Georeferencing Hydrologic and Hydraulic Control Points

in the Upper Mississippi Flood Studies

 Table of Contents

• Introduction
• Objectives
• Data Requirements
• Data Development
• GIS Procedure
• Developing a River Coverage
• Developing Gage Point coverages
• Producing Basemaps
• Conclusions
• Future Work
• References
• Data Dictionary

Click here to see a class presentation given on this project.

For additional information on the 1993 flood see these sites:

Following the flood and the significant damages caused by it, the Congress tasked the United States Army Corps of Engineers (USACE) to conduct a comprehensive, system-wide analysis to assess flood control and floodplain management in the areas that were flooded. USACE is the United States governmental organization responsible for the management and day to day operation of the nation's waterways. This responsibility includes oversight and permitting for locks, dams, lakes and rivers throughout the country.

USACE has two divisions that split primary responsibility for this area. The Mississippi Valley Division has responsibility for the entire length of the Mississippi River and some of her tributaries. The Northwestern Division controls the entire Missouri River and her associated tributaries. Each division is divided into a series of smaller organizations known as engineer districts which are geographically oriented. Specifically, the USACE districts are oriented along all of the nation's river basins. The geographical extent of these two arms of USACE are shown below.

Click here for a map of the USACE Mississippi Valley Division

Click here for a map of the USACE Northwestern Division

Two major studies are the result of this 1993 initiative. Hydrologic and hydraulic studies are being made of flood movement in the Upper Mississippi and Missouri River basins. The hydraulic study is applying the UNET model to simulate the movement of floods on the main stem Mississippi, Missouri and Illinois Rivers. The hydrologic study from HEC is focused on flood frequency analysis of stream gaging information from 30 gages on the Mississippi and Missouri rivers and some of their tributaries. These studies are currently each independently developing data and modelling the region.

The goal of the Mississippi Basin Modeling System (MBMS) was to produce a large-scale river model that produces uniform,system-wide unsteady data specifically designed for the Mississippi and Missouri Rivers and their tributaries. It is capable of handling both low flow data and routine day-to-day forecasting needs as well as simulation and forecasting of flood related events. MBMS is used to analyze and predict system-wide impacts of various alternative courses of action during flood events. An exciting application of this model is the real-time ability to take data and develop forecasts accurately and quickly for use during floods. The engineer districts which participated in the development of this tool included St. Paul, Rock Island and St. Louis Districts on the Mississippi River and Omaha and Kansas City Districts on the Missouri River. The model selected for use was the UNET model. The St. Paul District has implemented the use of this model and has used it since the 1997 flood to forecast the water surface elevations and assist in the mangement of the locks and dams at downstream locations. It has an interesting information paper available online to review current operations.

The second study is being developed by a technical advising group for USACE headed out of the Hydrologic Engineering Center in Davis, California. Its goal is to better estimate the flood-frequency curves for the Mississippi and her tributaries. The problem with modelling the flood frequency accurately arises from the tremendous sizes involved in this study. It drains all or part of 13 states and includes over 714,000 miles. The study area is approximately 57 percent of the Mississippi Basin and 23 percent of the area in the continental United States. The standard methods for drainage analysis are designed for much smaller areas and may not be appropriate for the large areas involved in this study. The methods and management of this study are currently in development and the ultimate tool is not yet complete.

The hydrologic and hydraulic studies should be linked because the design flows from the hydrologic study are inputs to the water surface elevations being determined in the hydraulic study. This linkage is complicated by the fact that the control points for the hydraulic model are in some cases at different locations than the stream gages of the hydrologic study, and because there are stream gages included in the hydrologic study which are not control points in the hydraulic model. A first step in linking these two studies is to georeference the hydraulic and hydrologic control points onto a common river database so that they can be linked by stationing along the river network.

This project is the initial step in my research effort to link the two studies of the Upper Mississippi Basin. My overall research goals are also available online.The goal of this term project is to georeference the gaging points for both studies. This would allow the studies to be related both geographically and spatially. To achieve this goal, four major subordinate tasks were necessary.

