&EPA
EPA/600/R-17/448F I November 2017 I www.epa.gov/research
United States
Environmental Protection
Agency
Procedures for Delineating
and Characterizing Watersheds for Stream
and River Monitoring Programs
n
Office of Research and Development
Washington, D.C.
-------
EPA/600/R-17/448F
November 2017
Procedures for delineating and characterizing watersheds for
stream and river monitoring programs
National Center for Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Washington, DC 20460
-------
Disclaimer
This document has been reviewed in accordance with U.S. Environmental Protection Agency policy and
approved for publication. Mention of trade names or commercial products does not constitute
endorsement or recommendation for use.
Abstract
Stream and river monitoring organizations often want to know the properties of the watersheds
draining to proposed or selected monitoring stations to help select candidate monitoring sites, classify
existing sites, or analyze data from existing sites. This manual describes procedures for delineating
watersheds at any point on a stream or river, then calculating a suite of watershed characteristics,
including land use composition, base flow, channel slope and sinuosity, watershed slope, and
enumeration of potential point source facilities of concern. It describes purpose-made ArcMap tools to
partially or fully automate all these procedures using the same interface as ArcMap geoprocessing tools.
The tools presented in this manual do not exhaustively characterize watersheds but simply cover some
basic watershed characteristics in which managers are often interested. If additional watershed
characteristics are desired, they can be added after the steps described in this manual. Both the tools
and this manual should increase the standardization, efficiency, and reproducibility of watershed
characterization within and between monitoring programs. These tools and manual were originally
developed for stream Regional Monitoring Networks (RMNs) but can be used for stream or river
monitoring programs at spatial scales from municipal to national.
Preferred citation:
Gibbs, DA; Bierwagen, B. (2017) Procedures for delineating and characterizing watersheds for stream and river
monitoring programs. (EPA/600/R-17/448F). Washington, DC: U.S. Environmental Protection Agency, Office of
Research and Development.
ii
-------
Contents
List of Figures iv
List of Abbreviations vi
Authors and Reviewers vii
1. Introduction 1
2. Preprocessing 3
3. Watershed delineation 7
3.1. For states with StreamStats batch processor available 8
3.2. For states without StreamStats batch processor available 16
4. Characterization of delineated watersheds 26
4.1. Process overview 26
4.2. Loading the tools into ArcMap 27
4.3. Characterization 29
4.3.1. Land use composition 30
4.3.2. Base flow 31
4.3.3. Channel slope and sinuosity 32
4.3.4. Watershed slope 43
4.3.5. Dams, mines, NPDES, and CERCLA site preprocessing 44
5. Technical support, updates, and known issues 48
5.1. Technical support 48
5.2. Updates 48
5.3. Known issues 48
6. Technical details of scripts 49
6.1. Land use composition 49
6.2. Base flow 50
6.3. Channel slope and sinuosity, pre-stream trace 50
6.4. Channel slope and sinuosity, post-stream trace 50
6.5. Watershed slope 52
6.6. Dams, mines, NPDES, and CERCLA site preprocessing 53
-------
List of Figures
Figure 1-1. Watershed delineation and characterization workflow 1
Figure 2-1. Required columns of initial site table 3
Figure 2-2. Site table with added fields 4
Figure 2-3. Excerpt of the NHDPIus Version 2 download page for the Mid-Atlantic 6
Figure 3-1. StreamStats flowline raster (light blue boxes) and NHD flowline 7
Figure 3-2. Reprojecting sampling stations into projection of StreamStats flowline raster 9
Figure 3-3. Entering editor mode for the sampling station shapefile
RMN_primary_secondary_[STATEABBRC]_reproj_aligned.shp 10
Figure 3-4. Three cases of sampling station-flowline raster alignment 11
Figure 3-5. Attribute table with alignment notes, original pour point locations, and aligned
pour point locations 12
Figure 3-6. Populating the latitude field of a point shapefile's attribute table 13
Figure 3-7. StreamStats batch processor upload interface 14
Figure 3-8. Sample StreamStats watershed delineation with StreamStats flowline raster 15
Figure 3-9. "Mosaic to New Raster" configuration for multiple DEMs in the same study area 16
Figure 3-10. Exporting flow accumulation raster to 32-bit signed TIF 18
Figure 3-11. Creating an attribute table for the flow accumulation raster 18
Figure 3-12. Converting the flow accumulation raster into a flowline raster 19
Figure 3-13. Examples of nested and unnested watersheds 22
Figure 3-14. Interface for the ArcMap "Watershed" tool 23
Figure 3-15. Incorrect watershed delineation by Watershed tool in ArcMap 24
Figure 4-1. Sample Results window output from land use composition tool 27
Figure 4-2. List of files included in folder 28
Figure 4-3. Displaying the RMN toolbox in ArcMap 29
Figure 4-4. Interface for land use composition tool 31
Figure 4-5. Interface for base flow tool 32
Figure 4-6. Schematic of the channel slope and sinuosity workflow 33
Figure 4-7. Interface for tool to do the first part of calculating channel slope and sinuosity 35
Figure 4-8. (a) Polylines of flowline raster with sampling station and aerial imagery
(b) Polylines of flowline raster with stream trace upstream and downstream to
the extent of the polyline flowlines 37
Figure 4-9. Example of a stream trace stopped at a lake 38
Figure 4-10. Sample of trace geodatabase with trace direction and notes 39
Figure 4-11. Interface for tool to calculate channel slope and sinuosity 40
Figure 4-12. Sample attribute table for
traces_with_slope_sin_[STATEABBREV]_m_[DISTANCE]_[YYYYMMDD_HH_MM_
SS] 41
Figure 4-13. Partial output of watershed delineation shapefile with channel slope and
sinuosity, calculated over 2,000 m stream traces wherever possible 42
iv
-------
List of Figures (continued)
Figure 4-14. Channel slope-sinuosity geodatabase with the polyline versions of the flowline
rasters created in Part 1 of this analysis, the manual stream traces created in
Part 2 of this analysis, and the files with slope and sinuosity calculated over
three stream distances 43
Figure 4-15. Watershed slope tool interface 44
Figure 4-16. Dams, mines, NPDES, and CERCLA site search support tool interface 46
Figure 4-17. Selecting point source facilities within watersheds 46
Figure 4-18. Example site counts and descriptions 47
v
-------
List of Abbreviations
CERCLA
Comprehensive Environmental Response, Compensation, and Liability Act
DEM
digital elevation model
FID
feature ID
GIS
geographic information system
LUC
land use code
NHD
National Hydrography Dataset
NLCD
National Land Cover Database
NPDES
National Pollutant Discharge Elimination System
RMN
Regional Monitoring Network
USGS
U.S. Geological Survey
vi
-------
Authors and Reviewers
Authors
David A. Gibbs, ORISE fellow, U.S. Environmental Protection Agency, Office of Research and
Development
Britta Bierwagen, U.S. Environmental Protection Agency, Office of Research and Development
Reviewers
U.S. EPA Reviewers
Ryan Hill, PhD—ORISE fellow, U.S. Environmental Protection Agency, Office of Research and
Development
Phil Morefield, U.S. Environmental Protection Agency, Office of Research and Development
External Peer Reviewers
Timothy Randhir, PhD—Professor, University of Massachusetts, Amherst
Michael Strager, PhD—Professor, University of West Virginia
Dale White, PhD—Water quality engineer, Ohio Environmental Protection Agency
Acknowledgments
We thank the Regional Monitoring Network partners for providing the data used to develop and test
these methods.
vii
-------
1. Introduction
Characterizing the properties of contributing watersheds—or the calculation of a suite of properties of
the drainage area for a location on a stream—has three main purposes: (1) to screen candidate
monitoring sites for some target status (e.g., whether they meet "reference site" status),1 (2) to classify
candidate or actual monitoring sites into various physical stream categories, and (3) to serve as
covariates in the analysis of biological community and hydrologic data from monitoring sites.
This manual explains how to use several purpose-made ArcMap tools to calculate watershed properties
commonly of interest to resource managers. These include (1) the fraction of land under each National
Land Cover Database (NLCD) land use within the whole watershed and within 1- and 5-km radii of the
sampling stations within the watershed; (2) the percentage of stream flow that comes from base flow at
the sampling station; (3) the channel slope and sinuosity at each station; (4) the slope of the watershed
(average, minimum, maximum, range, standard deviation); and (5) the number of dams, mines, National
Pollutant Discharge Elimination System (NPDES),2 and Comprehensive Environmental Response,
Compensation, and Liability Act (CERCLA)3 sites within each watershed. The end product is a watershed
delineation shapefile that has fields with all of the watershed properties examined. This shapefile can be
mapped or exported to other programs for additional analyses, such as the association between
macroinvertebrate communities and land use.
Before watersheds can be characterized, they need to be delineated (see Figure 1-1). Delineation is
based on the locations of stream sampling or monitoring stations relative to contributing land. Following
some preprocessing steps (see Section 2), this manual presents two ways to delineate watersheds
(see Section 3); the one that is used depends on the state or territory where the sampling stations are
located. Following delineation instructions, the characterization steps are presented (see Section 4).
These steps do not produce every characteristic in which resource managers are interested. If further
characterization steps are conducted after the steps shown in Figure 1-1, the process can be extended
using whatever additional geoprocessing tools are desired. Contact and help information are provided in
Section 5. More technical outlines of the characterization tools are provided in Section 6. These
methods are an improvement over the previous RMN delineation and characterization methods that
approximated characterizations by delineating to the next largest National Hydrography Dataset (NHD)
catchment. By only including the area that drains to the sampling stations, rather than "rounding" to the
next largest catchment, this method more precisely delineates, and therefore characterizes, watershed
properties.
Preprocessing
Delineate
watersheds
Land use
composition
Channel slope
and sinuosity
Watershed
slope
Dams, mines,
NPDES,
CERCLA sites
Figure 1-1. Watershed delineation and characterization workflow. Watershed
characterization steps are in gray.
^'Reference site" can be defined many ways, but, in this context, it refers to the highest quality available sites.
2U.S. EPA NPDES homepage: https://www.epa.gov/npdes.
3U.S. EPA CERCLA homepage: https://www.epa.gov/superfund/superfund-cercla-overview.
1
-------
Although this document was written to guide Regional Monitoring Network (RMN)1 partner agencies
(e.g., states, tribes, watershed commissions, EPA regions) in characterizing RMN watersheds, it can also
be used by other monitoring programs or groups interested in understanding their own particular
watersheds. The slight modifications other programs will need to make to these procedures are not
covered in this guide because every program will have its own variation. For example, this guide uses
the RMN watershed processing naming system. If this guide is being used for non-RMN programs, the
files described below can be renamed accordingly. Furthermore, all sites in the RMNs are designated
either primary or secondary sites. Primary sites are core, reference (i.e., least disturbed) sites where all
required parameters (biological, physical, habitat, and chemical) are collected. Secondary sites are
additional sites of potentially lower quality where fewer parameters are collected. Some states have
both primary and secondary sites; some have only primary sites. Both primary and secondary sites in a
state can be processed together (in a single file), if desired.
Watersheds in different states should be delineated and characterized (collectively called "processed")
separately. There are two main reasons for this. First, the recommended watershed delineation method
uses a different projection for geospatial files for every state (see Section 3.1), so each state needs to
have a separate sampling station shapefile in the proper projection for delineation. Second, processing
states' monitoring stations separately helps with tracking progress and record keeping. This document
describes how to delineate and characterize the monitoring stations in one state. If monitoring stations
from multiple states are being processed, the directions in this document should be repeated separately
for each states' stations.
The directions for running each tool note which ArcMap extensions and/or licenses are required beyond
ArcGIS Desktop Basic Version 10.3 or later. Most of the tools described in this guide require the Spatial
Analyst extension. One tool requires the 3D Analyst extension and the ArcGIS Desktop Advanced license.
The tools have been tested on ArcGIS Desktop 10.3, 10.4, and 10.5. This guide assumes users have basic
familiarity with ArcMap.
