USF Libraries
USF Digital Collections

A framework to develop an interactive web database for delivery of water resources field data over the internet

MISSING IMAGE

Material Information

Title:
A framework to develop an interactive web database for delivery of water resources field data over the internet
Physical Description:
Book
Language:
English
Creator:
Pujari, Swarna
Publisher:
University of South Florida
Place of Publication:
Tampa, Fla.
Publication Date:

Subjects

Subjects / Keywords:
ArcIMS
Access database
J2EE technologies
Linear correlation
XML
Web GIS
Dissertations, Academic -- Science Engineering -- Masters -- USF   ( lcsh )
Genre:
government publication (state, provincial, terriorial, dependent)   ( marcgt )
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Summary:
ABSTRACT: The objective of the research is to develop a user friendly framework for an internet GIS (Geographic Information Systems) application. The study aims at providing a map with GIS capabilities without requiring the clients (users of the website) having to install ArcGIS (a product of ESRI) software on their personal computers along with the facility to download hydrological data. It also aims at providing a methodology to synthesize data in case of data gaps. High resolution data was collected in a small watershed in west Central Florida to measure hydrologic storages and fluxes during 2001-2004. Meteorological, surface and groundwater data were collected at 5 to 10 minute intervals. A watershed database was constructed using Microsoft Access and was normalized to 3NF (third Normalization Form) for easy update. The databases available do not involve user-friendly tools to map data collection locations or to facilitate interactive delivery of data.In many fields like hydrology, it is important to provide spatial location of the data points with the data. In addition there is also a need for one platform wherein various user communities (students, planners, hydrologists etc) can be served. Thus there is a need for an online user-friendly environment, which allows for interactive delivery of data along with mapping and spatial analysis. To address this need an ArcIMS website was constructed. It includes ortho-photography of the site, which can be overlaid with the shapefile of the data locations thus giving the user reference orientation of the data locations. The user is allowed to download data in the form of text files based on the required temporal resolution and date ranges. Furthermore, users are provided with valuable data sets to parameterize or test hydrologic models applied to west-central Florida conditions.
Thesis:
Thesis (M.S.E.S.)--University of South Florida, 2005.
Bibliography:
Includes bibliographical references.
System Details:
System requirements: World Wide Web browser and PDF reader.
System Details:
Mode of access: World Wide Web.
Statement of Responsibility:
by Swarna Pujari.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains 85 pages.

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001670397
oclc - 62352595
usfldc doi - E14-SFE0001272
usfldc handle - e14.1272
System ID:
SFS0025593:00001


This item is only available as the following downloads:


Full Text

PAGE 1

A Framework To Develop An Interactive Web Database For Delivery Of Water Resources Field Data Over The Internet by Swarna Pujari A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in Engineering Department of Civil and Environmental Engineering College of Engineering University of South Florida Major Professor: Mark Ross, Ph.D. Kaushal Chari, Ph.D. Kenneth Trout, Ph.D. Paul Zandbergen, Ph.D. Date of Approval: July 19, 2005 Keywords: ArcIMS, Access Database, J2EE Technologies, Linear Correlation, Web GIS, XML Copyright 2005, Swarna Pujari

PAGE 2

DEDICATION This work is dedicated to Mrs. and Mr. Chiranjeevi Rao. Thank you for all the support that you have given me in my academic and professional pursuits.

PAGE 3

ACKNOWLEDGEMENTS I would like to thank Dr. Kenneth Trout, Dr. Ahmed Said and Nirjhar Shah for their frequent assistance in this research, and for their a dvice. I extend special gratitude for Dr. Mark Ross who had given me the oppor tunity to be part of the elite CMHAS group and for his constant guidance, commitment to quality research and importantly for the way he keeps people at ease when talking to him. Thank you to Dr. Kaushal Chari for his support and his willingness to help. I woul d like to acknowledge Dr. Paul Zandbergen for meeting with me and providing helpful in sights which paved the path for the entire work. Finally I would like to acknowledge some of friends Srinivas, Gazal and Vamsi for their help and encouragement.

PAGE 4

i TABLE OF CONTENTS LIST OF TABLES………………………………………………………………..............iv LIST OF FIGURES…………………………...…………………………………………..v ABSTRACT…………………………………………………………………………......vii CHAPTER 1. INTRODUCTION………………………………………………………....1 1.1 Background………………………………………………………...……………...1 1.2 Definition of Problem………………………..........................................................2 1.3 Objectives and Scope……………………………...................................................3 1.4 Need for an Interactive Environmen t for Hydrologic Databases over the Internet.……………………………………………………………………………4 CHAPTER 2. REVIEW OF EXISTING LITERATURE………….……………………...6 2.1 GIS on the Web……………………………………………………………………6 2.2 ArcIMS Solution for Flood Risk Lookup………………........................................9 2.3 ArcIMS Solution for Transportation……………………………………………..10 2.4 ArcIMS Solution for Crime Mapping System…...……………………………....11 2.5 ArcIMS Website for Displaying Inform ation about Emergency Facilities……...12 CHAPTER 3. MATERIALS AND METHODOLOGY……………..………………….13 3.1 ArcIMS Architecture……………………………………………………………13 3.1.1 Components Needed to Support ArcIMS………………………………….14 3.1.1.1 Web Server……………………………………………………….15 3.1.1.2 JavaVM…………………………………………………………..15 3.1.1.3 Servlet Engine……………………………………………………15 3.1.2 Arc IMS Components in the Business Logic Tier…………………………16 3.1.3 Managing an ArcIMS Site…………………………………………………16

PAGE 5

ii 3.1.3.1 Authoring Map Configuration Files……………………………...17 3.1.3.2 Publishing and Administ ering ArcIMS Services………………...17 3.1.3.3 Designing Web Pages……………………………………………17 3.1.4 The Presentation Tier, ArcIMS Client Viewer…………………………….18 3.2 Database Properties……………………………………………………………....18 3.2.1 Database Description………………………………………………………18 3.2.1.1 Data………………………………………………………………19 3.2.1.2 DataCollection………………………………………………...…20 3.2.1.3 DataCorrection…………………………………………………...22 3.2.1.4 DataType…………………………………………………………22 3.2.1.5 Relationships……………………………………………………..23 3.3 Customizing ArcIMS…………………………………………………………….25 3.3.1 ArcXML and its Structure…………………………………………………25 3.3.2 HTML Frame Structure……………………………………………………26 3.3.3 ArcIMS Param.js File...……………………………………………………28 3.4 Java Servlets and the HTML Files….....................................................................29 3.5 Synthesize Data using Multiple Imputation Techniques………………….……..30 3.5.1 Use of Simple Ratio Correlation to Impute the Missing Data…………..…31 3.5.1.1 Illustration of Simple Ratio Correlation Method………………...33 3.5.2 Use of EM Algorithm to do Multiple Imputation for Multivariate Normal Data……………………………………………………………….35 CHAPTER 4. RESULTS…………………………….…………………………………..39 4.1 Description of Toolbar…………………………………………………………...39 4.2 Home Page……………………………………………………………………….31 4.2.1 Text Area of the Home Page……………………………………………….42 4.3 Data Synthesis……………………………………………………………………45 CHAPTER 5. CONCLUSIONS ………………………………………………..……….50

PAGE 6

iii REFERENCES…………………………………………………………………………..53 APPENDICES………….……………...………………..…………………...…………..55 Appendix A: Servlet to Query and Do wnload Data from the Database…………56 Appendix B: Visual Basic Code fo r Linear Correlation Technique……………..67

PAGE 7

iv LIST OF TABLES Table 1.Table Depicting the Eight Different Ways a Dataset Containing Data for Three Variables Can Exist………………………………………………………………32 Table 2.Sample Dataset Used to Explain Met hodology of Data Imputation using Simple Ratio Correlation Method………………………………………………………..34 Table 3.Table Showing Artificially Removed Da ta Points and the Predicted Values for Well USF3…………………………………………………………………….....47 Table 4.Table Showing the Comparison Between the Removed and the Predicted Values in Case of Two Missing Values (PS43 and USF3) in a Record………………....48 Table 5.Statistics Gene rated by the F-test...………………………………………...……49

PAGE 8

v LIST OF FIGURES Figure 1. Site Location on the Map……………………………………………………….2 Figure 2. Architecture of ArcIMS 9.1……………………………………………….…...13 Figure 3. Components Outside of ArcIMS…..…………………………………………..14 Figure 4. Components of Business Logic Tier of the ArcIMS…………………………..16 Figure 5. Screenshot of the Data base Showing all the Tables…………………………...19 Figure 6. Screenshot of the Da ta Table in the Database..….……………………….……20 Figure 7. Screenshot of the DataCollection Table…….………………………………....21 Figure 8. Screenshot of the DataType Table………………………………………..…...23 Figure 9. Screenshot of the Relationships Table……………………………………..….24 Figure 10. The Hierarchy of the ArcXML Elements…………………………………….25 Figure 11. Figure Illustrating the Frame Stru cture Chosen for the ArcIMS Website……27 Figure 12. Plot for the Sensor 6 of the PS40 Well……………………………………….37 Figure 13. The Toolbar that is seen on the “Thesis” Website……...……………………39 Figure 14. Screenshot of the Home Page………………………………………………...42 Figure 15. Screenshot of the ‘Identify’ Tool in Use……..……………………….……...43 Figure 16. Screenshot of the ‘Query’ Tool in Use………………………………….……43 Figure 17. Screenshot of the Results of the ‘Query’ Tool…………………………….…44 Figure 18. Screenshot of the ‘Find’ Tool in Use…………………………..………….…44

PAGE 9

vi Figure 19. Screenshot of the Webpage th at is Displayed on Clicking Well PS41…..45

PAGE 10

vii A FRAMEWORK TO DEVELOP AN INTERACTIVE WEB DATABASE FOR DELIVERY OF WATER RESOURCES FIELD DATA OVER THE INTERNET Swarna Pujari ABSTRACT The objective of the research is to de velop a user friendl y framework for an internet GIS (Geographic Information System s) application. The study aims at providing a map with GIS capabilities without requiring th e clients (users of the website) having to install ArcGIS (a product of ESRI) software on their personal computers along with the facility to download hydrological data. It also aims at providing a methodology to synthesize data in case of data gaps. High resolution data was collected in a small watershed in west Central Florida to measure hydrologic storages and fluxes during 2001-2004. Meteorological, surface and groundwate r data were collected at 5 to 10 minute intervals. A watershed database was constructed using Microsoft Access and was normalized to 3NF (third Normalization Fo rm) for easy update. The databases available do not involve user-friendly t ools to map data collection lo cations or to facilitate interactive delivery of data. In many fields like hydrology, it is important to provide spatial location of the data points with the da ta. In addition there is also a need for one platform wherein various user communities (s tudents, planners, hydr ologists etc) can be served. Thus there is a need for an online user-friendly environment, which allows for interactive delivery of data along with mapping and spatial analysis. To address this need