• Research, acquire and develop the necessary geographic data on the river network in the Upper Mississippi and the associated gaging stations from both studies
• Create a coverage for the associated river network within the study area
• Create a coverage for each of the two sets of gages associated with the studies
• Produce a series of map products to display the georeferenced data

The first challenge was to get the data to create the basemap. A minimum of three data sets were necessary: geographic data of the region's rivers, geographic position for the stream gages referenced in the MBMS hydraulic study, and geographic position for the stream gages referenced in the HEC flood frequency study.

Each of the studies currently are using a schematic to spatially describe the locations. These schematics are not to scale and don't accurately reflect the relationships and distances between the gaging points. The basemaps being developed with this project would ideally replace these schematics to assist the visualization of the study area.

Click Here to view the MBMS schematic

Click Here to view the HEC study schematic

 The geographic data for the entire United States' river system is available from a numbers of sources in ARCINFO format. The EPA's RF1 file was chosen as the most complete current data set available. It is available covering the entire continental United States and also a subset is available clipped to only cover the Upper Midwest.

 Data Sources for River

• Full RF1/RF3 River Reach Files Website: http://nsdi.usgs.gov/nsdi/wais/water/rf1.html
• Clipped RF1 File for the Midwest Website: http://edcwww2.cr.usgs.gov/dslist.html

 The complete RF1 file is also available on the USACE-HEC Spatial Hydrology CD-ROM in ARC INFO format. A database developed by Aubrey Dugger (a graduate who studied at the University of Texas) was used as a start point for identifying the rivers of interest. Her coverage (called RF1SAST) contains the river name as an attribute for each river reach. The geographic extent of RF1SAST was generally the same area as the clipped RF1 coverage available on the SAST website.

 The other two data sets were developed based on stated gage locations from the two studies of interest. Both studies have a series of gages that are located along the major rivers in the study area. These gages are stream flow gages managed by either the USGS or the US Army Corps of Engineers. Note the locations of the various gages on the hydrologic study and the hydraulic study schematics.

The location of each gage was determined by checking the state USGS websites for each state in the study area. A location in both geographic terms (lat/long) and spatial terms (rivermile positioning) was determined to further identify the geographic location of the gages.The following states' USGS sites were used to ascertain the locations.

State USGS Sites for Upper Midwest

• Minnesota: http://wwwmn.cr.usgs.gov/rt-cgi/gen_tbl_pg
• Iowa: http://wwwdiaiwc.cr.usgs.gov/realtime/rt-data.html
• Missouri: http://wwwdmorll.er.usgs.gov/rts/rts_table.html
• Illinois: http://www-il.usgs.gov/nwis-w/IL/datasum.components/rtsumm.cgi
• Nebraska: http://www-ne.cr.usgs.gov/rt-cgi/gen_tbl_pg
• Wisconsin: http://wwwdwimdn.er.usgs.gov/realtime/rt_latest_table.html

Unfortunately, Iowa, Nebraska and Missouri do not have location data for their gages noted in geographic coordinates (lat/long) on their websites. For these states, contact directly with each state's USGS Water Resources staff was necessary to confirm geographic positioning for the sites. These contact individuals are listed below:

 USGS Points of Contact

 In order to confirm the geographic locations of the gages, a second check of the locations was conducted using the Water Control websites for the Corps of Engineer district offices in the Mississippi Basin. Rock Island, St. Paul, Kansas City and St. Louis Districts have websites with extensive coverage and data for all gages within their area of interest. Omaha District's water control website is under construction and was not used. The website data are listed below:

Corps of Engineers District Contact Data

An additional challenge was the definition of the location of each gage. The geographic location was relatively straightforward. However, to spatially define the location of each gage, the term "rivermile" was used throughout all of the references. A rivermile is notionally defined as the number of miles that the given location lies upstream from the mouth of a river. This definition does however allow some interpretation of the locational attributes for the "mouth" of the river. Because of this definitional question, the locations for some of the gages were not consistent with the remainder of the map. The engineer districts had a variety of interpretations of the meaning of a "rivermile".

Click here for a visual representation of the different ideas about rivermiles.

The entire development of this basemap can be done using ARCVIEW on a PC workstation. It can also be done using ARCINFO on a UNIX workstation.

