Jessica L. Watkins, University of Georgia
2 April 2009
For this project, I delineated watersheds for the Coweeta synoptic sampling area using ESRI’s ArcMap 9.2 mapping software and created a geodatabase for the watershed data using ESRI’s
ArcCatalog. The watersheds were created for the scientists working on the Coweeta synoptic sampling project and will be used for landscape characterization.
METHODS AND PROCEDURE
The methods in the watershed delineation process first involved obtaining a digital elevation model (DEM) of the synoptic sampling area were downloaded by Katherine Neder from USGS seamless website in June of 2008. DEMs contain valuable and necessary information on the topography of a ground surface needed for watershed delineation. Since DEMs usually contains sinks, I used the fill tool to get rid of them because, otherwise, portions of the desired watersheds would be isolated and incorrectly delineated. After the sinks in the DEM were filled, I created a flow direction grid to find the direction of water flow for the streams. Then, I used the flow direction grid to determine the flow accumulation and find the areas of highest accumulation and then added the synoptic points from Jack Webster to the map for analysis.
Zooming in closely, I made sure the synoptic sampling points were located along the high accumulation cells to prevent incorrect watershed delineation later on. In order to make corrections for those synoptic points not located on the accumulation flow, I created a shapefile for the new synoptic points to take into account the adjustments and then added a field for the attribute table of the new synoptic points to include the original synoptic sampling ID numbers from Jack Webster. This ID also identifies the delineated watersheds. The next step included delineating the watersheds using the watershed tool by inputting the flow direction, which contained the stream flow information, and the new synoptic points (lowest points of elevation). After the watersheds were created, I checked the raster properties to make sure that there were 57 watersheds in total to match with the 57 synoptic sampling points and then converted the raster to a shapefile. Next, I spatially joined the attributes from the original synoptic points to the watershed attributes. This completed the watershed delineation process.
I used three supplementary surface rasters to help with analysis and the watershed delineation process: hillshade, slope, and aspect. These were creating using the spatial analyst toolbar and inputting data into the various fields of each raster to be created. The hillshade, slope, and aspect rasters allow you to get an idea as to where the streams flow because of the shaded relief, the degree of slope, and also the direction of the slope displayed. Lastly, I created a geodatabase to include all of the necessary and important data needed for the synoptic watershed delineation. Two feature data sets were created to include the original base data and the rest of the files for delineation.
First, I set the working directory and output extent for the watershed delineation environment properties (Tools > Options > Geoprocessing > Environments). Under General Settings, I set
the current workspace and scratch workspace to: D:\synoptic_sampling\business_processes\
cwt_synoptic_watersheds.mdb\ and set the output extent to: Same as Layer “synoptic_dem”.
Under the Raster Analysis Settings, I also set the cell size to Same as Layer “synoptic_dem”
and then enabled the Spatial Analyst Extension so that I could begin the watershed delineation process. I added the most recent synoptic points from Jack Webster to an ArcMap Document and saved it as: D:\synoptic_sampling\business_processes\watershed_delineation\
Created a depressionless DEM
I made sure that the DEM had no sinks by using the Fill tool (Spatial Analyst Tools >
Hydrology > Fill). I set the input surface raster as the synoptic_dem grid and the output
Also, using the Spatial Analyst Tools (Spatial Analyst Tools > Hydrology > Flow
Direction), I created a flow direction raster and then set the input surface raster to the
synoptic_filled DEM that was just created and the output raster to: D:\synoptic_
Through the Spatial Analyst Tools, I created a flow accumulation grid (Spatial Analyst
Tools > Hydrology > Flow Accumulation). Then, I set the flow direction raster
synoptic_flowdir as the input and the output accumulation raster to: D:\synoptic_
changed the symbology of the flow accumulation to classified using 2 classes with a
break value at 5000 for the first class and change the colors for the two classes to display
the new flow accumulation with the appearance of a drainage network.
Pour Points (Synoptic points)
I added the streams shapefile for the synoptic study area to the map to compare the flow
accumulation network. The flow accumulation locates cells with high cumulative flow.
By zooming in closely I noted the location of the pour points (synoptic_points_original).
Pour points are supposed to be located in cells with high accumulation flow; otherwise,
they will result in incorrect watershed delineation and will be very small. In order to
make sure that our watersheds were delineated correctly, I created a new shapefile for the
pour points (synoptic_points_corrected) using the Editor Toolbar and Create New
Feature function to account for the adjustments and correct for any of the original points
that were not located within a cell on the accumulation flow network.