PAGE 11

viii an ArcIMS website was constructed. It incl udes ortho-photography of the site, which can be overlaid with the shapefile of the data locat ions thus giving the user reference orientation of the data locations. The user is allowed to download data in the form of text files based on the required temporal resolu tion and date ranges. Furthermore, users are provided with valuable data sets to parame terize or test hydrologi c models applied to west-central Florida conditions. The website has a feature to ge nerate synthetic data using linear correlation in cases where data are una vailable for the requested period of time. The website also serves the dual role of providing accessible surface and groundwater information to the public. It can also be us ed as a desktop geographic work tool for engineers, planners and developers, to help better understand the surface water, vadose zone and groundwater interaction. Hence this website is useful not only for professional hydrologists but also fo r graduate research. Key words: ArcIMS environment, Data Synthesis, Linear correlation method

PAGE 12

1 CHAPTER 1. INTRODUCTION 1.1 Background Two local agencies, Tampa Bay Water, th e quasi-governmental water wholesaler for west-central Florida and the Southwest Fl orida Water Management District, the state agency responsible for the regulation of surface and groundwater resources for westcentral Florida, funded the University of S outh Florida for a high -resolution and a more comprehensive data collection to understand and predict va rious hydrological processes. Data was collected for this study at a sm all watershed in west-central Florida (Figure 1). The study site has a high water tabl e, a shallow surficial aquifer. The site is also characterized by different land types like “Forest” (the land th at was covered with trees) and “Grass” (the land th at was covered with grass). Meteorological, surface water and groundwat er data was collected at 5 to 10minute intervals from the summer of 2001 to the summer of 2004. Meteorological data, collected at 5-minute intervals, include rainfa ll (tipping bucket), temperature, barometric pressure, solar radiation, wi nd speed and wind direction, humidity and pan evaporation. Surface water data collected were stream flow into and out of the watershed (10-minute intervals), runoff (5-minute intervals). surficial aquifer water levels and soil moisture data was also collected.

PAGE 13

2 Figure 1.Site Location on the Map of Florida. 1.2 Definition of Problem There are many existing websites from wh ich hydrologic data from all over the United States can be downloaded and used. The researchers and decision-makers usually come across missing data when trying to download hydrologic data pertaining to a specific location. In such a s ituation, the website users woul d definitely appreciate a Geographic Information System (GIS) map that could allow them to synthesize data to replace missing or bad data. ESRI has soft ware named “ArcIMS” for making web based GIS applications. Hence, the main idea of this work is to provide an ArcIMS website capable of doing rudimentary viewing, using the functionality such as zoom, pan, legend Hillsborough Data Site Location

PAGE 14

3 controls, identification of f eatures, printing of simple maps, querying of GIS data and also enabling various da tabase interactions. Connecting GIS data with tabular and timeseries data has limitless possibilities for water resources management. The idea of a “Virtual” map offers the potential to make the hydrologic database more accessible and current to all kind of users. 1.3 Objectives and Scope The purpose of this research is to deve lop an ArcIMS website through which the delivery of both map and data over the In ternet is more accessible, current and convenient. ArcIMS allows us ers to integrate local data sources with Internet data sources for display, query and analysis in a user-friendly web browser. Creating and maintaining a GI S would be less useful if it could not be used by everyone and from any place. The justification for its existence is the value derived from its use. The decision to concentrate on a web-accessible GIS site was the need for a widely accessible system with a common user interface and no software costs for the end user. The results of this research offers th e hydrologic community a better way of looking at and using hydrologic data. This work is divided into five chapters. Chapter 1 reviews background information for having a web based GIS, coup ling hydrological database with a “Virtual” map. Chapter 2 discusses in de tail some of the past studie s of coupling map with data using in the ArcIMS environment. Chapter 3 describes general methodology used in this research. Chapter 4 describes actual results of the ArcIMS applica tion. Finally, Chapter 5

PAGE 15

4 discusses factors affecting the performance of the method and summarizes the important conclusions and recommendations for fu ture work with this application. 1.4 Need for an Interactive Environment for the Hydrological Databases over the Internet The World Wide Web (WWW) serves as a means to provide large amounts of information to planners, decision makers a nd other potential users. In the past, online databases available did not provide user-friendl y tools that facilitated, interactive delivery of data along with maps. So the project aims at developing a website with a GIS map of the hydrological parameters and also allowi ng downloading the data for the hydrological parameters available on the map. This will prove useful not only for professional hydrologists but also for people involved in academia. This will also help the users to synthetically generate data in case of missing/bad data based on a linear correlation method that uses alternate hydrological stations’ data. ArcIMS allows users to integrate local data sources with Internet data sources for display, query and analysis in a user-friendl y web browser. ArcIMS uses java based components that communicate with and access, its services. The platform used to communicate with the ArcIMS application server is Java Server Pages (Sun Developer Network website, http://java.sun.com/produc ts/jsp, last refe rred on 6/10/2005) JSP is chosen due to the following reasons Simple yet robust Platform independent Multithreaded and gives high performance

PAGE 16

5 Object oriented, producing reusab le object oriented components Network oriented To achieve this objective, a two-tier clie nt server architecture is used. On the server side, the JSP tags programmatically establish communication with an ArcIMS application server via HTTP connection. The application server on receiving the request processes it and returns an a ppropriate response based on wh ich the application can react accordingly. On the client side the java vi ewer/browser uses Java 2 Applet for user interface components such as the map display, scale bar and legend. The user is provided with great flexibility in choosing how the data is presen ted. This is achieved by a variety of complex queries that retrie ve the required data from the database. In spite of all the flexibility offered, the user is not allowed to directly manipulate the original database or its contents in any manner. To conclude, the approach used in the project provides a new perspective for viewing and working with hydrological models.

PAGE 17

6 CHAPTER 2. REVIEW OF EXISTING LITERATURE A review is provided in this section a bout the origin of the web-based GIS and ArcIMS solutions for some of the problem s in the field of science are discussed. 2.1 GIS on the Web The Internet has undergone explosive growth during the past decades. The Internet began to be used extensively by the public, business, and society in 1995 (Castells 2001). In 1995, there were about 16 million users worldwide. Internet usage in 2000 was around 400 million people. The usage is currently around 888 million (InternetWorldStats.com, 2005). Internet was envisioned in the 1960’s as a U.S. military “fai l-safe” network. The first network was based on packet switchi ng communications technique and was called ARPANET as it was developed by the Advanced Research Projects Agency (ARPA) of the U.S. Department of Defense. ARPANET was developed and promoted initi ally by computer scientists working at research institutions and universities. The goal of the development was to generate a new kind of computer-based, digital comm unications network for non-military and noncommercial reasons. ARPANET became operati onal in 1975 when it was transferred to

PAGE 18

7 the Defense Communication Agency. The numbe r of Internet host computers reached 10,000 in 1987 (Branscomb 2003). Internet functions on three basi c principles (Castells 2001): Decentralized network structure, with no single “headquarters” that control the whole system. Distributed computing capabilities throughout many nodes of the network. Redundancy of functions and control of the network to minimize risk of disruption in service. The World Wide Web is one of the major and also rapidly developing applications of the Internet. The web is the collection of all information available anywhere within the Internet system. The technological basis for the Web is Hypertext Transfer Transport Protocol (HTTP), wh ich establishes comm unication between a browser (client) and a Web server (ser ver) computer. Hypertext Markup Language (HTML) forms the basis for writing docum ents and creating links on the web. The locations of the files and other resources on remote server computers are identified by a Uniform Resource Locator (URL). Client so ftware opens the window to the web. Web browsers allow the user to move through the Web with text, graphic, video, and audio capabilities. As Internet became a widely used tool, GIS over the Internet became an important component of Enterprise GI S implementation. Some of the reasons necessitating the Internet GIS is summari zed as follows (Goddard et al, 2002)

PAGE 19

8 The ability to extend the audience of GIS to a large circle of users so that a better return on investment could be possible through the collaborative use of data, services, and applications for multi-user, multi-agency, cooperative work. The enhancement in customer satisfac tion by providing automated information and related services based on visually referenced/spatially enabled databases. Communicating dynamic/interactive spatia l information and services to the users (i.e., real-time information dissemination, integration, communication) To support daily or emergency operations of the organizations in a timely and cost effective manner. Incorporation of complex spatial /statistical analysis functionalities, modeling and decision support (ie., da ta-to-models and data-to-in terpretation and others). The Web GIS technology paved path for be tter ways to share and display spatial information over the internet. The Web offers us er interaction, so that a distant user can access, manipulate, and display geographic databases from a GIS server computer. The Xerox web site, established in 1993 (mapweb.parc.xerox.com/map), is one of the first implementations of GIS over the web. The site had used pre-created image maps for each case scenario for the users to find the pa rticular html page containing the map for display. GIS capabilities on the World Wide We b, using the middleware program called Internet Map Server (IMS), was first comm ercially launched by Environmental Systems Research Institute Inc. (ESRI). Internet Map Server was an efficient way to make realtime GIS products available to a large population at minimum cost.

PAGE 20

9 ArcIMS is Internet-based GIS launche d by ESRI to allow for authoring and publishing a wide range of GIS maps, data and applications to users over the web. ArcIMS accommodates client and server technol ogies. It extends the web to serve spatial data and limited functionalities. ArcIMS could also work with a wide range of clients including desktop and mobile devices. The Ar cIMS technology thus helps in delivery of GIS maps over the internet, integrating data from multiple sources, scaling of the system to meet the user needs. 2.2 ArcIMS Solution for Flood Risk Lookup. After the devastation of Tropical storm “Allison”, the Federal emergency Management Agency and the Harris County Flood Control District joined forces to develop new flood insurance rate maps fo r Harris county, Texas (Jacob Spenn et al, 2004). These maps were made easily available to the public via the Internet. The public had a strong desire to know their specific flood risk and make comparisons to previous versions of these special flood hazard areas. In order to provide the data to the public in a timely and widely accessible manner, the maps were published via the Internet on an easy-to-use web based ESRI ArcIMS platform. The users could view the new floodplains and their current fl ood risk by locating properties by street address, zip code, or select by area. Apart from that the users can also set parameters like which layers of the GIS ma p are they interested in viewing during the request and at what pre-determined scale, for which a map output would be delivered.

