9 0 42 9686 13 8472 45 1448 107 zipcode 32 longitude 80 2700 deg latitude 25 3400 deg ter day hour min zonal meridional total m s dir deg OT 1 9 0 30 4650 21 1683 37 0973 235 3 B 3 B Bypassing Storms 3 For a storm that does NOT make landfall but passes close enough to a zip code to do damage a For each affected zip code find the peak OT wind and the date time of this peak wind speed b Find the track positions closest to the time of the peak zip code wind speed and choose the fix with the lowest central pressure Use the fix information to compute the maximum marine MA exposure wind speed in the storm c Label this storm as By Passing in the header and include the MA wind speed and date time d Include the zip code information identical to that done with the storms that make landfall This method will not include storms that bypass and then make landfall further along the track or landfall and then bypass further along the track DAVID 9 03 79 06 00 UTC bypass longitude 80 5000 deg latitude 28 8000 deg ter day hour min zonal meridional total m s dir deg MA 2 4 0 23 6102 34 6558 41 9340 145 OT 2 4 0 26 7334 23 0222 35 2802 130 zipcode 31 longitude 80 1000 deg latitude 25 5900
162. r the execution plan for a SQL statement They can be used as an alternative to stored outlines to permanently change the execution plan for a SQL statement When a hint is added to a SQL statement during tuning the tuning changes will take effect e Verify Optimal Join Techniques Some queries will perform faster with nested loop joins while others may work better with Hash joins or merge star joins In general it is better to use simple join whenever it is possible e Review Sub queries Every correlated and non correlated sub query should be examined to determine if the SQL query could be rewritten as a simple table joins Having shown the goals of SQL tuning the followed section is to tune the SQL statements for the database queries Oracle Corporation has developed a lot of utilities to facilitate the SQL tuning process In this project we mainly use the SQL Trace TKPROF and the Timing Environments Parameter for SQL tuning 5 A 18 Using the Timing Environments Parameter SQL timing environments parameter is used to record total time elapsed for a SQL statement For the purpose of testing we turn on this timing parameter and run the desired SQL statement Based on the total time used we change the structure of the SQL statement and run it again Then the two results are compared to decide which statement has a better performance Example Structure 1 SQL gt select first rows Year count Cyc
163. rated in the surface friction terms in the momentum equations which depend on the and are specific for the direction of storm translation which is aligned with the Y axis The wind field grid is then rotated so that the computational y axis coincides with the actual direction of motion of the cyclone center The wind field 3 B 6 thus far constructed Fig 2 usually shows the location of peak winds to be to the right or forward edge of the right rear quadrant of the cyclone EARTH RELATIVE TOTAL WINDS M S c 5 m s N Figure 3 2 3 Horizontal distribution of mean boundary layer wind speed m s relative to the earth for a Hurricane moving northward top of page at 5 m s Horizontal coordinates are scaled by the radius of maximum wind 3 2 3 1 Wind Model Parameters Following are the input parameters to the wind field model 3 2 3 1 1 Delta P Intensity parameter This is the difference between the central minimum sea level pressure and an outer peripheral pressure assumed to be 1012 hPa Intensity change is modeled by using the observed geographic probability distribution of six hour changes of central pressure as related to the relative intensity Potential intensity takes into account the concept of the hurricane as a heat engine constrained by the input sea surface and outflow upper troposphere temperatures Intensity change is limited so as to not exceed the maximum observed change for a partic
164. re displayed in the dynamically generated web page Based on these retrieved historical data the system also utilizes several stochastic probability distributions to fit the occurrence frequency 2 Simulation Selection JSP File Edit View Favorites Tools Help Q O x A f Jose Sy Favontes Q veia O Q amp Address amp http firene cs fiu edu 8888 FDOI AHO simuSelection jsp Annual Hurricane Occurrence Altlantic Basin Only Simulation Selection Page You have selected dataset 1900 2003 There are totally 104 pairs of data in that dataset statistic information about dataset 1900 2003 Mean Variance Standard Deviation 1 88 2 05 1 43 You can conduct your simulation now The simulation will be carried out using the best distribution model Number of Year for Simulation 1000 Submit Query Figure 2 1 17 Snapshot of the second web page for AHO The snapshot of the second web page is given in Figure 2 1 17 The upper part of the Simulation Selection Page has a table which contains some statistical features about the selected data set The lower part of the Simulation Selection Page offers the user a platform to compose and submit the simulation request The user can specify his desired number of years for simulation The simulation request is submitted to the system 2 A 24 Step 3 The system conducts the simulation with respect to the probability distribution that has the best goodnes
165. red Neutral or average Multi Decadal Warm and active Cold quiet See Table 2 1 2 for a listing of Multi Decadal year ranges ENSO EL Nino La Nina see Table 2 1 3 for a listing of El Nino and La Nina years Table 2 1 2 Listing of Multi Decadal year ranges and temperature Temperature Warm Temperature Cold 1870 1902 1903 1925 1926 1970 1971 1994 1995 2003 Table 2 1 3 Listing of El Nino and La Nina years El Nino Year La Nina Year 1925 1933 1929 1938 1930 1942 1940 1944 1941 1945 1951 1948 1953 1949 1957 1950 1963 1954 2 A 4 1965 1955 1969 1956 1972 1961 1976 1964 1977 1967 1982 1970 1986 1971 1987 1973 1990 1974 1991 1975 1993 1978 1994 1988 1997 1995 1998 1999 2000 2002 2 Based on the selection of year range inputted by the user from Step 1 the system queries the database and returns data of the years within the desired year range and their associated numbers of tropical cyclones occurrences Table 2 1 4 illustrates the content of the returned data Table 2 1 4 Matrix of Number of Hurricanes Per Year Year Y Y Hurricanes Ho H 1900 1 1901 3 Yi Hi Yn H NOTE In the initial development the model considered all the tropical cyclones which include all hurricanes and tropical storms that wer
166. refore the data file needs to be processed and two attributes have to be added into the FIX object in the database as follows Name Length rmax NUMBER 4 Crossing VARCHAR2 10 Following is the data format of the new text file Storm Name 4 Year Mo Dy Time Lat Lon Wsp Pmn RMW Cat Crossg NOT NAMED 3 1903 9 9 0600 21 4 72 4 50 0 O TSt NOT NAMED 3 1903 9 9 1200 21 8 73 4 50 0 O TSt NOT NAMED 3 1903 9 9 1800 22 2 74 0 50 0 O TSt NOT NAMED 3 1903 9 10 0000 22 6 dA T 335 0 O TSt NOT NAMED 3 1903 9 10 0300 22 9 75 0 60 0 O TSt ISLAND new NOT NAMED 3 1903 9 10 0600 23 2 75 3 60 0 O TSt NOT NAMED 3 1903 9 10 1200 23 8 76 0 65 0 O Hul NOT NAMED 3 1903 9 10 1800 24 0 76 5 70 0 O Hul NOT NAMED 3 1903 9 11 0000 24 4 7649 80 0 O Hul NOT NAMED 3 1903 9 11 0600 24 9 iD 859 0 O Hu2 NOT NAMED 3 1903 9 11 1000 25 03 78 1 85 0 0 Hu2 ISLAND new NOT NAMED 3 1903 9 11 1200 25 4 78 4 85 0 0 Hu2 NOT NAMED 3 1903 9 11 1800 25 8 J9l 85 0 0 Hu2 NOT NAMED 3 1903 9 11 2200 26 1 80 0 85 0 0 Hu2 LAND new NOT NAMED 3 1903 9 12 0000 26 4 80 3 75 0 O Hul NOT NAMED 3 1903 9 12 0600 26 9 81 2 65 0 O Hul Where the RMW column represents the rmax value and the Crossg column represents the crossing value And the word new next to the Crossg column indicates the new fixes The data processing steps are si
