GIS in Water Resources Term Project
CE 394 K3
Analysis using MIKE 11 Software
of Mill Creek
by Daniel B. Snead
Become familiar with flood modeling using software from the Danish Hydraulic Institute (DHI).
Develop a topographic model of a flood plain using DHI's MIKE 11 GIS software.
Develop a river network model of Mill Creek and its branches using DHI's MIKE 11 version 4.1 software.
Create flood model of Mill Creek basin using the MIKE 11 model.
US Army Corps of Engineers, Louisville District.
The Environmental Protection Agency's Basins web page.
The USGS Water Resources web page.
Gain a full understanding of DHI's MIKE 11 and MIKE 11 GIS software.
Generate a DEM flood model of the Mill Creek basin using MIKE 11 GIS.
Develop a Branch Route System (BRS) of the Mill Creek river network using MIKE 11.
Develop a flood inundation map of the Mill Creek basin in MIKE 11 GIS based on the January 21, 1959 flood of record (25-yr. return period).
Compare results of model to existing data from the 1959 flood.
The Louisville District of the U.S. Army Corps of Engineers currently does not use state-of-the-art flood plain mapping technology when analyzing the waterways within their area of responsibility. With the assistance of the Danish Hydraulic Institute (DHI) and the Center for Research in Water Resources (CRWR), I am creating a MIKE 11 model of Mill Creek located in Cincinnati, Ohio. The purpose is to provide the Louisville District a modeling system they can use for future analyses of flooding.
MIKE 11 is a powerful software tool that can be used for a wide range of flood and water quality modeling and studies. Models can range from hydrology and hydrodynamics-based, to advection dispersion, sediment transport, and rainfall runoff.
DHI’s MIKE 11 software is known worldwide. MIKE 11 is the
industry standard for hydraulics and water quality modeling for the United Kingdom
as well as other European countries. It is also used by Federal agencies in the
U.S., one example being the Corps of Engineers’ Rock Island District in
Illinois. They have used MIKE 11 to conduct a
river flooding and water quality study on the Des Plaines River outside the
City of Joliet.
Mill Creek is located in Butler and Hamilton Counties in southwestern Ohio. It flows from the southeastern part of Butler County in a southerly direction across Hamilton County and through the city of Cincinnati to its confluence with the Ohio River. The total fall in elevation from the headwaters of Mill Creek to the Ohio River is about 250 feet over an approximate distance of 28 stream miles, with an average gradient of 8.9 feet per mile. The Mill Creek watershed is in the northeastern finger of the Hydrologic Unit Code (HUC) #05090203. In 1997, the environmental interest group, American Rivers, designated Mill Creek as the most threatened urban stream in North America.
Flooding has been a major problem for Mill Creek for some time. The most damaging flood occurred in January 1959. Since then, there have been numerous headwater floods of lesser magnitude. Over bank flooding occurred in some areas as late as the spring of 1996.
Based on a Local Cooperation Agreement in 1975, construction by the Corps of Engineers was initiated in 1981 and eventually suspended in 1992 with approximately 50% of the construction complete. The construction was suspended for a number of reasons. The Assistant Secretary of the Army for Civil Works suspended the project because 1) there were problems in acquiring real estate and relocations for the remaining sections of the creek requiring construction, 2) project costs had soared over 126% of the authorized amount, 3) there was likely contamination of the water from non-point source pollutants from old landfills along some of the uncompleted portions of the reach, and 4) there were problems maintaining and operating the sections where construction was completed.
In 1997, a reevaluation study was performed. The effort showed that even with the partially completed plan in place that significant damage would occur from a flood with a 50% chance of occurrence. Total residual damage is estimated over $486 million for the 1% chance flood and over $910 million for the 0.2% chance flood. Total expected annual damage for the flood area is estimated over $32 million, almost 96% of the damage being commercial or industrial.
Although completion of the previous plan is economically feasible, the Corps of Engineers believe that a more cost effective and environmentally acceptable plan can be formulated. There is strong local support at the present time to address the environmental needs of Mill Creek.
The MIKE 11 floodplain model may provide the Corps of Engineers some additional
analysis of the flooding that may occur from Mill Creek. Additionally, the MIKE 11 software can also graphically represent the floodplain in such a way
that the public can understand its results easily. A successful MIKE 11 model of Mill Creek can also assist planners in evaluating the water quality
and environmental issues that still exist.
As stated previously, DHI has developed the MIKE 11 software as well as the MIKE 11 Geographic Information Systems (GIS) software. MIKE 11 GIS integrates the MIKE 11 river and floodplain modeling into the Environmental Systems Research Institute’s (ESRI) Arcview software. MIKE 11 GIS is suited as a spatial decision support tool for river and floodplain management.