PAGE 21

10 2.3 ArcIMS Solution for Transportation An interactive ArcIMS website was devel oped for the city of Boulder, Colorado transportation department (Laurence Ferguson et al, 2004). The ArcIMS website displays both the existing and proposed transportation geographic featur es and attributes (bike, pedestrian, transit, and auto). The website served the dual role of providing accessible transportation information to the public a nd also functioning as a desktop geographic work tool for engineers, pl anners and developers, to he lp them better understand the city’s transportation system, policy plans and infrastructure needs. A major aspect of the project was th e creation of both existing and proposed enterprise transportation GIS data (bike, street, and transit) to make it compatible with the ArcIMS web based application. This was accomp lished in ArcGIS by creating a series of Oracle feature data sets and feature classes that were based on a data schema of unique ID numbers which were depende nt on travel type (auto, bi ke, pedestrian and transit). Oracle-based feature data set was used to stor e GIS transportation data that was digitized and registered to the city’s aerial photography base map. Normalized data tables were created to store attribute info rmation that pertain to proposed and existing transportation features (i.e. type, location, year built, cost). By linking this attribute information table to a transportation GIS feature, Transportation Master Plan (TMP) website visitors were able to query and then parse th e responses to generate attrac tive web pages that displayed maps, PDF documents and photographs of propos ed and existing transportation features. The significant advantage to the enterprise GIS user community that came out of the TMP was a single table, or f eature class in GIS parlance, containing all transportation routes such as alleys, streets, main ro ads, trails, bikeways and green ways.

PAGE 22

11 2.4 ArcIMS Solution for Crime Mapping System. The Fairfax Crime Mapping Website was a joint venture between GIS division and police department at Fairfax County, Virg inia (Gao et al, 2004). The purpose of the site was to Increase awareness: an interactive, us er-driven web application would allow for the public to be more informed about crim e in a particular area. It would also allow the public to be more aware about what is bei ng done to address crime and traffic issues. Increase participation: an interactive, user-driven site coul d spark public support and activity to resolve issues or problems. Increase efficiency: a majority of the in formation listed in the goal has currently being provided by the Crime Analysis Un it. While important, this hampers the unit’s overall analytic function. An Internet site is an ideal means to reduce this burden for the Crime Analysis Unit. The mapping system had three-tier archit ecture. A Microsoft SQ L server database stored the crime incident data tables. The fr ont end (user interface) comprised of a query interface and interactive mappi ng. The query interface and criteria specified by the user determined the mapping results. Thus the Fairfax crime mapping website was successful in providing general information about poli ce service for an area based on the user entering a specific address.

PAGE 23

12 2.5 ArcIMS Website for Displaying Info rmation about Emergency Facilities. TNRIS (Texas Natural Resources Information system) has designed a web interface and database design for collecting and displaying information about emergency facilities like evacuation shelte rs, including location, status, p hotographs, type of facility, etc. (Brent Porter et al, 2004). This application used ArcIMS, SQL server to obtain information about shelters for visualizati on and logistics suppor t through a web based tabular interface and a custom ArcIMS application. Thus, the application helped local, regional and state officials to make better decisions about where and when to open a shelte r in the event of an emergency such as a hurricane.

PAGE 24

13 CHAPTER 3. MATERIALS AND METHODOLOGY 3.1 ArcIMS Architecture The software was designed to ease crea tion of maps, develop Web pages that communicated with the maps, and administer a Web mapping site. The software was also designed to be distributed acr oss a network and to be scalable as the demand for maps increases. ArcIMS has a multi-tier architecture consisting of presentation, business logic, and data tiers. In addition, ArcIMS has a set of applications for managing a Web mapping site. The illustration belo w provides an overview of the ArcIMS architecture. Figure 2.Architecture of ArcIMS 9.1 Presentation Tie r ArcIMS Business Logic Tier Web Serve r ArcIMS Application Server ArcIMS Spatial Server Data Tie r Data Sources ArcIMS Management Administrato r Designer Autho r

PAGE 25

14 The ArcIMS architecture can be described as follows: The presentation tier includes the ArcIMS client viewers for accessing, viewing, and analyzing geographic data. The components in the business logic tier are used for handling requests and administering the ArcIMS site. The data tier includes all data sour ces available for use with ArcIMS. The ArcIMS site management applications provide access to components in the business logic tier for authoring maps administering ArcIMS services, and designing Web sites. 3.1.1 Components Needed to Support ArcIMS ArcIMS is an Internet product that works in a Java™ environment. For ArcIMS to run correctly, supporting components are needed that are not part of ArcIMS. This includes a Web server, JavaVM (Java Virtual Machine), and a servlet engine (see Figure 3). Thus ArcIMS website is a final outcome of all the above discussed components working along with ArcIMS. Figure 3.Components Outside of ArcIMS (ada pted from ESRI website, www.esri.com)

PAGE 26

15 3.1.1.1 Web Server: A Web server handles requests from a clie nt using Hyper Text Transfer Protocol (HTTP). The Web server forwards a request to the appropriate a pplication and sends a response back to the requesting client. A We b server is not included with ArcIMS. 3.1.1.2 JavaVM: Many of the ArcIMS components are Ja va components and require a JavaVM (ArcIMS manual), which provides the basic application programming interface (API) for running these applications. Th e JavaVM is included in either the Java Runtime Environment (JRE) or Java Developer Kit (JDK). ArcIMS requires a JRE, which is installed if it is not already on the machine. It is possible to have more than one JavaVM on a machine, and ArcIMS can be made to use an existing JavaVM as long as it is compatible. 3.1.1.3 Servlet Engine : Servlets are Java programs written in a web application such that they would query the database and return results as require d to the client viewer Therefore, ArcIMS requires a servlet engine. A servlet engine is an extension to th e JavaVM and provides support for servlets through a se rvlet API. The servlet engine plugs into a Web server and provides the link between the JavaVM and th e Web server. A serv let engine is not included with ArcIMS.

PAGE 27

16 3.1.2 ArcIMS Components in the Business Logic Tier. An ArcIMS site is made of components in the business logic tier along with data in the data tier. This secti on discusses the business logic tier, and the data tier. The ArcIMS business logic tier contains the com ponents needed to run services and process requests and responses. Figure 4.Components of Business Logic Tier of the ArcIMS The components include the Applicat ion Server Connectors, the ArcIMS Application Server, and the ArcIMS Spatia l Server. The framework also requires the Web server, JavaVM, and the servlet engine. A shown in the Figure 4 an ArcIMS request is first handled by the Web server, passed through one of the connectors, and then forwarded to the ArcIMS Application Server. The Application Server, in turn, dispatches the request to an ArcIMS Sp atial Server for processing. 3.1.3 Managing an ArcIMS Site As previously mentioned, the business l ogic and data tiers make up an ArcIMS site. To access components in the busine ss logic tier, ArcIMS provides a set of management applications. Data in the data tier, although accessed by ArcIMS, is managed using other tools such as ArcGIS. Ar cIMS management consists of three tasks: Web Server Servlet Engine Java Connecto r ArcIMS Applicatio n Server ArcIMS Spatial Server

PAGE 28

17 3.1.3.1 Authoring Map Configuration Files: (Implementation tools used are ArcIMS Author, XML editor.) The first management task is to generate map configuration files. These files are written in ArcXML and are the input to ArcIMS services. ArcIMS Auth or is one tool that can be used to create a map configuration f ile. It can access shape files and some images formats. Once the layers are established, ArcIMS Author is used to define symbology, set scale dependencies, and defi ne other mapping parameters. Map configuration files can also be created and edited using an XM L editor. The procedure for creating map configuration files for ArcMap Image Servic es is the same as for Image or Feature Services. The authoring tool, however, is Ar cMap, which is used to define symbology, set scale dependencies, and define other ma pping parameters. The output file is in a binary format and ha s “.axl” extension. 3.1.3.2 Publishing and Administer ing ArcIMS Services: (Implementation tools used is ArcIMS Administrator) A second management task is to publish and administer services. Users can add, start, stop, and delete services using ArcIMS Administrator. The input to a service is the map configuration files already described. 3.1.3.3 Designing Web Pages: (Implementation tool used is ArcIMS Designer.) The final management task is to genera te a Web site using ArcIMS Designer. ArcIMS Designer leads the user through a series of panels for selecting which services to use, which page style to use, and which ope rations and functions will be available in a

PAGE 29

18 client Web browser. Designer has three optio ns: an HTML Viewer, a customizable Java Viewer, and a non customizable Java View er. The output from Designer is a group of HTML pages. The Web pages can be modi fied and enhanced to meet specific requirements of the web application. 3.1.4 The Presentation Tier, ArcIMS Client Viewer The presentation tier consists of client viewers for us ers to access, view, and manipulate geographic data. A t ypical client viewer includes a map and some method for interacting with the map. The client viewer can be generated using ArcIMS Designer. Clients can make a request to a service residing at an ArcIMS site which then processes the request and sends back the results. 3.2 Database Properties The project involves the colle ction of rainfall, groundwater elevations, runoff, stream flow, weather and ET data. All this data is uploaded into a relational database. The Database Management Software being used to maintain this relational database is the Microsoft Access 2002-2003 version. The database is regularly updated and modified as required. 3.2.1 Database Description The database is divided into a number of tables. The advantage in doing this is the ease of expansion. After the original database creation, a new data category can be added without requiring that all exis ting applications be modifie d. Relationships and integrity

PAGE 30

19 constraints have been defined between the ta bles. Integrity constrai nts provide a way of ensuring that changes made to the database by authorized users do not result in a loss of data consistency. A screen shot showing the va rious tables being used is displayed in the Figure 5. Figure 5.Screenshot of the Data base Showing all the Tables. A brief description of the different tables used in the database follows below. 3.2.1.1 Data This is the largest table and contains mo st of the information of the database (Figure 6). This table contains rainfall, weather, runoff, ET, streamflow and well elevation data. The table has the following columns: data collection ID

PAGE 31

20 date and time data value corrected data value ( the data values for the well elevation have a correction applied so that correspond to NGVD values) Figure 6. Screenshot of the Da ta Table in the Database. This table has a many-to-one relationship with the “Data Collection” table which is described a little later. This means that every instance in the “Data” table has to be related to one particular instance in the “Data Collection” table. 3.2.1.2 DataCollection This table contains all the description re quired for the instruments in the field. The table contains the information relating the da ta collection ID with the location ID and

PAGE 32

21 type ID. The table also account s for brief description about the instruments like the depth and elevation. It contains the following columns: ID Location ID Type ID Description Index Depth Elevation Figure 7.Screenshot of the DataCollection Table.