167. respective damages as follows Ds DMs DMs C AP x D D C DM AP x D Dar AP DM C AP x D Date 0 15 Apply the pro rata structure deductible Dsijk and limits LM to each of the cells of the damage Matrix DMijx Calculate the structure loss Lsijkn net of deductible and truncate it on the upside by LMj and on the downside by Dgijx Thus a vector Lii of insured losses is generated for property k Its elements are L ijkn If Lsijkn is 2 Linijk then Laijka Linijk If Esijkn is lt 0 then let Esijkn 0 16 Repeat step 15 for C AP and ALE Here these variables are means conditional on the wind speed Generate L Lap and Lar y 17 Next to get the expected insured loss for a given wind speed w multiply each element Lijkn of the vector Lijk by its corresponding probability Pii to compute Lijknw and then sum over the N intervals Steps 15 17 can be represented by 4 A 8 18 19 20 21 22 23 Expected Structure Loss E Ls Y DM Ds Ps Y LMsPs Ds Expected Content Loss E Le Y C D P Y LMP Expected Appurtenant Loss E Lar y AP Dar Par y LMarPar Expected ALE Loss E Late gt ALE Dare Pare e LMateP ate where Lijkwn LMijx if DMijn Dig 2 Linijks and if DM ijn Dsijk lt 0 then let DMijn Dsix 0 i e replace negative values of net of deductible loss with zero The same applies
168. resses the major classes used and their functionalities gt Client This class is a virtual class It refers to the user who uses this system No need to implement it LoginCheckBean This class is for the user login authorization purpose It gets the username ID and the associated password verifies the information with data stored in Oracle Database DSSelection SimuSelection This class is used to get the user s selection of year range It then passes control to the classes that can get data from database and do the simulation GetDBean This class is used to get hurricane data from Oracle database DataEntry This class is used to hold data records CalMVSBean This class is used to calculate statistic characteristics of a data set such as mean value and standard deviation Database This class is an abstract concept It includes all the systems that can provide database operations FitDistriBean The class is used to interface with the actual math model MathModel This C class using IMSL library functions to fit distribution and generate the simulation result It communicates with the Java main application using JNI interface IMSL Library This is a statistical and mathematic functions library provided by IMSL PlotSimulation This class is used to visualize the simulation result 2 A 15 MyPlot This class gets the simulation result and draws the simulation result figure
169. rily limited to a Technical information Methods processes formulae compositions systems techniques inventions machines computer programs and research projects b Business information Insurance data customer lists pricing data and financial data 2 L agree that I shall not during or at any time after the termination of my employment with the PHRLM use for others or myself or disclose or divulge to others including future employees any confidential information or any other proprietary data of PHRLM in violation of this agreement 3 That upon the termination of my employment from PHRLM a I shall return to the project manager all documents and property of PHRLM including but not necessarily limited to drawings blueprints reports manuals correspondence computer programs business data and all other materials and all copies thereof relating in any way to PHRLM or in any way obtained by me during the course of employment I further agree that I shall not retain copies notes or abstracts of the foregoing b This agreement shall be binding upon me and my personal representatives and successors in interest and shall inure to the benefit of PHRLM its successors and assigns Signed this on 20 MM DD YYYY Project Manager or Professor Employee Name Print Signature 7 A 8 Section 8 System Hardware and Software Configurations
170. rix for contents for structure of type i based on its construction date d If the construction date d is not available another set of vulnerability matrix is used The matrix is provided by the Engineering team and consists of the simulated probabilities for various content damage ratio intervals and wind speeds The row represents a given interval n of content damage ratios and the column represent a given wind speed w The interpretation of the cells values etc is similar to the description given above for structure damage matrix Although the content damage depends indirectly on structural damage there is no stipulated functional relationship between the two matrices and their damage intervals Select the AP and ALE damage matrices accordingly The Engineering team has generated independent matrices for AP and ALE based on indirect relationships between structural damage and both ALE and AP From the insurance policy file get the property value Vig its policy limits LMijx and its deductible Dijk The limit LM is the default value of the property k default is V LM if value is not available Value is contingent on the type of policy specified and is either replacement cost or actual cash value replacement cost minus depreciation Select the damage vector for the observed wind speed Apply the damage ratio vector X to the property k of type i in zip code j For each damage interval n calculate the damage DM Vi Xx X Thus a N
171. s Figure 3 2 2 Polar coordinate system for solving equations of motion Implementation proceeds according to the following steps First based on the input parameters namely the radius of maximum winds the central pressure and the Holland B parameter radial profiles of the radial and tangential winds are calculated based on a stationary cyclone over open water to provide an envelope with which to set the size of the cyclone vortex The wind field produced by these profiles is radially symmetric Azimuthal variation is introduced thru the use of two form factors The form factors multiply the radial and tangential profiles described above and provide a factorized ansatz for both the radial and tangential storm relative wind components Each form factor contains three constant coefficients which are variationally determined in such a way that the ansatz constructed satisfies as far as its numerical degrees of freedom permit the scaled momentum equations for the storm relative polar wind components The azimuthal variable 9 has its usual mathematical meaning such that increases from left to right with the rectangular X axis aligned 180 0 and the Y axis aligned 270 90 with Y increasing in the direction of storm translation The translational motion of the storm is vectorially added to the storm relative wind components in order to obtain the earth relative wind field The translational motion of the storm is incorpo
