A Comparative study on Stream Network Delineation for Upper, Lower Dolores Basin, Colorado-Utah using Arc-Hydro and TauDEM

SEKAR RAMU- GIS IN WATER RESOURCES, UTAH STATE UNIVERSITY, FALL 2008

The contents of the project are listed and linked as below

Introduction

Objectives

Study Area

GIS Data & Data Sources

Methodology

· Stream Network Delineation Using Arc-Hydro

· Stream Network Delineation Using TauDEM

Results and Discussion

Summary and Conclusions

References

Introduction

Watershed management requires physiographic information such as watershed slope, configuration of channel network, location of drainage divide, channel length and geomorphologic parameters viz. relative relief, shape factor, circulatory ratio, bifurcation ratio, drainage density for watershed prioritization and implementation of soil and water conservation measures. Traditionally, these parameters are obtained from topographic maps or field surveys.

There exist several methods for watershed delineation. Hand delineation based on the contour information depicted on topographic maps has been used for a long time. Even with the advent of GIS technology, this method is often still used prior to creating a digital watershed dataset. While this manual method can result in accurate delineations, it is a time-consuming and expensive task. The availability of digital topographic maps, in the form of Digital Raster Graphic (DRG) data, has improved the process but this method can also be slow and costly. This led to the advent of Digital Elevation Model based watershed and stream network delineation. So, over the past two decades the physiographic and geomorphologic parameter has been increasingly derived from digital representation of topography, generally called the DEM (Moore et al., 1991; Martz and Garbrecht, 1992). The automated derivation of topographic watershed data from DEMs is faster, less subjective and provides more reproducible measurements than traditional manual techniques applied to topographic maps (Tribe, 1992).

Advances in computational power and growing availability of spatial data have made watershed-based analysis more systematic and meaningful. Watersheds are appropriate spatial units for managing environmental problems and reflect the hydrological responses of the delineated spatial unit. The key property of watershed delineation is the creation of a watershed boundary. The watershed boundary uniquely defines the land area from which the surface water drains to the watershed outlet. The advent of spatial data in the form of Digital Elevation Models (DEM), Triangulated Irregular Network (TIN) and Digital Line Graphs (DLG) has led to the use of improved tools for watershed management and hydrologic modeling. Geographic Information System (GIS) software has made the task of spatial data management much easier, interactive and informative. In this context, this term project titled: A Comparative study on Stream Network Delineation for Upper, Lower Dolores Basin, Colorado-Utah using Arc-Hydro and TauDEM focuses on delineating the stream network using Arc GIS 9.3. The objectives of the project are listed as below.

Objectives

To Build a base-map for Upper and Lower Dolores basin

To build a stream network for Dolores basin using Arc Hydro toolbar and TAUDEM

To Calculate and compare stream length and drainage area for Dolores basins(upper, lower)

To Calculate and compare the drainage density using Arc Hydro, TauDEM for the study area

Study Area

Nestled amidst the San Juan Mountains in western Colorado is the magnificent Dolores River.  Running more than 170 miles, the Dolores is one of the longest and most ecologically diverse river basins in America, containing a wide range of native wildlife, plants, and habitat. The Upper and Lower Dolores basins are sub-basins of upper Colorado basin, which itself is a sub-region of Colorado basin. Elevations in the Dolores Basin range from about 14,200 feet near the Dolores River headwaters, to 4,100 feet at its confluence with the Colorado River in Utah. The terrain of the Dolores Basin consists of high plateaus with deeply incised canyons and dry arroyo. The geographical extent of the Dolores basin is 37.3830 to 38.873⁰ North and 107.866⁰ to 109.264⁰ W. The UTM zone code is 13. The Geographical location is shown in the following figure.

Figure 1. The Study Area

The Base map for Upper and Dolores basin with their USGS gauging stations are shown in the following figures.

Figure 2. Upper Dolores Basin with USGS Gauging Sites

Figure 3. Lower Dolores Basin with USGS Gauging Sites

GIS Data and Data Sources

The data requirement for this project is Digital Elevation Model and Hydrography data set for the selected upper and lower Dolores basin. These data were requested and downloaded from the National Hydrogrphy Data set web-site (http://nhdgeo.usgs.gov) and National Elevation Data set (http://seamless.usgs.gov) through the internet server.

