Developing an Arc Hydro Dataset
By Daniel Obenour, 12/06/02
Introduction:
An Arc Hydro Dataset is an organized collection of hydrological information. All information (data) included in the dataset must be developed in specific ways in order to take full advantage of the functionality of the Arc Hydro system. This development process is the focus of my term project.
Developing raw data into a coherent Arc Hydro Dataset can be a daunting task. The amount of data alone may often seem overwhelming. Also, the tasks required to prepare, organize, and create relationships within the data can seem complex and unintuitive. For these reasons, I have created a set of instructions to help streamline the data development process. My goal is to create an instruction set that takes the user through the data development steps in a logical order, while explaining the reasoning behind the various tasks that must be performed. When using the instruction set, the user will hopefully determine that creating an Arc Hydro dataset is a reasonable and interesting process.
The completed instruction set can be viewed only by downloading it from the link listed directly below. The instruction set includes 13 sections, each of which pertain to a specific part of the data development process. The report included on this web page describes these 13 sections. This report will be of particular value to those who want a relatively conceptual, non-technical view of what goes into creating an Arc Hydro dataset. It can also be used to enhance the instruction set, especially for users who are less familiar with GIS and/or Arc Hydro.
The Instruction Set:
Download the Instruction Set in MS Word format
Download an Enhanced Instruction Set (with graphics) in MS Word format
Table of Contents:
1. Acquiring the Required Input Data
4. Loading Monitoring Locations into HydroJunction
5. Breaking HydroEdges at Locations of Interest
6. Creating WshOutlet and WbdOutlet Features
7. Loading Outlets into HydroJunction
8. Preparing Data for the Arc Hydro Network
9. Creating the Arc Hydro Network
11. Implementing the Arc Hydro Schema
12. Assigning Arc Hydro Attribute Values
1. Acquiring the Required Input Data:
The first step in creating an Arc Hydro dataset is to collect GIS data, commonly represented as feature classes. Feature classes are collections of geometric objects (points, lines, or polygons) that share common themes and common attribute types. All Arc Hydro projects will require a feature class representing river reaches. Furthermore, most projects will require feature classes representing monitoring points, watersheds, and waterbodies. In addition to this, other feature classes can be added at the designer's discretion.
A typical resource for hydrological data for the United States is the National Hydrography Dataset (NHD). In fact, the NHD includes everything necessary for a simple Arc Hydro model. NHD data is arranged by Hydrologic Unit Code (HUC) and is available from the NHD website: http://nhd.usgs.gov/.
Considerable data preparation is often required before data sets can be effectively loaded into the Arc Hydro database. Most importantly, every type of data must be condensed into a single GIS shapefile. Therefore, if the data is split into separate files, based on geographic location, the designer must merge these files together. Differentiation between data located on and off the river network must also be made. For instance, the user will probably need to select waterbodies located along the river network and save the selection as a new watebody shapefile.
Merging river reach shapefiles
In ArcCatalog the designer will need to create a personal geodatabase in which all of the Arc Hydro objects (feature classes, tables, etc.) will be stored. During this process, it is important to select a correct projection system and spatial reference frame. The projection system should be one that will accurately represent the areas of watersheds. The spatial reference frame should be large enough to encompass all features that might be added to the model.
A Geodatabase with typical Arc Hydro feature classes as shown in ArcCatalog
GIS terminology can seem a little strange to those who are unfamiliar with it. Listed below is the data structure found in a personal geodatabase. Terms are listed from the smallest to largest database element.
It is also important to take note of Arc Hydro terminology. Below is a short list of the primary Arc Hydro feature classes and what types of data they represent.
HydroEdge == river reaches
HydroJunction == points of interest that fall along river reaches
4. Loading Monitoring Locations into HydroJunction
At this point, the designer creates the new feature class named "HydroJunction". This feature class will eventually represent all locations of interest along the river network. The first locations of interest to be added are stream monitoring sites. The designer can make use of the tools found in ArcMap to accomplish this. As the points are loaded into HydroJunction, they must be snapped to the nearest river line. Therefore, the exact locations of the HydroJunctions may be slightly different than the monitoring points they represent (Hydrojunctions must lie directly on top of river lines, while the monitoring points lie at their true location).
HydroJunctions and associated MonitoringPoints
5. Breaking HydroEdges at Locations of Interest
The designer also needs to create HydroJunctions that represent watershed and waterbody outlets. This is a somewhat complex process and compromises Sections 5, 6, and 7 of the Instruction Set. This first section describes how to break the HydroEdges at locations of interest (watershed and waterbody outlet locations). We are doing this because we will eventually create junctions at the endpoints (broken edges) of all HydroEdge features.
HydroEdge broken at watershed boundary
6. Creating Watershed Outlet and Waterbody Outlet Features
Once the HydroEdges have been broken, we apply a temporary geometric network in order to create network junctions at the broken ends of each HydroEdge segment. Network junctions at the borders of watersheds will be used to create the watershed outlet junctions. However, as noted in the figure below, sometimes the locations of these junctions must be modified. For instance, each watershed should have only one outlet, and it should be located on watershed's primary river.