172. s a percentage of exposure for structure type i in zip code j use E Li Eto EM Repeat steps 5 through 23 for all property types i 1 I to get the Expected Aggregate Loss and Expected loss for all property types in zip code j Sum the E ALjj across all property types i to get the Expected Aggregate Loss for all exposure in zip code j E AL gt ElaL Sum 0 to get o the variance for all exposure in zip code j Pick another zip code and repeat steps 4 through 28 to generate E AL for all zip codes Sum across the zip codes j 1 J to get the Expected Aggregate Loss for insurance company m E ALm for company m Y E AL J 1 J Sum o to get O7 the variance for insurance company m Pick another insurance company m and repeat steps 1 through 30 Sum across the insurance companies to get the Overall Expected Loss The Non Parametric Algorithm for Generating Scenario Based Expected Loss Costs In this section we develop the algorithm for estimating expected loss costs for a given scenario Typically the scenario refers to a particular hurricane with a given set of characteristics Hence both the exposure data and the wind speeds by zip code are given observed data The damage matrices as before are modeled Most of the steps in this algorithm are the same as in the prior section 1 2 3 4 Start with a particular insurance company m Next pick a residential policy exposure un
173. s of fit and the number of simulated years that is determined by the user in the previous step A series of years and their associated number of hurricane occurrences in that year are generated There is a new page for the simulation purpose in that page the simulation result is plotted to offer better visual effect The result is visualized 100 pairs of data per screen the user can use the forward button to browse more and use the backward button to go back Two different types of plots are supported bar chart and line chart Figure 2 1 18 2 1 19 illustrate respectively the snapshots of the bar plot example and line plot example 2 Simulation Result Plot JSP Aaa Ele Edt yew Favorites Tools Help r Qo x 2 JD sen Be ravortes Ama O 2 0 Jas 3 Address E http firene cs fiu edu 8888 FDOL AHO plotSimulation jsp Be Simulation Result Plot Page You want to conduct your simulation in the following way Distribution model Poisson The number of years to simulate 1000 Dataset based on 1900 2003 k value 19 9 The simulation has completely successfully The simulation result is plotted as follows Annual Hurricane Occurrence Result M ouch tee Betts simulation graph o 5 Mi 20 25 30 35 40 45 50 55 60 65 Simulation Number 20030czZ 0500 cI e gt 80 85 90 95 Figure 2 1 18 Example of bar plot of the simulation result 2 A 25 Simulation Result
174. s pro 14 6 1 6 27 shift m 6 2 0 0 8 suv pro 31 10 1 9 51 tek m 7 2 0 0 9 thinner pro 51 12 1 12 76 track pro 69 15 1 16 101 udvs pro 14 6 1 6 27 usadv m 7 5 0 4 16 usg pro 28 9 2 8 47 usnoadv m 6 6 0 4 16 vghgen pro 19 5 1 5 30 zmar2zot pro 12 6 1 5 24 6 E 3 Engineering Module Filename Source Comment Both Blank Total Site BuiltiContUtilities Validation Prog 112704 m 405 96 15 176 692 Site Built Final_ VM Plot Prog 101704 m 140 21 5 58 224 Site BuiltMatrix Weight Prog 1212005 final m 923 76 101 296 1396 Site BuiltVulnerability Fragility Plot Prog 111904 type1 m 386 22 6 85 499 Site Built Vulnerability Prog 020405 m 1108 373 14 382 1877 Montecarlo Codes Filename Source Comment Both Blank Total capacity manuf house m 54 28 7 19 108 capacity opening m 103 8 6 16 133 capacity r2w m 61 17 6 7 91 capacity roofcover m 17 4 2 7 30 capacity sheathing m 17 4 2 8 31 capacity wall m 118 20 8 21 167 capacity wall sheathing m 15 6 3 6 30 debris model input m 29 91 2 5 127 missile impact m 40 21 1 8 70 pressures m 6 3 0 7 16 r2w conn uplift m 76 4 5 5 90 r2w conn uplift hipb638 m 127 15 2 22 166 r2w conn uplift hipb644 m 127 17 2 20 166 r2w conn uplift hipe038 m 125 15 2 23 165 r2w conn uplift hipe044 m 129 18 2 20 169 redist gable m 12
175. s will be suspended after 30 minutes or other specified period of inactivity and require the password to be reentered Successful logons should display the date and time of the last logon and logoff All user logons will be recorded for future audit For detailed information please check the following documents 7 A 3 7 2 FIU SCS Computer and Networking Security Procedures Manual Draft Revised 08 23 2002 I Responsibilities and Scope of Work The role of our system administrators is to provide technical support for our diverse network and computing systems technical consulting services for faculty and researchers and education for users on the use of our systems System administrators are responsible for the day to day operation and maintenance of our systems and networking environment which include but not limited to Operating system installation configuration updates security monitoring and automation of services The systems administers goal is to provide a reliable state of the art computing environment for instructional and research use The following positions are assigned the computer and networking security responsibilities for the School of Computer Sciences computer and networking facilities 1 Associate Director for Computing Responsible for the policy and procedures established by the School of Computer Science to assure the security of employee and student information and intellectual proper
176. serID PaseWD LOGIN Wrong user name password Elvas Figure 3 3 11 Login webpage shows the inputted user ID or password is wrong 3 C 17 3 3 5 2 WSC Page If the login is successful the user can see the web page named Service Selection Page as shown in Figure 3 3 12 To view the WSC use case page from the drop down list select Wind Speed Correction and click Go button task page Microsoft Internet Explorer a x File Edit View Favorites Tools Help Kal Back gt amp A Gb search gFavortes media lt 4 Eh HO E gt Address l http firene cs fiu edu 8888 FDOItest FDOI FdoiLogin jsp Go Links Y search web gt E2 NEW Toolbar Update Service Selection Page Please choose an online service Vind Speed Correction E Done a Internet Figure 3 3 12 Service selection page for WSC Several steps need to be followed to accomplish the task of Wind Speed Correction User can select two input methods a Input from file this is the use case four output b Manual Input If user want to take the input from file 1 Click on From File radio button 2 Select the input data set from an available data set 3 Click on submit If the user wants to enter the input manually first click on Manual Input radio button Step 1 Then the users need to input the wind field data fields Input data sets from wi
