Explore the cloudburst issue

On the evening of July 2, 2011, downtown Copenhagen was hit with 136 millimeters of rain (nearly 5 inches) in an hour and a half.[1] The overwhelmed sewer system backed up, causing thousands of basements to be flooded with a mixture of rain and wastewater. In the aftermath of the storm, insurance companies paid about $1 billion in compensation to some 90,000 policyholders.[2]

The Danish Meteorological Institute defines an extreme rainfall event, or cloudburst, as 15 millimeters of rain (about 0.6 inches) in a 30-minute period.[3] The Copenhagen downpour qualified with buckets of water to spare.

Present, future, past

Standard climate models predict a rainstorm of such intensity once in a thousand years.[4] Yet a number of other extreme rain events have recently occurred in the same area. In August 2010, 99 millimeters of rain fell on Copenhagen within a few hours. Statistically, this was a 50-year event.[5] In August 2013, the town of Høng— some 80 kilometers west of the capital—received 15 millimeters of rain in 20 minutes. The nearby village of Rye received 46 millimeters in an hour.[4] On August 31, 2014, Copenhagen was drenched again with 119 millimeters in the span of a few hours. This was a 480-year event.[4] It seems that climate change has upset standard predictive models. Cloudbursts in Denmark should probably now be expected with greater frequency.

Both the city of Copenhagen and the entire country of Denmark are actively preparing for future cloudbursts. At national and municipal levels, plans being considered or implemented include new surface channels to lead rainwater away from buildings; upgraded sewer systems; and pumping stations to pump excess water to the sea. The City of Copenhagen's Cloudburst Management Plan 2012 considers these issues in more detail.

A simple but critical planning precaution is not to build in places where water flows in but doesn't flow out. Low-lying areas—variously called depressions, sinks, or hollows—are everywhere. They come in all shapes and sizes. On farmland and in parks, fields that look flat may actually contain shallow depressions that trap rainfall.

Soccer field
A soccer field in Copenhagen after a rainfall. Image © Thomas Balstrøm.

Some residential areas are situated in places where the low-lying landscape goes unnoticed under dry conditions—it only becomes apparent when so much rain falls that that the ground can't absorb it and the sewer system can't take it away. Roads and other types of infrastructure, such as subway stations or railroad tracks, are also vulnerable.

Sunken taxi
A flooded highway underpass in Copenhagen after the August, 2014, cloudburst. Image © Johanna Sørensen.

Almost all low areas, whether on cultivated or developed land, pose some risk. On farmland, they may be a threat to crops and equipment. In residential areas, unless buildings are constructed on high bases or pillars, they lead to flooding.

Structures that are not located in depressions may still be at risk. If so much rain falls that a depression overflows at its so-called pour point, adjacent areas may be flooded.

Pour point sketch

Historically, the risks of building in low-lying terrain weren't foreseen. Perhaps they couldn't have been. But in Denmark, at least, several developments are located on land that was drained after the 1850s. Such areas were once wetlands or small lakes and became fine agricultural land when ditched or reclaimed. If the drainage was efficient, these areas may have gone without flooding for long periods in the absence of heavy rain. Unfortunately, over the past decades, many residential developments were constructed in such areas. Today, these are the low-lying areas where water pools during cloudbursts. Homeowners living on such abandoned farmland face the challenge of frequent floods.

Map comparison
Left: a topographic map from 1860. Green areas are low-lying meadows and blue lines are drainage ditches. Right: the same area in 1990 is a residential development with no farmland left. The map area is approximately 1.2 square kilometers. This image contains data from the Danish Geodata Agency.

Bluespot maps

The combination of the GPS system and lidar technology has made it possible in recent years to create detailed, high-resolution landscape models. Two Danish companies, NIRAS and COWI, have led the way in creating bluespot maps from such models.

Although "bluespot" is simply another term for a depression or sink in the landscape, it puts a special emphasis on flood risk: a bluespot is an area likely to fill or overflow in a cloudburst, endangering structures located within or near it.

Bluespots
A bluespot map of the low-lying area shown above. This image contains data from the Danish Geodata Agency.

In these lessons, you'll create bluespot maps similar to those used by the Danish government. The method for making the bluespot map is presented in the form of a geoprocessing model: a sequential chain of GIS analysis tools. You will not build the model yourself, but you will explore it, run it, and map the results. In fact, two models are presented: a basic model that identifies bluespots (Lesson 3) and an extended model that also calculates the volumes of bluespots and the amounts of rainfall that would be required to fill them (Lesson 4).

The primary value of both models is in the identification of low-lying areas. For planners and developers, bluespot maps can be used to make informed decisions about where not to build or where development should be accompanied by special landscape architecture. In areas that are already developed, bluespot maps can be used to prioritize areas for better climate-proofing in some of the ways discussed earlier.

Bluespot maps can be tailored to other uses as well. For example, the Identify Bluespot Fill Up Values model calculates the varying depths of each bluespot across its area. This information could be used to create "rubber-boot" maps for search-and-rescue crews, helping them distinguish areas that can be reached on foot from those that require boats.