Links to online help pages are provided for activities or data sets with which users may be less familiar. If
the URLs change overtime and links break, all of the items can be located through internet searches.
Throughout this guide, file names are written in bold and parts of file names that are variable are in all
capitals inside brackets (e.g., [STATEABBRV]_landuse_[YYYYMMDD].shp).
1U.S. EPA (2016) Regional monitoring networks (RMNs) to detect changing baselines in freshwater wadeable
streams. (EPA/600/R-15/280). Washington, DC: Office of Research and Development, Washington. Available online
at https://cfpub.epa.goy/ncea/global/recordisplav. cfm?deid=307973.
2
-------
2. Preprocessing
Before a watershed can be delineated, some preliminary processing must be done to the sampling
stations, which are the basis for watershed delineation and characterization.
1. Sampling stations that are to be characterized must have the following basic information
(see Figure 2-1):
a. Station ID (rename as field "OrgStatnID")
b. Waterbody name (rename as field "Waterbody")
c. Responsible agency (rename as field "Entity")
d. Whether each site is primary, secondary, or under consideration (rename as field
"Status")
e. Additional information (rename as field "Notes")
FID
OrgStatnID
Waterbody
Entity
Status
Notes
202
SF 1
Sipsey Fork
AL DEM
primary
201
BRSL 3
Brushy Creek
AL DEM
primary
205
66g_WRD773
Jones Creek
GA DNR
primary
203
HURR 2
Hurricane Creek
AL DEM
primary
223
EC071F29
Hurricane Creek
TN DEC
primary
219
SV 634
Crane Creek
SC DHEC
primary
207
66d 44 2
Coleman River
GA DNR
primary
206
66d WRD768
Charlies Creek
GA DNR
primary
204
3890_1
Fightingtown Creek
TVA
primary
Figure 2-1. Required columns of initial site table.
NOTE: If the stations are in a shapefile (*.shp), proceed to Step 2. If the stations are in another spatial
format (e.g., a geodatabase, or in *.kmz format), they must be converted to a shapefile. If they
are in a nonspatial format (e.g., Excel spreadsheet), they must have latitudes and longitudes and
be imported1 as a shapefile. Regardless of the route taken, the station shapefile should be
named RMN_primary_secondary_[STATEABBRV].shp.
NOTE: The shapefile name RMN_primary_secondary_[STATEABBRV].shp assumes that both primary
and secondary sites are included in the shapefile. If only one type of site is included, change the
name of this shapefile and all derivative shapefiles accordingly.
2. Three more fields should be added to RMN_primary_secondary_[STATEABBRV].shp and
populated (see Figure 2-2):
importing Excel spreadsheet as shapefile: http://support.esri.com/en/technical-article/000Q12745.
3
-------
a. The abbreviation of the state in which the stations are located ("StateAbbrv"; text data
type, 10 characters). For sites monitored by non-state organizations, put the state
where the site is located, not the name of the monitoring organization.
b. The stations' latitude and longitude ("OrignlLat" [double data type], and "OrignlLong"
[double data type], respectively). Latitude and longitude can be populated by
right-clicking on the field name ("OrignlLat" and "OrignlLong"), clicking on "Calculate
Geometry," and selecting "Latitude" or "Longitude," as appropriate.1
FID
OrgStatnID
Waterbody
Entity
Status
Notes
StateAbbrv
OrignlLat
OrignlLong
202
SF 1
Sipsey Fork
AL DEM
primary
AL
34.28558
-87.3991
201
BRSL 3
Brushy Creek
AL DEM
primary
AL
34.3307
-87.2862
205
66g_WRD773
Jones Creek
GA DNR
primary
GA
34.60201
-84.1512
203
HURR 2
Hurricane Creek
AL DEM
primary
AL
34.91799
-86.133
223
EC071F29
Hurricane Creek
TN DEC
primary
TN
34.91799
-S6.133
219
SV 634
Crane Creek
SC DHEC
primary
SC
34.9235
-83.0793
207
66d 44 2
Coleman River
GA DNR
primary
GA
34.95203
-83.5166
206
66d WRD768
Charlies Creek
GA DNR
primary
GA
34.95895
-83.5716
204
3890_1
Fightingtown Creek
TVA
primary
GA
34.9851
-84.3851
Figure 2-2. Site table with added fields.
3. A text field should be created in RMN_primary_secondary_[STATEABBRV].shp called
"StationID" (25 characters). This should be populated with the original station IDs
("OrgStatnID"). In "StationID", any station ID more than 25 characters should be shortened
to 25 characters or less, and any special characters except underscores (e.g., dash, slash,
comma, period) should be replaced with another character. For example, "East Fork First
Fork Sinnemahoning Creek" might become "E Fk 1st Fk Sinnemahoning" and "B-099.7"
might become "B_099_7". Shortening the station IDs and removing special characters is
necessary for using the delineation and characterization tools.
4. Add seven empty fields to RMN_primary_secondary_[STATEABBRV].shp.
a. A text field (200 characters) called "Align_Note".
b. Two double data type fields called "PourPtLat" and "PourPtLong". These will store the
latitude and longitude of the sampling stations once they have been aligned with rasters
used for delineation (see Section 3).
c. A text field (50 characters) called "RMNRegion". This allows sites to be sorted by RMN
region once sites from multiple RMN regions are combined after characterization.
d. A text field (10 characters) called "HUC8". This allows analysis of sites based on the
8-digit hydrologic unit code (HUC8).
Calculate latitude and longitude of points:
http://desktop.arcgis.eom/en/arcmap/10.3/manage-data/tables/calculating-area-length-and-other-geometric-pro
perties.htm.
4
-------
e. A long-integer field called "COMID". This field will be used to associate each sampling
station with a unique stream reach identifier used in the NHD, which allows RMN sites
to be connected to properties calculated for NHD reaches, such as the Index of
Watershed Integrity.1
f. A double data type field called "Area_km2". This will store the total area of the
watershed delineation in square kilometers.
NOTE: Because watershed processing often involves changing stations' IDs and coordinates, the
original station IDs and coordinates should be maintained in the station shapefiles in
"OrignlLat", "OrignlLong", and "OrgStatnID" fields. Doing so will help match reformatted station
names and locations to the original names used in other files.
5. Populate the "RMNRegion" field with the region each station is in. Examples are
"Northeast", "Southeast", and "Region 7".
6. Download an HUC8 boundary shapefile or geodatabase for the area covered by
RMN_primary_secondary_[STATEABBRV].shp from the U.S. Geological Survey (USGS)
Watershed Boundary Dataset website2 and add it into your ArcMap Table of Contents.
Populate the "HUC8" field by either manually entering the HUC8 in which each sampling
station is located or spatially joining3 the HUC8 file to the sampling station shapefile and
then using "Calculate Attribute" to copy the HUC8 values into the HUC8 field of
RMN_primary_secondary_[STATEABBRV].shp.
7. Download a shapefile of the EPA Level 3 and Level 4 ecoregions.4 Spatially join the
ecoregions to the sampling stations and delete all fields except "US_L4CODE",
"US_L4NAME", "US_L3CODE", "US_L3NAME", "NA_L3CODE", "NA_L3NAME",
"NA_L2CODE", "NA_L2NAME", "NA_L1C0DE", and "NA_L1NAME" (i.e., delete fields that do
not have ecoregion names or codes). No preprocessing is required before the spatial join.
8. Download NHDPIus Version 2 (NHDPIusV2) digital elevation models (DEMs)
("NEDSnapshot.7z") and flowlines ("NHDSnapshot.7z") from the NHDPlusV2 homepage5 for
all land areas likely to be covered by watershed delineations (see Figure 2-3). Some states
may have their own high-resolution DEMs, but for consistency among states, the
NHDPIusV2 DEMs should be used. NHD flowlines are useful for "ground-truthing" the
streams identified during watershed delineation in ArcMap. NHD flowline files will be
referred to as NHDFIowline.shp. DEMs will be used during both watershed delineation and
characterization, so do not delete them until characterization is complete and checked. For
large states with monitoring stations spread out over a wide area, it may be necessary to
1lndex of Watershed Integrity: https://cfpub.epa.gov/si/si public record report.cfm?djrEntrvld=309175.
2USGS HUC8 files: ftp://rockvftp.cr.usgs.gov/vdeliverv/Datasets/Staged/Hydrographv/WBD/.
3Spatial join: http://pro.arcgis.com/en/pro-app/tool-reference/analysis/spatial-ioin.htm.
4EPA Level 3 and Level 4 Ecoregions:
https://www.epa.gov/eco-research/level-iii-and-iv-i
5NHDplusV2 homepage: http://www.horizon-systenis.com/NHDPlus/NHDPlusV2 data.php.
5
-------
download DEMs from several NHD subregions; each region should have just one flowline
shapefile.
Preprocessing is now complete and you can delineate the watersheds for these sampling stations.
File
Name
HTTP
Download
FTP
Download
0release_notes_VPU02.pdf
NHDPIusV21_MA_02_02a_CatSeed_01.7z
NHDPIusV21_MA_02_02a_FdrFac_01.7z
NHDPIusV21_MA_02_02a_FdrNull_01.7z
N H D PI usV21_M A_02_02a_Fi 11 ed Areas_01. 7z
NHDPIusV21_MA_02_02a_Hydrodem_01.7z
N H D PI usV21_M A_02_02a_N EDSra pshot_01. 7z
NHDPlusV21_MA_02_02b_CatSeed_01.7z
NHDPIusV21_MA_02_02b_FdrFac_01.7z
NHDPIusV21_MA_02_02b_FdrNull_01.7z
NHDPIusV21_MA_02_02b_FilledAreas_01.7z
NHDPIusV21_MA_02_02b_Hydrodem_01.7z
NHDPIusV21_MA_02_02b_NEDSnapshot_01.7z
HTTP
HTTP
HTTP
FTP
FTP
FTP
FTP
FTP
FTP
FTP
FTP
FTP
FTP
FTP
FTP
FTP
HTTP
HTTP
HTTP
HTTP
HTTP
HTTP
HTTP
HTTP
HTTP
HTTP
Figure 2-3. Excerpt of the NHDPIus Version 2 download page for the Mid-Atlantic.
Many NHD regions are divided into multiple subregions, which makes file sizes
smaller; the Mid-Atlantic has two subregions (a and b). Each subregion has its own
DEM file, among other files. The DEMs ("NEDSnapshot") for the Mid-Atlantic region
are boxed. Each region has one NHD flowline file called "NHDSnapshot".
6
-------
3. Watershed delineation
Delineating watersheds is the precursor to characterizing them. This manual describes two watershed
delineation methods: (1) the USGS StreamStats1 web tool and (2) a series of geoprocessing tools in
ArcMap. Both methods allow users to delineate watersheds from whatever points on streams they
choose (i.e., users are not limited to delineating watersheds in increments of NHD catchments, the
nearest USGS gage, etc.). Whenever possible, delineation using the first method (StreamStats) is
preferable to delineation using ArcMap's tools because of its simplicity, lower time and technical
requirements, and lower risk of delineation errors.
Both methods require aligning sampling stations to flowline rasters, which are the rasterized depictions
of the paths along which water accumulates (i.e., where streams and rivers occur according to DEMs).
Flowline rasters will generally closely match their corresponding NHD stream flowlines (see Figure 3-1).
Aligning sampling stations to flowline rasters tells StreamStats and ArcMap which stream raster cells
flow to which sampling stations, allowing the programs to delineate the correct watershed boundaries
for the given sampling stations.
Figure 3-1. StreamStats flowline raster (light blue boxes) and NHD flowline (dark blue
lines).
1USGS StreamStats home page: https://water.usgs.gov/osw/streamstats/.
7
-------
3.1. For states with StreamStats batch processor available
StreamStats is an online service that delineates watersheds given one or more points on streams and
calculates various hydrologic and drainage area properties. States partner with the USGS to develop a
StreamStats application for their state. While most states have fully functional StreamStats applications,
a few do not (as of October 2017).1 For states with functional StreamStats, Version 3 has a batch
processing feature to which users can submit a point shapefile and receive a geodatabase containing
watershed delineations for those points. Delineation procedures are the same for all states with a
StreamStats application.