177. shes a connection with the database Step 6 WindSpeedCalc jsp requests WSCCalVamphBean to query the database based on data passed from JSP Step 7 WSCCalVamphBean queries the database which returns a ResultSet to the BEAN Step 8 WSCCalVamphBean calls its findCol method Step 9 WSCCalVamphBean queries the database which returns a ResultSet to the BEAN Step 10 WindSpeedCalc jsp requests to perform Vamph calculations Step 11 WindSpeedCalc jsp requests the results for Zoo Uo Ua Va and Vamph and WSCCalVamphBean returns the required data Step 12 WindSpeedCalc jsp notifies that it is okay to close the database connection Step 13 WSCCalVamphBean closes the database connection 3 C 16 3 3 5 Implementation of WSC Currently the implementation for Use Case five WSC has been finished The demo is online at http www cs fiu edu PHRLM 3 3 5 1 Login page Users need a username and a password to access the FIU IHRC Public Hurricane Risk and Loss Model Following is a snapshot of the login web page FIU IHRC Public Hurricane Risk and Loss Model User Login Page UserID PassWD LOGIN Elvas Figure 3 3 10 Login webpage for FIU IHRC PHRLM If the username password is wrong an error message will be displayed User will be required to input the username and password again to enter FIU IHRC Public Hurricane Risk and Loss Model Relogin Page U
178. sis Time Threat Area Only Description The end user enters a range of years and the system generates the following 1 A probability distribution for SGT Storm Genesis Time 2 Genesis time of simulated hurricanes generated in Use Case One l The end user enters a year range from the following selections 1851 2003 1900 2003 1944 2003 Multi Decadal ENSO NOTE Neutral Years All non ElNino and non LaNina years are considered Neutral or average Multi Decadal Warm and active Cold quiet See Table 2 1 1 for a listing of Multi Decadal year ranges ENSO EL Nino La Nina see Table 2 2 2 for a listing of El Nino and La Nina years Table 2 2 1 Matrix of Multi Decadal year ranges and temperature Temperature Warm Temperature Cold 1870 1902 1903 1925 1926 1970 1971 1994 1995 2003 Table 2 2 2 Matrix of El Nino and La Nina years El Nino Year La Nina Year 1925 1933 1929 1938 1930 1942 1940 1944 1941 1945 1951 1948 1953 1949 1957 1950 1963 1954 2 B 3 1965 1955 1969 1956 1972 1961 1976 1964 1977 1967 1982 1970 1986 1971 1987 1973 1990 1974 1991 1975 1993 1978 1994 1988 1997 1995 1998 1999 2000 2002 2 Based on the user input from step 1 the system queries the database and the query results contain fix data for all the hurricanes The query results
179. storm tracks for simulated storms based on data obtained from Use Case II and stochastic algorithms Use Case IV Wind Field Generation Use Case IV is used to generate wind fields for storms based on the data generated in Use Case IV for the year range specified by the user Use Case V Wind Speed Correction Use Case V is used to refine open terrain wind speed produced by the hurricane wind model with respect to the actual terrain based on land use land cover Use Case VI Wind Speed Probability Use Case VI is used to calculate the probabilities of the 3s gust wind speeds affecting each of the zip codes Use Case VII Insurance Loss Module Use Case VIT is used to calculate the expected loss values 1 A 3 C Use Case Diagram ho A j TN xdi bh p A A EOS P j NC AES AnnualHurricaneOccurence p y Scientist StormTrack Scientist j Pd A 7 E E lt lt include gt __ i e gi 777 lt lt include gt gt X P DN Tem P p PE F N P di F NG Q A BR P4 H n a BS Pa WindFieldModel CJ gt s A P gt M gt h P d pi lt A lt lt include gt gt it P d sl F i K TT StormGenesis Time p Statistician N b e include WindSpeedCorrection d Y lt lt include gt gt di h gt x P Y d gt ES WindSpeecP robability DamageProcess A Statistician lt lt include gt lt lt include gt gt T LossProc
180. storms e Initial storm location motion and intensity PDFs e Storm motion and intensity change PDFs e Diagnostic output file 3 1 2 2 The storm track generator STORMGEN STORMGEN generates the stochastic tracks based on the PDFs derived by GENPDF The initial conditions may either be sampled from the initial storm location motion and intensity PDFs or taken from observed initial conditions Both these input data are created by GENPDF The model uses a 1 hour time step which requires interpolation of the 6 hour report changes used in the storm motion change and intensity PDFs Currently storm motion is persisted during 6 hour intervals and the pressure is linearly interpolated The basic flow of the model is as follows 1 If using specified initial conditions read in initial storm location date motion and intensity If using random initial conditions read in storm genesis time see Use Case for Hurricane Genesis SGT and sample initial storm location motion and intensity PDFs Add a uniform random term equal to the width of the location PDF bin size so that the storm may form anywhere within the bin 2 Sample storm parameters Rmax and Beta 3 Update storm position using current motion 3 A 4 4 If at 6 hour interval sample new motion and intensity change Pressure tendency is interpolated to one hour tendency 5 Determine if landfall or currently over land If yes decay the storm using the decay mode
181. t LATITUDE_DEG LONGITUDE_DEG MAX WINDSPEED MPS 5 A 12 MIN_PRESSURE_MB stage fix_id obsid_seq nextval After finishing the data loading all the constraints and data references will be enabled 5 5 2 New Data Loading Since rmax dat contains only the updated or supplemented information for the hurricanes stored in the database the new data loading process is different from the original data loading process Basically two tables in database need to be altered atmosevent_list and stormfix_list The updating steps are discussed as follows IL Create a temporary table oldstormfix_list by copying all the data from table stormfix list Create table oldstormfix list as select from stormfix list III Create a new data type NEWFIX which has two more attributes rmax crossing than FIX IV Replace table stormfix list by using new data type NEWFIX instead of the original data type FIX V Copy all the data in table oldstormfix list to table stormfix list The values are set as NULL for rmax and crossing insert into stormfix list fix id when t at time event id fixobj select fix id when t at time event id newfix c fixobj latitude deg c fixobj longitude deg c fixobj max _windspeed_mps c fixobj min pressure mb null c fixobj stage null null from oldstormfix list c VI Get the according fix id in table stormfix list for e
182. t are used as input in the insured loss model Damage matrices are provided for building structure contents appurtenant structures and additional living expenses A separate damage matrix is provided for each construction type But within a certain range of building ages a particular construction type will have the same damage matrix across all the zip codes in the same region The cells of the matrix provide probabilities of damage outcomes for a given wind speed The damages are specified in intervals or classes of ratios The row represents a given interval of damage ratios and the column represent a given wind speed In practice the damage probabilities are assigned to the mid point of the interval of damage ratios The probabilities of all possible damage outcomes must add up to 1 Therefore the sum of the cells in any given vector column for a wind speed add up to 100 It should be noted that both the damages and wind speeds are initially specified as a set of discrete points If needed one can interpolate to get a rough continuous function by using either some standard smoothing techniques e g by defining the jump of the distribution function and using it with a kernel function and optimal bandwidth to estimate a smooth PDF or by specifying an empirical set of ranges or intervals where each interval has an associated probability The latter method is used by the engineering component and its output is specified as a set of damage ratio i