The National Hydrography Dataset (NHD) is a comprehensive set of digital spatial data that contains information about surface water features such as lakes, ponds, streams, rivers, springs and wells. The resolution of the data set is 1:100,000. Within the NHD, surface water features are combined to form "reaches," which provide the framework for linking water-related data to the NHD surface water drainage network. These linkages enable the analysis and display of these water-related data in upstream and downstream order.

The NHD is based upon the content of USGS Digital Line Graph (DLG) hydrography data integrated with reach-related information from the EPA Reach File Version 3 (RF3). The downloaded data for the upper and lower Dolores basin are shown in the following figures.

Figure 4. Digital Elevation Model (DEM) with NHD flow lines for Upper Dolores Basin

Figure 5. Digital Elevation Model (DEM) with NHD flow lines for Lower Dolores Basin

(The Elevation data is the newer higher resolution of 1 arc second (1") NED product that is recently available for both basins. The Projected DEM is shown in figure 4 and 5)

Methodology

The stream network delineation was carried out using Arc GIS 9.3 with help of extension toolbars Arc Hydro and Terrain Analysis Using Digital Elevation Model (TauDEM). The methodology behind the two methods is described as follows.

a. Stream Network Delineation using Arc Hydro

Arc Hydro is a geospatial and temporal data model for water resources that operate within ArcGIS. Arc Hydro has an associated set of tools that populates the attributes of the features in the data framework, interconnect features in different data layers, and support hydraulic analysis and simulations (David R.Maidment, 2006).

The principal functions that are available in Arc Hydro to delineate the stream network and watershed is listed and described as below.

1. DEM Reconditioning

The DEM Reconditioning function (DEM Manipulation menu) modifies Digital Elevation Models (DEM) by imposing linear features onto them (burning/fencing) by means of processing the “Raw DEM" Grid and the” Agree DEM" Grid.

2. Fill Sinks

The Fill Sinks function (DEM Manipulation menu) fills sinks in a grid. If a cell is surrounded by higher elevation cells, the water is trapped in that cell and cannot flow. The Fill Sinks function modifies the elevation value to eliminate these problems. This function takes as input a DEM grid, which can be either an unprocessed DEM or a preprocessed DEM with DEM Reconditioning function.

3. Flow Direction

The Flow Direction function (Terrain Preprocessing menu) takes a grid ("Hydro DEM" tag) as input, and computes the corresponding flow direction grid ("Flow Direction Grid" tag). The values in the cells of the flow direction grid indicate the direction of the steepest descent from that cell.

4. Flow Accumulation

The Flow Accumulation function (Terrain Preprocessing menu) takes as input a flow direction grid ("Flow Direction Grid" tag). It computes the associated flow accumulation grid ("Flow Accumulation Grid" tag) that contains the accumulated number of cells upstream of a cell, for each cell in the input grid.

5. Stream Definition

The Stream Definition function (Terrain Preprocessing menu) takes a flow accumulation grid ("Flow Accumulation Grid" tag) as input and creates a Stream Grid ("Stream Grid" tag) for a user-defined threshold. This threshold is defined either as a number of cells (default 1%) or as a drainage area in square kilometers.

6. Stream Segmentation

The Stream Segmentation function (Terrain Preprocessing menu) creates a grid of stream segments that have a unique identification. Either a segment may be a head segment, or it may be defined as a segment between two segment junctions. All the cells in a particular segment have the same grid code that is specific to that segment.

7. Catchment Grid Delineation

The Catchment Grid Delineation function (Terrain Preprocessing menu) creates a grid in which each cell carries a value (grid code) indicating to which catchment the cell belongs. The value corresponds to the value carried by the stream segment that drains that area, defined in the input Link grid.

8. Catchment Polygon Processing

The Catchment Polygon Processing function (Terrain Preprocessing menu) takes as input a catchment grid (‘Catchment Grid" tag) and converts it into a catchment polygon feature class ("Catchment" tag). The adjacent cells in the grid that have the same grid code are combined into a single area, whose boundary is vectorized

The functions that are available in TauDEM to delineate the stream network are described as below.

b. Terrain Analysis Using Digital Elevation Models (TauDEM)

There are two (Basic Grid analysis and Network delineation) set of functions are available in TauDEM toolbar. Using Basic Grid Analysis, the terrain preprocessing such as fill pits, flow direction (D₈ and D_inf), contributing area (D₈ and D_inf) could be performed. After the terrain preprocessing the network can be delineated in six ways that are listed as follows.