The first ojective of this project was to create a coverage of the rivers of interest in the study area. The coverage was developed using the files acquired from the sources noted in Data Development. The first step was to open a view and add the theme containing the RF1 data using the Add Theme tool. This is a very large data set and takes a few seconds to open.

Next, the data set RF1SAST was added to the view.  RF1SAST was then highlighted and using the table icon, its attribute table was opened. A query that identifies the various rivers from the complete coverage must then be constructed. The rivers identified for this project are the Mississippi River and her tributaries, the Missouri, Illinois, Rock, Des Moines, Minnesota and Iowa Rivers. These rivers contain all of the gage locations identified for either study. Using the query tool,, a query was constructed that identified the rivers of interest noted above. The method for using the query tool to deliniate data from a set is outlined in Exercise 4 of CE397:GIS in Water Resources.

This query separated the river reaches from the selected rivers from the balance of the data set. If the data set didn't accurately identify the river reaches by their respective rivers, the rivers were identified using the Select Feature tool. This method required an atlas and a steady hand. Each river is traced by selecting river reaches one at a time and comparing to the map reference. This method is much more time consuming.

Using the Create Shapefile command in ARCVIEW, a new coverage containing only the rivers of interest was created, saved and added to the view. Unfortunately, RF1SAST contains some minor gaps in data coverage. Overlaying two sets of data, RF1 and the newly created coverage can create a new more complete coverage. Using the reaches determined above as a guideline, the Select Featuretool was used to identify the complete set of reaches for all the rivers in the basin. A new shapefile was created, saved as MISSISSIPPI and added to the view using the Create Shapefile command. This coverage was a subset of the RF1 riverreach file containing only rivers that were selected above.

a. Convert to Decimal Degrees.

If the coordinates are in geographic degrees, minutes, and seconds then it is necessary to convert to decimal degrees since this is the form needed to work within Arc/Info. Use the following equation for the conversion:

Decimal Degrees = Degrees + minutes/60 + seconds/3600 (the west longitude is negative)

Do the conversion using a spreadsheet and save it for further conversions. See an example spreadsheet below.

b. Create a table for Gaging Sites. Using the Table/New command, new tables were made for both sets of gage data. These tables were used to link the numerical designation for the gages with the geographic data. Each study's gage locations were compiled into tables using the ARCVIEW table editor by adding records and fields. The completed table consisted of a numerical designation and the latitude and longitude in decimal degrees for each gage.

Screen Capture of Initial .dbf file for the MBMS gages

c. Convert to Point Coverage. The resulting tables were converted to point coverages using a AVENUE script developed by Seann Reed (a member of Dr. Maidment's research group). The procedure for converting the tables to a point coverage is also available in Exercise 4.This script imports the latitude and longitude from a series of data points in a table and develops a simple point coverage. It maintains six significant figures thereby retain maximum data resolution. It produces a point coverage in any required conic projection. It was slightly modified to project the coverage in the Albers Equal Area projection rather than the Texas State Plane projection. An excellent resource that contains other prepared scripts for a variety of purposes can be found in the Environmental Systems Research Institute (ESRI) website. The attribute table for the point coverage consists of the same fields as noted above. The projection used for this data layer was the same as the RF1 data. The parameters for the Albers Equal area conic projection were:

 The attribute tables for the point coverage was then edited using the Table/Start Editing and Edit/Add Field commands to add the following fields: Name, State, River, RF1 River Reach designation, Rivermile location and USGS/Corps of Engineers Gage Number.  The data was then added using the EDIT tool which enables the user to edit each individual data cell.

Screen Capture of Attribute Table for one of the point coverages generated

The attribute tables were then exported to EXCEL and can be viewed as Tables 1 and 2.

 The data then must be reviewed for accuracy. Primary focus must be on ensuring geographic accuracy for the gages. In addition, the placement of the gages was checked to see if they were coincident with the arc coverage of the rivers.

The point coverage can also be created using ARCINFO in addition to ARCVIEW.

Creating point coverage of gaging stations from latitude-longitude coordinates

To make the point coverage compatible with other maps used in this study, the map was projected into Albers system of coordinates as follows . The projection parameters used are standard ones for USGS maps except that the Albers rather than Lambert projection was used to preserve correct surface area througout the study region.