Assumptions and limitations of the models

A number of assumptions and limitations affect the precision and certainty of the model results. Some of these factors, such as the water balance equation that determines how much water can be drawn off by the sewer system and natural processes, are discussed in Lesson 4. Although the model has undergone a critical review, errors may be present. If such errors are discovered, please inform the author.

Neither the author nor Esri may be held responsible for conclusions drawn from the analysis. In particular, the results may not be used to blacklist properties. At most, they should be used in careful discussions among planners to focus attention on areas that may face drainage problems in the event of a cloudburst.

Next, you'll open a map document and take a closer look at the Gentofte study area.


Work with the project data

Previously, you learned about recent cloudbursts in Denmark and how bluespot maps identify areas vulnerable to flooding. Next, you'll download the data you need for the project and explore the study area: the municipality of Gentofte (Jen - Toft) near Copenhagen.

Gentofte

Old maps reveal that many of the built-up areas in Gentofte were filled with water in earlier times; some thousands of years ago they were highly active river channels transporting melting glaciers to the sea. Your starting data will be a high-resolution digital elevation model (DEM) of the area and a layer of building footprints.

Download the data

The models and data for the project are stored on ArcGIS Online.

  1. Go to the Find Areas at Risk of Flooding in a Cloudburst group.
  2. Click the thumbnail image of the Cloudburst Models and Data item to download the data.
    Note:

    The data download size is about 40 megabytes.

    The data is downloaded in a file called Cloudburst.zip to your default downloads folder. (For example, to C:\Users\<profile>\Downloads.)

  3. Extract the data to your C:\ drive or a folder of your choice.

Start ArcMap

You'll confirm that you have a recent version of ArcMap and access to the ArcGIS Spatial Analyst extension in the next steps.

  1. Start ArcMap.
    Note:

    If you don't have ArcMap, you can get it by purchasing ArcGIS Desktop from the Esri Store.

  2. If the ArcMap - Getting Started window opens, click Blank Map to select it and click OK.

    Getting Started window

    Note:

    The Getting Started window contains a blank map template. (It also shows recent maps if you've made any.)

    A blank untitled map document opens. Before you continue, you'll confirm that you have the proper software configuration.

  3. At the top of the ArcMap window, on the main menu, click Help and choose About ArcMap.

    About ArcMap

  4. Confirm that your version of ArcGIS Desktop is 10.3 or later.
    Caution:

    If your ArcMap version is earlier than 10.3, you won't be able to do the lessons.

  5. Click OK in the About ArcGIS for Desktop window.
  6. On the main menu, click Customize and choose Extensions.
  7. In the Extensions window, click the check box next to Spatial Analyst and close the Extensions window.

    Extensions

    Caution:

    If the Spatial Analyst extension isn't available, you won't be able to do the lessons.

  8. Optionally, maximize the ArcMap application window.

Make a folder connection to your project data

You access maps and data through folder connections in the Catalog window.

  1. On the upper right side of the ArcMap application window, click the Catalog tab.
    Tip:

    Alternatively, on the ArcMap main menu, click Windows and choose Catalog.

  2. Click the Auto Hide pin to keep the window open.

    Default Catalog window

  3. On the Catalog window toolbar, click the Connect to Folder button.

    Connect to Folder button

  4. In the Connect to Folder window, browse to the location of the Cloudburst folder (for example, Computer > C: > Cloudburst). Click the folder to select it and click OK.
  5. In the Catalog window, expand the Cloudburst folder.

    The Cloudburst folder contains a folder, four geodatabases, and four toolboxes.

  6. Expand the BluespotModels_Metric toolbox.

    Expand Cloudburst folder

    The toolbox contains two geoprocessing models.

    • Identify Bluespots finds bluespots in the study area and selects buildings within or adjacent to them. The results show, at a very general level, which structures face increased flood risk in a cloudburst. The model does not attempt to quantify this risk. You'll run this model in Lesson 3.
    • Identify Bluespot Fill Up Values does the same analysis and also calculates the volume of each bluespot and the area of its local watershed or catchment (the basin that drains water to it). From this information, the model calculates how much rain it would take to fill the bluespot to capacity. This model allows for some ranking of flood risk. Other factors being equal, bluespots that need less rainfall to fill up pose a greater flood risk to buildings. You'll run this model in Lesson 4.

    The contents of the geodatabases and toolboxes are described in the following table:

    NameContents

    Maps and Layers

    Map document for visualizing input data and results; layer file for symbolizing results.

    Inputs.gdb

    Starting data for models.

    Outputs_Bluespots.gdb

    Empty. Holds outputs of the Identify Bluespots model.

    Outputs_BluespotsFillUp.gdb

    Empty. Holds outputs of the Identify Bluespot Fill Up Values model.

    ResourceData.gdb

    Additional feature and raster data to explore and analyze.

    BluespotModels_Metric.tbx

    Geoprocessing models used in Lessons 3 and 4.

    BluespotModels_Metric_NoBuildings.tbx

    Model versions for input datasets where building footprints are not available.

    BluespotModels_US.tbx

    Model versions for input datasets that use United States measures.