MIKE 11 GIS is a fully integrated extension of Arcview. Flood depths and water levels are represented as layers of data in Arcview and is easily related and analyzed together from other data from other flood management components through its graphical user interface (GUI).
MIKE 11 is a modeling package for the simulation of surface runoff, flow, sediment transport, and water quality in rivers, channels, estuaries, and floodplains. The most commonly applied hydrodynamic (HD) model is a flood management tool simulating the unsteady flows in branched and looped river networks and quasi two-dimensional flows in floodplains. MIKE 11 HD, when using the fully dynamic wave description, solves the equations of conservation of continuity and momentum (known as the 'Saint Venant' equations). The solution to the equations are based on the following assumptions:
The equations used are:
A: flow area, (m²)
q: lateral inflow, (m²/s)
h: stage above datum, (m)
C: Chezy resistance coefficient, (m½/s)
R: hydraulic or resistance radius, (m)
I: momentum distribution coefficient
The four terms in the momentum equation are local acceleration, convective acceleration, pressure, and friction (Source: MIKE 11 online help).
In MIKE 11 a network configuration depicts the rivers and floodplains as a system of interconnected branches. Flood levels and discharges (h and Q) are calculated at alternating points along the river branches as a function of time. It operates on the basic information from the river and floodplain topography, to include man-made features and boundary conditions.
MIKE 11 GIS provides a bi-directional interchange between MIKE 11 and Arcview. MIKE 11 GIS can extract cross-sectional profiles of the river channels and area-elevation curves from a Digital Elevation Model (DEM) and export that data into a MIKE 11 cross-sectional database. MIKE 11 GIS can then import simulated water levels and discharges from a result file created from a MIKE 11 simulation. Based on the discrete information, MIKE 11 GIS creates a grid-based water surface, compares it with the generated DEM, and produces floodmaps. MIKE 11 simulation results can also be displayed as graphs and longitudinal profiles using the MIKE 11 GIS GUI.
MIKE 11 GIS Interface Schematic.
The main outputs in MIKE 11 GIS are inundation maps presented as depth, duration, or comparison maps. It
can also present flood comparison maps, comparing the difference between two
flood maps to illustrate impacts or changes from proposed construction of
detention basins, levees, weirs, etc.
The data came from a number of sources. The Louisville District provided a 1° Seamless DEM of the Mill Creek Watershed, created by the United States Geologic Survey (USGS) for a nominal fee. Pete Andrysiak projected the grid into Ohio State Plane projection using Arcinfo and used the DEM for creating a HEC-HMS rainfall runoff model of the Mill Creek watershed. Using the DEM grid Pete used, I converted the grid file to a shape file and used it for my project.
From the EPA BASINS web page, the reach of Mill Creek was extracted from the RF3 file and was used to develop the geographic link between the MIKE 11 GIS and MIKE 11 models.
To create a model based on historic data, flow data from the January 1959 25-yr flood event was obtained from USGS gage station #03259000 (Carthage gage station) just upstream of the model area. The data was extracted from 17 January 1959 to 03 February 1959 from the USGS Water Resources of the United States web page.
Data from the Louisville District: Demgrid – 1° Seamless DEM of the Mill Creek area
Data from Pete Andrysiak: Theme3 – Shape file converted from a DEM grid representation of the Mill Creek Watershed
Data from the USGS Basins web page: 05090203_rf3 - RF3 file of HUC #05090203
Data from the USGS Water Resources web page: Carthagedata.xls – spreadsheet of flow data from the Carthage gage station, converted to m3/s
MIKE 11 data
The initial data required for establishing the MIKE 11 model was 1) the river reach or reaches, and 2) the cross-section data depicting the river channel. Fortunately, the Corps of Engineers’ Louisville District had surveyed cross-section data. This data was provided as HEC-2 files. HEC-2 software is an early version of the Hydrologic Engineer Center’s (HEC) HEC-RAS software (RAS stands for River Analysis System).
For the purpose of analysis, the Louisville District divided Mill Creek into three separate sections in HEC-2. The HEC-2 files once converted into HEC-RAS, provided the cross-section data needed for the MIKE11 model, known as geometry files (file with extension .G01). The HEC-2 data used for this project was the second section of Mill Creek upstream of its confluence with the Ohio River.
Data provided by the Louisville District: millnat2.txt- a text file generated from HEC-2 depicting the second section of Mill Creek
The project required a various amount of software for use on a Windows-based Personal Computer. For the purposes of research, DHI donated the entire MIKE 11 software package to CRWR for a limited time. The overall cost of the MIKE 11 package is $11,000 commercially. All other software has been provided on the computers either at the Civil Engineering department’s Learning Resource Center (LRC) or at CRWR. CRWR has the only working MIKE 11 software package on the UT campus.