PAGE 33

22 This table has a one-to-many relationship with the “Data”, “Site Maintenance” and “Moisture Data” tables. That is, a partic ular instance in the “D ataCollection” table may have many related instances in any of the above mentioned tables. It also has a many-to-one relationship with the “DataTyp e” table. That is, all instances in the “DataCollection” table are re lated to one particular entr y in the “DataType” table. 3.2.1.3 DataCorrection This table contains the correc tion to be applied to the we lls so that wells have an elevation that corresponds to the NGVD elev ation. This table contains the following columns: data collection ID starting date amount of correction to be applied to the well data collected from the field. (The amount of correction to be applied is calcu lated with the help of manual readings collected from the field) 3.2.1.4 DataType This table contains the relationship between the ID and the Data Type. It has only two columns: ID Type A screen shot of the “DataType” table with a few of its rows is displayed in Figure 8.

PAGE 34

23 Figure. 8 Screenshot of the DataType Table. This table has a one-to-many relationship with the “DataCollection” table. That is, one instance in the “DataType” table may ha ve more than one related instance in the “DataCollection” table. 3.2.1.5 Relationships A number of relationships have been defined between the tables. These relationships are displayed in the screen shot shown in Figure 9. All instances of data collections can be classified in to one of the categories specifi ed in the “Data Type” table. Therefore, there exists a one-to-many rela tionship between the “Data Type” and “Data Collection” table. The location specified in th e “Data Collection” table has to be one of the instances specified in the “Location” ta ble. So, there is a one-to-many relationship between the “Location” and “Data Collection” table. E ach instance of the “Data Collection” table has with it associated mu ltiple instances of “SiteMaintanance” and

PAGE 35

24 “MoistureData” tables. So there is a one-to-many relationship between the “DataCollection” table and “SiteMantaina nce” and “MoistureData” table. Integrity Constraints have also been de fined. This ensures that change s made to the database by authorized users do not result in a loss of data consistency. Figure 9. Screenshot of the Relationships Table

PAGE 36

25 3.3 Customizing ArcIMS 3.3.1 ArcXML and its Structure In ArcIMS, the type of Extensible Markup Language (XML) that is used is called ArcXML. XML allows developers to create their own set of elements that define established functionality. An XML parser is software that can read and interpret XML requests and then perform the correct acti on. Another way of authoring a map for the ArcIMS website would be by understanding th e ArcXML and using it to alter the map configuration file to meet th e website requirements. Thus some of the ArcXML elements and their hierarchy is defined below. The client requests are coded in ArcXML and parsed on the server computer. In addition, all server respon ses are returned in ArcXML format. Figure10.The Hierarchy of the ArcXML Elements. In reference to the Fi gure 10, the following is the ArcXML elements. ARCXML CONFIG REQUEST RESPONSE MARKUP MAP ENVIRONMENT PROPERTIES WORKSPACES LAYER

PAGE 37

26 All ArcXML code begins with the ARCX ML element. The sub elements of the ARCXML element are CONFIG, REQUE ST, RESPONSE, or MARKUP, and define the type of the ArcXML statement. The CONFIG element indicates that it is a map configuration file, defining the map properties and data layers. The CONFIG element has MAP and ENVIRONMENT elements as sub elements. The ENVIRONMENT element defines the la nguage being used for this viewer. The MAP element indicates that there is information on the definition of the map. The MAP element has three key sube lements: PROPERTIES, WORKSPACES, and LAYER. The PROPERTIES element describes th e general properties of the map. The WORKSPACES element lists the type and lo cation of the data layers for the map. The LAYER elements define how each layer in the map will be drawn. 3.3.2 HTML Frame Structure HTML frame provides the skeleton for th e HTML viewer. Frames are defined within a html file called view er.htm. (Refer to next page)

PAGE 38

27 Figure 11.Figure Illu strating the Frame Structure C hosen for the ArcIMS Website. The following HTML file obtains the above frame structure for the ArcIMS website. left.ht m top.htm right.htm bottom.htm

PAGE 39

28 FRAMESET elements control rows and colu mns. Each frameset contains one or more frames and each FRAME element references one HTML file. In the above HTML code the first FRAMESET element defines three rows. Below that element there are two subservient FRAME elements and one FRAMES ET element, for a total of three. Each FRAME element takes a name attribute, which specifies how Web pages may refer to that frame, and an “src” attribute that indica tes which HTML page is to fill the available space. 3.3.3 ArcIMSParam.js File ArcIMSParam.js is a parameter file in the HTML viewer provided by the ArcIMS. The ArcIMSParam.js file has a va st scope for customization. Depending on website requirements the ArcIMSParam.js file can be modified. For the “Thesis” website the file was modified to allow for the addi tion of a new tool (D ownloadData) in the toolbar and also to add hyperlink f acility for the “Well Locations” Layer. A simple description of how the new tool was added into the toolbar. Declaration of new variable called “u seDownloadData” and initializing the variable to true. Creating a button for the tool in the t oolbar. To do this the file called “toolbar.htm” was modified. Defining the functionality of the tool lik e what it has to do when the button is clicked. The functionality is defined in a file called “aimsClick.js” file.

PAGE 40

29 Providing the Actual definition of the function for the Download Data button in the file called “aimsQuery.js” file. The hyperlink facility was provided to th e “Well Location” layer in the following way: A new field was added to the attribute table of the “Well Location” layer to contain the urls of the pages of the website. In the “ArcIMSParam.js “ file the ActiveLayerIndex is set to the “Well Location “ layer. useHyperlink variable is set to true in the “ArcIMSParam.js” file. The Hyperlink layers array is set to contain the “Well Location” layer. The Hyperlink Fields array is se t to contain the name of the field in the attribute table of the “Well Location” layer which contains the urls of the web pages. The code added to all these files is pr ovided for reference in the appendix. 3.4 Java Servlets and the HTML Files Java Servlets are programs that run on a Web server and build Web pages. If the data being shown on the web pages does not change (static) then simple HTML web pages would be sufficient, but normally a ny web application would demand a dynamic web page. Building Web pages on the fly is useful (and commonly done) for a number of reasons:

PAGE 41

30 The Web page is based on data submitted by the user. The data changes frequently.(ex. Weather Reports) The Web page uses information from corpor ate databases or othe r such sources. To be a servlet, a class should extend HttpServlet and override doGet or doPost (or both), depending on whether the data is being sent by GET or by POST These methods take two arguments: an HttpServle tRequest and an HttpServletResponse. The HttpServletRequest has methods that let us ers find out about incoming information such as FORM data. The HttpServle tResponse has methods that let users specify the HTTP response line (200, 404, etc.), response header s (Content-Type, Set-Cookie, etc.), and, most importantly, lets users obtain a PrintWrite r used to send output back to the client. For the “Thesis” website the java servlets re quired to query the data base were built in an IDE called “NetBeans”. The Web Server used to execute these servlets was Apache Tomcat 4.1 version. The java servlets written to query the “cmhas database” along with the html pages is provided in the appendix. 3.5 Synthesizing Data using Multiple Imputation Techniques Imputation, the practice of 'filling in' missi ng data with plausible values, is an attractive approach to analyzing incomplete data. It apparently solves the missing-data problem at the beginning of the analysis. Mo st statistical packages do not handle missing data well. Cases with missing values are typi cally discarded, resulting in substantial loss of information and perhaps biasing the results in unpredictable ways. One of the important aspects of this thesis is to explore diffe rent ways in which the imputation can be done. An important fact or to be addressed here is the ease and

PAGE 42

31 availability of the recommended data imputation techniques. Thus the use of commercially available software was avoide d. Depending on the data distribution, thus two methods are suggested and described. The first method uses the concept of multivariate normality to impute the missing da ta while the second method is based on using the correlation of the similar land use sens ors which are expected to behave in sync. The following section describe the two methods in detail 3.5.1 Use of Simple Ratio Correlation to Impute the Missing Data This technique is less rigorous and based mo re on empiricism than statistics. It is well recognized that the for same precipitation th e moisture at same depths for the similar land use will show more strong correlation in the values than the sensors at different depths or in different land covers. For this particular reservoir site the soil moisture sensors located at PS43, USF1 and USF3 all ha ve a pasture grass as land cover and are at similar elevation. Statistical summary from the NORM freeware st atistical software (Schafer, 1997) on these wells show a high co rrelation between their moisture records. Similarly soil moisture sensors located at PS40, PS41, and PS42 have a forest land cover and hence they showed strong correlation in their recorded values Therefore the wells within the similar land cover were grouped together. This property of correlation was thus used to impute the missing values; the following paragraph desc ribes the procedure used for data imputation. Data for different sensors for each well was listed and the missing values flagged as -999.99. A VB code was written which, for th e available data finds the ratio of the values between each well, hence for each set of available data three ratio (between first

PAGE 43

32 and second, second and third and first and thir d) values were found and stored in three arrays. As the resolution of th e available data is 5 min, ther e are 288 data records for each 24 hour period. Assuming that co rrelation between the values is stronger for past 24 hours than the previous timed records onl y 288 records were stored and used for calculating the missing values. The assumption that the correlation is stronger for the past 24 hour period is based on the fact that the field data that is being us ed in this research has temporal variation. For instance consider a situation where the previous week was a total dry period and the present period is wet and trying to use data from the dry period for the correlation would result in wrong pred iction and hence the window has been set to 24 hour period. The program scans the data a nd stores all three ra tios using a moving window array where the last 288 ratios are stor ed and past records are discarded. As soon as a missing value is found the program looks a nd identifies the pattern (see Table 1) i.e. only one out of three values is missing or pair of values is missing. In an event if all three values are missing the program is incapable of imputing and it leaves the missing values as it is. The next paragraph discusses the wa y in which the data can be found missing for three variables while the next deal with the way the program handles these forms. Table1. Table Depicting the Ei ght Different Ways a Datase t Containing Data for Three Variables Can Exist First Form 0 0 0 Second Form 0 0 1 Third Form 0 1 0 Fourth Form 1 0 0 Fifth Form 1 1 0 Sixth Form 1 0 1 Seventh Form 0 1 1 Eighth Form 1 1 1