1. Use of Existing streams

2. DEM Curvature based method

3. Contributing Area threshold

4. Grid Order threshold

5. Area and slope threshold

6. Length and Area threshold

For this project, the first three methods are selected to delineate the stream network. The concepts behind these methods are described as follows.

1. Use of Existing streams

To use this method existing streams need to have been "burned in" by using an enforced flow path grid (fdr) with "Fill Pits" and "D₈ Flow Directions" functions. The stream network raster is then defined from this flow direction. The D₈ flow direction grid from Arc Hydro can be used as an input. (David G.Tarbaton, 1991)

2. DEM Curvature based method

The DEM is first smoothed by a kernel with the weights at the center, sides and diagonals as specified.  The Peuker and Douglas (1975) method) is then used to identify upwards curved grid cells.  This method flags the entire grid, then examines in a single pass each quadrant of 4 grid cells and unflags the highest.  The remaining flagged cells are deemed "upwards curved" and if viewed resemble a channel network, although sometimes lacking connectivity, or requiring thinning.  The thinning and connecting of these grid cells is achieved here by computing the contributing area using only these upwards curved cells.  An accumulation threshold on the number of these cells is used to map the channel network. . (David G.Tarbaton, 1991)

3. Contributing Area Threshold

A threshold on the contributing area (in number of cells) computed by the D8 (suffix ad8) method is used to delineate streams. (David G.Tarbaton, 1991).

Constant Drop Analysis

The drop analysis for above three methods was done by means of automatically selecting the threshold value between 5 to 500 to determine the drainage density for the upper and lower basin. The procedure suggested in Tarboton et al. (1991) is to select the smallest threshold for which the absolute value of the t statistic is less than 2.  This selects the highest resolution network consistent with the "constant drop law". 

Then the calculated drainage density was compared with the NHD method and Arc Hydro method. The formula for calculating drainage density is:

Total Stream Length (km)

Drainage Density (km^-1) = ---------------------------------

Total Drainage Area (km²)

The results of delineated stream network and drainage density for the upper, lower Dolores basin is shown as follows.

Results and Discussion

The Arc GIS created delineated stream network using above methods for Upper Dolores basin is shown below.

Figure 6. NHD Flowline Network for Upper Dolores

Figure 7. Arc Hydro Network for Upper Dolores

Figure 8. TauDEM Existing stream Network for Upper Dolores

Figure 9. TauDEM Curvature based Stream Network for upper Dolores

Figure 10. TauDEM Contributing Area threshold network for Upper Dolores

The ArcGIS created delineated stream network for Lower Dolores basin using five methods are given below.

Figure 11. NHD Flowline Stream Network for Lower Dolores

Figure 12. Arc Hydro Delineated Stream Network for Lower Dolores

Figure 13. TauDEM Existing Stream Network for Lower Dolores

Figure 14. DEM Curvature based stream Network for Lower Dolores Basin

Figure 15. TauDEM Contributing area threshold Network for Lower Dolores

The observed data for all the above methods are tabulated below to analyze the drainage density of upper and lower Dolores basin.

Table 1. Channel Length and Drainage Area analysis for Upper Dolores Basin

	NHD Flow lines	Arc Hydro Network	TauDEM –Existing streams	TauDEM –DEM curvature-	TauDEM-Contributing area threshold
Channel Length(km)	3736.9	2416	3692.5	3594	3638
Drainage Area(km^2)	5637.4	4489	5592.5	5584	5580

(t=5 km² for all TauDEM method)

Table 2. Channel Length and Drainage Area analysis for Lower Dolores Basin

	NHD Flow lines	Arc Hydro Network(t=20 km^2)	TauDEM –Existing streams(t=5 km^2)	TauDEM –DEM curvature(t=6 km^2)	TauDEM-Contributing area threshold(t=5 km^2)
Channel Length(km)	1636.5	1236	1583	1486	1502
Drainage Area(km^2)	2379.7	2012	2331.5	2228	2221