To use the batch delineation2 feature of StreamStats, sampling stations must be aligned to StreamStats
flowline rasters, the rasterized depiction of where the streams are. (Refer to the StreamStats website for
details on how USGS created StreamStats' flowline rasters.) StreamStats flowline rasters do not always
match the exact course of rivers and streams but they are generally close; regardless, sampling stations
must be aligned to the flowline rasters for delineation to work. Each state with a functioning
StreamStats batch processor has its own flowline raster in a state-appropriate projection.
1. Download the StreamStats flowline raster(s) for the state(s) with which you are working.3
2. Each state's StreamStats flowline raster is in a different, state-appropriate projection.
Because sampling stations must be in the same projection as their state's flowline raster,
the sampling station shapefile for each state must be reprojected4 into the StreamStats
raster's projection (see Figure 3-2). The reprojected sampling stations should be named
RMN_primary_secondary_[STATEABBRV]_reproj.shp. If the sampling stations' shapefiles
are not reprojected to the flowline rasters, misalignment of the sampling stations to the
flowline rasters and incorrect delineations could occur.
^he status of StreamStats state applications is here: httpi//water.usgs.gov/osw/streamstats/ssonline.html.
2USGS StreamStats batch feature: http://streamstatsags.cr.usgs.gov/ss bp/.
3StreamStats flowline raster download:
4Reprojection: http://desktop.arcgis.eom/en/arcmap/10.3/tools/data-managetTient-toolbox/proiect.littTi.
8
-------
4 Project
I ° Ifall S3 I
Input Dataset or Feature Class
| RMN_primary_secondary_MD
Input Coordinate System (optional)
"3
NAD_1983_StatePlane Deiaware_FFS_0700
Output Dataset or Feature Class
onitoring_Networks\GIS\Mid-AtlanticVWN _primary_secondary_MD_reproj.shp
i Output Coordinate System
Geographic Transformation (optional)
a
0
s
a
| I Preserve Shape (optional)
Maximum Offset Deviation (optional)
OK ) ) Cancel | | Environments... | [ Show Help »
Spatial Reference Properties [ S3
XY Coordinate System [z Coordinate System |
Tf: w | Type here to search ~ Q. jgL'j | ~
B Favorites
© NAD_1983_UTM_Zcme_18N
B Q Geographic Coordinate Systems
E £3 Projected Coordinate Systems
B IS Layers
B Q NAD_1983_AI bers
& md_str900
B © N AD_1983_StateP I a n e_Del awa re_RP S_0700
Current coordinate system:
NAD_1983_Albers
Authority: Custom
Projection: Albers
False_Easting: 0.0
False_Northing: 0.0
Centrai_Meridian: -96.0
Standard_Parallel_l: 29.5
Standard_Parallel_2: 45.5
Latitude_Of_Origin: 23.0
Linear Unit: Meter (1.0)
OK | | Cancel |
Figure 3-2. Reprojecting sampling stations into projection of StreamStats flowline
raster. The desired projection is that of "md_str9QQ'".
3. Copy the reprojected sampling station shapefile, naming it
RMN_primary_secondary_[STATEABBRV]_reproj_aligned.shp. The sampling stations in this
shapefile will eventually be aligned to their flowline raster and submitted to StreamStats for
delineation.
4. Load RMN_primary_secondary_[STATEABBRV]_reproj_aligned.shp and its corresponding
flowline raster into ArcMap's Table of Contents. Load the study area's NHD flowlines into
the Table of Contents. Add a basemap with aerial imagery for further visualization of rivers
and streams, if desired.
5. Enter editor mode1 and make
RMN_primary_secondary_[STATEABBRV]_reproj_aligned.shp editable (see Figure 3-3).
Zoom to the first station in the attribute table of the shapefile.
1Editor mode:
http://desktop.arcgis.eom/en/arcmap/10.3/manage-data/editing-fundamentals/starting-an-edit-session.htm.
9
-------
Start Editing
This map contains data from more than one database or folder,
Please choose the layer or workspace to edit,
^va_str900.vat
RMN_primary_secondary_VA_reproj_aligned
Source
1,3 C:yjsersVdgibbspocurnents\Projects'flegio...
l3 c : Vjsers 'vdgibbs '(documents projects V"egiona.,,
Type
Shapefiles / dBase Files
Arclnfb Workspace
About editing and workspaces
OK
Cancel
Figure 3-3. Entering editor mode for the sampling station shapefile
RMN_primary_secondary_[STATEABBRC]_reproj_aligned.shp. The two files shown in
the window are the two files present in ArcMap's Table of Contents.
6. Align each sampling station to the flowline raster by moving the sampling stations to the
appropriate pixel of the flowline raster (see Figure 3-4). Sampling stations can be moved1 by
clicking on and dragging them to the desired location. There are four general situations for
the alignment of sampling stations to flowline rasters:
a. A sampling station could be located directly on the flowline raster (see Figure 3-4a). The
station does not need to be moved in order to align it with the flowline raster
(see Figure 3-4b).
b. A sampling station could be located off of the flowline raster but in a location where it is
easy to tell where it should be on the flowline raster to within a few pixels (see Figure
3-4c). The station should be moved to a plausible flowline raster pixel (see Figure 3-4d).
Do not worry about aligning to the exact stream pixel that corresponds with where
sampling occurs.
1Moving features: http://webhelp.esri.com/arcgisdesktop/9.3/index.cfm?TopicName=Moving features.
10
-------
c. A sampling station could be at the confluence of multiple flowlines and it is not clear
whether it is on one of the branches or downstream of the confluence (see Figure 3-4e).
Clarification about the location of the station should be sought from site information
databases, if available (e.g., USGS site inventory website1), or from the source agency,
and the station moved accordingly (see Figure 3-4f). Whether the station is on one of
the branches or below the confluence can dramatically affect the watershed's
boundary.
d. A sampling station is not clearly near any flowline raster. Check the NHD flowlines in the
area to see if NHD shows any rivers or streams nearby. Clarification about the location
of the station should be sought from the source agency and the station moved
accordingly. It may be that the coordinates provided by the source are incorrect or that
the watershed is very small (less than 900 pixels, or 810,000 m2 when each pixel is
30 x 30 m). If the latter, you can manually delineate using the DEM and NHD flowlines.
.
%
a
-V
•
¦
c
¦
V
1
b
Q.
¦
_¦
¦
J
jr
_¦
¦
i
i
- f
Figure 3-4. Three cases of sampling station-flowline raster alignment. Green circle is
the sampling location as provided by the source agency. Purple triangle is the aligned
location. In (a), the station is already on the flowline raster (b). In (c), the station is
very close to the flowline raster and can be aligned with the flowline raster without
additional information (d). In (e), the station is at a confluence and additional
information is needed to determine whether sampling occurs on one of the branches
or below the confluence (f).
1USGS site inventory: https://waterdata.usgs.gov/nwis/inventorv.
11
-------
7. Any steps needed to align each station with its flowline raster should be noted in the
"Align_Note" field of the attribute table of
RMN_primary_secondary_[STATEABBRV[_reproj_aligned.shp (see Figure 3-5), especially if
the location of the sampling station relative to the flowline raster needs clarification from
the source agency. Include the name of the database or person that clarified the location of
the sampling station and the date it was clarified. Frequently save1 your edits (changes to
point locations and new text in "Align_Note" field).
mm
PourPtLat
Align note
PourPtLong
QritplLat
OrignlLorw
i Wcvatf
j Mc.-ed
i Wc.-ed
j Didnt need to neve
j Confluence Meghan Latly 13,'16/16
j l.'c.ed
Meghan La fly i3,'17/16) said they
:l 474758
41 Sic72f
~il 545*12
iZ 02E-322
1.1.410127"
~41.-6C295
~<1.9iC&29
-71 £ "5 >1n
2^'zin
-72 13K27
-72 S322?7
-72 22*123
-72
-72 !;228&2
41.47482
" 41.9287
41.946389
42 025:22
1,1.410047""
41 4-SC303"
i1.&4€285
-71.82424
-73.276 5
-72.639167
-7233BJ57'
"-72.32£s25
-72 3242S5
-72 82S1€7
Figure 3-5. Attribute table with alignment notes, original pour point locations, and
aligned pour point locations.
8. Add NHDFIowline.shp from the NHDPIus download to the Table of Contents. Zoom to the
first station in the RMN_primary_secondary_[STATEABBRV[_reproj_aligned.shp attribute
table. Fill in the "COMID" field for each station based on the nearest NHD COMID.
9. Save your edits again and quit editor mode, then use the "Calculate Geometry"2 option for
fields in the attribute table to populate the "PourPtLat" and "PourPtLong" fields with the
latitudes and longitudes of the sampling stations (now the pour points for the watersheds)
(see Figure 3-6). At this point, the station shapefile will have four populated coordinate
fields: "OrignlLat", "OrignlLong", "PourPtLat" and "PourPtLong" (see Figure 3-5).
1Saving edits:
http://desktop.arcgis.eom/en/arcmap/10.3/manage-data/editing-fundamentals/stopping-an-edit-session-stopping
-editing-.htm.
Calculating feature geometry:
http://desktop.arcgis.com/en/arcmai:
perties.htm.
12
-------
Calculate Geometry
Property:
Y Coordinate of Point
Coordinate System
o Use coordinate system of the data source:
| GCS: North American 1983
Use coordinate system of the data frame:
PCS: NAD 1983 StatePlane Connecticut FIPS 0600 Feet
Units:
Decimal Degrees
I I Calculate selected records only
About calculating geometry | Cancel
Figure 3-6. Populating the latitude field of a point shapefile's attribute table. Repeat
for the longitude field.
10. Once all stations in a state have been aligned with the StreamStats flowline raster, that
state's stations can be submitted to the StreamStats batch processor (see Figure 3-7).
StreamStats' "Local ID Field" should be the shapefile's "StationID". Depending on your
objective, it is not necessary to have StreamStats compute any basin characteristics or flow
statistics because each state's applications calculate different parameters, and therefore,
the StreamStats-generated statistics are not comparable across states.
13
-------
c o A Secure | https://streamstatsags.cr.usgs.gov/ss_bp/
UUJAI..UJ-L LVUUl.
StreamStats Batch Processing Tool
This tool produces shapefiles that contain the delineated basins, basin characteristics, and flow statistics for multiple sites requested at once by users.
Before this tool can be used, the the points of interest will likely need to be edited in GIS so that they are coincident with the stream grid used by
StreamStats for delineations and saved to a shapefile. Click here to select and download the stream grid for your area of interest. The number of points in
the shapefile generally should not exceed 200. Insert the filenames for the shapefile of snapped points of interest below. The batch process will delineate
the drainage areas and if checked, will compute basin characteristics, and/or estimate flow statistics by use of regression equations for the selected points.
The user will be notified by email where to pick up the results when they are done.
State
Abbrev:
Local ID Field: Station ID
: ct_ ¦
Enter email Address for completion notification:
org
Delineate
Compute Basin Chars ~ Compute Flow Stats
Select the 4 files to upload a shapefile
.SHP file
.DBF file
.PRJ file
.SHX file
Help
RMN_primary_secondary_CT_FIPS_aligned.shp
RMN_primary_secondary_CT_FIPS_aligned.dbf
RMN_primary_secondary_CT_FIPS_aligned.prj
RMN_primary_secondary_CT_FIPS_aligned.shx
Submit to Queue
Figure 3-7. StreamStats batch processor upload interface.