Arcview version 3.1 from ESRI
Arcview extensions required:
Spatial Analyst – used to convert grid DEM to a shape file
Projector! – to project shape files into Ohio State Plane projection
CRWR Vector – created by CRWR, used to add x, y coordinates to a digitized point theme of the Mill Creek reach, for importing into MIKE 11
MIKE 11 GIS extension for Arcview provided by DHI
MIKE 11 version 4.1 provided by DHI
MIKE View version 1.44 provided by DHI, used for viewing MIKE 11 simulations
HEC-RAS version 2.1/2.2 – software by HEC, to generate .G01 files
Hec2m11.exe - Visual Fortran program created by Mr. Stefan Szylkarski, DHI’s San Francisco Office, used to convert HEC-RAS .G01 files to MIKE 11 cross-section files
MS Excel - used to manipulate and/or convert data
Converting the DEM Grid to a Shape File
Since MIKE 11 GIS and MIKE 11 are separate entities, the models can be created simultaneously. I have separated the two processes for ease of understanding. I wanted to define the study area first, so I started with establishing the MIKE 11 GIS model.
Since MIKE 11 works with x, y, and z coordinates in meters, I used the Projector! extension to project the RF3 file into Ohio State Plane coordinates. This was done in Arcview without the MIKE 11 GIS extension. The DEM grid had already been converted to the Ohio State Plane using Arcinfo for Pete’s project.
1° Seamless DEM of Mill Creek Provided by the Louisville District
I first started by opening Arcview from the Windows Menu. I clicked on Extensions under the File menu. I checked the Spatial Analyst, Projector!, and CRWR Vector extensions:
The next step was to convert the grid DEM to a shape file, using the Convert to Shapefile command under the Theme menu. Since the grid was a floating-point grid, the elevations were converted to integer format using the Map Calculator command, found under the Analysis menu. Choosing the Arithmetic calculation, the suffix .Int was added to [Demgrid] and the Evaluate button was pressed:
Once the new grid was created, it was converted to a shape file using the Convert to Shapefile command:
The model area was cut out of the created shape file, Theme3.shp. The new shape file was saved as Sect3.shp:
I saved the project before exiting, calling it Millcrkav.apr. Upon completion of creating a shape file depiction of the modeled area’s DEM, the data was ready for MIKE 11 GIS.
Defining the Topographic DEM in MIKE 11 GIS
I opened MIKE 11 GIS from the Windows menu. Once in MIKE 11 GIS it automatically asked me to set the working directory. I set mine to d:\dano. There were two buttons at the top left of the window, the first for the DEM module and the second for the Flood Management module. I entered the DEM module by clicking on the blue grid button at the top of the window:
The button opened a view called DEMView. Once opened, I added sect3.shp to the view:
In MIKE 11 GIS, the topographic DEM is created from a scatter point shape file and a user-inputted grid mesh. To create the scatter point shape file, sect3.shp was converted to an XYZ text file using the DEM menu:
The ‘level’ field of the XYZ file was chosen as ‘value’ from the attributes of sect3.shp and the export option was chosen as ‘export center points’. By choosing ‘value’ for the ‘level’ field and the ‘export center points’ as an option, I created a text file of X, Y, and Z coordinates at the interpolated center of each polygon in the shape file. Once the XYZ file was created, the scatter point shape file was converted from the text file, also from the DEM menu:
The scatter point shape file created is shown in the next
The next step was to define the DEM grid mesh. From the DEM menu, I clicked on New definition file:
A window appeared that allows the user to define the extent of the grid and the grid cell sizes. For this project, I used 30-meter grid cells. Using various combinations for grid extent based on the model area, the number of grid cells and grid extent was derived. After using some combinations, I created a definition file for the section of Mill Creek used in the model:
The output file name will be the new shape file and image file created from the grid, called dem3.shp. Using the DEM menu again, I clicked on Save definition file as and saved it as sect3osp3.inp. Then I clicked on Load definition file to load the saved definition file. The DEMView now showed the grid mesh and the point shape file:
I clicked on Make grid from the DEM menu and after an interpolation process the following DEM was created:
The generated DEM now appeared in a new view automatically called Generated DEM, and was called dem3.shp. An image file of the DEM was also created, called dem3.bil. The legend editor was modified using Green monochromatic to define the DEM's topography, so it could be seen easier. The legend file was saved as legend1.avl. The DEM for the project area has now been defined and created.
Integrating floodplain features into the DEM
The DEMs generated in MIKE 11 GIS do not always contain the detailed information of certain features within the floodplain. Features such as embankments, roads, lakes, urban areas, levees, etc. can change the characteristics of a floodplain and are very important to integrate into the model. If data exists of such features, those features can be defined into the DEM. Although it is not the model area, the following DEM is an example of how to integrate features into the topography.
Using the MIKE 11 GIS menu item Tools, features can be built into the DEM as layers represented by individual themes, or shape files. These shape files can be created as polygons or polyline themes, depending on the type of feature. For this example, I integrated the topography of the lakes included in the studied area as a polygon theme.