183. t the user of the application can query the Database and get the statistic reports from the Database In this project an interface is provided for the users to select any data series from five data sets The application then retrieves data from Oracle Database uses two probability distributions to fit the selected data set and returns the fit result to users The simulation results are plotted as graphs 5 2 Data Modeling Since the data modeling is the most important part of the system s development process the characteristics of data captured during data modeling are crucial in the design of database programs and other system components The facts and rules captured during the process of data modeling are essential to assuring data integrity in an information system Data rather than processes are the most complex aspects of many modern information systems and hence play a central role in structuring system requirements An Object Relational Model is based on the traditional Oracle Relational Database and is extended to include Object Oriented concepts and structures such as abstract data types nested tables and varying arrays In this project we use the Object Oriented concept due to the following reasons 1 Object Reuse Creating Object Oriented Database objects will facilitate the reuse of the Database objects 2 Standard Adherence If multiple applications or tables use the same set of Database objects a standard must be creat
184. terrain friction velocity m s U a Actual terrain friction velocity m s Va Surface wind speed for actual terrain m s Vamph above with english units of statute miles per hour Vimph 1 Second wind gust speed miles per hour V3mph 3 Second wind gust speed miles per hour Site directory El Done Internet Figure 3 3 15 Snapshot of the result web page for WSC 3 3 5 3 Exception Handling Users may make some error inputs The JSP webpage can catch the exceptions and show the error messages Here we show some examples After user fills in the set number blank and clicks the Set button the system will check if the inputted value is an integer If no the corresponding error message will be generated and shown under this blank as shown in Figure 3 3 16 Zi Roughness Calculation Microsoft Internet Explorer Ele Edt view Favorites Tools Help Qa x ZG Osee Jernes Qe O 2 3 As Wind Field Roughness Calculation How many sets C Poseo input toa Zip Code Wind Speed m s Wind Direction Roughness length m Input From Wind Model Zip Code Zip Wind Speed Vo Surface wind speed for open terrain produced by the wind model m s Wind Direction Wd Surface wind directian Deg from North Roughness Length Zoo Roughness length m for open terrain 0 03 m Site directory Documentations FDOI Publications Demo Usecasel Demo Usecase2 Engineering Module QUA So
185. tervals based on the storm category For the landfall fix get the landfall location and time 9 Calculates the storm translation speed in m s spdmsx and bearing bearx based on the fix data 10 Smoothens translation speed and bearing clock wise angle from north on hourly grid 11 Evaluates smoothed translation speed spdms and bearing bear at fix times using Cubic spline interpolation 12 Evaluates smoothed track positions elonk nlatk and Rmax rmwk minute by minute 13 Finally saves track related quantities for use by other procedures as D 9o trackc idl bear Bearing at each fix cpr center pressure at each fix day day of each fix elon longitude of each fix elonk longitude of storm at each minute fbr f bar at each fix hdb Holland B at each fix ktime array from 0 to last minute of storm track step 1 Isflg land sea flag of each fix minz min of each fix nlat latitude of each fix nlatk latitude of storm at each minute pdf delta p of each fix rmwk R max at each minute rmx Rmax of each fix spdms translation speed of the storm at each fix ii THINNER PRO This module reduces the number of fixes depending upon the storm intensity 1 Locate landfalls sea falls and minimum pressure point 2 Classify fixes by category 3 Thin out the fixes as below i Category 0 Select fixes in 8 hour separations ii Category 1 Select fixes in 6 hour separations 111 Category 2 Select fixes in 4 hour separatio
186. the imported schema and compared then with the original data file Three main tables have been checked by this way and two of them were found correctly imported But for the third one stormfix_list there is a problem with one of the attributes Some values of that attribute are not consistent with the original data So we double checked the database and realized that the problem is due to the format of the original file After changing the program the needed data can be extracted correctly 5 8 Queries 5 8 1 Change the Query Based on the New Schema Once the new schema has been successfully migrated onto the new database server the next step is to provide the database queries based on new schema Since the original queries are based on the old schema we need to revise the original queries according to the new schema The following is the query for the old schema Select Year count 1 From select to char s when t yyyy Year from fdoifiu stormfix list s where s for event basin 1 and when t between 01 JAN 1851 and 31 DEC 2000 and s fixobj stage like H or s fixobj stage Tropical Storm group by to char s when t yyyy event id group by Year order by Year In the old schema the storm category is represented by string instead of category id For example the string Hurricane or Tropical Storm was used to record the type of tropical cyclones But in the new schema the numbered id is used to categorize th
187. the useful data and to remove the unwanted data or format symbols For Table atmosevent list which records the high level information for all storms we need to extract the following corresponding data fields from the original data file 1 Storm number 2 Begin date of that storm or hurricane 3 Type of the storm or hurricane The type of the hurricane or storm is based on a category criterion which is calculated by converting the maximum wind speed of each storm to its corresponding storm category according to some criteria 5 A 10 We use a C program to retrieve the data and then categorize the storm type based on its maximum wind speed Table stormfix_list stores the detailed information about each storm or hurricane For example it records how many days a storm lasts the exact latitude and longitude the wind speed and the central pressure at different fix point of each day We therefore need to obtain this information from the original data file A java program is developed to achieve this goal In order to make sure the extracted data consistent with the original data file we have done a lot of checking either manually or by programs 5 4 2 New Data Processing On 04 24 03 we received a new data file rmax dat which contains Rmax value for each fix and the crossing point for specific points In addition some intermediate fixes which are not in the HURDAT database have been included The