    BluespotModels_US_NoBuildings.tbx

    Model versions for input datasets that use United States measures and do not have building footprints.

Open the Gentofte map

Before running the models, you'll familiarize yourself with the study area.

  1. In the Catalog window, expand the Maps and Layers folder. Double-click the Gentofte.mxd map document to open it.
    Tip:

    If you're prompted to enable hardware acceleration, click Yes.

  2. Optionally, unpin the Catalog window.
  3. On the ArcMap main menu, click Bookmarks and choose Study Area.

    Gentofte map

    Buildings in Gentofte municipality are drawn in purple. The municipal boundary is outlined in red.

  4. Zoom to the Ordrup bookmark.

    Ordrup

    The study area is highly developed but includes many water bodies, parks, and green spaces.

  5. Optionally, on the Tools toolbar, click the Pan tool and the Fixed Zoom In and Fixed Zoom Out buttons to move around the study area.

    Pan tool

  6. When you're ready, zoom to the Study Area bookmark.
  7. In the Table of Contents, check the box for the DEM layer to turn it on.

    DEM

    The digital elevation model (DEM) displays elevation values for the study area. In the Table of Contents, you see that elevation values range from a low of -3.23 meters to a high of 56.58 meters. The values are symbolized with a brown color ramp. The lighter the shade of brown, the higher the elevation.

  8. Zoom to the Gentofte Lake bookmark.
  9. On the Tools toolbar, click the Identify tool.

    Identify tool

  10. Click any location on the map.

    The Identify window opens and identifies the Municipal Boundary layer, the top-most layer in the Table of Contents. This isn't the layer of interest. You want to explore the DEM layer.

  11. At the top of the Identify window, change the Identify from setting to DEM.
  12. On the map, click a few locations in Gentofte Lake. (Drag the Identify window out of the way if necessary.)

    Identify DEM

    Wherever you click, there is an elevation value in the DEM layer. As you click the lake, you should see consistent values of 16.82 (meters) because the lake surface has constant elevation.

  13. Click some other locations around the edges of the lake.

    The results vary depending on where you click. If you click a building, the value represents the ground elevation at that location.

  14. Close the Identify window and zoom to the Study Area bookmark.

View the DEM metadata and layer properties

The DEM is a raster dataset; in other words, a grid of cells or pixels, like a digital image. With ArcGIS hydrology tools, you can analyze the DEM to find bluespots, calculate their size and volume, and delineate areas that contribute water flow to them in a cloudburst.

To study the location of bluespots within Gentofte, it is necessary to analyze the DEM to a wider extent than the municipal boundary because some bluespots and their local watersheds may lie partly in neighboring municipalities. To ensure that all bluespots and watersheds are identified correctly, the DEM for Gentofte municipality was expanded by a 5-kilometer buffer zone.

To learn the specific properties of this DEM, you can look at its metadata and layer properties.

  1. In the Table of Contents, right-click the DEM layer name, point to Data, and choose View Item Description.

    DEM Item Description

    Under the Description heading, you see that the DEM resolution is 1.6 meters and the vertical accuracy is 0.05 meters. This means that each cell in the dataset represents 1.6 square meters of ground and each cell's elevation value is accurate to within 0.05 meters. This high-resolution data was collected with lidar.

  2. Close the item description.
  3. In the Table of Contents, right-click the DEM layer name and choose Properties.
  4. In the Layer Properties dialog box, click the Source tab. Scroll down to the Spatial Reference heading.

    DEM Spatial Reference

    The DEM dataset, like the other lesson data, is projected to the Universal Transverse Mercator system and based on the ETRS 1989 datum (the most frequently used spatial reference in Denmark). This spatial reference system is used for the input and output data in the models.

  5. Close the Layer Properties dialog box.
  6. On the Standard toolbar, click the Save button to save the map.

Next, you'll explore and run the Identify Bluespots model. If you're experienced with ModelBuilder and familiar with hydrology tools, or if you're working in a classroom setting with an instructor, you may prefer to go directly to Assess flood risk to buildings.


Find bluespots and affected buildings

Previously, you looked at the Gentofte study area. In this lesson, you'll examine and run the Identify Bluespots geoprocessing model. This model analyzes the DEM with hydrology tools to find bluespots. It then compares the locations of bluespots to the locations of existing buildings and selects buildings that are situated inside or adjacent to bluespots. In general, these buildings are at greater risk of being flooded in a cloudburst.

A geoprocessing model is a collection of input data and tools organized as a workflow and run as a single operation.

Open the model for editing

Models are saved in toolboxes. You can work with a model either by editing or opening it. Editing the model opens the ModelBuilder window. In the ModelBuilder window, you interact fully with the model: you can build, configure, view, test, and run it. Opening the model opens a dialog box that lets you run the model as a geoprocessing tool. Next, you'll work with the model in the ModelBuilder window.

  1. If necessary, start ArcMap and open the Gentofte map document.
    Tip:

    If you're starting a new ArcMap session, the map document may appear in the ArcMap - Getting Started window as a recent map.

  2. Open the Catalog window. If necessary, expand the Cloudburst folder and the BluespotModels_Metric toolbox.
  3. Right-click the Identify Bluespots model and choose Edit.