Modification of network junctions to accurately represent watershed outlet locations
(black arrows represent river flow direction, varying shades of green represent different watersheds)
In a similar fashion, network junctions will also be used to create waterbody outlet junctions. Once all of the watershed and waterbody outlets are created, they will be exported as their own unique feature classes: WbdOutlet and WshOutlet. Like the network created in this section, these feature classes are only temporary and may be discarded after the completion of Section 7.
A Waterbody outlet point at the downstream end of a reservoir
7. Loading Outlets into HydroJunction
The outlet feature classes are transformed into HydroJunctions using the same method described in Section 4 (for loading monitoring points). However, unlike with MonitoringPoint, once the outlet feature classes are loaded into HydroJunction they may be discarded. As HydroJunctions, these points will be used to create links between the river network and their respective watersheds and waterbodies.
A symbolic view of Waterbody Outlets (blue) and Watershed Outlets (red) being loaded into the HydroJunction feature class.
Attribute table for the HydroJunction Feature Class. Note that the Ftype field keeps track of what
feature type each HydroJunction represents.
8. Preparing Data for the Arc Hydro Network
Even data acquired from well established sources like the NHD may have errors or missing data. It is important to correct any such errors before applying the final Arc Hydro Network (see Section 9). One common error, illustrated in the figure below, is that there may be gaps in the HydroEdge feature class. The temporary network (created in Section 6) proves to be a valuable tool in determining where many of these errors are located. For instance, gaps can easily be found by performing connectivity tests on a geometric network. However, once all corrections have been made, the network created in Section 6 should be deleted in order to make way for the final Arc Hydro Network discussed in the following section.
Procedure for correcting an error in the HydroEdge (NHD river reach) feature class
9. Creating the Arc Hydro Network
The final Arc Hydro Network creates a well-defined topology (relationships based on connectivity) between HydroEdges and HydroJunctions. This means that every feature knows what other features are connected to it. For instance, ArcGIS can now be used to trace paths between any two network locations.
HydroEdges and HydroJunctions: the two components of the final Arc Hydro Network
Correct network flow direction must be assigned to every HydroEdge feature. There are a couple of different methods for assigning directions, and when used correctly, the process proves to be fairly simple. Once properly set, flow directions greatly enhance the functionality of the Arc Hydro Network. When flow directions are set, the designer can select a given network feature and know what HydroEdges and HydroJunctions fall upstream and downstream of it.
Properly assigned flow directions
Upstream trace (red) from a given network location (green)
Downstream traces on a network
11. Implementing the Arc Hydro Schema
The primary purpose of the Arc Hydro Schema is to create important relationships between different feature classes. For instance, we know that certain HydroJunctions represent certain Watersheds. Also, if we have time series data, then that data is related to certain MonitoringPoint features. The schema creates these relationships and creates attribute fields within feature classes that are necessary to the relationships' functionality .
File structure required for an "Arc Hydro Framework with Time Series" schema
12. Assigning Arc Hydro Attribute Values
As mentioned in Section 11, the schema will create many attribute fields within the various feature classes. These attribute fields are created empty and need to be assigned values using the the Arc Hydro toolset. Some of these values are necessary in order for the schema relationships to be complete. Other values are simply informational, such as attribute fields containing the areas of the watersheds or the distances HydroJunctions are located from the basin outlet.
Map showing the distance upstream of HydroJunctions from the basin outlet
These distances were calculated using the Arc Hydro toolset
As with any project, it is important to clean up your data and leave good documentation. Cleaning up can be accomplished by deleting extraneous attribute fields and feature classes. Documenting can be performed by adding metadata to your geodatabase and feature classes.
This term project is just one part of a larger research endeavor aimed at creating a functional Arc Hydro Model for the Llano River basin in Texas. One of the future project goals is to create an Arc Hydro system for displaying real-time rainfall data. This data would be acquired through NEXRAD radar stations and from Lower Colorado River Authority (LCRA) gaging stations. Some development of the NEXRAD rainfall data has already been performed, and is included in this report because it is believed to be a subject of particular interest. The image linked below demonstrates an example of this progress.
A Storm Moving Through the Llano Basin, (click on thumbnail to view larger image).
This project has been made possible by a number of people and organizations. First of all, I would like to thank the Lower Colorado River Authority (LCRA) for funding my ongoing Arc Hydro research involving the the Llano River basin. I would also like to thank Dan Yates, my contact person at the LCRA, for providing me with all of the data that I have so far required.
I would also like to thank my co-workers at the Center for Research in Water Resources (CRWR) for providing me with valuable data and helpful advice. Specifically, I would like to thank Oscar Robayo, who provided me with considerable guidance during the development of the Instruction Set. Finally, I would like to thank Dr. Maidment and the CRWR crew for creating the class exercises that proved invaluable to me throughout this project.
I have used one major reference for this project:
Arc Hydro: GIS for Water Resources. David R. Maidment, Editor. ESRI Press, Redlands California: 2002.