In order to further organize the data, I created an “Id” field in the attribute table of the
synoptic_points_corrected and gave integer values to the original synoptic points accordingly (i.e. a corrected pour point was given an Id of 3 because it was the corrected point for the original synoptic sampling point Id of 3). In order to make sure that the watersheds would not be cut short, I reset the output extent (Tools > Options >
Environment > General Settings > Output Extent), but this time using the filled raster,
synoptic_filled. Under the Raster Analysis settings, I also changed the cell size to Same
as Layer “synoptic_filled”. I did the same again for the extent options and under the Spatial Analyst tool set them to Same as Layer “synoptic_filled” for the analysis extent
(Spatial Analyst > Options > Extent >Analysis Extent) and analysis cell size (Spatial
Analyst > Options > Extent >Cell Size). I set the cell size to 9.3526343.
The next step was to finally delineate the watersheds by using the Watershed Tool from
ArcToolbox (ArcToolbox > Spatial Analyst Tools > Hydrology > Watershed). I
input the synoptic_flowdir raster as the flow direction, the synoptic_points_corrected
shapefile as the feature pour point data, and I set the output raster to D:\synoptic_
The new watershed raster was added to the data frame. I then converted the watershed raster file to a polygon shapefile using the Conversion Tools (ArcToolbox >
Conversion Tools > From Raster > Raster to Polygon). I set the synoptic_sheds_
raster grid as the input raster to be converted and the output polygon features to: D:\synoptic_sampling\business_processes\cwt_synoptic_watersheds.mdb\ synoptic_shed_delineation\synoptic_sheds_nodata.
Joining attribute tables
Using Spatial Join (ArcToolbox > Analysis Tools > Overlay > Spatial Join), I created
a new watershed shape file (synoptic_watersheds) that combined the attributes from the
synoptic_points_original to the synoptic_sheds_nodata attribute table. I set the target
feature for joining the data to synoptic_sheds_nodata and the join feature that the
dataset was obtained from to synoptic_points_original. I set the output to
D:\synoptic_sampling\business_processes\cwt_synoptic_watersheds.mdb\synoptic_ shed_delineation\synoptic_watersheds and kept all targeted features.
I created three surface rasters for reference and analysis of watersheds using the Spatial Analyst Toolbar:
; Hillshade (Spatial Toolbar > Spatial Analyst > Surface Analysis > Hillshade)
When prompted with the hillshade window, I used the raster with the sinks filled
in, synoptic_filled, as the input surface, set the output cell size to 9.3526343,
and left everything else with the DEFAULT settings. I set the output raster set to:
; Slope (Spatial Toolbar > Spatial Analyst > Surface Analysis > Slope)
When prompted with the slope window, I used the raster with the sinks filled in,
synoptic_filled, as the input surface, set the output cell size to 9.3526343, and
left everything else with the DEFAULT settings. I set the output raster to:
; Aspect (Spatial Toolbar > Spatial Analyst > Surface Analysis > Aspect)
When prompted with the aspect window, I used the raster with the sinks filled in,
synoptic_filled as the input surface, set the output cell size to 9.3526343, and
the output raster as: D:\synoptic_sampling\business_processes\cwt_synoptic_
Synoptic Personal Geodatabase
In order to organize the data, I created watershed delineation personal geodatabase called
synoptic_watersheds.mdb in ArcCatalog and was saved to D:\synoptic_sampling\
business_processes\cwt_synoptic_watersheds.mdb. I added two feature data sets and
imported the files used during the watershed delineation process accordingly:
; Synoptic_shed_basedata – This feature data set contains the base data shapefiles
for the watershed delineation.
Synoptic_boundary_lt – This is the boundary that was used for clipping o
the DEM for the synoptic sampling study area.
o Synoptic_points_original – These are the original synoptic sampling
points from Jack Webster.
o Synoptic_streams – This is the stream shapefile for the synoptic sampling
; Synoptic_shed_delineation – This feature data set contains the shapefiles used to
delineation the synoptic sampling watersheds.
o Synoptic_points_corrected – These are the new synoptic sampling points
corrected to make sure in the path of high cumulative flow.
o Synoptic_sheds_nodata – This is the watershed shapefile converted from
the watershed raster missing the data from the original synoptic points.
o Synoptic_watersheds –This is the shapefile created after a spatial join
was made between the original synoptic points attribute table to the
; Synoptic_aspect (Raster) – This is the aspect slope raster that identifies the slope
; Synoptic_dem (Raster) – This is the original raster prior to filling in the sinks. ; Synoptic_filled (Raster) – This is the raster after the sinks have been filled in. ; Synoptic_flowacc (Raster) – This is the flow accumulation grid which gives
contains the values for water accumulation in the synoptic area. ; Synoptic_flowdir (Raster) – This is the flow direction raster contains the data for
the direction of stream flow.
; Synoptic_points_2009_02_23 (CSV File) – This contains the original
information with the newest synoptic points from Jack Webster. ; Synoptic_shade (Raster) – This is the hillshade raster which contains shaded
; Synoptic_sheds_raster – This is the watershed raster created from the watershed
; Synoptic_slope (Raster) – This is the slope raster which contains the slope
degrees in the synoptic area.