    Edit model

    The ModelBuilder window opens. The data and tools in the model are represented as colored geometric shapes, which are connected to build a workflow.

    The basic working unit of a model is a process. A process is an input dataset (blue) connected to a tool (yellow) connected to an output dataset (green). Input and output model elements are variables because you can open their properties and change their pathnames.

  4. At the top of the ModelBuilder window, click the Windows menu and choose Overview.

    Overview window

    An Overview window opens and you can see what portion of the full model diagram is visible in the ModelBuilder window. You can leave the Overview window open, or close it if you prefer.

    Conceptually, the model does the following three main things:

    • It finds bluespots on the DEM.
    • It processes this result and the Buildings layer so the data is in the proper format for making a spatial selection.
    • It selects the buildings on the map that lie within or adjacent to bluespots.

    Model three parts

    The basic function of each tool is labeled on the model. For a more detailed breakdown of the workflow, read the model overview.

    Tip:

    To navigate the model, you can use the scroll bars or pan and zoom tools in the ModelBuilder window. You can also drag the blue rectangle in the Overview window.

Review model element settings

Before running the model, you'll look at some of the settings for model elements that affect their behavior and output.

  1. On the left side of the model, right-click the DEM input data element and choose Open.

    Open DEM

    The DEM dialog box shows you the path to the dataset on disk. If you extracted the data to your C:\ drive, your path will match the one shown. If you extracted it to a different location, your path will be different.

    DEM path
    Tip:

    To use a different DEM as input, browse to it from the dialog box.

  2. Click Cancel to close the DEM dialog box.

    The DEM element is connected by arrows to two tools: the Fill tool and the Fill (2) tool. These are two instances of the same tool (the model requires that they have unique names). The gray labels under the tools describe what they do.

  3. Right-click the Fill tool and choose Open.

    Fill dialog box

    The input dataset is the DEM. The output dataset name is SmallSinksFilled. The output path is set with an inline variable: %Output Workspace% (explained below). The inline variable allows you to define a pathname once and apply it to all output datasets in the model. If you decide to change the output path, you only need to change it in one place.

    In addition to input and output data variables, models can also have value variables. These are usually tool parameters that you might change if you run the model with different data. For example, the Z limit parameter on the Fill tool is set with a variable rather than with a specific value. Whatever value is set in the Vertical accuracy (meters) variable is the value that the tool will use.

  4. Click Cancel to close the Fill dialog box.
  5. Right-click the Output Workspace element and choose Open.

    Output Workspace

    This is the variable that defines the output path for all datasets in the model. Wherever its name (%Output Workspace%) is used instead of an explicit pathname, the output location will be replaced by the pathname stored in the variable. If you change this pathname, it will be changed for all output datasets that reference the variable.

  6. Close the Output Workspace dialog box.
  7. Open the Vertical accuracy (meters) element.

    Vertical accuracy

    This is the variable that defines the Z limit value for the Fill tool. It is set to 0.05 meters, which is the appropriate value for the Gentofte DEM, as you saw in the DEM item description in Lesson 2.

  8. Close the Vertical accuracy (meters) dialog box.

    If you want to run the model as a geoprocessing tool, you designate variables as model parameters (identified by the capital letter P).

    Parameters

    If the model is run as a geoprocessing tool, these variables will be exposed on the tool dialog box.

  9. Move the ModelBuilder window so you can see the Catalog window.
  10. Right-click the Identify Bluespots model and choose Open.

    The Identify Bluespots tool dialog box opens with four parameters. Notice that they correspond with the variables in the model that are marked with a capital letter P. These are values that a user can change before running the model as a geoprocessing tool.

  11. Click Cancel to close the tool.
  12. In the model window, right-click the DEM element.

    Model Parameter

    Elements are set as model parameters by toggling the Model Parameter menu command.

  13. Click some white space on the model to close the menu.
  14. Open the Small Sinks Filled element.

    Small Sinks

    Output data variables hold a pathname and file name.

  15. Close the Small Sinks Filled dialog box.
    Tip:

    You can also see the values of tools and variables by hovering over them with the Select tool.

  16. Right-click the Small Sinks Filled element.

    Small Sinks menu

    By default, output datasets, except for the last output of the model, are flagged as intermediate. Intermediate data is data that you don't need once you have the model results. When you run the model as a geoprocessing tool, intermediate data is deleted automatically. When you run the model from the ModelBuilder window, you can delete intermediate data from the Model menu.

    The Small Sinks Filled dataset is not flagged to be added to the display. If it were, a layer would be added to the ArcMap Table of Contents when the model finished running.

  17. If necessary, click some white space on the model to close the menu.

Review model environment settings

Apart from the settings for individual elements, there are environment settings that apply to the entire model.

  1. At the top of the ModelBuilder window, click the Model menu and choose Model Properties.
  2. On the Identify Bluespots Properties dialog box, click the Environments tab.

    Environment settings have been made for Output Coordinates, Processing Extent, and Raster Analysis.

  3. Click the Values button.
  4. In the Environment Settings dialog box, click the Output Coordinates heading to expand it.