I digitized the polygon theme (or lake theme) using the lakes defined in the rf3osp.shp (rf3 file projected in Ohio State Plane) file. If a polygon theme already existed for the lakes, it could be opened in DEMView. Unfortunately, the lakes in the rf3 shape file existed as polylines. To create the polygon-lake theme, I added the rf3osp.shp file to DEMView. Using the same method in the previous step, I also generated a DEM called Demtest1.shp of the entire Mill Creek reach to ensure I obtained all the required lakes in the study area:
Using the Select feature button in Arcview, I selected all the polylines making up the lakes in the area, and saved the shape file as milllakesosp1.shp. Opening the MIKE 11 GIS Tools menu, clicking on Create theme, and selecting polygon theme, I digitized the lakes. I also ensured the polygon tool button was pressed. The new polygon theme was saved as polygon1.shp.
Elevations were assigned to each lake. They were assigned by selecting the Tools menu and the assign elevations field. Each lake was highlighted yellow with a menu prompt:
The elevation assigned to each polygon in the theme is horizontal. There is also a buffer zone prompt. When specifying a buffer zone, a zone outside the perimeter of the polygon is created. The grid cell elevation in the buffer zone is assigned the polygon elevation. Inside each polygon the specified elevation is assigned to a grid. Along the digitized polygon line elevation is assigned at every line vertex (mouse click). The polygon step field is used to increase the number of points along a polygon line in which the elevation is assigned. It is recommended to use a polygon step equal to the grid cell size.
Ensuring the XYZ point theme pthucsosp3.shp is in DEMView, I moved the polygon theme to the top of the legend and loaded the previously saved grid mesh, sect3osp.inp. Clicking on re-define grid in the DEM menu, the mesh appeared in DEMView. By zooming into a specific area, you can see the differentiation of the point theme, polygon theme, and grid mesh:
Finally, clicking on Make grid, MIKE 11 GIS ran an interpolation and created a new shape and image file of the study area. I inserted the Mill Creek reach from the rf3 file, rf3osp.shp, and the lake theme, polygon1.shp, to analyze the new topography:
The DEM changed, integrating the topography of the five
lakes into the overall topography. When generating a DEM, more time spent
scrutinizing its accuracy as well as including specific data of existing
features will create a more accurate model.
MIKE 11 comprises of nine different editors in which data can be implemented and edited independently of other information. Once the information has been inputted into each specific editor, the MIKE 11 Simulation editor integrates the separate files to create a simulation.
Creating the River Network using the Network Editor
The Network editor is central to the MIKE 11 system and displays information from all the other editors. The editor presents data in two views, the Graphical View and Tabular View. The Graphical View shows a graphical depiction of the river network and the Tabular View contains the properties of the network. Both views can be edited according to each user’s requirements.
Two things were required prior to opening the MIKE 11 interface. The first was to convert the River Stations established for Mill Creek used in the HEC-2 text file, millnat2.txt, to Chainages in MIKE 11. The second requirement was to import the attributes of the Arcview theme, Millreachosp.shp, into the MIKE 11 Network editor. River Stations in the United States are points of reference along the length of a river reach and is in units of feet. River Stations start from the most downstream point of the reach and increase as you continue upstream. Chainages in MIKE 11, on the other hand, are in units of meters. To convert the River Stations to Chainages, the River Station values were extracted from the millnat2.txt HEC-2 file. The River Stations were imported into MS Excel.
The Louisville District established River Station #102400 as the first river station in Mill Creek, just 305 meters upstream of its confluence with the Ohio River. Using the first Chainage # as 305 at River Station #102400, I converted all of the River Stations to Chainages by using simple interpolation in Excel:
The river network was defined in MIKE 11 by digitizing a point shape theme of the Mill Creek reach, copying each point’s x and y coordinates, and pasting them into MIKE 11’s Tabular View of the Network editor.