PAGE 44

33 With reference to the Table 1, the “0” re presents missing data and “1” represents good data. So as seen from the table the data can be missing in all the variables (First Form) or the data can exist for variable 1 and variable 2 and be missing for variable 3 (Fifth Form). Similarly the data can exist for all the three variables of the dataset (Eighth Form). If only one value is missing, the program fr om the two arrays of ratios finds two average values of ratio one with each known well. Using the two average values and the two available data two values for the missi ng well are computed. The average of these two computed values thus gi ves the missing value which is used to replace the missing value. However if two values are missing (say for example first and third wells are missing) the average of the ratio between the first and the second and second and the third is calculated and then multiplied by th e observed value of the second well the get the imputed values for both the well. In all the cases only the past 288 record s are used and if any missing data is observed then that record is completely skipped. Though the code written is for three wells it be easily expanded to included mo re wells if they are found to be correlated 3.5.1.1 Illustration of Simple Ratio Correlation Method As an example of how the technique works lets consider a sample data with three variables (Table 2). From the ta ble it can be seen that the da ta is missing for all the three basic patterns i.e. one value mi ssing, two values missing and all th ree values missing ( -999.99 represent missing data). Now for all the available data the ratio between the

PAGE 45

34 variable is calculated (as can be seen from the table 2) and stored in an array (NC represents that the ratios are not calculated). Table2. Sample Dataset Used to Explain Me thodology of Data Imputation using Simple Ratio Correlation Method Variable 1 Variable 2 Variable 3 Ratio1 (Var1/Var2) Ratio2 (Var2/Var3) Ratio3 (Var1/Var3) 30.73 34.17 33.93 0.90 1.01 0.91 30.74 34.17 33.93 0.90 1.01 0.91 30.73 34.16 33.93 0.90 1.01 0.91 30.73 34.17 33.93 0.90 1.01 0.91 30.74 34.16 33.93 0.90 1.01 0.91 30.74 34.17 33.93 0.90 1.01 0.91 Case 1 30.74 34.17 -999.99 0.90 NC NC 30.74 34.17 33.94 0.90 1.01 0.91 Case 2 30.74 -999.99 -999.99 NC NC NC 30.74 34.17 33.93 0.90 1.01 0.91 30.74 34.17 33.93 0.90 1.01 0.91 30.74 34.16 33.94 0.90 1.01 0.91 30.74 34.17 33.93 0.90 1.01 0.91 30.74 34.17 33.94 0.90 1.01 0.91 30.73 34.17 33.93 0.90 1.01 0.91 Case 3 -999.99 -999.99 -999.99 NC NC NC 30.73 34.17 33.94 0.90 1.01 0.91 In a record if one variable is missing, say for example Variable 3 is missing (Case 1) than the average of values recorded befo re the incomplete record under ‘Ratio2’ and ‘Ratio 3’(see Table 2) are cal culated ,for the sample data they turn out to be 1.01 and 0.91 respectively. The available values i.e. 30.74 (variable1) and 34.17 (variable 2) are then divided by these ratios to get tw o values of Variable3. The tw o values obtained are then averaged to get the final imputed value for variable3 i.e. 33.93. In case of two values missing in a record for example Variable 2 and Variable 3 (Case 2) the average of values recorded be fore the incomplete r ecord under Ratio1 and

PAGE 46

35 Ratio3 are calculated. For the sample da taset they turn out to be 0.90 and 0.91 respectively. The observed value i.e. Variab le1 is then divided by the two ratios to get imputed values for Variable2 and Variab le3 which are 34.17 and 33.93 respectively. For Case 3 when all the variables are absent in a record the values are not imputed and the record is flagged as completely missi ng. An important thing to note in the whole process is that the ratios are recorded only for the observed values and the imputed values are never used to find the ratios. 3.5.2 Use of EM Algorithm to do Multiple Im putations for Multivariate Normal Data An important component of the thesis website is to complete the missing soil moisture data by synthesizing it using statistical methods. These methods device the most expected values of the missing data. Fr ee software called NORM (Schafer, 1999) was used to generate the missing data using a multivariate approach. It can be safely assumed the reading of the soil moisture sensor along a particular probe can be easily taken to be dependent on one another, which imply that the reading of one sensor can be a good indicator of the value of the other sensors. This appro ach thus allows the soil moisture variation along a particular probe to modeled using a multivariate model. NORM assumes a multivariate normal model and suppor ts a large variety of transformations to convert non-normal data to become almost nor mal. Even though the software is designed for multivariate normal data, Schafer (1997) has demonstrated that it makes sense to use normal model for multiple imputation (MI) even for non-normal data, as the process is quite forgiving for reasonable departures from normalities.

PAGE 47

36 The soil moisture data for each individual sensor obtained from the field shows a partial log-normal behavior and, using the above rational, can be modeled using the normal MI model. The data for soil moisture is transformed usi ng the logit model (as supported in NORM); this not only allows the data to become almost normal but also constraints it within pre-specified physical bounds (if any). As the soil moisture data can never be lie outside the range of 0-40% in this area these valu es were defined as the range for the ‘LOGIT model’. Figure 12 shows the da ta for well PS40 (sensor #6) before and after the transformation. The Fi gure 12 clearly shows the cha nge in the behavior from non-normal to almost normal, increasing the confidence in the normal model. Similar results were obtained for diffe rent wells and sensors. The following paragraphs describe the basic steps followed in using NORM and the statistical tools which the software uses in each step (for more details the soft ware Help manual should be referred). From the database the soil moisture data for all the eight sensors for each well was obtained as a single text file, with only th e eight variables i.e. the eight sensor values. The missing values were flagged by ‘-999.99’ so that the software can distinguish between the observed and the missing values. Each column of the data file is associated with a variable and each row with data valu es. After the data is imported, an appropriate transformation is applied, for this par ticular case, as di scussed above LOGIT transformation with pre-specifi ed ranges is used. As the da ta is not truly normal the rounding was reduced to tenth of a decimal place so as to take care of some of the error that may arise due to this non-normality. The data is then analyzed using the EM (Expectation Maximization) algorithm which is a two step process of determining most likely values of parameters of the multivariate normal model; Parameter such as means,

PAGE 48

37 variances, and covariance are calculated for the entire dataset using values from the cases that are partially incomplete. Running the EM algorithm helps in estimating parameters for the multivariate dataset which comes out to be useful in the data augmentation step where the missing values are computed. Figure 12. Plot for the Sensor 6 of the PS40 well Following the EM algorithm the parameters are then used for another step called Data Augmentation (DA). DA is an iterativ e simulation technique, a special kind of Markov chain Monte Carlo (MCMC). In DA ther e are three types of quantities: observed data, missing data, and parameters. The missing data and parameters are unknown. DA alternately performs the following steps: I-step: Impute the missing data by drawing them from their conditional distribution given the observed da ta and assumed values for the parameters P-step: Simulate new values for the parameters by drawing them from a Bayesian posterior distribution give n the observed data and the most recently imputed values for the missing data.

PAGE 49

38 Alternating between these tw o steps sets up a Markov ch ain that converges to a stationary distribution, the joint distribution of the missi ng data and parameters given the observed data. DA bears a strong resemblance to the EM algorithm, and may be regarded as a stochastic version of EM. It is usef ul for multiple imputation of missing data. By running DA for a large number of cycles, and st oring the results of a few I-steps along the way (with enough cycles in between to en sure independence), one obtains proper multiple imputations of the missing data. The data is imputed from the software a nd is automatically transformed back and stored in the separate file which is a comp lete data set including the observed and the missing data (replaced with new imputed values) The data files obtained are again combined with the date and time fields and are imported back into database. This data can then be accessed using the synthesized data link from the website

PAGE 50

39 CHAPTER 4. RESULTS 4.1 Description of the Toolbar The toolbar is the leftmost frame of the first page of the “Thesis” website. The toolbar is illustrated in Figure 13. Figure 13. The Toolbar that is seen on the “Thesis” Website. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

PAGE 51

40 The buttons on the toolbar are: 1 Toggle between Legend and Overview map Button: To toggle between the overview map and the Legend. 2 Toggle Button: To toggle the overview map 3 ZoomIn Button: This bu tton zooms in to the map 4 ZoomOut Button: This button zooms out to the map 5 ZoomToFullExtent: This button z ooms to full extent of the map 6 ZoomToActiveLayer: This button zooms to full extent of layer that is set active 7 BackToLastExtent: This button zooms in or out depending on what was the previous operation on the map 8 Pan: This button pans the data in the map by dragging the display in any direction with the mouse 9 PanToNorth: This button pans the data in the map in the North direction 10 PanToSouth: This button pans the da ta in the map in the South direction 11 PanToEast: This button pans the data in the map in the East direction 12 PanToWest: This button pans the data in the map in the West direction 13 Hyperlink: This button gets some information a bout the data point (here the well selected) and also leads to the us er interface where the clients can query the database for information about that data point 14 Identify: This button gets the attr ibute information about the geographic feature or the data point selected

PAGE 52

41 15 Query: The user can build spatial queri es on the data in the map using the Query tool 16 Find: This tool finds a geographical feature or a da ta point with the given attribute 17 DownloadData: This tool when used provides the user with a few links which would allow them to download some of the useful data of the CMHAS database 18 Measure: This tool measures distances on the map 19 SetUnits: This tool sets the units for the map 20 Buffer: This tool creates buffer pol ygon to a specified distance around the input feature. 21 SelectByRectangle: This tool selects the features on the map by constructing a rectangle that intersects the features you wanted to select 22 SelectByLine/Polygon: The functionali ty of this tool is same as the previous tool except that either a line or a polygon should be constructed to intersect the features that needed to be selected 23 ClearSelection: This tool clears the selection that has already been made 24 Print: This tool prints the map 4.2 Home Page The home page of the “Thesis” website co mprises of the actual map of the project site in the center, the legend to the utmost right and the toolbar in the leftmost part of the webpage (Figure 14).

PAGE 53

42 Figure 14.Screenshot of the Home Page. 4.2.1 Text Area of the Home Page: The results of the tools li ke Identify, Query, Find, DownloadData are displayed in the text area of the home page. The results of each of these tools are depicted in the screenshots which follow. The Figure 15 below shows the results of the “identify” tool when clicked on the data point called “PS41”.