    Output Coordinates

    Output datasets created by the model will be projected to the same coordinate system as the input dataset. For the Gentofte data, this will be UTM Zone 32 North based on the ETRS 1989 datum.

    Tip:

    If you use your own data with the model, a good way to make environment settings is to specify a layer or dataset that has appropriate values. For example, clicking the Output Coordinate System arrow presents a list of layers in the open ArcMap document. Choosing one of these layers applies its coordinate system to the environment setting. (You can also browse to a dataset on disk.) In the same way, you can make other environment settings by choosing layers, datasets, or model variables from lists.

  5. Click the Output Coordinates heading again to collapse it.
  6. Click the Processing Extent heading to expand it.

    Processing Extent

    All raster datasets created by the model will have the bounding coordinates shown in the boxes. (These are the bounding coordinates of the DEM dataset.) Their lower left corners will also snap, or align, to the lower left corner of the DEM.

  7. Click the Processing Extent heading again to collapse it.
  8. Click the Raster Analysis heading to expand it.

    Raster Analysis

    Raster datasets created by the model will have a cell size of 1.6 meters. This setting, together with the Processing Extent settings, ensures that all raster datasets in the model will align and have spatially coincident cells. This is important for calculations involving multiple rasters.

    Before running this model with different input data, you would need to change the environment settings appropriately for your data.

  9. Close the Environment Settings dialog box and the Identify Bluespots Properties dialog box.

Run the model

Now you'll run the model. When the model is finished, two layers will be added to the open Gentofte map document: a layer of bluespots and a layer of buildings that lie within or adjacent to bluespots.

  1. On the ModelBuilder toolbar, click the Validate Entire Model button.

    Validate button

    Validating the model confirms that all processes are ready to run.

    A model process (and, by extension, the model as a whole) can be in one of four states. If a process is ready to run, its elements are connected by arrows and filled in with solid colors. If a process is not ready to run, the tool and output elements are white. This means that some required information, such as a tool parameter, is missing. While a process is running, the tool element is red. After a process has been run, its tool and output elements are shadowed. (If the model has been run previously, validation prepares it to run again.)

  2. Click the Run button.

    Run button

    As the model runs, a window shows its progress. The tool that is currently running turns red in the ModelBuilder window. (The Fill operation takes a few minutes.)

    Process states

    When the model is completed, a message appears on the progress window.

    Progress window

  3. Close the progress window.

    In the ModelBuilder window, the processes are shadowed, showing that they have been run.

  4. On the ModelBuilder toolbar, click the Save button. Minimize the ModelBuilder window or move it away from the ArcMap window.
  5. Open the Catalog window and expand the Cloudburst folder if necessary.
  6. Right-click Outputs_Bluespots.gdb and choose Refresh. Expand the geodatabase if necessary.

    Outputs_Bluespots geodatabase

    The output data has been written to the Outputs_Bluespots geodatabase, as specified in the model's Output Workspace variable. The two important datasets are Bluespots and BuildingsTouchBS (Buildings Touching Bluespots). These show you where there are bluespots in Gentofte and which buildings are inside or adjacent to them.

    In ArcMap, two new layers appear at the top of the Table of Contents: Buildings Touching Bluespots and Bluespots.

    Table Of Contents

  7. If the two new layers are not added to the map automatically, drag Bluespots and BuildingsTouchBS onto the map from the Outputs_Bluespots geodatabase. Rename the BuildingsTouchBS layer Buildings Touching Bluespots.

Map results

You'll look at the Buildings Touching Bluespots attribute table to see how many buildings in the study area are at risk. Then you'll symbolize the map layers to draw attention to the buildings that touch bluespots. You can symbolize the map any way you want; these instructions are just suggestions. The dark gray canvas basemap is used to minimize background distraction, but if you want to explore the area closely, you may prefer to use an imagery basemap.

  1. In ArcMap, zoom to the Study Area bookmark if necessary.
  2. In the Table of Contents, right-click the Buildings Touching Bluespots layer and choose Open Attribute Table.

    Table

    Over 12,000 buildings lie within or adjacent to a bluespot. If you open the attribute table for the Buildings layer, you'll see that there are about 31,000 buildings in Gentofte. This means that almost 40 percent of the buildings have some level of flood risk in a cloudburst.

  3. Close the attribute table.
  4. At the bottom of the Table of Contents, right-click the Basemap layer and choose Remove.
  5. Remove the DEM layer as well.
  6. Turn off the Buildings layer but don't remove it.
  7. On the Standard toolbar, click the arrow next to the Add Data button and choose Add Basemap.

    Add Basemap

  8. In the Add Basemap window, click the Dark Gray Canvas basemap and click Add. Close the Geographic Coordinate Systems warning.
  9. In the Table of Contents, click the symbol for the Buildings Touching Bluespots layer to open the Symbol Selector.
  10. In the Symbol Selector, change Fill Color to Yucca Yellow and Outline Width to 0. Click OK.