The second requirement was initiated by creating a digitized point theme in Arcview along the reach of Mill Creek:
Points were added by using the Create a Point button in Arcview. With the Mill Creek reach Millreachosp.shp displayed, I inserted a point at each vertex along the polyline theme, which created a point theme. Using the Zoom-in button, I noticed that the reach’s curves are short straight lines connected together. I saved the point theme as reachpoints.shp. Once saved, I used the CRWR Vector extension in Arcview to add x and y coordinates to the attributes of the reachpoints.shp table:
Before opening up MIKE 11, I needed to use MS Excel to open the reachpoints.shp table so I can copy and paste the x and y point coordinates created in Arcview:
I next created the Mill Creek network in MIKE 11. Opening MIKE 11 4.1 from the Windows menu, I opened a new file in MIKE Zero, and the following menu popped up:
As you can see, a number of different editors can be opened from this menu. I clicked on the River Network editor and clicked on Tabular View from the View window. The Tabular View editor was displayed. I copied and pasted the x and y coordinates of the points into the Tabular View from Excel. The Graphical View automatically displayed the points. To define the project area, user defined settings for the Chainages were inputted under the Branch menu in the Tabular View. The project’s area Chainage ran from Chainage #18131 to Chainage #32024:
Once the Chainage area was defined and the points were inputted into the Tabular View, the network was created using the Define Branch tool in the Graphical View. Holding down the left mouse button and running the Define Branch icon across the points in the Graphical View from upstream to downstream, the network was created. The created network looked like the Mill Creek reach theme in Arcview. The network file was saved as millsect3-2a.nwk11:
There are number of ways to extract the information of the reach from Arcview and import it into MIKE 11. Another method would be to scan the image of the reach and input it into the Network editor using the Layers menu.
Inputting River Cross Sections Using the Cross Section Editor
If the DEM is accurate enough river cross sections can be imported from MIKE 11 GIS into MIKE 11 using the XStool in MIKE 11 GIS. The generated DEM did not contain accurate topography of the river reach, but the Louisville District’s millnat2.txt HEC-2 file had surveyed data of river cross sections.
Mr. Stefan Szylkarski from DHI created a Visual Fortran executable program called Hec2m11.exe to convert a HEC-RAS geometry file to a MIKE 11 cross-section file. First, the HEC-2 file was imported into HEC-RAS 2.1. A new project was created in HEC-RAS called millcrk1.prj. From the File menu in HEC-RAS, the Import HEC-2 data command was highlighted and the millnat2.txt file was selected and imported into HEC-RAS. Making sure the units used in HEC-RAS were SI units, the geometry text file created was millcrk2.go1. Notice the difference in formatting of a MIKE 11 cross section file and a HEC-RAS geometry file for the same cross section (River Station #132435):
Comparison of MIKE11 cross section file to a HEC-RAS Geometry file.
To make the HEC-RAS file readable for MIKE 11, the Hec2m11.exe program converted the text accordingly. It is an MS-DOS executable program, inputting the HEC-RAS geometry file (millcrk2.go1) and MIKE 11 cross-section data output file (mill23m11.txt).
By creating a new file in Mike Zero and choosing Cross sections, the Cross Section
editor interface appeared. By selecting Import under the File menu, and choosing Input Raw
Data, mill23m11.txt was imported into
the Cross Section editor.
River Station #163925 from Mill Creek in HEC-RAS.
Conversion of River Station #163925 to Chainage #19058 in MIKE11.
Chainages for each cross section were inputted manually in the MIKE 11 Cross Section editor interface. The Cross Section editor displayed the graphical representation of the cross section along with the tabular data.
Bed resistance factors and channel markers (markers that differentiate the river channel from the flood plains in the cross section) can be modified in the Cross Section editor for each cross section. The roughness parameter in the Bed Resistance Property Page of the Hydrodynamic editor is factored by the resistance factor inputted in the Cross Section editor.
Creating Time Series Files using the Boundary editor and Time Series editor
To create time series and boundary conditions for the river network, I inputted a water level (h, in meters) downstream and flow rate (Q, in m3/s) upstream as the boundary conditions. I used hydrodynamic parameters to simulate the flow in the river channel. Optional parameters include rainfall runoff, advection dispersion, and sediment transport.
First, a hotstart simulation was created to establish steady-state flow in the channel before inputting the Q and h boundary conditions for a flood event. The boundary values for Q and h in the hotstart file do not have to be precise to depict the average flow in the channel, as long as the values are reasonable to allow the file to reach a steady state.
Using MIKE 11’s Time Series editor, I inputted a constant value for Q (upstream boundary, chainage #32024.32) as 1.5 m3/s and a constant value for h (downstream boundary, chainage #18131.23) as 177 meters (water surface elevation) from 12:00 pm on 16 January 1959 to 12:00 pm on 18 January 1959, using a 1-minute time step. Both time series files were saved as TS2_32024.dfso and TS2_18131.dfso, respectively. The diagram below depicts the time series file TS2_32024.dfso in the Time Series editor, showing the time series graph and tabular data simultaneously:
Once the time series data was inputted, the files were used as the boundary conditions in the Boundary editor. Each boundary chainage was inputted into the Boundary editor, and the upstream chainage Q and downstream chainage h items were selected from the time series files. Q and h items were selected from specific time series files by 1) hitting the Browse button to select the time series file, and then 2) hitting the Items button to select which item (Q or h) from the available items in the time series file:
The boundary file was saved as Bndsect3-2.bnd11. Additional time series and boundary files were created for the actual flood event. Upstream boundary flow data was inputted into a time series file from the Carthagedata.xls spreadsheet, depicting historic flow data measured by the Carthage gage station just upstream of the model area. Since there was no data on downstream h levels, a downstream h level file was assumed based on calculated h values from steady state flow data. The time series ranged from 12:00 pm on 18 January 1959 to 12:00 pm on 03 February 1959.