188. their associated wind probabilities per each zip code using Vulnerability Matrices to calculate the expected losses 4 1 1 Design Requirements Name Insurance Loss Model Use Case Description The traditional actuarial method in loss estimation is typically parametric and involves fitting some distribution to the number of claims typically Poisson and the amount of the losses A variety of distributions are available for fitting losses e g Lognormal Weibull Pareto gamma Burr mixture of distributions most with the preferred two parameters but a few with three parameters to be estimated There are several techniques to estimate the parameters percentile matching method of moments minimum distance minimum chi square and maximum likelihood The models are validated or accepted using both statistical and non statistical criteria If more than one model are acceptable then a ranking of the models is in order Models can be ranked and selected by using e g large maximum likelihood function value small chi square goodness of fit test statistics small Kolmogorov Smirnov test statistics large p value from chi square goodness of fit test minimum cdf MSE or LAS etc Though not necessary ranking is often done by using the same method employed for estimating parameters Once the loss distribution has been selected and its parameters estimated and validated it is rather easy to use and a variety of hypotheses can be tested For example
189. their related Julian date the first fixed time and so on The system uses the specific stochastic approaches to fit the storm genesis time based on the historical data retrieved from the Oracle database according to the user s year range selection Then the system generates a sequence of genesis time for the simulated hurricanes produced in Use Case One Annual Hurricane Occurrence Figure 2 2 16 and 2 2 17 depict the snapshot of some final result after running SGT The first 100 storm genesis time obtained was displayed in a table for both debugging and demonstrating purpose 2 B 22 3 SGT Distribution and Simulation JSP Microsoft Internet Explorer File Edit view Favorites Tools Help Qe x e JO search Shp Favortes Meda 2 a ix Las Address amp l http firene cs fu edu 8888 FDOI SGT SGTsimulation jsp Storm Genesis Time Atlantic Basin Oniy SGT Distribution and Simulation Page You have selected dataset 1900 2003 There are 195 records retrieved by your query SGT simulated value table 2631 3 3722 1195 7 4858 2952 11 772 2329 15 2485 2981 19 2875 4092 23 2722 1220 27 3565 4066 31 3134 Figure 2 2 16 Snapshot of the result page for Use Case Two 2 SGT Distribution and Simulation JSP Microsoft Internet Explorer File Edit View Favorites Tools Help Qu la 2
190. tion JDBC JDBC is a Java program that provides a way for the user to invoke SQL statements to access the database JDBC API is used to build the communication between the Java program and the database server Multiple database drivers for connecting to different databases are supported by JDBC Actually JDBC technology allows users to access virtually any tabular data source from the Java programming language It provides cross DBMS connectivity to a wide range of SQL databases 1 A 6 Through JDBC API developers can take advantage of the Java platform s Write Once Run Anywhere capabilities for industrial strength cross platform applications that require access to enterprise data With a JDBC technology enabled driver a developer can easily connect all corporate data even in a heterogeneous environment JNI Java is one of the most popular languages with strong support for web application however the math model is implemented using C for the sake of speed and the stronger functionalities supported in the IMSL library C version To bridge the gap between the Java application and the math model the JNI is employed JNI stands for Java Native Interface JNI is a standard programming interface for writing Java native methods and embedding the Java virtual machine into native applications The primary goal is binary compatibility of native method libraries across all Java virtual machine implement
191. tion collecting additional log information correlating security data and providing recommendations as directed by Group Manager II Definitions Computer Account A username and password credential used to identify an authorized user of FIU SCS computer and networking resources Unauthorized Use term used to describe when an unauthorized person utilizes computer and or networking resources restricted by FIU SCS Authentication The process of providing correct computer account credentials to obtain access to FIU SCS computer or networking resources Security incident Any unauthorized utilization of FIU SCS computer and networking resources Mission Critical Server Any computer server which provides the majority of SCS users computer services which if down would result in 8 hours of lost user productivity III Policies 1 All computer and networking resource usage on the FIU SCS network must be authenticated 2 Each computer user must be assigned one unique computer account Exceptions may be made in order to manage software hardware services but account ownership is documented 3 Critical computer systems and networking devices are to be monitored regularly to insure security is maintained 4 Root access to the primary trusted system goedel cs fiu edu is by permission of the Assoc Director for Computing only All work on the primary trusted system must be conducted via the sudo
192. tion for system dumps on your operating system This value can be modified with ALTER SYSTEM SET USER DUMP DEST newdir This is a system parameter 5 A 21 Step 2 Enable the SQL Trace Facility Enabling the SQL Trace Facility for Current Session To enable the SQL trace facility for our current session we use the following command ALTER SESSION SET SQL_TRACE TRUE Alternatively one can enable the SQL trace facility for a session by using the DBMS_SESSION SET_SQL_TRACE procedure To disable the SQL trace facility we use the following command ALTER SESSION SET SQL_TRACE FALSE The SQL trace facility is automatically disabled for the tuning session when the application disconnects from Oracle Step 3 Format Trace Files with TKPROF TKPROEF accepts as input a trace file produced by the SQL trace facility and produces a formatted output file Once the SQL trace facility has generated a number of trace files we can e Run TKPROF on each individual trace file producing a number of formatted output files one for each session e Concatenate the trace files and then run TKPROF on the result to produce a formatted output file for the entire instance TKPROF does not report COMMITs and ROLLBACKs that are recorded in the trace file The syntax for TRPROF is as follows TRPROF lt input _tracefile gt output filename Explain user password Step 4 Interpret TKPROF OutputTabular Statistics TKPRO