    Yucca Yellow

  11. For the Bluespots layer, change Fill Color to Steel Blue and Outline Width to 0. Click OK.

    Steel Blue

  12. Open the Symbol Selector for the Municipal Boundary layer and click Edit Symbol.
  13. In the Symbol Property Editor dialog box, click Outline.
  14. In the Symbol Selector, scroll down in the list of symbols and click the Boundary, City symbol.

    Boundary

  15. Change the color to Gray 10% and Width to 1.5. Click OK.
  16. Click OK on the Symbol Property Editor and click OK on the Symbol Selector.
  17. In the Table of Contents, drag the Municipal Boundary layer above the Buildings Touching Bluespots layer.

    Final Table Of Contents

  18. On the main menu, click File and choose Save As.
  19. Browse to your Cloudburst > Maps and Layers folder. Save the map as Gentofte Bluespots.

    Final map

  20. Restore the ModelBuilder window and save the model.
  21. Close the ModelBuilder window.

A casual look at the map shows you that buildings touching bluespots are located throughout Gentofte municipality, not concentrated in any particular area. Keep in mind that flooding may affect other kinds of infrastructure as well as buildings. You may want to add some datasets from the ResourceData geodatabase, such as roads and railroads, to see where they lie in relation to bluespots.

The Identify Bluespots model does not evaluate levels of risk to buildings. However, not all bluespots pose the same risk. How quickly a bluespot fills and overflows in a cloudburst depends on its depth, its area, and the size of the catchment, or local watershed, that contributes rainfall to it.

Next, you'll run a more sophisticated version of the model that estimates how much rainfall would fill each bluespot to its pour point. This information will help you make a better estimate of the risk to buildings.

To run the model with different input datasets, choose the appropriate model from those provided. Versions of both models have been created for use with United States standard units. Versions have also been created for scenarios in which building footprints are not available.

Here are the basic steps you need to complete before running the model:

  • Point the input layers to your own datasets.
  • Set the output workspace variable.
  • Set the vertical accuracy variable. (To get the vertical accuracy, you'll need to check the metadata of your DEM and perhaps do additional research to find out how the source data was created.)
  • Check the model environment settings to make sure the output coordinates, processing extent, and raster analysis settings are appropriate.

Assess flood risk to buildings

Previously, you ran a model to identify bluespots. The model not only identifies bluespots on a DEM, it also calculates how much rainfall is needed to make each bluespot fill up in a cloudburst. This knowledge allows you to make a better estimate of flood risk: a building located in a bluespot that fills up quickly is at greater risk than a building in a bluespot that fills up slowly.

The model is based on the hydrologic concept that every sink (bluespot) in the landscape has a catchment or local watershed: the area that contributes flow to that bluespot and no other. By calculating the volume of a bluespot and the area of its watershed, you can determine how much precipitation is needed to fill the bluespot: it's the volume divided by the area. For example, if a bluespot's volume is 500 m3 and its watershed is 10,000 m2, the precipitation needed to fill the bluespot to its pour point is 500 m3 / 10,000 m2 = 0.05 m = 50 mm.

In fact, not all the precipitation that falls in the watershed flows into its bluespot because perfect run-off conditions don't exist. However, run-off conditions in a cloudburst are pretty close to perfect. The water balance equation P = I + E + Ao + Au + Q states that precipitation (P) is equal to the interception caught by vegetation (I) plus evaporation (E) plus surface run-off (Ao) plus soil infiltration and sewer systems (Au) plus deposition in local reservoirs (Q). In this context, local reservoirs means bluespots.

In a cloudburst, interception, evaporation, and soil infiltration can be assumed to be zero. The maximum capacity of Danish sewer systems in residential areas is approximately 40 millimeters of rainfall per day. Keeping the focus on 1 hour of precipitation, and making the generous assumption that the daily capacity could actually be handled within 1 hour, the value for Au can thus be set to 40. Surplus runoff (Ao) will not be a factor in the equation until after the bluespots fill up. Therefore, for the purposes of calculating the fill up value, the formula can be simplified to P = 40 + Q or Q = P – 40 millimeters per hour. If 90 millimeters of rain falls in 1 hour, the sewer system will carry away 40 millimeters, while 50 millimeters will flow into the bluespot—filling it either partially or entirely. If the bluespot is filled to its pour point, the runoff (Ao) will enter the next downstream sink, lake, river, or sea.

While this model presents a better evaluation of risk, it still leaves out many factors. For example, the contribution of surplus runoff to downstream bluespots is not considered. Neither is the base elevation of buildings within bluespots. For example, if a building is located near the bottom of a bluespot, it may be flooded before the bluespot fills to capacity. Structural factors, such as whether buildings have basements or elevated foundations, are also not considered.

Open the model for editing

You'll work with the model in the ModelBuilder window.

  1. If necessary, start ArcMap and open the Gentofte map document (not Gentofte Bluespots).
  2. Open the Catalog window. If necessary, expand the Cloudburst folder and the BluespotModels_Metric toolbox.
  3. Right-click the Identify Bluespot Fill Up Values model and choose Edit.

    The model opens in the ModelBuilder window. You can only see part of the model.