Flow Data from the Carthage Gage Station (Source: USGS Water Resources Web Page).
MIKE 11 Time Series File with inputted Flow Data.
The time series files were saved as TS2_32024-2.dfso and TS2_18131-2.dfso. The boundary file was created the same way as the hotstart boundary file, and was saved as Bndsect3-3.bnd11.
Inputting Hydrodynamic Parameters using the Hydrodynamic editor
The final data required to run the hotstart and flood simulations was the hydrodynamic parameters. The parameters are created in the HD Parameter editor in MIKE 11. For this model, the only parameter changed from the default values was the initial water levels. If using the rainfall runoff, advection dispersion, and/or the sediment transport models, additional parameter files must be created for each. The initial elevation across the reach was established as 172 meters. The HD parameter file was saved as HDsect3-2.hd11 and was used as the HD parameter file input for the hotstart simulation. The hotstart simulation was used to establish the initial conditions for the flood simulation.
Creating a Hotstart file for Initial Conditions
Now that we created the required files to run a simulation, the hotstart file was run initially to establish steady-state conditions for the flood simulation. A new simulation file was created, hydrodynamic model was checked, and input files were inserted into the MIKE 11 Simulation editor:
The network file, millsect3-2a.nwk11; the cross section file, MillXsect3-2.xns11; the boundary file, Bndsect3-2.bnd11; and the HD file, HDsect3-2.hd11 were inserted into the simulation using the Simulation editor under the Input tab. Under the Simulation tab, simulation start and end times, time step, and parameter type(s) were inserted:
The simulation start and end times were set at 12:00 pm on 16 January 1959 and 12:00 pm on 18 January 1959. Time step was set at 1-minute and type of initial conditions was set as HD parameter file.
Results were set as filename hotstart.res11 with a frequency of 60. The frequency established how many time steps would be saved in the simulation file. For instance, with hotstart.res11 every 60th time step was saved. Since the time step was set at 1 minute, each hour’s results were saved in the simulation file. The simulation file was saved as results3-2.sim11.
Under the Start tab, any discrepancies in the input files were listed in the Validation messages box. If there weren’t any discrepancies, then the simulation was ready to start. Some errors initially occurred during the running of the simulation, which were listed on Word Pad after the simulation ran.
After correcting the errors and the simulation was successfully executed, the hotstart.res11 file was checked in MIKE View to ensure its validity.
The simulation file for the flood event was created the same way as the hotstart file. Under the Input tab, the boundary file was changed from the hotstart simulation to Bndsect3-3.bnd11 since the boundary conditions were based on separate time series files. Inputs under the Simulation tab were modified based on new start and end times, a new time step, and new initial conditions created from the hotstart result file, hotstart.res11:
The flood simulation started at 12:00 pm on 18 January 1959 and ended at 12:00 pm on 3 February 1959 with a time step of 10 minutes. The simulation file was saved as results3-2b.sim11. The results file was given the name, sect3-4b.res11. Errors were checked under the Start tab, and the simulation was started:
Once completed, the results were checked in Mike View.
Checking Results in MIKE View
Upon opening the MIKE View interface from Windows, the sect3-4b.res11 file was opened from the File menu. Once opened, MIKE View displayed the horizontal view of the reach. Using the Plot longitudinal profile tool and clicking on the reach on the horizontal view, the longitudinal profile was displayed. The profile depicted the extent of the reach with included cross sections with the initial water level values:
Using the Run Animation pull-down menu , the simulation for the duration of time during the flood event was run. The following depicts the MIKE View animation by developing a GIF animation for purposes of displaying on the Internet. 31 time steps were used from the simulation:
Additional data can be viewed in MIKE View: 1) water levels at a cross section, 2) time series plots of water level and flow at points along the reach, and 3) empirical values of water level and flow at a specific point along the reach at a specific point in time.
Once the MIKE 11 simulation was created, the calculated data was imported in MIKE 11 GIS using the Branch Route System. Upon completion of Generating a DEM, the DEM Module was closed and the Flood Management Module was opened. Once opened, the Specify MIKE 11 GIS data menu automatically appeared:
The DEM inputted into the menu was
the generated Topographic DEM, dem3.shp. The grid input
file was the grid mesh used to develop the topographic DEM, sect3osp3.inp. And finally, the MIKE 11 version 4.x was
clicked and the Ok button was pressed.