PAGE 54

43 Figure 15.Screenshot of ‘Identify’ Tool in Use The Figure 16 below shows how the query is bu ilt using the “Query” tool to select the data points on the map which are located in “Grass”. The results are displayed in Figure 17. Figure 16.Screenshot of the ‘Query’ Tool in Use

PAGE 55

44 Figure 17.Screenshot of the Results of the ‘Query’ Tool. The Figure 18 below shows the use of the “Find” tool to search for the given attribute among the layers on the map. Figure 18. Screenshot of the ‘Find’ Tool in Use

PAGE 56

45 The Figure 19 below has a screenshot of the webpage that is displayed when you click on a particular well. Figure 19. Screenshot of the Webpage th at is Displayed on Clicking Well PS41 4.3 Data Synthesis The soil moisture data for the reservoir was analyzed first for normality and though it generally followed a log normal trend it was not perfectly log normal and for several sensors the trend deviated significantl y. As part of this research the data was analyzed using two methods discussed earlier in the Methods and Methodologies section. The data imputed using the statistical soft ware did not reproduce the smooth natural variation of the system becau se the statistical method gene rated random numbers to fit

PAGE 57

46 the mean matrix and the variance covariance matrix. Hence the approach of finding the correlation between the sensors was used to do the imputation of missing data. The wells in the grassland and forest were segregated as two different groups and the VB code written (attached in the appendix) was used to synthesize the data. To test the accuracy of the method some of the known data was artificially removed and marked as missing to allow the comparison with th e predicted and the know values. Table 3 summarizes the results obtained when one of the wells in the data set was removed, it can be clearly seen that the predicted and the ac tual values vary by less than 1%. Similarly data values from two wells were removed and compared against the predicted values again the correlation seem to give a good re sult with the average difference of 1%. Thus, the confidence on the scheme is incr eased. The same method was used to impute the data at all the other well records as well as water table elevation records.

PAGE 58

47 Table 3. Table Showing Artificially Removed Da ta Points and the Predicted Values for Well USF3 (all Data is Percent Soil Water Content for the Second Sensor from Land Surface of Each Well) To quantify the results produced by the linear correlation method F-test (MS Excel) has been performed on the original data and the predic ted data. The results of the f-test shows that there was no significant di fference between the original data and the predicted data (see Table 4). The Table 5 s hows the statistics gene rated by the F-test based on the data in Table3. PS43 USF1 USF3 Data Removed Data Predicted 30.87 34.26 -999.99 34.04 34.00878 30.98 34.33 -999.99 34.12 34.10642 31.02 34.34 -999.99 34.15 34.13552 31.03 34.35 -999.99 34.16 34.14809 31.04 34.36 -999.99 34.17 34.16063 31.04 34.37 -999.99 34.16 34.16755 31.05 34.36 -999.99 34.17 34.16996 31.05 34.37 -999.99 34.18 34.17734 31.09 34.41 -999.99 34.24 34.22188 31.18 34.47 -999.99 34.33 34.30412 31.21 34.49 -999.99 34.36 34.33404 31.21 34.50 -999.99 34.39 34.34321 31.22 34.51 -999.99 34.39 34.35754 31.23 34.51 -999.99 34.39 34.36562 31.23 34.52 -999.99 34.39 34.37304 31.23 34.53 -999.99 34.40 34.38012 31.23 34.53 -999.99 34.41 34.38204 31.23 34.53 -999.99 34.42 34.38355 31.24 34.53 -999.99 34.42 34.39048 31.23 34.54 -999.99 34.43 34.39142 31.24 34.54 -999.99 34.43 34.39847 31.24 34.54 -999.99 34.43 34.39999 31.25 34.54 -999.99 34.44 34.40668 31.24 34.54 -999.99 34.45 34.40238

PAGE 59

48 Table 4. Table Showing the Comparison between the Removed and the Predicted Values in Case of Two Missing Values (PS43 and USF3 ) in a Record (all Data is Percent Soil Water Content for the Second Sensor from Land Surface of Each Well) PS43 USF1 USF3 Values Removed Predicted Values PS43 USF3 PS43 USF3 -999.99 35.31 -999.99 32.45 35.62 32.47 35.56 -999.99 35.31 -999.99 32.45 35.62 32.47 35.56 -999.99 35.31 -999.99 32.44 35.62 32.47 35.56 -999.99 35.31 -999.99 32.45 35.62 32.47 35.56 -999.99 35.31 -999.99 32.44 35.62 32.47 35.56 -999.99 35.30 -999.99 32.44 35.62 32.47 35.55 -999.99 35.30 -999.99 32.44 35.62 32.47 35.55 -999.99 35.30 -999.99 32.45 35.62 32.47 35.55 -999.99 35.29 -999.99 32.44 35.62 32.46 35.54 -999.99 35.30 -999.99 32.44 35.62 32.47 35.55 -999.99 35.30 -999.99 32.45 35.62 32.47 35.55 -999.99 35.30 -999.99 32.44 35.62 32.47 35.55 -999.99 35.30 -999.99 32.44 35.62 32.47 35.55 -999.99 35.29 -999.99 32.44 35.61 32.46 35.54 -999.99 35.30 -999.99 32.45 35.62 32.47 35.55 -999.99 35.30 -999.99 32.44 35.62 32.47 35.55 -999.99 35.29 -999.99 32.45 35.62 32.46 35.54 -999.99 35.30 -999.99 32.45 35.61 32.47 35.55 -999.99 35.30 -999.99 32.45 35.61 32.47 35.55 -999.99 35.30 -999.99 32.45 35.59 32.47 35.55 -999.99 35.29 -999.99 32.44 35.60 32.46 35.54 -999.99 35.30 -999.99 32.45 35.60 32.47 35.55 -999.99 35.29 -999.99 32.44 35.59 32.46 35.54 -999.99 35.30 -999.99 32.45 35.59 32.47 35.55 -999.99 35.30 -999.99 32.45 35.60 32.47 35.55 -999.99 35.30 -999.99 32.44 35.59 32.47 35.55 -999.99 35.29 -999.99 32.45 35.60 32.46 35.54 -999.99 35.29 -999.99 32.44 35.59 32.46 35.54

PAGE 60

49 Table 5. Statistics Generated by the F-test. F-Test Two-Sample for Variances Variable 1 Variable 2 Mean 34.32405 34.29764038 Variance 0.017325 0.014716825 Observations 26 26 df 25 25 F 1.177245 P(F<=f) one-tail 0.34323 F Critical one-tail 1.955447 The table above shows the result of the Ftest done for the data values given in Table 3. The result further proves that though the method is more or less empirical it is really useful for the data imput ation especially for the wells located at the reservoir site.

PAGE 61

50 CHAPTER 5. CONCLUSIONS In this research a framework for presen ting the water resources field database online has been developed. The “Thesis” websit e is a unique way of presenting data as it involves an integration of th ree different methodologies, na mely providing a map with GIS capabilities without requiri ng the clients (users of the website) having to install ArcGIS software on their personal comput ers, providing a comprehensive watershed database simultaneously maintaining the s ecurity of the database, and developing a methodology to synthesize data in case of data gaps. The importance of the work can be briefly stated as follows: The map having comprehensive information of the project site is provided in order to fill the need of allowing sp atial analysis and mapping along with downloading data for a particular data point. The users of the website are not required to have ArcMap installed on their PC’s. This means that providing an ArcIMS cu stomized map allows the users to enjoy the GIS functionality while saving them from buying a GIS license. Thus, having an ArcIMS map is quite economical. The website makes it easy to access the watershed database by providing an easyto-use and understand user interface, b ecause many people are not comfortable in

PAGE 62

51 querying the data from the database as such. However useful a database is, it is less useful if users cannot get information out of it ea sily. A web application is like a “ Cherry on the Sundae” providing access to the knowledge locked inside the database. Very few sites provide comprehensive dataset in the field of hydrology. This makes the website a thoroughly useful tool for researchers in multi-disciplinary water resource sciences such as So il Physics, hydrological modeling. The graduate students worldwide can use data for their research. The website, besides providing all the above discussed uses, maintains and preserves the database integrity. Providing the users with the capability of filling data gaps allows the water resource scientists to get a complete da taset for the period queried which is again an important requirement for many peopl e in the field of Water Resources to have a complete record for their research. In the course of creating the “Thesis” web application many issues have been encountered. Some to be consider ed in future studies include. Planning is the key to good application development. Before the web application is developed it is always necessary to ta lk out the requirements needed form the website which would save lot of “reworki ng” from efficient user friendly web application.

PAGE 63

52 A database should be recognized as not just a repository of data but as a mutable business process which constantly needs to be reassessed to determine if it still meets the needs. Planning for web application development is equal (if not greater) in importance to coding. While mediocre code can be adjusted to make it run better, web application not developed to meet the need s of the users requires that the whole system be redesigned. By keeping these thoughts and ideas in mind an ArcIMS web application can be implemented effectively while helping to impr ove efficiency and productivity at a utility. For future study I make the following recommendations: Upgrading the database to SQL server or Oracle would increase the security of the database and the users of the website can be provided with facility to make their own queries. All the sites of CMHAS can be integr ated together as a single system. Use more sophisticated software like SAS and methods like the propensity score method (which is still in experimental st age) to come up with the missing values to fill data gaps.

PAGE 64

53 REFERENCES Branscomb,L.M. 2003. Science in 2006, re vised. American Scientist 91, p. 250. Castells, M. 2001. The internet galaxy: Reflec tions on the Internet, business, and society. Oxford University Press, Oxford, U.K. (hinted in internet chapter.) Cornelis, K., P.Guido, D.Salvucci, and D.Entekhabi. 1999. Groundwater-surface Water interaction and the climatic spatial patterns of hillslope hydrological response. Hydrology and Earth System Sciences 3, (Sep):375-384. de Louw, P. G. B., J. L. va n der Meij and, R. J. Stuurman. 2000. Experiences with high frequency ground-water level monitoring; Fourth USA/CIS joint conference on Hydrologic issues of the 21st century; ecology, envir onment and human health. Hydrological Science and Technology 16, no. 1-4:13-23. Gao,C. and,Dave Cook. 2004. Development of a Crime Mapping System using .net and ArcIMS. ESRI User Conference. Godddard, S., S.Zhang, D.Lytle and, S. Anthony. 2002 A Software Architecture for Distributed Geospatial Decision Support System, available at www.diggov.org/libr ary/pdf/goddard.pdf Jacob, S., W.Meyer, J.S.Grounds and, T.Buehlman. 2004. ArcIMS Solutions for Harris County Residents’ Flood Risk Look-up. ESRI User Conference. Laurence, F. 2004. City of Boulder, Colo rado, Transportation Information System: Customized ArcIMS Website. ESRI User Conference. Porter,B. and, C Williams. 2004. Dynamic Additi on and Updates to Spatial Data Through the Internet. ESRI User Conference Rubin, D.B. (1996) Multiple imputation after 18+ years (with discussion). Journal of the American Statistical Association 91, 473-489. Sentek. 2004, (September 15, 2004); http://www.sentek.com.au/home/ Schafer, J.L. (1999) NORM: Multiple imputa tion of incomplete multivariate data under a normal model, version 2. Software for Windows 95/98/NT, available from http://www.stat.psu.edu/~jls/misoftwa.html.