11. Up to several hours after the request is submitted, StreamStats will send an e-mail with a
link to a database that contains the delineation feature class. Download the database to the
appropriate directory on your computer. The relevant feature class is called
GlobalWatershed[STATEABBREV]. Open this in ArcMap and confirm the following: (1) the
number of watersheds in the output files equals the number of stations submitted to
StreamStats, (2) the watershed boundaries visually make sense when overlaid with the
StreamStats flowline raster (see Error! Reference source not found.), and (3) sampling
stations near confluences include only the desired stream reach(es). Two common reasons
for watersheds missing from the StreamStats feature class are that some StationlDs in the
sampling station shapefile include special characters other than underscores, and that the
stations are not coincident with the StreamStats flowline raster.
12. Table join1 the sampling station shapefile to the delineation feature class using the
"StationID" and "Name" fields, respectively. ("Name" is what StreamStats uses as the
unique ID field; its entries should be the same as the "StationID" entries in the sampling
station shapefile.) Export the joined station-delineation feature class to a new shapefile. This
will permanently associate the delineations with their sampling stations' attributes.
¦'Table join:
http://desktop.arcgis.eom/en/arcmap/10.3/manage-data/tables/ioining-attributes-in-one-table-to-another.htm.
14
-------
Figure 3-8. Sample StreamStats watershed delineation with StreamStats flowline
raster. Sampling station is the purple triangle.
13. Use "Calculate Geometry" on the field "Area_km2" to populate the field with the area of
each watershed in square kilometers in the new shapefile.
14. Although having a single DEM for all the RMN watersheds in each state is not part of
watershed delineation, it is necessary for every state during watershed characterization and
is best done at this point. To get a single DEM that covers all the RMN watersheds in a state,
load the delineation shapefile (see Step 12) and the NHD DEMs that are necessary to cover
the watersheds into the Table of Contents. Merge them using the "Mosaic to New Raster"
tool (see Figure 3-9). "Number of Bands" should equal 1 and "Pixel Type" equal
"32_BIT_FLOAT". Running this for large NHD subregions can take several hours.
At this point, that state's delineations are completed and watershed characterization can begin.
15
-------
Mosaic To New Raster
r^l fallal
Input Rasters
m
0
a
a
v )elev_cm_ma02a
^ >elev_cm_ms05d
<*>elev_cm_tn06a
Output Location
C:\Jsers\dgibbspocuments\GIS general layers VJHDplusV21_DEM
Raster Dataset Name with Extension
va_rmn_demcm
Spatial Reference for Raster (optional)
Pixel Type (optional)
32_B1T_FL0AT
Cellsize (optional)
Number of Bands
Mosaic Operator (optional)
Mosaic Colormap Mode (optional)
FIRST
OK | [ Cancel | [ Environments... ] [ Show Help » ]
Figure 3-9. "Mosaic to New Raster" configuration for multiple DEMs in the same study
area.
3.2. For states without StreamStats batch processor available
For stations in states that do not have the StreamStats batch processor, watersheds can be delineated
using a series of geoprocessing tools in ArcMap. The methods below are essentially those found on
many websites (e.g., Tufts University1 and Trent University2 sites), but modified slightly for RMNs. They
deviate most from typical delineation procedures when any watersheds being delineated are nested
within others being delineated.
1. Activate the Spatial Analyst and 3D Analyst extensions in ArcMap under
Customize > Extensions and check Spatial Analyst and 3D Analyst. You will need these
extensions for watershed delineation without StreamStats.
2. It is necessary to have one DEM that covers the entire area covered by the watershed
delineations. If multiple subregions of NHD DEMs are necessary to fully cover the areas that
will be in the delineations, merge them using the "Mosaic to New Raster" tool3 (see Figure
3-9). "Number of Bands" should equal 1 and "Pixel Type" should equal "32_BIT_FLOAT".
1http://sites.tufts.edu/gis/files/2013/ll/Watershed-and-Drainage-Delineation-bv-Pour-Point.pdf.
2http://www.trentu.ca/librarv/sites/default/files/documents/WatershedDelineation 10 2.pdf.
3Mosaic to new raster:
http://desktop.arcgis.eom/en/arcmap/10.3/tools/data-management-toolbox/mosaic-to-new-raster.htm.
16
-------
Running this tool for large NHD subregions can take several hours. Because the extent of the
watersheds is not yet known, deciding which DEM subregions to include involves some
guesswork.
3. Use the "Flow Direction" tool1 on the DEM covering the study area to create a raster
showing which direction water flows at each pixel.
4. Use the "Sink" tool2 to identify sinks in the flow direction raster covering the study area.
Sinks are cells that have lower elevation than all surrounding cells; water cannot flow
outward from such cells and they cause erroneous delineations. NHD DEMs should not have
any sinks in them; sinks will already have been "filled" before the DEMs were posted. If no
sinks are identified in this step, skip to Step 7.
5. If Step 4 identified any sinks, use the "Fill" tool3 on the DEM covering the study area to
remove sinks in the study area. This tool may take a lot of time and memory to run.
6. If the study area DEM had any sinks and had to be filled, use the "Flow Direction" tool
(see Step 3) on the filled DEM (see Step 5) to get a corrected flow direction raster.
7. Use the "Flow Accumulation" tool4 on the flow direction raster (from Step 3, if no DEM
filling necessary; from Step 6, if DEM filling necessary) to create a raster called
[STATEABBRV]_fill_accum showing how many pixels drain to each downslope pixel in the
study area.
8. Export the flow accumulation raster [STATEABBRV]_fill_accum to a TIF with
"32_BIT_SIGNED" pixels called [STATEABBRVl_fill_accum.tif (see Figure 3-10).
1Flow direction: httpi//pro.arcgis.com/en/pro-app/tool-reference/spatial-analyst/flow-direction.htm.
2Sink: httpi//pro.arcgis.com/en/pro-app/tool-reference/spatial-analyst/sink.htm.
3Fill: httpi//desktop.arcgis.com/en/arcmap/10.3/tools/spatial-analvst-toolbox/fill.htm.
4Flow accumulation: http://pro.arcgis.com/en/pro-app/tooj-reference/spatial-anajyst/fjow-accumulation.htm.
17
-------
Copy Raster 1 ~ | [ E] 11 ES [
Input Raster
| C:\Users\dgibbs\Documents\Projects\Regional_Monitoring_Networks\GIS\Region7\MO_watershed_deliri\mo_fill_accum
Output Raster Dataset
C:\psers\jdgibbs documents \ProjectsViegional_Monitoring_Netvv,orks1\GIS Region 7,^10jA|atershed_delin Vno_fill_accum.tif j[^|
Configuration Keyword (optional)
Ignore Background Value (optional)
NoData Value [optional)
-3.402823e+0 38
~ Convert 1 bit data to 8 bit {optional)
D Colormap to RGB (optional)
Pixel Type (optional)
32_BU_SIGNED ~
~ Scale Pixel Value (optional)
O RGB To Colormap (optional)
OK [ | Cancel j | Environments,.. | | Show Help >> |
Figure 3-10. Exporting flow accumulation raster to 32-bit signed TIF.
9. Use the "Build Raster Attribute Table" tool1 to create an attribute table for the TIF flow
accumulation raster [STATEABBRVl_fill_accum.tif (see Figure 3-11).
Build Raster Attribute Table
1 a || B II S3
Input Raster
1
| mo_fill_accum.tif
~
O Overwrite (optional)
OK J | Cancel | (Environments.,. | j Show Help >> j
Figure 3-11. Creating an attribute table for the flow accumulation raster.
10. Use the "Extract by Attributes" tool2 (Spatial Analyst extension required) to convert the flow
accumulation raster into a flowline raster. Inputting 900 for the "Where clause" field will
produce a raster in which the flowlines drain more than 900 pixels (810,000 m2 for
30 x 30 m pixels) and generally provides well-defined flowline rasters for headwaters
(see Figure 3-12). Flowline rasters with a minimum drainage area of 900 DEM pixels results
1Build raster attribute table;
http://desktop.arcgis.eom/en/arcmap/10.3/tools/data-management-toolbox/build-raster-attribute-table.htm.
2Extract by attributes: http://pro.arcgis.com/en/pro-app/tool-reference/spatial-analvst/extract-bv-attributes.htm.
18
-------
in a reasonable number of streams for these purposes. Change the color of the flowline
raster to blue. This flowline raster is a rough, less processed version of the flowline raster
downloaded from StreamStats. The flowline raster should approximate the NHD flowlines.
\ Extract by Attributes
_=JD=>
Input raster
>
| ne_fill_accum.tif
d
&
Where dause
"Value' >900
1=1
SQL
Output raster
C: \Users\dgibbs documents projects V^egional_Monitoring_Net,Alorks\GIS Region7VJE_watershed_delin Vie_flow_raster_900. tif
ft
-
OK | | Cancel | | Environments.., | | Show Help » |
Figure 3-12. Converting the flow accumulation raster into a flowline raster.
11. Reproiect1 RMN_primary_secondary_[STATEABBRV].shp to the projection of the DEM/flow
accumulation grid, naming it RMN_primary_secondary_[STATEABBRV]_reproj.shp. Not
reprojecting the sampling station shapefiles to the projection of the flowline rasters could
lead to misalignment of the sampling stations to the raster and incorrect watershed
delineations.
12. Copy the reprojected sampling station shapefile from Step 11, naming it
RMN_primary_secondary_[STATEABBRV]_reproj_aligned.shp. The sampling stations in this
shapefile will be aligned to their flowline raster (from Step 9) for delineation. Load the study
area's NHD flowlines into the Table of Contents. Add a basemap with aerial imagery for
further visualization of rivers and streams, if desired.
13. Enter editor mode2 and make
RMN_primary_secondary_[STATEABBRV]_reproj_aligned.shp editable (see Figure 3-3).
Zoom to the first station in the shapefile.
14. Align each sampling station to the flowline raster by moving the sampling stations to the
appropriate pixel of the flowline raster (see Figure 3-4). Sampling stations can be moved3 by
clicking on and dragging them to the desired location. There are four general situations for
the alignment of sampling stations to flowline rasters:
1Reproject: http://desktop.arcgis.eom/en/arcmap/10.3/tools/data-management-toolbox/proiect.htm.
2Enter editor mode:
http://desktop.arcgis.eom/en/arcmap/10.3/manage-data/editing-fundamentals/starting-an-edit-session.htm.
3Moving features: http://webhelp.esri.com/arcgisdesktop/9.3/index.cfm?TopicName=Moving features.
19
-------
a. A sampling station could be located directly on the flowline raster (see Figure 3-4a). The
station does not need to be moved in order to align it with the flowline raster
(see Figure 3-4b).
b. A sampling station could be located off of the flowline raster but in a location where it is
easy to tell where it should be on the flowline raster to within a few pixels (see Figure
3-4c). The station should be moved to a plausible flowline raster pixel (see Figure 3-4d).
Do not worry about aligning to the exact stream pixel that corresponds with where
sampling occurs.
c. A sampling station could be at the confluence of multiple flowlines and it is not clear
whether it is on one of the branches or downstream of the confluence (see Figure 3-4e).
Clarification about the location of the station should be sought from site information
databases, if available (e.g., USGS site inventory website1), or from the source agency,
and the station moved accordingly (see Figure 3-4f). Whether the station is on one of
the branches or below the confluence can dramatically affect the watershed's
boundary.
d. A sampling station is not clearly near any flowline raster. Check the NHD flowlines in the
area to see if NHD shows any rivers or streams nearby. Clarification about the location
of the station should be sought from the source agency and the station moved
accordingly. It may be that the coordinates provided by the source are incorrect or that
the watershed is very small (less than 900 pixels, or 810,000 m2 when each pixel is
30 x 30 m). If the latter, you can either manually delineate using the DEM and NHD
flowlines or repeat Steps 10-13 using a smaller watershed threshold size (e.g.,
450 pixels).
15. Any steps needed to align each station with its flowline raster should be noted in the
"Align_Note" field of the attribute table of
RMN_primary_secondary_[STATEABBRV[_reproj_aligned.shp (see Figure 3-5), especially if
the location of the sampling station relative to the flowline raster needs clarification. Include
the name of the database or person that clarified the location of the sampling station and
the date it was clarified. Frequently save2 your edits (changes to point locations and new
text in "Align_Note" field).