Creating the Branch Route System (BRS)
The next menu that appeared determined the MIKE 11 input into MIKE11 GIS. This was where the Branch Route System (BRS) was defined:
The "Convert Mike 11 network file to Branch Route System (BRS)" was checked and the network file used was identified was millsect3-2a.nwk11. The MIKE11 MSD path was chosen as sect4b.msd. The sect4b.msd file was a folder automatically created in MIKE 11 from the simulation. The MSD folder stands for MIKE 11 Simulation Data. The MSD contained the data of h- and Q-point values for each time step used in the simulation.
The BRS is a network of lines used to geographically locate and display MIKE 11 model features such as branches, h-points, and Q-points. It is conceptually similar to a MIKE 11 network, except that it is geo-referenced using polyline shape files. Chainages are assigned at nodes (end points of each line) along a branch route and used to dynamically locate MIKE 11 features. The BRS mapped MIKE 11 branch networks to the topography of the DEM.
Once the "Ok" button was pressed on the menu, the following view appeared:
With the MIKE 11 model geo-referenced, the next step was to import the MSD data into MIKE 11 GIS.
Importing Q- and h-Points from MIKE 11
The BRS was used to determine the geographic location of each h- and Q-point within the model area. Using the BRS Tool menu, Calc. MSD h/Q points location tool was chosen:
Once clicked, MIKE 11 Q- and h-points were added to the View as point themes. Qpoints.txt and hpoints.txt were added as attribute tables to the themes. The attribute tables contained data like coordinates, branch name, and chainage of each point. The following shows the Q- and h-point themes added to the view:
Extracting Simulation Results from Different Points in Time
The final step before creating the flood maps was to join the h- and Q-points with the simulation data (MSD). The Time Series button on the tool bar was pressed to import the MIKE 11 simulation data from the MSD directory. Q- and h-values at each time step were added as additional fields to the Q- and h-point attribute tables.
As shown above, water surface elevations at each time step were added as additional fields to the attributes table of Hpoints.txt.
The MIKE 11 GIS Flood Management Module can generate and display flood maps using the DEM, BRS, and MIKE 11 data now stored in the h-points attributes table. The properties (extent) of flood map were initially defined using the Properties tool from the FM Tool menu:
The Properties menu allows the user to define the flood map’s legend label in the view, minimum flood depth used for interpolating the flood maps, and monitor resolution. For this model, I used the default values above.
After the flood map properties were specified, the Inundation Map tool was used from the FM Tool menu. From the drop down list the "Dynamic and user defined CBL" was picked:
CBL stands for channel boundary lines. By selecting the "Dynamic and user defined CBL" MIKE 11 GIS extracted the position and the water level for every h-point from the h-point theme. This information created an index file, which for each DEM grid cell contained a list of direction and distance to all h-points. The index file was important and reduced the time required for the interpolation process considerably.
A water level surface grid was interpolated based on the DEM, the index file, and the MIKE 11 results stored in the h-points attributes table. The interpolation in all grid cells was based on a distance-weighted interpolation of the nearest h-point found in each quadrant. The interpolation method was the same used for generating DEMs.
MIKE 11 GIS then asked to specify an output file name for the flood inundated map data. I chose fpsect3a.fim (the suffix .fim stands for flood inundation map). From the time step selection, I chose "21jan1959 12:00" and pressed "Ok":
MIKE 11 GIS interpolated the flood map. When the interpolation was complete, the flood map info file was displayed. For reference purposes, the file showed the information relating to the flood map:
Viewing the Floodplain at a Specific Point in Time
Selecting the Import/Display flood map button from the menu bar, the fpsect3a.fim file was chosen from the pop-up menu:
The FLOOD MAP theme was automatically joined with the DEM theme and a default legend was applied. The DEM shape file, dem3.shp, was added to the view. The following flood map was created from Arcview’s Layout menu for the time step of 21 January 1959 at 12:00:
Flood map animation was created directly in MIKE 11 GIS. By using the "Watch" button from the toolbar, "Prepare video animation " was selected from the pop-up menu. The time steps were then selected from the pull-down "Select item(s)" menu. I chose seven time steps and saved the file as flood1.flc. The animation can be played from the "Watch" button. Video player software called Autodesk Animator Player ran the animation as a separate executable program. Below depicts the animation as a GIF for purposes of displaying it on the Internet:
In conclusion there are many advantages to using DHI’s MIKE 11 software for flood analysis. As with many modeling techniques, there are some limitations.
MIKE 11 Advantages
Easy to use. Anyone with GIS experience that understands basic hydraulics and hydrology will find the user interface for MIKE 11 simple. The graphical editor interfaces allow you to visualize editing for increased accuracy.
Easy to expand model capabilities. Because of MIKE 11’s expandability, rainfall runoff, advection dispersion, and sediment transport models can easy be integrated into an existing model for future analysis. Additionally, DHI has successfully created a flood forecasting extension for the MIKE 11/MIKE 11 GIS environment.