PAGE 65

54 Schafer, J.L. (1997) Analysis of Incomplete Multivariate Data Chapman & Hall, London.

PAGE 66

55 APPENDICES

PAGE 67

56 Appendix A: Servlet to Query and Do wnload Data from the Database. This is a sample servlet to query the database for soil moisture data and this can be taken as a guideline for writing servle ts to query data for other hydrological parameters from the database. import java.sql.*; import java.util.*; import java.math.*; import java.io.*; import java.net.*; import java.text.*; import javax.servlet.ServletConfig; import javax.servlet.ServletException; import javax.servlet.http.HttpServlet; import javax.servlet.http.HttpServletRequest; import javax.servlet.http.HttpServletResponse; import java.lang.*; /** * @author nshah5 @version */ public class NewWellServlet extends HttpServlet { /** Initializes the servlet. */ /** Destroys the servlet. */ /** Processes requests for both HTTP GET and POST methods. @param request servlet request @param response servlet response */ protected void processRequest(HttpSe rvletRequest request, HttpServletResponse response)

PAGE 68

57 Appendix A: (Continued) throws ServletException, IOException { response.setContentType("text/html"); PrintWriter out = response.getWriter(); ResultSet recResults = null; Statement STrecResults = null; String wellName=request.getParameter("wellID"); String position=request.get Parameter("wellNumber"); String D1=request.getParameter("Day1"); String M1=request.getParameter("Month1"); String Y1=request.getParameter("Year1"); String D2=request.getParameter("Day2"); String M2=request.getParameter("Month2"); String Y2=request.getParameter("Year2"); String interval=re quest.getParameter("Interval"); ResultSetMetaData meta; String [] ColumnName =null; int DisColumns = 0; int unDisCol = 0; String temp; String [] tempColumnName; String sLine = ""; String sQuery=""; int d1val=0; int d2val=0; int m1val1=0; int m1val=0; int m2val1=0; int m2val=0; int y1val1=0; int y1val=0; int y2val1=0; int y2val=0; int dataID=0; if (wellName.equals IgnoreCase("PS40")) { if (position.equals IgnoreCase("first")) dataID=36; else if (position.equals IgnoreCase("second")) dataID=37; else

PAGE 69

58 Appendix A: (Continued) if (position.equalsIgnoreCase("third")) dataID=38; else if (position.equals IgnoreCase("fourth")) dataID=39; else if (position.equa lsIgnoreCase("fifth")) dataID=40; else if (position.equa lsIgnoreCase("sixth")) dataID=41; else if (position.equals IgnoreCase("seventh")) dataID=42; else if (position.equa lsIgnoreCase("eigth")) dataID=43; } else if (wellName.equa lsIgnoreCase("PS41")) { if (position.equals IgnoreCase("first")) dataID=26; else if (position.equals IgnoreCase( "second")) dataID=27; else if (position.equa lsIgnoreCase("third")) dataID=28; else if (position.equals IgnoreCase("fourth")) dataID=29; else if (position.equa lsIgnoreCase("fifth")) dataID=30; else if (position.equa lsIgnoreCase("sixth")) dataID=32; else if (position.equals IgnoreCase("seventh")) dataID=33; else

PAGE 70

59 Appendix A: (Continued) if (position. equalsIgnoreCas e("eigth")) dataID=34; } else if (wellName.equals IgnoreCase("PS42")) { if (position.equals IgnoreCase("first")) dataID=18; else if (position.equals IgnoreCase("second")) dataID=19; else if (position.equa lsIgnoreCase("third")) dataID=20; else if (position.equals IgnoreCase("fourth")) dataID=21; else if (position.equa lsIgnoreCase("fifth")) dataID=22; else if (position.equa lsIgnoreCase("sixth")) dataID=23; else if (position.equals IgnoreCase("seventh")) dataID=24; else if (position.equa lsIgnoreCase("eigth")) dataID=25; } else if (wellName.equals IgnoreCase("PS43")) { if (position.equals IgnoreCase("first")) dataID=10; else if (position.equals IgnoreCase( "second")) dataID=11; else if (position.equa lsIgnoreCase("third")) dataID=12;

PAGE 71

60 Appendix A: (Continued) else if (position.equals IgnoreCase("fourth")) dataID=13; else if (position.equa lsIgnoreCase("fifth")) dataID=14; else if (position.equa lsIgnoreCase("sixth")) dataID=15; else if (position.equals IgnoreCase("seventh")) dataID=16; else if (position.equa lsIgnoreCase("eigth")) dataID=17; } else if (wellName.equalsIgnoreCase("USF1")) { if (position.equals IgnoreCase("first")) dataID=45; else if (position.equals IgnoreCase( "second")) dataID=46; else if (position.equa lsIgnoreCase("third")) dataID=47; else if (position.equals IgnoreCase("fourth")) dataID=48; else if (position.equa lsIgnoreCase("fifth")) dataID=49; else if (position.equa lsIgnoreCase("sixth")) dataID=50; else if (position.equals IgnoreCase("seventh")) dataID=51; else if (position.equa lsIgnoreCase("eigth")) dataID=52;

PAGE 72

61 Appendix A: (Continued) } else if (wellName.equalsIgnoreCase("USF3")) { if (position.equals IgnoreCase("first")) dataID=54; else if (position.equals IgnoreCase("second")) dataID=55; else if (position.equa lsIgnoreCase("third")) dataID=56; else if (position.equals IgnoreCase("fourth")) dataID=57; else if (position.equa lsIgnoreCase("fifth")) dataID=58; else if (position.equa lsIgnoreCase("sixth")) dataID=59; else if (position.equals IgnoreCase("seventh")) dataID=60; else if (position.equa lsIgnoreCase("eigth")) dataID=61; } if (interval.equalsIgnoreCa se("5 Minute Interval")) sQuery = "select from Mo istureData where Moisture Data.DataCollectionID=? AND MoistureData.DateTime>? AND MoistureData.DateTime? AND HourlyMoistureData.DateTime? AND DailyMoistureData.DateTime
PAGE 73

62 Appendix A: (Continued) try{ d1val=Integer.parseInt(D1); d2val=Integer.parseInt(D2); m1val1=Integer.parseInt(M1); m1val=m1val1-1; m2val1=Integer.parseInt(M2); m2val=m2val1-1; y1val1=Integer.parseInt(Y1); y1val=y1val1-1900; y2val1=Integer.parseInt(Y2); y2val=y2val1-1900; } catch(NumberFormatException numberException){ numberException.printStackTrace(); System.exit(1); } Connection con=null; java.sql.Timestamp startDate; java.sql.Timestamp stopDate; PreparedStatement ps1 = null; //Mark this end java.util.Date myDate = new java.util.Date(); SimpleDateFormat df = new SimpleDateFormat("MM/dd/yyyy"); String path = "c:/temp"; String fileName = "temp.TXT"; File file= null; FileOutputStream fo = null; PrintStream ps = null; try { file = new File(path,fileName); if(file.exists()) {

PAGE 74

63 Appendix A: (Continued) file.delete(); file.createNewFile(); } fo = new FileOutputStream(file); ps = new PrintStream(fo); }catch(IOException exp){ System.out.println("IO Exceptio n: +exp.toString() ); } //Connecting to the Database try{ Class.forName("sun.jdbc.odbc.JdbcOdbcDriver"); con=DriverManager.getConnection("j dbc:odbc:ThesisDatabase","",""); } catch (SQLException sqlException) { sqlException.printStackTrace(); System.exit(1); } catch(ClassNotFoundExcep tion classNotFound) { classNotFound.printStackTrace(); System.exit(1); } //Mark this start try { startDate= new java.sql.Timestamp(y1val,m1val,d1val,0,0,0,0); stopDate = new java.sql.Timestamp(y2val,m2val,d2val,0,0,0,0); ps1 = con.prepareStatement(sQuery); ps1.setInt(1,dataID); ps1.setTimestamp(2, startDate); ps1.setTimestamp(3, stopDate); } catch(IllegalArgumentException illegalArgument){

PAGE 75

64 Appendix A: (Continued) illegalArgument.printStackTrace(); System.exit(1); } catch (SQLException sqlException) { sqlException.printStackTrace(); System.exit(1); } //Mark this end int columns =0; try { recResults = ps1.executeQuery(); meta = recResults.getMetaData(); columns = meta.getColumnCount();columns = meta.getColumnCount(); tempColumnName = new String[columns]; if(recResults.ne xt()) { //if_1 for(int n=0;n
PAGE 76

65 Appendix A: (Continued) sLine = sLine + "'"+tempColName+"'" + "; } } ps.println(sLine); }while(recResults.next()); } //end if_1 } catch (SQLException sqlException) { sqlException.printStackTrace(); System.exit(1); } finally{ try{ recResults.close(); recResults = null; } catch(Exception exception) { exception.printStackTrace(); System.exit(1); } } out.println(""); out.println(""); out.println("Servlet"); out.println(""); out.println(""); out.println(""); out.println(""); out.println(""); } /** Handles the HTTP GET method.