  4. At the top of the ModelBuilder window, click the Windows menu and choose Overview.

    An Overview window opens and you can see what portion of the full model diagram is visible in the ModelBuilder window. You can leave the Overview window open, or close it if you prefer.

    Overview

    This model accomplishes more than the Identify Bluespots model, so of course it's longer. After finding bluespots, it calculates their volumes and watersheds. Then it uses this information to determine the amount of rainfall needed to fill each bluespot. The model may seem less complicated when you see that many of its processes are table operations: adding fields, joining fields, and calculating field values.

    The basic function of each tool is labeled on the model. For a more detailed breakdown of the workflow, see the model overview.

    Tip:

    To navigate the model, you can use the scroll bars or pan and zoom tools in the ModelBuilder window. You can also drag the blue rectangle in the Overview window.

Run the model

This part of the lesson assumes that you are familiar with ModelBuilder functionality. If you are not familiar with variables, inline variable substitution, model parameters, model environment settings, and so on, see the previous lesson.

  1. On the ModelBuilder toolbar, click the Validate Entire Model button.

    Validate button

  2. Click the Run button.

    Run button

    As the model runs, a window shows its progress. When the model is done, a message appears on the progress window.

    Progress window

  3. Close the progress window.
  4. On the ModelBuilder toolbar, click the Save button. Minimize the ModelBuilder window or move it away from the ArcMap window.
  5. Open the Catalog window and expand the Cloudburst folder if necessary.

    The output data has been written to the Outputs_BluespotsFillUp geodatabase.

  6. Right-click Outputs_BluespotsFillUp.gdb and choose Refresh. Expand the geodatabase if necessary.

    Outputs

    In ArcMap, three new layers were also added to the Gentofte map document: a layer of bluespots, a layer of buildings that touch bluespots, and a layer of bluespots that touch buildings.

    Table Of Contents

  7. If the new layers are not added to the map automatically, add them from the Outputs_BluespotsFillUp geodatabase. The dataset names are as follows:
    • BSPolyDissolved
    • BuildingsTouchBS
    • BSTouchBuildings

Map results

You'll look at the fill-up values for the bluespots and symbolize the Bluespots Touching Buildings layer on the basis of these values. The symbology will indicate higher levels of risk for bluespots that fill up more quickly. This part of the lesson assumes that you are familiar with using ArcMap.

  1. In ArcMap, zoom to the Study Area bookmark if necessary.
  2. Open the attribute table for the Bluespots Touching Buildings layer.

    Table

    The Volume field contains the volume of each bluespot (in cubic meters). The FillUp field contains the amount of rainfall (in millimeters) needed to fill each bluespot in a cloudburst.

  3. Sort the FillUp field in ascending order.

    In general, the smaller the fill up value, the greater the risk. Bluespots with extremely low fill up values, however, are usually small and shallow. They might overflow so quickly into the next downstream bluespot that it would be hard to notice them.

  4. View the statistics for the FillUp field.

    Statistics

    The Statistics window shows the minimum and maximum values. On the Frequency Distribution chart, you can see that almost all the fill-up values are at the very low end of the range.

  5. Close the Statistics window and close the attribute table.
  6. In the Table of Contents, remove the following layers: Basemap, DEM, Buildings, and Bluespots.
  7. Turn off the Buildings Touching Bluespots layer but leave it in the map.
  8. Add the Light Gray Canvas basemap. Close the Geographic Coordinate Systems warning.
  9. Drag the Municipal Boundary layer above the Bluespots Touching Buildings layer.
  10. Open the Symbol Selector for the Municipal Boundary layer and click Edit Symbol.
  11. Change the symbol outline to the Boundary, City symbol. Change its color to Gray 60% and make its width 1.5.

    New symbols

    The bluespot fill-up values don't tell the whole story, of course. As discussed earlier, the sewer system will carry away a significant amount of overflow. In this model, it's assumed the system can handle 40 millimeters of rain in an hour, although this assumption is probably too generous. If you run the model with your own data, consider your local sewer system's capacity (if applicable) to estimate a reliable value for this component.

    Note:

    The 40 millimeter value represents maximum efficiency. That value may not be attainable by all parts of the system. Furthermore, the system may be impaired by pre-existing blockages or blockages caused by debris flow during a cloudburst.

    You'll symbolize the Bluespots Touching Buildings layer by importing symbology from a layer file.

  12. Open the layer properties for the Bluespots Touching Buildings layer. On the Symbology tab, click Import.
  13. On the Import Symbology dialog box, browse to the Cloudburst > Maps and Layers folder. Add Adjusted Fill Up Values.lyr.
  14. On the Import Symbology dialog box, click OK.
  15. On the Import Symbology Matching dialog box, confirm that Value Field is set to FillUp and click OK.

    Import Symbology

    On the Symbology tab of the Layer Properties dialog box, note that the values are labeled 40 millimeters higher than the actual range values from the attribute table.

    Bluespots symbology

    The labels for the fill-up values have been adjusted by 40 millimeters to account for the sewer system capacity. (In other words, values of 0 - 20 mm in the FillUp field will be labeled as 40 - 60 mm when the symbology is applied.) You could also add a field to the Bluespots Touching Buildings layer and calculate its values to [FillUp] + 40.