16. Add NHDFIowline.shp from the NHDPIus download to the Table of Contents. Zoom to the
first station in the RMN_primary_secondary_[STATEABBRV[_reproj_aligned.shp attribute
table. Fill in the "COMID" field for each station based on the nearest NHD COMID.
1USGS site inventory: https://waterdata.usgs,gov/nwis/inventory.
2Saving edits:
http://desktop.arcgis.eom/en/arcmap/10.3/manage-data/editing-fundamentals/stopping-an-edit-session-stoppir
-editing-.htm.
20
-------
17. Save your edits again and quit editor mode, then use the "Calculate Geometry"1 option for
fields in the attribute table to populate the "PourPtLat" and "PourPtLong" fields with the
latitudes and longitudes of the sampling stations (now the pour points for the watersheds)
(see Figure 3-6). At this point, the station shapefile will have four populated coordinate
fields: "OrignlLat/' "OrignlLong/' "PourPtLat/' and "PourPtLong" (see Figure 3-5).
At this point, the sampling stations have been aligned to flowlines and everything is prepared for
delineating the watersheds. The procedures take two routes here, depending on whether any sampling
stations are nested inside each other. Nesting occurs when one RMN sampling station is upstream of
another one (see a), as opposed to independent branches (see b). Nested sampling stations produce
nested watersheds. Delineation is simpler if no watersheds are nested. You can map the sampling
station shapefile against the flowline raster to check for nested sampling stations.
Calculating feature geometry:
http://desktop.arcgis.com/en/arcmai:
perties.htm.
21
-------
Figure 3-13. Examples of nested and unnested watersheds. Green circles are sampling
stations and blue lines are flowline rasters, (a) Example of triply nested watersheds.
The watershed outlined in black is the outermost, the watershed outlined in red is the
middle level, and the watershed outlined in brown is the innermost, (b) Example of
two non-nested, adjacent watersheds (light and dark gray polygons).
22
-------
If no watersheds are nested:
18. Use the "Watershed" tool1 on the filled flow direction raster and the aligned pour point
shapefileto delineate watersheds. For the "Pour point field", select "FID" (see Figure 3-14)
(the "feature ID", an auto-assigned unique identifier for each feature). This will produce a
raster in which each pixel found to be in a watershed has the same value as the FID. Each
watershed will have a different value pixel according to its pour point FID.
\ Watershed
0 S3
Input flow direction raster
| wv_f i 1 l_d i r
d
£5
Input raster or feature pour point data
| RMN_primary_secondary_WV_aligned
jd
Pour point field [optional)
FID
~
Output raster
d cumen ts projects V? eqional_Monitoring_Net'Alork5 \GI5 \Mid-A tiantic\WV_Yva tershed_deliny.'v_delinras t
OK
Cancel
Environments..
Show Help >>
Figure 3-14. Interface for the ArcMap "Watershed" tool.
19. Check that each sampling station has a corresponding watershed delineation raster. A
sampling station missing a delineation raster may not have been properly aligned to its
flowline raster.
20. Use the "Raster to Polygon" tool2 on the watershed delineation raster (see Step 19) to
convert the delineation raster to delineation polygons. Use the "Simplify" option. This tool
will output a polygon shapefile in which each feature is one watershed boundary.
21. Table join the aligned sampling station shapefile to the delineation polygon shapefile using
the "FID" field in the station shapefile and the "GRIDCODE" field in the polygon shapefile.
22. Export3 the joined station-delineation shapefile to a new shapefile. This will permanently
associate the delineations with their sampling stations' attributes. This shapefile will be used
for watershed characterization. Check the delineation polygons against the flowline raster
(see Figure 3-15), NHD flowlines, and DEM to make sure they make sense. Correct the
1Watershed delineation: http://pro.arcgis.com/en/pro-app/tool-reference/spatial-analvst/watershed.htm.
2Raster to polygon: http://pro.arcgis.com/en/pro-app/tool-reference/conversion/raster-to-polygon.htm.
Exporting a shapefile to a new shapefile: http://support.esri.com/en/technical-article/000004655.
23
-------
delineation boundaries as needed in an edit session. For example, addend move2
delineation vertices to extend the boundary around the outsides of the flow/lines, tracing
along an estimated drainage divide using a DEM for support.
Figure 3-15. Incorrect watershed delineation by Watershed tool in ArcMap. The
delineation raster (green) does not extend to the northern ends of the flowline raster
(blue lines). The sampling station is black circle.
23. Use "Calculate Geometry" on the field "Area_km2" to populate the field with the area of
each watershed in square kilometers in the new shapefile.
This concludes the procedures for delineating non-nested watersheds using ArcMap. Watershed
characterization can begin (see Section 4).
Nested sampling stations make the "Watershed" tool produce erroneous results because each pixel
output from the "Watershed" tool can only be assigned one watershed value and nested sampling
stations, by definition, have more than one watershed value at each pixel within the inner watershed.
Thus, nested sampling stations cannot be delineated during the same run of the "Watershed" tool; they
must be delineated with separate runs of the "Watershed" tool to get complete delineations of each
nested watershed. The process below essentially delineates every watershed separately (regardless of
whether or not it is nested), then converts each one to a watershed polygon and merges all those
polygons into one shapefile.
xAdd vertices in edit mode:
http://desktop.arcgis.eom/en/arcmap/10.3/manage-data/editing-existing-features/adding-a-vertex-rnanuallv.htm.
2Move vertices in edit mode:
http://desktop.arcgis.eom/en/arcmap/10.3/manage-data/editing-existing-features/moving-a-vertex-bv-dragging-it
.htm.
24
-------
If any watersheds are nested:
18. Use the Model Builder tool called Watershed_delin_by_each_feature_20160824to get a
separate delineation raster for each sampling station with each station's FID at the end of its
name. Check that each sampling station has a corresponding watershed delineation raster.
A sampling station missing a delineation raster may not have been properly aligned to its
flowline raster.
19. Use the Model Builder tool called Watershed_rasters_to_polygons_20160824to convert
each watershed raster into its own watershed polygon. These watershed polygons will be
named with the FID of the sampling station IDs they correspond to.
20. Use the "Merge" tool1 to combine all the individual watershed polygons into a single
polygon shapefile. The names of the polygons in this shapefile will be the FID of the
sampling stations.
Proceed through Steps 21 through 23 for non-nested watersheds to finish delineating the nested
watersheds. This concludes the procedures for delineating nested watersheds using ArcMap. Watershed
characterization can begin (see Section 4).
1Merge: http://prq.arcgis.com/en/pro-app/tooj-reference/data-management/merge.htm.
25
-------
4. Characterization of delineated watersheds
Once all the target watersheds have been delineated, watershed characterization can begin. This
manual describes five characterization steps (see Figure 1-1). Using the automated tools described here
standardizes processing among agencies and users, reduces human error, expedites processing, and
makes results more reproducible. If the tools cannot be run on your computer, this guide also describes
the intent and output of the tools so the processes can be implemented by alternative means (manually,
modifying the scripts, etc.). (Descriptions of the Python scripts can be found in Section 6.)
4.1. Process overview
Corresponding watershed delineation and sampling station shapefiles are input to tools that calculate or
help calculate the following watershed characteristics: (1) the fraction of land under each NLCD land use
within the whole watershed and within 1 km and 5 km of the sampling stations within the watershed
(i.e., the entirety of the watershed that is within 1 km and 5 km of the sampling station); (2) the
percentage of stream flow that comes from the base flow fraction at the sampling station; (3) the
channel slope and sinuosity upstream and downstream of the station; (4) the slope of the watershed
(average, minimum, maximum, range, standard deviation); and, (5) the number of dams, mines, NPDES,
and CERCLA sites within each watershed. The processes that calculate 3 and 5 involve a few minutes of
manual processing per station; the processes that calculate 1, 2, and 4 are completely automated. The
output of all processes should be checked.
In each process, calculated watershed characteristics become new fields that are added to the attribute
table of the state's watershed delineation shapefile. To facilitate data analysis and sharing, it is
recommended that the output shapefile from each step be used as an input to the next step. This
workflow accumulates watershed property information in the delineation shapefiles. For example, if the
delineation shapefile with land use composition is input into the base flow tool, the resulting delineation
shapefile will have both land use composition and base flow information. Using the workflow shown in
Figure 1-1, the shapefile output from the dam/mine/NPDES/CERCLA step will have fields with the output
of all the preceding steps. Additional characterizations can be added after these and the results
appended as new columns.
Below are a few general notes about using these tools to help with characterization:
• The sampling stations and watersheds must have the same names in their respective
shapefiles, although their field names in their respective shapefiles can have different
names (e.g., "StationID" in the station shapefile and "NAME" in the delineation shapefile).
This condition should be met if the watershed preprocessing and delineation procedures
described in Sections 2 and 3, respectively, are followed.
• The interfaces of the tools default to "StationID" and "NAME" for the names of the fields for
the sampling stations and watershed names in the sampling station and watershed
shapefiles, respectively. Those should be the names for those fields if the preprocessing and
delineation procedures above are followed. The sampling station and watershed names
must not have any special characters except for underscores (_). The tools do not work with
26
-------
any other special characters, which should have been removed during predelineation
processing (see Section 2, Step 3).
• Most of the tools require the Spatial Analyst extension. One requires the 3D Analyst
extension and ArcGIS Desktop Advanced. Which extensions and licenses are needed for
each tool is noted below the tool's input files.
• Each tool displays information about its progress in the "Results" window1 as it runs
(see Figure 4-1). This information is useful for identifying problems if a tool stops
prematurely or produces unexpected results.
• Many of the tools automatically append a timestamp in the format of
"_YYYYMMDD_HH_MM_SS" to the end of output files. While the time a tool is run does not
affect the output, adding this unique identifier to the filename prevents the output from
accidentally being overwritten when the tool is rerun with the same data.
• Each tool has further documentation in the user interface, as well as commented code.
Results
X
~ Ofi Current Session
>
B % Land use composition
El O Inputs
El Environments
El 53 Messages
Ij] Executing; Land-use-composition RMN_primary_secondary_KS_RPS_aligned StationID rmn_primary_secondary_ks_wati
[T] Start Time: Sat Apr 0110:35:09 2017
E
Running script Land-use-composition...
£
NOTE: This tool may take up to 2 minutes to process each very large watershed.
Ij] Reprojecting sampling station and watershed delineation files to land use projection:
Reprojections complete
£
Creating buffers of 1000 and 5000 Meters around each sampling station:
QD Buffers created
0
Clipping 5000 Meters buffers to corresponding watershed delineations:
QD Clipping 5000 Meters buffer to Buck Creek watershed:
(T) Buffer clipped
(JD Clipping 5000 Meters buffer to Fox Creek watershed:
QD Buffer clipped
Figure 4-1. Sample Results window output from land use composition tool.
4.2. Loading the tools into ArcMap
The tools have two components: (1) the ArcPy scripts (*.py), and (2) the Regional Monitoring Network
toolbox (Regional_monitoring_network.tbx) (see Figure 4-2). The former are the actual directions for
1Results window:
http://desktop.arcgis.eom/en/arcmap/10.3/analvze/executing-tools/using-the-results-window.htm.
27
-------
the tools while the latter tells ArcMap how to create the standard ArcMap geoprocessing interface for
the tools. Keep these components in the same folder on your computer.
Name
_| B ase_fI ow_20160630. py
_| La nd_use_comp_20160601.py
3 NPDES_mines_dams_CERClA.201601101.py
Regional_monitoring_network.tbx
3) Slope_sinuosity_post_stream_trace_20160915.py
jj) Slope_sinuosity_pre_stream_trace_20160906.py
_j] Watershed_slope_20161005.py
Figure 4-2. List of files included in folder. There is one toolbox (*.tbx) and six tools (*.py).