193. to C AP and ALE Expected Loss E L E Ls E Lc E Lar E Luz Repeat step 10 through 18 for all the wind speeds to generate a row of expected insured loss for all wind speeds Multiply the Expected Loss E Lijxw for a given wind speed by the probability of the wind speed pw Next sum over all wind speeds to get the property k Expected Loss Ww Property k Expected Loss E Lix z E Liw x Pw w Steps 7 through 20 are repeated for all dwellings of type i in zip code j to generate E Lijk for all properties k 1 K The expected losses are then summed to get the Expected Aggregate Loss for property type i in zip code j K ExpectedAggregateLoss E Lij X E Lix Variance will now need to be computed empirically since all the terms in the calculations for the expected losses are correlated We compute the variance as follows The variance for all dwellings of type i in zip code j will be o 1 K EU y El f To get the Expected Loss mean loss for structure type i in zip code j the E AL is calculated as the weighted average of the Expected Loss of all properties of type i The weight is the relative value of the structure V gt Vi or relative exposure of the structure LM X LM 4 A 9 25 26 27 28 29 30 31 gt Expected Loss for property type i in zip code j y va Via m k 1 To estimate the expected loss a
194. to database schema 00005 06 25 1851 M 1 1 SNBR 1 NOT NAMED XING 1 SSS 1 00010 06 25 0 0 0 0 0 0 0 0 285 965 70 Ox 0 0 0 0 00015 HRBTX1 00020 07 05 1851 M 1 2 SNBR 2 NOT NAMED XING 0 SSS 0 00025 07 05 0 0 0 Ox 0 0 0 0 222 976 80 0x 0 0 0 0 00030 HR 00035 07 10 1851 M 1 3 SNBR 3 NOT NAMED XING 0 SSS 0 00040 07 10 0 0 0 O 30 0 0 0 120 600 50 0 0 0 0 0 00045 TS 00050 08 16 1851 M 12 4 SNBR 4 NOT NAMED XING 1 SSS 3 00055 08 16 134 480 40 0 137 495 40 0 140 510 50 0 144 528 50 0 00060 08 17 149 546 60 0 154 565 60 0 159 585 70 0 161 604 70 0 00065 08 18 166 625 80 0 169 641 80 0 172 660 90 0 176 676 90 0 00070 08 19 180 693 90 0 184 711 70 0 189 726 60 0 194 743 60 0 5 A 7 075 080 085 090 95 100 105 110 orPOROROUNODOFOWO OYA OA O oO0oO0O0O0O0O0O0O0O0O00O0O0OO00O0O0OoOoOo o 115 8 20 199 8 21 226 8 22 250 8 23 274 0 8 24 307 8 25 340 8 27 428 0 0 0 0 08 26 378 0 0 0 H RAFL3 GA1 759 70 0 205 776 70 0 212 790 70 0 219 804 814 60 0 232 825 60 0 239 836 70 0 244 843 849 80 0 256 855 80 0 262 860 90 0 268 863 865 100 0 280 866 100 0 285 866 100 0 296 861 851 90 0 316 841 70 0 325 830 60 0 334 814 800 40 0 348 786 40 0 358 770 40 0 368 751 736 40 0 389 718 40 0 400 700 40 0 413 668 633 40 0 445 602 40 0 464 572 40 0 485 542 There are three basic types of data lines
195. ts even if there is only one statement in the loop 6 A 6 6 1 6 5 While Statements A while statement should be written as follows while condition DoSomething 6 1 6 6 Try catch Statements A try catch statement should follow this form try catch Exception e OR try catch Exception e finally 6 1 7 White Space 6 1 7 1 Blank Lines Blank lines improve readability They set off blocks of code which are in themselves logically related Two blank lines should always be used between e Logical sections of a source file e Class and interface definitions try one class interface per file to prevent this case One blank line should always be used between e Functions methods e Logical sections inside a method to improve readability Note that blank lines must be indented as they would contain a statement This makes insertion in these lines much easier 6 A 7 6 1 7 2 Inter term spacing e There should be a single space after a comma or a semicolon Example Use TestMethod a b c or TestMethod a b c Don t use TestMethod a b c e Single spaces surround operators except unary operators like increment or logical not Example Use a b Don t use a b Use for inti 0 i lt 10 i Don t use for int i20 i lt 10 i or for int i 0 i lt 10 i 6 1 8 Naming Conventions 6 1 8 1 Naming Guidelines e Use Cam
196. ty to minimize loss of staff and student productivity due to computer and networking security violations and educate staff and students on best practices to secure their critical data Consults with the School Director and SCS faculty on computer and networking security requirements and directs the development of the policy and procedure needs with the Systems and Networking Group Manager Reports to the SCS director and other University Management security violations and liaisons with law enforcement should the violation require such interaction 2 Systems and Networking Group Manager Responsible for the engineering of computer and networking security services for the School of Computer Science The Group Manager establishes day to day procedures necessary to maintain computer and networking security for the School Makes recommendations to the Assoc Director on policy and procedures and deploys commercial open source or in house developed technology to implement computer and networking security policies The Group Manager will liaison with other technology groups on campus to coordinate security efforts 3 Systems Networking Administrator Responsible for day to day monitoring of security reports and logs and responds to security alerts as indicated in the SCS Computer and Networking Security Procedures Manual The administrator reports security anomalies to Group Manager and conducts 7 A 4 security investiga
197. ular geographic region When a storm center crosses the coastline landfall the intensity change follows a pressure decay model discussed below If the storm moves back over the sea the former intensity change model is reinstated 3 B 7 3 2 3 1 2 R max Radius of Maximum Wind The radius of maximum wind is determined from a distribution of values as a function of po and latitude where po is the central minimum sea level pressure A log normal distribution is assumed for R with a mean value determined as a function of Ap and Latitude The relationship between R and Ap and latitude shows much scatter but a generalized linear model for the natural log of R r2 0 212 provides a useful estimation Ap 1012 p In R max 2 0633 0 0182Ap 0 00019008Ap 0 0007336Lat 3 Where is a normal random variable with a mean of zero and a variance of 0 169 Equation 3 describes the mean of the log normal distribution of R in nautical miles When a simulated storm is close enough to land to become a threat an R value is randomly chosen given the Ap and Latitude R is computed at each time step but the random error term is computed only once for each landfall Rmax in nautical miles is calculated as follows Rimax Ri e 3 2 3 1 3 Pressure Profile amp Holland B The symmetric pressure field p r is specified as R max ba EF P r po Ape 4 where po is the central minimum sea level
198. utes Since the storm is moving it will affect one zip code for a variable time But we initialize zuvzip for the worst case nzip Number of zip codes werzipx holds the maximum wind per each zip Calculate all time series time k is incremented in steps of kinc from zero to kmax elc longitude of the storm center at each time step nlc latitude of the storm center at each time step rmw radius of maximum wind at each time step Determine which zip codes will be affected by the storm At time k the storm can affect several zip codes in its vicinity and the affected area depends on Rmax MAP_2POINTS is used calculate the distance from the center of the storm to each of the zip codes This is done at each time step Then REACH is used to calculate the reach of the storm at that particular time step Storm reach is calculated in terms of RMW If the calculated reach is less than 12 5 that calculated value is taken as the reach Other wise 12 5 is considered as the storm reach If at least one of the zip codes is affected by the storm generate relevant portion of gridded field for current time k unow value of u at this time at each grid point vnow value of v at this time at each grid point ds 10 11 Evaluate marine windfield components at admissible zip code centroids First use LLTOXY latitude amp longitude information of the storm center and zip code centroid to calculate the x y distance between