PAGE 77

66 Appendix A: (Continued) @param request servlet request @param response servlet response */ protected void doGet(HttpServletRequest request, HttpServle tResponse response) throws ServletException, IOException { System.out.println("Inside do get method"); processRequest(request, response); } /** Handles the HTTP POST method. @param request servlet request @param response servlet response */ }

PAGE 78

67 Appendix B: Visual Basic Code for Linear Correlation Technique This is the visual basic code written for synthesizing data using the linear correlation technique. 'Code for imputation of da ta using linear correlation 'Initializing variable Option Base 1 'Initializing counter for number of ratios recorded Dim cnt12 As Long, cnt13 As Long, cnt23 As Long 'Dynmaic Arrays to stor e ratio for each pair Dim r12() As Single, r13() As Single, r23() As Single Dim t1 As Single, t2 As Single, t3 As Single Private Sub Command1_Click() 'Opening dialog box to select the file with missing data cmddlg.ShowOpen r = cmddlg.FileName Open r For Input As #1 'Opeining dialog box to specify the file to save the imputed data cmddlg.ShowSave s = cmddlg.FileName Open s For Output As #2 top: 'Scanning the whole data file Do Until EOF(1) 'Reading data from the file Input #1, t1, t2, t3 'Identifying the missing patterns and sending the control accordingly If t1 = -999.99 And t2 <> -999.99 And t3 <> -999.99 Then GoTo missing_1 Pattern 011 If t2 = -999.99 And t1 <> -999.99 And t3 <> -999.99 Then GoTo missing_2 Pattern 101 If t3 = -999.99 And t1 <> -999.99 And t2 <> -999.99 Then GoTo missing_3 Pattern 110 If t1 = -999.99 And t2 = -999.99 And t3 <> -999.99 Then GoTo missing_12 Pattern 001 If t1 = -999.99 And t3 = -999.99 And t2 <> -999.99 Then GoTo missing_13 Pattern 010 If t3 = -999.99 And t2 = -999.99 And t1 <> -999.99 Then GoTo missing_23 Pattern 100 If t1 = -999.99 And t3 = -999.99 And t2 = -999.99 Then GoTo missing_123 'Pattern 000

PAGE 79

68 Appendix B: (Continued) 'Increasing counter for a complete record cnt12 = cnt12 + 1 cnt13 = cnt13 + 1 cnt23 = cnt23 + 1 'Keeping data for 24 hours 'For all the three dynamic arrays If cnt12 <= 288 Then ReDim Preserve r12(cnt12) r12(cnt12) = t1 / t2 Else For j = 1 To 287 r12(j) = r12(j + 1) Next j r12(288) = t1 / t2 End If If cnt13 <= 288 Then ReDim Preserve r13(cnt12) r13(cnt12) = t1 / t3 Else For j = 1 To 287 r13(j) = r13(j + 1) Next j r13(288) = t1 / t3 End If If cnt23 <= 288 Then ReDim Preserve r23(cnt12) r23(cnt12) = t2 / t3 Else For j = 1 To 287 r23(j) = r23(j + 1) Next j r23(288) = t2 / t3 End If Write #2, t1, t2, t3 Loop Close #1, #2 Exit Sub

PAGE 80

69 Appendix B: (Continued) 'CALCULATING MISSING VALUES 'Calculating vlaues for 1st sensor missing missing_1: If cnt23 <= 288 Then ReDim Preserve r23(cnt12) r23(cnt12) = t2 / t3 Else For j = 1 To 287 r23(j) = r23(j + 1) Next j r23(288) = t2 / t3 End If sum1 = 0 If cnt12 < 288 Then For i = 1 To cnt12 sum1 = sum1 + r12(i) Next i avg1 = sum1 / cnt12 Else For i = 1 To 288 sum1 = sum1 + r12(i) Next i avg1 = sum1 / 288 End If sum1 = 0 If cnt13 < 288 Then For i = 1 To cnt13 sum1 = sum1 + r13(i) Next i avg2 = sum1 / cnt13 Else For i = 1 To 288 sum1 = sum1 + r13(i) Next i avg2 = sum1 / 288 End If 'Finding average for the two values to determine the final value

PAGE 81

70 Appendix B: (Continued) t1 = (avg1 t2 + avg2 t3) / 2 Write #2, t1, t2, t3 GoTo top 'Calcualting values for the second sensor missing_2: If cnt13 <= 288 Then ReDim Preserve r13(cnt12) r13(cnt12) = t1 / t3 Else For j = 1 To 287 r13(j) = r13(j + 1) Next j r13(288) = t1 / t3 End If sum1 = 0 If cnt12 < 288 Then For i = 1 To cnt12 sum1 = sum1 + r12(i) Next i avg1 = sum1 / cnt12 Else For i = 1 To 288 sum1 = sum1 + r12(i) Next i avg1 = sum1 / 288 End If sum1 = 0 If cnt23 < 288 Then For i = 1 To cnt23 sum1 = sum1 + r23(i) Next i avg2 = sum1 / cnt23 Else For i = 1 To 288 sum1 = sum1 + r23(i) Next i avg2 = sum1 / 288 End If

PAGE 82

71 Appendix B: (Continued) t2 = (t1 / avg1 + avg2 t3) / 2 Write #2, t1, t2, t3 GoTo top 'Calcuating values for the third sensor missing_3: If cnt12 <= 288 Then ReDim Preserve r12(cnt12) r12(cnt12) = t1 / t2 Else For j = 1 To 287 r12(j) = r12(j + 1) Next j r12(288) = t1 / t2 End If sum1 = 0 If cnt13 < 288 Then For i = 1 To cnt13 sum1 = sum1 + r13(i) Next i avg1 = sum1 / cnt13 Else For i = 1 To 288 sum1 = sum1 + r13(i) Next i avg1 = sum1 / 288 End If sum1 = 0 If cnt23 < 288 Then For i = 1 To cnt23 sum1 = sum1 + r23(i) Next i avg2 = sum1 / cnt13 Else For i = 1 To 288 sum1 = sum1 + r23(i) Next i avg2 = sum1 / 288 End If

PAGE 83

72 Appendix B: (Continued) t3 = (t1 / avg1 + t2 / avg2) / 2 Write #2, t1, t2, t3 GoTo top 'Calculating missing values for 1st and 2nd sensor missing_12: '1st data sum1 = 0 If cnt13 < 288 Then For i = 1 To cnt13 sum1 = sum1 + r13(i) Next i avg1 = sum1 / cnt13 Else For i = 1 To 288 sum1 = sum1 + r13(i) Next i avg1 = sum1 / 288 End If t1 = (t3 avg1) '2nd data sum1 = 0 If cnt23 < 288 Then For i = 1 To cnt23 sum1 = sum1 + r23(i) Next i avg1 = sum1 / cnt23 Else For i = 1 To 288 sum1 = sum1 + r23(i) Next i avg1 = sum1 / 288 End If t2 = (t3 avg1) Write #2, t1, t2, t3 GoTo top 'Calculating missing values for 1st and 3rd sensor missing missing_13: '1st data sum1 = 0

PAGE 84

73 Appendix B: (Continued) If cnt12 < 288 Then For i = 1 To cnt12 sum1 = sum1 + r12(i) Next i avg1 = sum1 / cnt12 Else For i = 1 To 288 sum1 = sum1 + r12(i) Next i avg1 = sum1 / 288 End If t1 = (t2 avg1) '2nd data sum1 = 0 If cnt23 < 288 Then For i = 1 To cnt23 sum1 = sum1 + r23(i) Next i avg1 = sum1 / cnt23 Else For i = 1 To 288 sum1 = sum1 + r23(i) Next i avg1 = sum1 / 288 End If t3 = (t2 / avg1) Write #2, t1, t2, t3 GoTo top 'Calcualting values for the 2nd and 3rd sensor missing missing_23: sum1 = 0 '1st data If cnt12 < 288 Then For i = 1 To cnt12 sum1 = sum1 + r12(i) Next i avg1 = sum1 / cnt12 Else For i = 1 To 288 sum1 = sum1 + r12(i) Next i

PAGE 85

74 Appendix B: (Continued) avg1 = sum1 / 288 End If t2 = (t1 / avg1) '2nd data sum1 = 0 If cnt13 < 288 Then For i = 1 To cnt13 sum1 = sum1 + r13(i) Next i avg1 = sum1 / cnt13 Else For i = 1 To 288 sum1 = sum1 + r13(i) Next i avg1 = sum1 / 288 End If t3 = (t2 / avg1) Write #2, t1, t2, t3 GoTo top 'Keeping all the values as missing missing_123: Write #2, t1, t2, t3 GoTo top 'Program End End Sub


xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001670397
003 fts
005 20051216093350.0
006 m||||e|||d||||||||
007 cr mnu|||uuuuu
008 051128s2005 flu sbm s000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0001272
035
(OCoLC)62352595
SFE0001272
040
FHM
c FHM
049
FHMM
090
TA145 (Online)
1 100
Pujari, Swarna.
2 245
A framework to develop an interactive web database for delivery of water resources field data over the internet
h [electronic resource] /
by Swarna Pujari.
260
[Tampa, Fla.] :
b University of South Florida,
2005.
502
Thesis (M.S.E.S.)--University of South Florida, 2005.
504
Includes bibliographical references.
516
Text (Electronic thesis) in PDF format.
538
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
500
Title from PDF of title page.
Document formatted into pages; contains 85 pages.
520
ABSTRACT: The objective of the research is to develop a user friendly framework for an internet GIS (Geographic Information Systems) application. The study aims at providing a map with GIS capabilities without requiring the clients (users of the website) having to install ArcGIS (a product of ESRI) software on their personal computers along with the facility to download hydrological data. It also aims at providing a methodology to synthesize data in case of data gaps. High resolution data was collected in a small watershed in west Central Florida to measure hydrologic storages and fluxes during 2001-2004. Meteorological, surface and groundwater data were collected at 5 to 10 minute intervals. A watershed database was constructed using Microsoft Access and was normalized to 3NF (third Normalization Form) for easy update. The databases available do not involve user-friendly tools to map data collection locations or to facilitate interactive delivery of data.In many fields like hydrology, it is important to provide spatial location of the data points with the data. In addition there is also a need for one platform wherein various user communities (students, planners, hydrologists etc) can be served. Thus there is a need for an online user-friendly environment, which allows for interactive delivery of data along with mapping and spatial analysis. To address this need an ArcIMS website was constructed. It includes ortho-photography of the site, which can be overlaid with the shapefile of the data locations thus giving the user reference orientation of the data locations. The user is allowed to download data in the form of text files based on the required temporal resolution and date ranges. Furthermore, users are provided with valuable data sets to parameterize or test hydrologic models applied to west-central Florida conditions.
590
Adviser: Mark Ross.
653
ArcIMS.
Access database.
J2EE technologies.
Linear correlation.
XML.
Web GIS.
0 690
Dissertations, Academic
z USF
x Science Engineering
Masters.
773
t USF Electronic Theses and Dissertations.
4 856
u http://digital.lib.usf.edu/?e14.1272