Time series integration into GIS. Unlike other models, once a simulation has been successfully developed in MIKE 11, the results can be simulated spatially in Arcview at various points in time, from one results file.
Unsteady flow analysis. Since MIKE 11 is based on the St. Venant equations, unsteady flow analyses as well as steady flow analyses can be simulated.
Powerful Analysis Tools. With the incorporation of MIKE View in MIKE 11, numerous results graphs and simulations can be executed, increasing further the model’s analysis capabilities.
Spatial Analyst capability. Since MIKE 11 GIS does not have the Spatial Analyst extension incorporated in its software, it limits its capability to use grid coverage input for generating a DEM. Most of the United States DEM data is provided by the USGS in grid format, requiring data manipulation prior to using MIKE 11 GIS’s DEM Module.
MIKE11 developed using SI units. Converting English units data to SI units for use in the model created some difficulties. When providing results data to customers in the U.S., a conversion back to English units would be required. DHI is creating a new version with data in English units for broader use in the U.S.
DEM Accuracy. Highly accurate DEMs are required when developing flood maps in MIKE 11 GIS. Obviously the more accurate the DEM, the more accurate the flood map is. Topographic data with the accuracy of 30-meter grid cells or within the accuracy of a Triangulated Irregular Network (TIN) would be required for accurate flood map development.
Further model development. Additional development of the MIKE 11 Mill Creek model is still required.
Inputting Manning’s n values. The HEC-RAS geometry files have Manning’s n values for bed resistance for each cross section. In MIKE 11, Manning’s n, Manning’s M, or Chezy’s C can be used for bed resistance. Unlike HEC-RAS, the bed resistance is inputted in the MIKE 11 Hydrodynamic parameters editor and factors in which to multiply those values are inputted in the MIKE 11 Cross section editor. Mill Creek has variations in channel and bank surface types throughout its entire reach. The Manning’s n values could change the simulation results dramatically.
Integrate bridges into the MIKE 11 model. Unlike HEC-RAS, bridges are inputted into MIKE11 separate from the cross sections. Bridge dimensions are inputted as separate points along the reach’s extent. Numerous bridges need to be inputted into the model.
Compare to historic data. The MIKE 11 model needs to be compared to historic data of the 1959 flood for accuracy. Since limited modeling techniques existed in 1959, a way would be to compare water surface elevations at specific points along Mill Creek with MIKE 11 simulation results.
Develop Rainfall Runoff Model from Pete Andrysiak’s HEC-HMS Results. Incorporate the MIKE 11 Rainfall Runoff parameter based on Pete’s HMS results. The Louisville District could then use the model to compare flood reduction projects to existing conditions.
Expand Model Extent to entire Mill Creek Reach. Once all of the HEC-RAS cross section files can be converted into MIKE 11, expand the model extent to the entire Mill Creek reach.
Use Flood Duration Maps to Develop Flood Damage Assessment Tool. For future research, develop a spatial tool to assess flood damage based on flood duration. The MIKE 11 model could be the interface required to accomplish this.
(1) EPA BASINS Internet Site: http://www.epa.gov/ost/basins/
(2) Kjelds, J. T., “Flood Plain Management, Integrating Flood Models with GIS”, April 1998.
(3) Kjelds, J. T. and Jorgensen, G. H., “Flood Watch: GIS Based Management System for Flood Forecast Applications”, 1997.
(4) MIKE 11: A Modeling System for Rivers and Channels, Short Introduction Guide to Getting Started Tutorial, (March 1999), Danish Hydraulic Institute.
(5) MIKE 11 GIS: A Flood Modeling and Management Tool, Reference and User Manual (May 1998), Danish Hydraulic Institute.
(6) MIKE View: A Results Presentation Tool for MOUSE and MIKE 11, User Manual and Tutorial, (August 1998), Danish Hydraulic Institute.
(7) Roberson, J. A. and Crowe, C. T., Engineering Fluid Mechanics, (1985: Houghton Mifflin).
(8) USGS Water Resources Internet Site: http://water.usgs.gov/
Danish Hydraulic Institute – For providing the software package to CRWR for research purposes, especially to Mr. Jesper Kjelds, Director, U.S. Office, who answered any and all questions I had regarding the MIKE 11 software. Additional thanks to Mr. Stefan Szylkarski, DHI San Francisco Office, for providing the Visual Fortran program.
Center for Research in Water Resources (CRWR) – Thanks for providing a computer platform for the MIKE 11 software at the Pickle Research Center.
Corps of Engineers Louisville District – Thanks for answering all questions regarding the model area, providing the DEM, HEC-2 data, and other miscellaneous information.
USGS – Thanks for the Gage data and DEM.
Pete Andrysiak (Apache-6 in a previous life) – Thanks for 1) integrating me into the Mill Creek Project, and 2) our continuous cross-talk about the Mill Creek study.