199. ves out a listo User selects a gt System queries database f gt dataset b PIE get o pov 559 Si x d NT S E A W PA System estimates the N distribution model of System gives out a list simulated events N p User selects a N events d The selected events No lt gt Yes system triggers AHO N system queries database to get generate a set of simulated selected simulation system displays N systemsaves N system generates the SGT S generted SGT to generated SGT 1 _ the selected simulated eo M lm i x y x Figure 2 2 8 Activity Diagram for SGT 2 B 15 2 2 3 6 Sequence Diagram Similar to use case one we will give out the sequence diagrams for the three major activities in use case two which are login simulation and generate genesis time Because the login process is the same for all use cases the login process is the same as in use case one A Login Process client Client checkerBean database loginCheckBean Database login rtegisterDriver gt E getConnection lt executeQuery gt lt getStatus lt gt Figure 2 2 9 Sequence diagram for login process e Stepl The user enters its username and password in the web browser and click login button e Step 2
200. wind speeds are in units of miles per hour and should already have been corrected for roughness 3 Order according to the peak maximum sustained wind speeds of all storms affecting the zip code from low to high 4 Count the number of instances of maximum wind speeds within each wind speed band and above the mid point of each band 5 Estimate P probability of the maximum wind speed being less than mid point v as P Count V lt v N Where N Number of Storms affecting that Zip code Estimate P z lt V lt x P Count z lt V x N 6 Estimate A total of storms affecting that zip code total of years in simulation A N total of years in simulation 7 Compute the probabilities P V v 21 e e Also P V v 1 e e P coo f n E A e 8 P z lt V lt x pi apy ee 8 PC po n 1 This above equation is going to require a program however as n gets large the Poisson probabilities should start to go to 0 and therefore the sum might not be too formidable So the maximum limit of n is taken as 13 3 D 8 3 4 3 4 Class Diagram WP S8nStart_storm 8nEnd_storm nY ears S3WSArr JSP interface BSicount zip BSibuild zip array BSiwp max windspeeds BSOrder Max WindSpeeds BSiSort Max WindSpeeds DSibin decider BSICalcPVxz DSibistr Bands Figure 3 4 3 Class Diagram for WSP 3 4 3 5 Class Description WP cl
201. x1 damage vector DMijx is generated for property k This vector is associated with the observed wind speed ijknw ijnw For the observed wind speed estimate the row vector of wind conditional mean content damages where each element is the mean content damage for the given wind speed mean Cj LM X mean Cinw ratio 4 A 11 12 13 14 15 16 17 For the observed wind speed estimate the row vector of wind conditional mean AP damages where each element is the mean AP damage for the given wind speed mean AP4 LM xmeanAP ratio For the observed wind speed estimate the row vector of wind conditional mean ALE damages where each element is the mean ALE damage for the given wind speed mean ALE LM X mean ALE ratio ijkw ijnw Using the wind conditional mean structural damage DMijx and combining it with the wind conditional mean C mean APjjxy and mean ALEj calculate the deductibles Ds Dc Dap Dare on a pro rata basis to the respective damages as follows D DMs DMs C AP x D D C DM AP x D Dar AP DM C AP x D Date 0 Apply the pro rata structure deductible Dsijx and limits LM to each of the cells of the damage Matrix DM Calculate the structure loss Lj net of deductible and truncate it on the upside by LMijx and on the downside by Dsijk Thus a vector Lsijx of insured losses is generated for property k Its elements are
202. y all needed parameters and their values in a parameter file Storing the parameters in a file allows the parameters to be easily modified or reused which is the recommend method for invoking Export We create the parameter file using the DOS text editor as follows FILE dba dmp the name of the Exported dump file OWNER czhang02 we export the schema from czhang02 s account GRANTS y exports objects grants ROWS y rows of table data are exported COMPRESS y compress the exported file log dbaemp save export reports and error information to file dbaemp 5 A 14 Step 5 Invoking the Export Utility In our case we use the User mode to export the entire schema from Chengcui s account on the FDOI to HLDP As described above the parameter file method was used to invoke the export utility Execute the following command in DOS gt exp username password PARFILE params dat Import the Schema Through the above 5 steps we successfully export the entire schema from Chengcui s account The next step is to use the Import utility to read dba dmp file into the HLDP account Step 1 Verify that we have the required access privileges To use Import you must have the CREATE SESSION privilege on an Oracle database To Import tables owned by another user the IMP_FULL_DATEBASE role has to be granted to the user who will perform the export Step 2 Prepare the parameter file We specify all needed par
203. y distribution generator GENPDF This component derives the probability distribution functions PDFs from the historical record HURDAT that are subsequently used by the STORMGEN track generator The PDFs are conditional probabilities as they depend on location time of season and other parameters The PDFs are empirical in that they are obtained by discrete binning The following PDFs are derived Initial storm speed Initial storm direction Initial storm intensity pressure Change in storm speed Change in storm angle Change in storm intensity pressure and relative intensity The bin size and location of these PDFs are defined in a header file genpdf h which is used by both GENPDF and STORMGEN The bins may be linearly on nonlinearly spaced A mapping function is available which allows nonlinear mapping so that higher resolution of a particular parameter may be obtained Storm genesis is defined to occur when a storm first enters or appears within the threat area and has a minimum wind speed of 64 kt The threat area is described in Section 2 1 The HURDAT database contains a variety of storm report types R extratropical L low D depression S subtropical W wave tropical pressure reports tropical wind reports All non tropical storm reports E L D W S are excluded in the intensity PDFs Pressure reports are used whenever available If a pr

Public Hurricane Risk and Loss Model Primary Document Binder

Contents

Download Pdf Manuals

Related Search

Related Contents