  16. Click OK on the Layer Properties dialog box.

    Bluespot risk

    During the Copenhagen cloudburst of July, 2011, 136 millimeters of rain fell in 1.5 hours. That equates to about 90 millimeters in an hour. If a comparable event occurred again, bluespots in the two or three highest risk categories (red, dark orange, orange) would fill up.

  17. Drag the Buildings Touching Bluespots layer above the Bluespots Touching Buildings layer and turn it on.
  18. Change its symbol outline width to 0. Use the default fill color or choose another color.
  19. Zoom in on different parts of the map to compare buildings and bluespots.

    Buildings

    You'll see several areas that look susceptible to flooding. You'll also see areas where there are buildings but no apparent bluespots. There are, in fact, bluespots in those locations, but they are not symbolized because their fill-up values are more than 140 millimeters adjusted. It's very unlikely that these bluespots would overflow, even in a cloudburst.

    Note:

    The calculations assume all bluespots are empty, but this is not always the case. Some bluespots are permanent water bodies, such as Gentofte Lake.

  20. Zoom to the Study Area bookmark.
  21. Save the map as Gentofte Bluespots Fill Up in your Cloudburst > Maps and Layers folder.

You may want to do further analysis to find how many buildings touch bluespots of different risk levels. For example, you can select bluespots in the highest risk category with the attribute query "FillUp >= 0 AND FillUp <= 20" on the Bluespots Touching Buildings layer. You can then use Select By Location to select buildings that intersect the selected set of bluespots.

You may also want to add the Watersheds layer to examine the relationship between bluespots and the areas that contribute flow to them. A good way to symbolize the Watersheds layer is with the Discrete Color renderer.

Cloudbursts affect low lying parts of the infrastructure as well as buildings. Although the model doesn't use infrastructure as an input, you can easily add a layer of roads or railroads to the map. Feature classes of roads and railroads are included in the ResourceData geodatabase.

Flooded road
A flooded road in Copenhagen after the August, 2014, cloudburst. Image © Johanna Sørensen.

The model should be globally applicable because it's based on a simple investigation of the terrain surface and the locations of buildings. An underlying assumption is that the DEM generates a reliable flow direction surface; this assumption is better founded for higher resolution DEMs.

To run the model with different input datasets, choose the appropriate model from those provided. Versions have been created for use with United States standard units. Versions have also been created for scenarios in which building footprints are not available.

Here are the basic steps you need to complete to use the model with your own data:

  • Point the input layers to your own datasets.
  • Set the output workspace variable.
  • Set the vertical accuracy variable. (To get the vertical accuracy, you'll need to check the metadata of your DEM and perhaps do additional research to find out how the source data was created.)
  • Check the model environment settings to make sure the output coordinates, processing extent, and raster analysis settings are appropriate.

The Identify Bluespot Fill Up Values model demonstrates a workflow for identifying landscape sinks (bluespots) and making a quantitative assessment of flood risk to buildings in the event of a cloudburst.

The model uses ArcGIS geoprocessing tools to derive bluespots and their local watersheds. Fill-up values are calculated by dividing a bluespot's volume by the area of its watershed. Assumptions are also made about how much overflow water can be handled by the sewer system. These values and assumptions should be viewed with some degree of caution.

Are the model results sufficient to identify critical flood-risk thresholds for individual buildings? The answer is both yes and no. In the absence of building attributes, the best you can do is establish a worst-case scenario by assuming that the critical flood level for a building is at its base elevation. This is in fact true for many residences, warehouses, supermarkets, and office complexes. However, it's not true for all buildings. Actual water entry levels may be higher than assumed for buildings with high foundations or buildings that are raised above the ground. Conversely, the level may be lower than assumed for buildings with basements. (The Danish Building and Dwelling Register has information about whether buildings have basements, but that information is not included in the Buildings attribute table. It does not have information about base heights for buildings in general.) Incorporating building attributes into the model would improve the results.

Another element of uncertainty is that the water entry level for a building depends on the building's exact vertical position within a bluespot. Other factors being equal, a building at or near the bottom of a bluespot will be flooded sooner than a building higher up on its slope.

As discussed in the introduction to this lesson, perfect runoff conditions are rare in real life, but in a cloudburst, basic hydrologic assumptions can be relaxed. The normal infiltration capacity of soils becomes irrelevant, and sewer systems may reach maximum capacity very quickly. When this happens, precipitation will turn into rapid overland flows that fill up bluespots partially or entirely. (Note that the models presented here don't consider diversion of surface runoff through drainage ditches, tubes, or other channels.)

Some enhancements to the Identify Bluespots Fill Up Values model can be considered. An examination of surface solidity, especially of whether large parts of local watersheds are paved or not, could improve the risk assessment for individual buildings. (More paved surface means faster runoff.) A raster dataset of solid surface percentage for the Gentofte study area is included in the ResourceData geodatabase. A study of slopes and flow lengths within watersheds would also be relevant to determining which buildings would be affected first in a cloudburst.

You can find more lessons in the Learn ArcGIS Lesson Gallery.