1. If the tool folder is zipped, unzip it. The folder inside is called "RMN GIS tools".
2. Copy "RMN GIS tools" folder with the RMN geographic information system (GIS) tools inside
(*.py scripts and Regional_monitoring_network.tbx) to a local directory of your choice.
3. Navigate to that folder in ArcCatalog or the Catalog window of ArcMap. If
Regional_monitoring_network.tbx is not visible in the folder that it was moved to,
right-click on that folder and refresh it (see Figure 4-3).
4. Expand Regional_monitoring_network.tbx. Several scripts (e.g., -• 'Watershed slope j should
appear underneath (see Figure 4-3). You should be able to successfully run these scripts
now.
28
-------
Catalog
•j- - * ft his * I a i *s: i a
Loc&Svi! ^ 3e0onai nBrttninc netowwk.'ea:
El EG RegicnS
H fcD Region7
H £~ RMN 61S tori-;
El £j Southeast
« Example WHteran
Q Ndtl Nrtmjfk R(
(Si NJ.PA.WtBCjM
0 Nv firjl RMf'i !
QJ NY.firal.RMN s
Of] NV_s01nplin9.it
3 RMN.seitct.sW
3 fiMNLsitsjprore
£1 £j£ ftMhJ watoshK
01 RMN watershed
Cl 19 Watershed.char
0Q WatenhodLcharJ
W [ffl Wjte»sbKl_cltar.
Q'j VjV dclin.mid
El Lake RMNs
ID £j Mid-Atian:ic sites
S3 E^< National Networked Refcerce Watersheds
i
Copy
X
Delete
Ren tir*
J
Refresh
Cd'rt..
New
f
Add
~
5sve As
•
3
Item DettrpUon,.,
rf
Pecperti©-
EC1701G5.rn»d
rw.d
Kn.rrad
£)2.rro«d
ixd
c :t_dd i n_vs_ N * I D .cat ch men t_2C 15
Kectj^in_v3_NHD_catihment_Ii2J.5
d _u «_£« IX fifll p 611
B Q RMN GIS tools
~ l§l Regional_monitoring_network,tbx
X Base flow analysis
.4* Channel slope and sinuosity analysis, tool 1- before manual stream tracing
4* Channel slope and sinuosity analysis, tool 2- after manual stream tracing
4* Dam-mine-NPDES-CERCLA counts and properties
4" Land use composition
4^ Watershed slope
Watershed_delin_by_each_feature_201 &0S24
b Watershed_rasters_to_polygons_20160824
Figure 4-3. Displaying the RMN toolbox in ArcMap. To refresh, right-click on the
toolbox (a). The contents of the toolbox are shown in (b).
The tools have the standard ArcMap geoprocessing tool interface.
Each input field has help text associated with it, revealed by clicking the Help >> button in the
bottom right of the tool. Right-clicking on each tool in the catalog and clicking on "Item Description" will
also provide more information.
4.3. Characterization
If you have not already done so, activate the Spatial Analyst and 3D Analyst extensions in ArcMap under
Customize > Extensions and check Spatial Analyst and 3D Analyst. You will need them to run some of the
watershed characterization tools.
Below are descriptions of the five characterization processes.
29
-------
4.3.1. Land use composition
Objective:
To calculate the fraction of each land use within: (1) each RMN watershed, (2) 1,000 m of each
sampling station in its watershed, and (3) 5,000 m of each sampling station within its watershed.
NOTE: If other distances are desired, the values can be changed in the script.
Data use:
Land use is one of the primary characteristics for assessing whether a watershed is a "reference"
watershed. Reference watersheds have low fractions of developed and agricultural land and high
fractions of natural land. This step of screening allows assessment of what fraction of each RMN
watershed is "natural" versus "disturbed" within certain distances of RMN sampling stations.
Different levels of "developed" (i.e., land use codes (LUC) 22, 23, 24) or agricultural land uses (i.e.,
LUC 81 and 82) can be summed to produce totals for different disturbed categories.
Required input files:
1. Sampling stations aligned to flowline raster:
RMN_primary_secondary_[STATEABBRV]_reproj_aligned.shp
2. Watershed delineations with added characterizations thus far
3- NLCD land use raster1 (land use codes must be the same as the 2006/2011 NLCD codes2).
You do not need to do any processing of the downloaded NLCD raster before running this
watershed characterization tool. The tool will handle differences in projections between the
raster and shapefiles.
NOTE: This tool requires the Spatial Analyst extension (for the "Extract by Mask" command).
Output:
1. Watershed delineation shapefile with one field added for each land use-area extent
combination (whole watershed, 1,000 m, and 5,000 m). All added fields have the fraction of
that land use at that extent. Added fields are named according to the following scheme:
LU[LUC]_[DISTANCE] (bracketed text is variable), where [LUC] is the NLCD land use code
(e.g., 11, 72) and [DISTANCE] is 1,000, 5,000, or "whole." Thus, the field titled "LU71_whole"
shows the fraction of each watershed comprised of NLCD land use 71 and "LU11_1000"
shows the fraction of each watershed within 1,000 m of the sampling stations comprised of
NLCD land use 11.
Procedures:
Fill out the tool interface as each field directs (see Figure 4-4). See the tool help for more
information.
1National Land Cover Database (NLCD): https://www.rr
2NLCD 2006/2001 codes: http://www.mrlc.gov/nlcdll
30
-------
Land use composition
Sampling stations
| RMN_primary_secondary_MD_FIPS_aligned
&
Sampling station names
| StationID
Delineated watersheds
-1
| GlobalWatershedMD
Watershed names
NAME
Land use raster
-1
| nlcd_2011 _landcover_201 l_edition_2014_10_10. img
Output file name
MD_LU
Output location
| jmentsV5rojectsVlegional_Monitoring_Networks\GISVvlid-AtlanticV-and_use_composition |
£3
~ X
Land use composition
Analyzes the percent composition of land uses of entire watersheds and
within 1000 and 5000 m of sampling stations within watersheds (these
distances can be changed in the script). This tool outputs a watershed
polygon shapefile with fields containing the fraction of each watershed
under each land use and the fraction of the area within the specified
distances of the sampling stations under each land use. (For smaller
watersheds, the land use fractions within 5000 m and/or 1000 m of the
sampling station and for the whole watershed may be the same.) The
land use file used must be NLCD 2011.
This tool requires a Spatial Analyst extension to run.
OK Cancel | | Environments... | | << hide Help | Tool Help
Figure 4-4. Interface for land use composition tool. NOTE: tool interfaces may not
appear exactly as shown in this manual due to subsequent modification.
4.3.2. Base flow
Objective:
To calculate the base flow index at RMN sampling stations (i.e., at the mouth of each watershed).
Data use:
Base flow index (percentage) indicates what percentage of the streamflow at that location
originates from base flow. The index can be useful for determining how consistent streamflow is at
that location and how changes in precipitation are likely to impact stream hydrology. Streams with
high base flow index are more resistant to precipitation variability. Different fish and
macroinvertebrate taxa have varying sensitivities to hydrological alteration.
Required input files:
1. Sampling stations aligned to flowline raster:
RMN_primary_secondary_[STATEABBRV]_reproj_aligned.shp
2. Watershed delineations with added characterizations thus far
3. USGS base flow raster1 (scale of 0-100). You do not need to do any processing of the
downloaded base flow raster before running this watershed characterization tool. The tool
will handle differences in projections between the raster and shapefiles.
NOTE: This tool requires the Spatial Analyst extension (for the "Extract Values to Points" command).
1USGS base flow raster: https://water.usgs.gov/GIS/metadata/usgswrd/XML/bfi48grd.xml.
31
-------
Output:
1. Watershed delineation shapefile with a single field added ("BASE_FLOW") that contains the
value of the base flow index raster beneath the sampling station.
Procedures:
Fill out the tool interface as each field directs (see Figure 4-5). See the tool help for more
information.
Base flow analysis — ~ X
Base flow analysis
Extracts the percent of flow at each sampling station that
comes from base flow (ground water) and transfers those
values to a watershed delineation shapefile in a field called
"BASE_FLOW". A USGS baseflow raster can be downloaded
from:
https://water-Usos.Qov/GIS/metadata/usQswrd/XM L/bfi48Qrd.xml
This tool requires a Spatial Analyst extension to run.
< >
Tool Help
Figure 4-5. Interface for base flow tool.
4.3.3. Channel slope and sinuosity
Objective:
To calculate the channel slope and sinuosity within user-specified distances upstream and
downstream of RMN sampling stations. Channel slope is the change in elevation along a given
stream channel length divided by the channel segment length. Channel sinuosity is the length of a
segment of channel divided by the straight-line distance from one end of that channel segment to
the other.
Data use:
Channel slope and sinuosity are useful for classifying streams, which is one of the key potential
classification parameters used for RMN stations. They are also useful for screening out channels that
do not match RMN stream type targets.
Procedures overview (see Figure 4-6):
This characterization uses three processes, each with its own inputs and outputs:
1. Run the first custom tool to prepare to create stream traces (trace preprocessing),
Sampling stations
| RMN_primary_secondary_MD_FIPS_aligned
zl
t=3
Sampling station names
1 StationID
-1
Delineated watersheds
| MD_LU_20160613
zi
£3
Watershed names
I NAME
v|
Base flow raster
| bfi48grd
Output file name
MD_LU_BF
Output location
| ;V)oajmentsV5rojectsV*egional_Monitoring_Networks\GIS^'1id-AtlanticV-U-i-base_flow |
£5
OK Cancel Environments... << Hide Help
32
-------
2. Manually create stream traces, and
3. Run the second custom tool to calculate channel slope and sinuosity from the stream traces.
Calculating channel slope and sinuosity requires having polylines that follow the mainstem of
streams at specific distances upstream and downstream from the sampling stations (i.e., stream
traces). Creating those stream traces is better done by humans than a by computer algorithm
because tracking the mainstem involves extensive situational awareness. Thus, these procedures
describe manual stream tracing.
The two custom tools developed for channel slope and sinuosity allow users to manually trace each
stream once, then calculate channel slope and sinuosity over different distances (trace lengths)
simply by running the second tool on the stream traces again with a different trace length input
(explained further below). This promotes experimenting with slope and sinuosity over different
channel lengths to see what lengths best characterize these channel properties. Total trace lengths
of 1,000 and 2,000 m are standard for RMN sites.
NOTE: The first custom tool requires the Spatial Analyst extension (for the "Extract by Mask"
command). The second custom tool requires the 3D Analyst extension (for "Interpolate Shape"
command) and the ArcGIS Desktop Advanced license (for "Split Line at Point" command).
preprocessing
tool
stream tracing
3. Slope and
sinuosity
calculator tool
Figure 4-6. Schematic of the channel slope and sinuosity workflow. Shapes with
rounded corners are files; shapes with sharp corners are processes (manual or
automated).
Part 1 input and output files (manual stream tracing precursor tool):
Input:
1. Sampling stations aligned to flowline raster:
RMN_primary_secondary_[STATEABBRV]_reproj_aligned.shp
2. Flowline raster
33
-------
Output:
1. Geodatabase (slope_sin_[STATEABBRV]_[YYYYMMDD_HH_MM_SS].gdb) with two feature
classes: flowlines rasters within a 2,000-m radius of sampling stations converted into
polyline feature class (flow_polylines_[STATEABBRV]_[YYYYMMDD_HH_MM_SS]), and an
empty feature class with fields useful for creating stream trace polylines
(stream_traces_full_length_[STATEABBRV]_[YYYYMMDD_HH_MM_SS]).