Table 3. Calculated Drainage Density for Upper Basin

	NHD Flow lines	Arc Hydro Network	TauDEM –Existing streams	TauDEM –DEM curvature	TauDEM-Contributing area threshold
Drainage Density(km^-1)	0.66	0.53	0.66	0.64	0.65

Table 4. Caculated Drainage Density for Lower Basin

	NHD Flow lines	Arc Hydro Network	TauDEM –Existing streams	TauDEM –DEM curvature	TauDEM-Contributing area threshold
Drainage Density(km	0.68	0.61	0.67	0.66	0.67

(t=threshold level)

From the above results it could be observed that the TauDEM existing streams method estimate the drainage density value very close to the NHD flow line method than any other method for both upper and lower Dolores basin. The TauDEM contributing area threshold method estimate the drainage density slightly better than the TauDEM Curvature method for upper Dolores basin when the threshold value taken as same(t= 5 km²). The Arc Hydro estimate the drainage density value so lower than the NHD flow line and TauDEM methods. This may be due to the threshold value considered for the Arc Hydro method is higher (Arc Hydro default value of 1% of total drainage area) than the TauDEM methods. So, a reduction in threshold value may improve the Arc Hydro Drainage density for both basins.

The following table was generated to verify the drainage densities obtained in TauDEM using the stream drop analysis by means of selecting the threshold value automatically between 5 to 500 and t-statistics.

Table 5. Comparison between TauDEM derived drainage density with stream drop analysis values for Upper Dolores Basin

	Existing streams	DEM curvature	Cont.area threshold
Drainage Density(km-1)from Length and area	0.66	0.64	0.65
Drainage Density from Stream Drop Analysis (km-^1)	0.67	0.63	0.65

Table 6. Comparison between TauDEM derived drainage density with stream drop analysis values for Lower Dolores Basin

	Existing streams	DEM curvature	Cont.area threshold
Drainage Density(km-1)from Length and area	0.67	0.66	0.67
Drainage Density from Stream Drop Analysis (km-^1)	0.67	0.65	0.68

From the above comparison that the estimated drainage density values seem correct as the values are close to each other.

Summary and Conclusions

From the above results it could be concluded that the Arc Hydro delineated network varies from NHD flow line and TauDEM methods. To improve the Arc Hydro network drainage density value, the threshold (t) value could be reduced as it may improve the channel length for the given area.

The TauDEM existing stream method delineates the network close to NHD flow line. Considering the above said points, it is obvious that the each method generate different stream network, it is important for the user to highlight the method used to delineate the particular stream network.

Considering the TauDEM and Arc Hydro, TauDEM has advantages as it provides options to select threshold values from a whole range of values. But, in Arc Hydro every threshold input value terrain preprocessing needs to be performed sequentially leads to time consuming. The TauDEM also has “DO ALL” function for quick terrain preprocessing. Beside these, both methods have some inaccuracy in delineating the stream network as it may affect the drainage density of the study area. Finally, It is also imperative to consider the resolution of the DEM as it may affect the stream network delineation significantly.

References

Martz, L.W. and Garbrechet, J. 1992. Numerical Definition of Drainage Network and Subcatchment Areas from Digital Elevation Models. Computers and Geosciences. 18(6):747-761.

Moore, I.D., Grayson, R.B. and Ladson, A.R.1991. Digital Terrain Modelling: A Review of Hydrological, Geomorphological and Biological Applications. Hydrological Processes 5(1): 3-30

Tribe, A. 1992. Automated Recognition of Valley Heads from Digital Elevation Models. Earth Surface Processes and Landforms 16(1):33-49

Tarboton, D.G., Bras, R.L. and Rodrigues, I.1991. On the Extraction of Channel Network from Digital Elevation Data. Water Resources Research 5(1): 81-100.

Davi R. Maidment, 2002. Arc Hydro –GIS for Water Resources.

Internet References

http://seamless.usgs.gov/

http://nhd.usgs.gov/data.html

http://hydrology.neng.usu.edu/taudem/

http://www.engineering.usu.edu/dtarb/

------------------------------------------------------ ---------------------------------------------------------------------