Projecting the point coverage of USGS river flow stations from Geographic system into Albers coordinates

A text file can be created using the text editor window from the DEC command bar.

Once completed it was saved as albers.prj The file gages1 was then projected into Albers Equal Area projection using the command:

Arc: project cover gages1 gages2 albers.prj

Either of these two methods should be satisfactory to complete the development of the point coverages.

The map products are produced and refined using the Layout portion of ARCVIEW. This option allowed the manipulation of the coverage to add text and illlustratory graphics to add to the clarity of the product.

The map products can be enhanced by using the Theme/Auto-Label command. This command allows the user to identify each piece of data using one of the field in its respective attribute table. In the map products produced, the gages are labeled both by the name field and the ID # field.

The resulting map products can be viewed here. These products show the gage locations for both the MBMS hydraulic study and the HEC hydrologic study for the region.

The third map product displays both sets of gage locations to aid in comparison of distance and position and can be viewed here.

This term project provided a outstanding start to my research effort. The successful production of a series of georeferenced map products for the two studies of the Mississippi Basin will provide a good starting point for developing a tool to link these two studies. The hydraulic and hydrologic modelling of the Mississippi River and her tributaries are naturally inter-related. The flow that results from the UNET model should be related to the flow frequencies generated by the hydrologic model proposed by the TAG advisory group.

The need for both hydraulic and hydrologic studies to converge is a basic one. If the result from one series of models does not approach the other, the results might not support each other. Because the Corps of Engineers is sponsoring both studies, the ultimate goal must be a model that mirrors the impacts of possible flooding in the Upper Mississippi basin -- both hydrologically and hydraulically. Similarly, the amount of runoff that arrives at each incremental gaging station should be consistent with the modeled flow characteristic of the river or stream in which the water is flowing.

Data could be organized in a manner that could allow the study of subregions within the overall study region. Given the large ground area covered by the study and the inherent difficulty of data analysis due to the large drainage area involved, a series of incremental analyses might be possible using a relational database. A relatively detailed database has been developed here at the University of Texas focusing on drainage area delineation and precipitation characteristics. These databases could easily be integrated into any resulting relational database to assist in data analysis and synthesis.

Geographic Information Systems offer a dynamic and evolving opportunity for the presentation and display of hydraulic and hydrologic information. ARCVIEW and ARCINFO offer a great platform with which to model the real-world situations in the Mississippi Basin. The one-to-one relationship of data to georeferenced map objects in a spatial database offers dramatic possibilities for this model. The real-world requirement to manage the traffic and water levels in the Mississippi Basin will be a challenge for GIS efforts in the Corps of Engineers for some time, but the studies that are being developed at both the district level and the Corps research centers such as HEC and the Waterways Experimental Station are both promising and significant.

I plan to continue to expand this model in cojunction with DR. Maidment and his research group. Primary goals will be:

In addition to the online resources and e-mail correspondence noted above in the Data Development portion of this paper, the following references were used in developing the concepts for this project.

ESRI, 1994. Avenue: Customization and Application Development for ARCVIEW. Redlands, Ca.

Gee, D.M and Tseng, A. G., 1996, Unsteady Flow Model for Forecasting Missouri and Mississippi River, Proceedings RIVERTECH '96 1st International Conference on New/Emerging Concepts for Rivers, September 22-26, 1996, Chicago, Illinois, 869-876 International Water Resources Association, Urbana, IL.

HEC, 1997, Flow Frequency Analysis Methods For the Upper Mississippi Basin (Draft). Hydrologic Engineering Center, U. S. Army Corps of Engineers. Davis, California, December 1997

Maidment, David A.,1992, A Grid-Network Procedure for Hydrologic Modelling. Stanford University, Stanford, CA.

USACE, 1995, Floodplain Management Assessment of the Upper Mississipi and Lower Missouri Rivers and their tributaries - Main Report. U.S. Army Corps of Engineers. Washington, D.C., June 1995.

In addition, I have compiled the web resources used on a online link page accessble here.

Web page authored by: Ben Bigelow
Last Revised on 18 May 1998
Questions? Send email to my mailbox at bjb2@aol.com or bbigelow@mail.utexas.edu

Back to Ben's Home Page