Part 2 input and output files (manual stream tracing):
Input:
1. Sampling station aligned to flowline raster:
RMN_primary_secondary_[STATEABBRV]_reproj_aligned.shp
2. Polylines of flowline raster, created in Part 1
(flow_polylines_[STATEABBRV]_[YYYYMMDD_HH_MM_SS])
3. Empty feature class with useful fields for tracing, created in Part 1
(stream_traces_full_length_[STATEABBRV]_[YYYYMMDD_HH_MM_SS])
4. Aerial imagery
5. NHD flowlines
Output:
1. Feature class in geodatabase populated with one upstream stream trace feature and one
downstream stream trace feature for each sampling station, starting at the sampling
stations (stream_traces_full_length_[STATEABBRV]_[YYYYMMDD_HH_MM_SS])
Part 3 input and output files (slope and sinuosity calculator tool):
Input:
1. Sampling stations aligned to flowline raster:
RMN_primary_secondary_[STATEABBRV]_reproj_aligned.shp
2. Watershed delineations with added characterizations thus far
3. Manually created stream traces, created in Part 2
(stream_traces_full_length_[STATEABBRV]_[YYYYMMDD_HH_MM_SS])
4. DEM to the extent of the study area
Output:
1. Feature class in the trace geodatabase with additional values used to calculate slope and
sinuosity for each trace
(traces_with_slope_sin_[STATEABBRV]_[DISTANCE]_[YYYYMMDD_HH_MM_SS]).
2. Watershed delineation shapefile with fields added for: channel sinuosity ("sinDISTANCEm"),
channel slope("slpDISTANCEm"), notes ("SSNts[DISTANCE]") (empty), meters of elevation
34
-------
change over the distance ("mChng[DISTANCE]"), and the trace length over the distance
("trc[DISTANCE]").
Procedures by part:
Part 1: Run the first custom too! to prepare to create stream traces.
1. Fill out the interface for the tool called "Channel slope and sinuosity analysis, tool 1- before
manual stream tracing" as each field directs (see Figure 4-7). See the tool help for more
information. Running this tool creates two files inside the geodatabase named
slope_sin_[STATEABBRV]_[YYYYMMDD_HH_MM_SS].gdb. One file is ail flowline rasters
within a 2,000-m radius of each sampling station converted into polylines
(flow_polylines_[STATEABBRV]_[YYYYMMDD_HH_MM_SS]); these polylines are used for
stream tracing. The other is an empty feature class
(stream_traces_full_length_[STATEABBRV]_[YYYYMMDD_HH_MM_S5]) that you will
populate with stream traces during the manual stream tracing procedure (see Part 2). This
tool creates this second feature class so that all stream trace files have fields with the same
names and properties. This automatic naming helps to standardize procedures across users
and agencies. The flowline raster can either be downloaded from StreamStats (preferred) or
created in Step 9 of Section 3.2.
¦3? Channel slope and sinuosity analysis, tool 1- before manual stream tracing
Sampling stations
| RMN_primary_secondary_MD_FIPS_aligned
"3
State abbreviation
Flowline raster
| md_str900
Output location
l_Monitoring_Networks\GISV^id-Atiant'cY-U+6F+channel_slope_sin
Channel slope and sinuosity analysis, tool 1 - before manual
stream tracing
Process overview:
The general steps for calculating channel slope and sinuosity are: (1) run
"Channel slope and sinuosity analysis, tool 1- before manual stream tracing",
(2) manually trace stream mainstems upstream and downstream of monitoring
stations using a feature class output from (1). and (3) run "Channel slope and
sinuosity analysis, tool 2- after manual stream tracing" on a feature class
modified in (2).
More about this tool:
This tool performs preliminary processing to prepare users to manually trace
stream segments in order to get channel slope and sinuosity around sampling
stations. It produces two output feature classes in a geodatabase: (1) it turns
the flowline raster within 2000 m of all input sampling stations into a polyline
feature class to make stream tracing easier ("flow_polylines_..."), and (2) it
creates an empty feature class with standardized fields that can be populated
with stream traces ("stream_traces_fulljength_..."). The polyline shapefile
helps users trace stream segments because users can snap the trace to the
polyline vertices. This should be more accurate than tracing flowline rasters
Cancel
<< Hide Help
Tool Help
Figure 4-7. Interface for tool to do the first part of calculating channel slope and
sinuosity (before manual stream tracing).
Part 2: Manually create stream traces.
1. Load stream_traces_full_length_[STATEABBRV]_[YYYYMMDD_HH_MM_SS],
RMN_primary_secondary_[STATEABBRV]_reproj_aligned.shp,
flow_polylines_[STATEABBRV]_[YYYYMMDD_HH_MM_SS], and basemap aerial imagery
35
-------
into the Table of Contents. These are the files necessary for doing stream tracing.
Additionally, NHD flowlines may be helpful and can also be added to the Table of Contents.
2. Make stream_traces_full_length_[STATEABBRV]_[YYYYMMDD_HH_MM_SS] editable and
open the "Create Features" window1 under the Editor > Editing Windows menu. Select
stream_traces_full_length_[STATEABBRV]_[YYYYMMDD_HH_MM_SS] in the "Create
Features" window and open its attribute table.
3. Zoom to the first sampling station in the sampling station shapefile. It might help to change
the sampling stations' symbology to something large and visible, like C*5. Make the
symbology of flow_polylines_[STATEABBRV]_[YYYYMMDD_HH_MM_SS] thick and visible,
like Make the symbology of
stream_traces_full_length_[STATEABBRV]_[YYYYMMDD_HH_MM_SS] somewhat thinner
and a different color, like — . Make sure that
stream_traces_full_length_[STATEABBRV]_[YYYYMMDD_HH_MM_SS] is above
flow_polylines_[STATEABBRV]_[YYYYMMDD_HH_MM_SS] in the Table of Contents so that
traces are visible on top of the flowlines as they are created. ArcMap may look like Figure
4-8a.
4. Click on stream_traces_full_length_[STATEABBRV]_[YYYYMMDD_HH_MM_SS] in the
"Create Features" window, snap the vertex held by the mouse to the sampling station, and
click there. If you do not have snapping enabled, then enable2 it. (NOTE: It is critical that all
stream traces—both upstream and downstream—start exactly at sampling stations and that
all sampling stations have one upstream and one downstream trace emanating from them.)
Then trace the flowline one direction from the sampling station for at least 1,500 stream
meters, if possible (see Figure 4-8b). The stream trace feature class will update each trace's
length when you finish that trace.
1Create features:
http://pro.arcgis.com/en/pro-app/help/
2Enable snapping:
http://desktop.arcgis.eom/en/arcmap/10.3/manage
g^htm-
-and-3d-features.htm.
36
-------
Figure 4-8. (a) Polylines of fiowline raster with sampling station and aerial imagery.
(b) Polylines of fiowline raster with stream trace upstream and downstream to the
extent of the polyline flowlines (2,000 m each direction).
NOTES on stream tracing:
• Whenever the fiowline does not match the actual stream according to aerial imagery, trace
the actual stream, not the flow polyline. The flow polyline is just a convenience to make
tracing faster and more consistent when the trace matches the stream.
• Tracing must start exactly at the sampling stations and work outwards (i.e., the starts of the
stream traces must be snapped to the aligned sampling station). Traces cannot start away
from sampling stations and end at them.
• You can snap the vertices of the stream trace to the fiowline to ensure that the trace
matches the fiowline. In general, the flowlines should correctly follow actual streams if the
flowlines are derived from the StreamStats fiowline rasters. If the flowlines were created
from a DEM using ArcMap geoprocessing tools, flowlines may not follow actual streams as
well as they do for the StreamStats flowlines so you may need to compare them more
closely with aerial photographs and NHD flowlines and adjust accordingly.
• To reduce the risk of error, it is recommended you trace sites sequentially (i.e., trace both
directions at one site before moving to the next) and consistently trace one direction at
each site before tracing the other direction (e.g., always trace upstream at a site before
tracing downstream).
• Stop tracing streams whenever they reach a lake, dam, a coastline, visibly larger river, or
anything else that would change channel slope or sinuosity dramatically (see Figure 4-9). For
37
-------
example, channel sinuosity is meaningless inside lakes or ponds. Use aerial imagery to look
for these disruptions.
• When tracing upstream, it may be difficult to tell which branch is the mainstem. Generally,
choose the branch that appears to have more stream length based on the raster flowlines
and NHD.
• Make sure that stream traces are sequentially numbered in the trace feature class. There
should be no gaps in the OBJECTID field. Gaps in the OBJECTID field can be caused by
creating a trace, then deleting it. If any gaps do appear, one way to remove them is to save
the edits to your traces, leave editing mode, copy the feature class into a shapefile, edit the
duplicate ID numbers in the shapefile to be sequential, and copy the shapefile back into a
new stream trace feature class.
• If the sampling station is located at some discontinuity in the stream channel (e.g., on a
confluence with larger river, at the coast, entrance to or exit from a lake, etc.), include a
token trace (even 1 m) in the direction of the disruption. For the slope and sinuosity
calculations to work, there must be one trace for each direction at each station no matter
how short it is.
• A good magnification for tracing is between 1:8,000 and 1:12,000.
• Save your stream traces frequently within editor mode.
strea mjtra ces_f u I l_l en gth_M E_20160906_10_50_40
Shape_Length
Direction
Notes
OBJECTID
1878.006515
Polyline ZM
Down
None
Polyline ZM
None
Polyline ZM
Down
None
2108.197933
Polyline ZM
1864.514733
Down
Polyline ZM
2137.420207
Down
Polyline ZM
Polyline ZM
Polyline ZM
1641.094932
Down
None
2217.558405
None
1801.644846
Down
Polyline ZM
Polyline ZM
987.201603
Down
Beyond 980 m is open water. Stopped trace there.
H < 1 ~ H \m\m (0 out of 14 Selected)
strea m_tra ces_f u 11J en gth_M E_20160906_10_50_40
Figure 4-9. Example of a stream trace stopped at a lake. The green circle is the
sampling station, the black line is the downstream trace stopping at the lakeshore, the
red line is the upstream trace, and the thin blue lines are the polyline flowlines. The
last visible row of the attribute table notes that the trace stopped at a lake.
38
-------
5. After you complete each trace, note whether that trace went "Up" or "Down" in the
"Direction" field of stream_traces_full_length_[STATEABBRV]_[YYYYMMDD_HH_MM_SS]
In the "Notes" field, note whether you had to stop tracing within the flowlines provided by
the pre-trace tool due to coastline, lakes, confluences with visibly larger rivers, and so on.
This will inform subsequent users why some traces are shorter than expected (see Figure
4-10).
42
Polyline ZM
514.468246
Down
Lake begins 520 m downstream of station. Stopped trace there.
43
Polyline ZM
246.088837
Up
Lake 250 m upstream of station. Stopped trace there.
44
Polyline ZM
1539.437574
Down
Lake 1550 m downstream of station. Stopped trace there.
45
Polyline ZM
1424.4S5222
Up
None
46
Polyline ZM
2307.522281
Down
None
47
Polyline ZM
1725.354816
Up
None
4S
Polyline ZM
2411.322419
Down
None
49
Polyline ZM
808.470326
Up
Traced to upstream end of StreamStats stream grid
Figure 4-10. Sample of trace geodatabase with trace direction and notes.
6. Once all traces have been completed (twice the number of sampling stations), save the edits
to your trace feature class and exit editor mode.
7. Confirm that each sampling station has one upstream trace and one downstream trace and
that the traces are numbered sequentially.
NOTE: The slope and sinuosity calculator will not work properly if traces are not numbered sequentially
and each station does not have two traces snapped to it.
Part 3: Run the second custom tool to calculate channel slope and sinuosity from the stream traces.
1. Fill out the interface for the tool called "Channel slope and sinuosity analysis, tool 2- after
manual stream tracing" as each field directs (see Figure 4-11). If the RMN watersheds are
contained within one NHD subregion, input that DEM. If the RMN watersheds span multiple
NHD subregions, the input DEM was composited in Section 3.1, Step 14 (for StreamStats
delineations) or Section 3.2, Step 1 (for ArcMap delineations). Standard values for "Trace
length in either direction" for the RMN are 500 and 1,000 m. The tool's default vertical units
are centimeters because the NHD DEM units are centimeters.
39
-------
Channel slope and sinuosity analysis, tool 2- after manual stream tracing
Sampling stations
| RMN_primary_secondary_MD_FIPS_aligned
d
Sampling station names
| StationID
-1
Delineated watersheds
| M D_LU_BF_20160630_18_10
zi
£5|
Watershed names
I NAME
-1
State abbreviation
| MD
Trace length in each direction fm)
1000 1
Stream traces
| stream_traces_full_length_MD_201609G6_12_25_25
zi
£^|
DEM
J elev_cm_ma02a
d
£5
DEM vertical units
| Centimeters
Output file name
| MD_LU_BF_chan_slope_sin
Output location
| C:1ijUsers\dgibbspocumentsV5rojectsV:tegional_Monitoring_Network|
£31
Channel slope and sinuosity analysis, tool 2- after manual
stream tracing
Process overview:
The general steps for calculating channel slope and sinuosity are: (1) run
"Channel slope and sinuosity analysis, tool 1- before manual stream tracing",
(2) manually trace stream mainstems upstream and downstream of monitoring
stations using a feature class output from (1), and (3) run "Channel slope and
sinuosity analysis, tool 2- after manual stream tracing" on a feature class
modified in (2).
More about this tool:
This tool calculates the channel sinuosity and slope of stream traces in a
feature class in a geodatabase. The user can select the distance upstream and
downstream over which slope and sinuosity will be calculated.
The combined length of the upstream and downstream traces at each station
should be greater than twice the user input 'Trace length in each direction",
which represents how over how many meters upstream and downstream should
be used for slope and sinuoaity calculations. If not enough stream was traced
for any station, the tool will run to completion, outputting a delineation file that
uses trace lengths shorter than the desired distance for those stations. There
must be one trace upstream and one downstream snapped to each
monitoring/sampling station in the trace feature class.
Input files are: full length stream trace feature class, sampling stations,
watershed delineations, and DEM. The starts of the stream traces must be
snapped to their sampling stations (i.e., the upstream ends of the downstream
traces and the downstream ends of the upstream traces must be snapped to
Cancel
<< Hide Help
Tool Help
Figure 4-11. Interface for tool to calculate channel slope and sinuosity (after manual
stream tracing).
NOTE: This tool looks for adequate "Trace length in each direction" in both directions (upstream and
downstream of the sampling station). If there is not enough trace length in one direction, it will
add length to the other direction to make up the difference, if there is not enough trace length
in both directions combined, the ArcMap Results display will tell users, and the output field
"trc[DISTANCE]m" wili display the deficient trace length for that feature. For example, if the
input "Trace length in each direction" is 500 m and a station has traces of 472 m and 976 m, the
resulting traces will be 472 m and 528 m. If the traces are 482 m and 405 m, all of both traces
will be used for the channel slope and sinuosity calculations but the total trace length will only
be 887 m, not 1,000 m.
2. This tool produces two output files.
a. One is a feature class in the geodatabase called
traces_with_slope_sin_[STATEABBREV]_m_[DISTANCE]_[YYYYMMDD_HH_MIVI_SS].
For each watershed, this line feature class provides information on: how long the trace
is (field "Shape_Length"), the starting and ending X and Y coordinates, the straight-line
distance from the start to the end of the trace (field "strgt_dist"), the channel sinuosity
over that distance (field "sinuosity"), the starting and ending elevations of the trace in
meters (fields "startElevM" and "endEievM", respectively), the change in elevation over
the trace (field "elevChng_m"), and the channel slope (field "chan_slope"). This file
contains fields useful for quality control and confirming the tool's output (see Figure
4-12). The most useful field to examine for quality control is "Shape_Length"; if it does
40
-------
not equal the total desired trace distance for a station, something is amiss and the
relevant traces need to be investigated. Moreover, the "strgt_dist" should always be
less than the "Shape_Length" and the latter divided by the former should equal the
sinuosity. "elevChng_m" should be the difference between "startElevM" and
"endElevM", and one-one-hundredth of the difference between "startElev" and
"endElev" (elevations in centimeters).
b. The other output file is a shapefile in the user-specified output folder. This file has all
the fields of the input watershed delineation shapefile with channel slope and sinuosity
(fields "sin[DISTANCE]m" and "slp[DISTANCE]m", respectively), the change in elevation
over the trace (field "mChnglOOO"), the trace length (field "trc[DISTANCE]m"), and a
notes field (field "SSNts[DISTANCE]m"),
tra
ces_with_slc
)pe_siri_ME_m_l
]00_20160919_1'
l_10_35
StationID
Shape_Length
start_x
start_y
end_x
end_y
strgt_dist
~
17
1000.000051
1999102.3866
2700570.1472
1998261.4962
2701081.4296
984.127206
345
999.999871
1985495.8101
2676848.7826
1935271.9076
2677719.4068
898.954256
56817
1000.000148
2076386.5486
2649092.6315
2076578.869
2650024.9422
951.940322
57011
1000.000042
2076395.8101
2666241.3108
2076945,8917
2666830.4701
806.038738
57065
1000.000004
2174766.2588
2697688.1818
2174863.9062
2698628.6448
945.51873
626
999.999834
2115238.6785
2671573.8553
2115757.765
2671668.6786
527.676276
MEDEP 57
999.999952
2050058.7293
2726589.6429
2050380.9324
2727450.2199
918.916542
sinuosity
startElev
endElev
startElevM
endElevM
elevChng_m
chan_slope
1.016129
19226.28
18820.4
192.2628
188.204
4.0588
0.004059
1.112403
22261.06
21782.9
222.6106
217.829
4.7816
0.004782
1.050486
3151.96
3134.47
31.5196
31.3447
0.1749
0.000175
1.240635
7559.01
6879.99
75.5901
68.7999
6.7902
0.00679
1.057621
6204.29
2572.95
62.0429
25.7295
36.3134
0.036313
1.895101
3671.37
3643.39
36.7187
36.4339
0.2848
0.000285
1.08S238
15811.22
14977.89
158.1122
149.7789
8.3333
0.008333
Figure 4-12. Sample attribute table for
traces_with_slope_sin_[STATEABBREV]_mjDISTANCE]_[YYYYMMDD_HH_MM_SS].
3. Check the "trc[DISTANCE]m" field in the output shapefile to make sure that all traces are
within rounding error of the desired trace distance (e.g., if the "Trace length in each
direction" input was 500, "trc[DISTANCE]m" values should be between 999.9 and 1,000.1). If
any values are less than twice the input value, check whether the upstream and
downstream traces for that sampling station are collectively less than the trace length
objective. If there are any surprising results, also confirm that the starts of the traces are
snapped1 to their sampling stations.
1Details on snapping points:
http://desktop.arcgis.eom/en/arcmap/10.3/manage-data/editing-fundamentals/about-snapping.htm.
41
-------
4. If you want to check any distance measurements output by the tool, make sure that the
dataframe's projection1 is the same as that of the DEM. Otherwise, the distances found by
+ +
the Ruler "iM* tool will not match the tool's outputs. You can change the dataframe
projection in the dataframe properties menu. (NOTE: The channel slope-sinuosity tool
calculates planar distances.)
5. If it is not possible to collectively trace upstream and downstream far enough to provide the
total desired distance (e.g., upstream encountered the end of the flow grid and downstream
encountered a lake), note in the "SSNts[DISTANCE]m" field of the output shapefile that the
total trace was expected to be short and that the slope and sinuosity calculations are over a
shorter distance than the other values (see Figure 4-13, red box). This will assist other users
who subsequently use the watershed delineation shapefile.
sin2000m
slp2000m
trc2000m
mChng2000
SSNts 2000m
2.008035
0.010032
1999.999987
20.0632
None
1.136444
0.003476
2000
6.9515
None
1.088207
0.004117
2000.000073
8.2339
None
1
1.172475
0.048824
2000.000022
97.6477
Downstream trace was 80 m to a lake; upstream trace was 1920 m
|
r
—
1.193147
0.004882
1999.999981
9.7641
1.168206
0.010667
1999.999987
21.334
Downstream trace was 20 m to a lake; upstream trace was 1980 m
I
1.270575
0.054683
1999.999938
109.3666
None
1.038684
0.118255
2000.00005
236.5102
None
i
1.382398
0.04496
1362.334479
61.2506
Total trace length is 1362 m because downstream trace reaches a tlake in 15 m and upstream trace reaches end of StreamStats grid
I
Figure 4-13. Partial output of watershed delineation shapefile with channel slope and
sinuosity, calculated over 2,000 m stream traces wherever possible. Note in the final
row (red box) that the total trace length was only 1,362 m because of obstacles to the
traces upstream and downstream. In the two rows with blue boxes, the total trace
lengths are 2,000 m but the notes field says that they are not evenly split between
upstream and downstream traces.
6. If it is not possible to trace in one direction far enough to provide the desired distance but it
is possible to trace the other direction far enough to compensate, note in the
"SSNts[DISTANCE]m" field that the slope and sinuosity calculations used traces of unequal
length (i.e., did not use equal length upstream and downstream traces) (see Figure 4-13,
blue boxes). This will assist other users who subsequently use the watershed delineation
shapefile.
7. If you want to calculate channel slope and sinuosity over another trace length, repeat Part 3
with the new value in the "Trace length in each direction" input field. If you want the new
channel slope and sinuosity values in the same watershed delineation file as the previous
values, use the output delineation shapefile from the first run as the input delineation
shapefile from the second run. As long as the "Trace length in each direction" field is
different, the tool will create new fields in the output shapefile. It will also create a new
feature class in the geodatabase with the new desired trace distance and the unique time
1Dataframe projection: http://support.esri.com/technical-article/0000Q5357.
42
-------
stamp (see Figure 4-14)
(traces_with_slope_sin_[STATEABBREV]_m_[DISTANCE]_[YYYYMMDD_HH_MM_SS]).
B l3 slope_si n_NY_20160906_10 _51_58,gdb
Qflow_polylin es_N Y_20160906_10_51_58
til strearn_traces_full_length_NY_20160906_10_51_58
0 traces_with_slope_sin_NY_m_1000_20160919_14_56_29
£3 tra c es_with_sl o p e_si n_N Y_m_2000_20160930_16_41_20
.-H traces with slope sin NY m 3000 20161005 10 04 26
Figure 4-14. Channel slope-sinuosity geodatabase with the polyline versions of the
flowline rasters created in Part 1 of this analysis, the manual stream traces created in
Part 2 of this analysis, and the files with slope and sinuosity calculated over three
stream distances (500 m, 1,000 m, and 1,500 m each direction).
4.3.4. Watershed slope
Objective:
To calculate the mean, minimum, maximum, range, and standard deviation of watershed slopes.
Watershed slope is the slope of the land in each watershed, as opposed to the slope of the stream
channel.
Data use:
These watershed slope values, especially the average, can help with classifying watersheds by their
slopes, in addition to the values calculated by the channel slope tool. For example, an analysis may
be done on only low-gradient watersheds.
Required input files:
1. Sampling stations aligned to flowline raster:
RMN_primary_secondary_[STATEABBRV]_reproj_aligned.shp
2. Watershed delineations with added characterizations thus far
3. DEM to the extent of the study area
NOTE: This tool requires the Spatial Analyst extension (for "Slope" and "Zonal Statistics as Table"
commands).
Output:
1. RMN watershed delineation shapefile with five fields added with the following values for
each watershed: slope mean, minimum, maximum, range, and standard deviation (fields
"avg_ShdSlp", "min_ShdSlp", "max_ShdSlp", "rng_ShdSlp", and "std_ShdSlp", respectively).
43
-------
Procedures:
Fill out the tool interface as each field directs (see Figure 4-15). If the RMN watersheds are
contained within one NHD subregion, input that DEM. If the RMN watersheds span multiple NHD
subregions, the input DEM was composited in Section 3.1, Step 14 (for StreamStats delineations) or
Section 3.2, Step 1 (for ArcMap delineations).
Watershed slope — ~ X
Watershed slope
Calculates the minimum, maximum, average, range, and
standard deviations of slopes within each provided watershed
using the user-supplied DEM. Watershed slope parameter
fields are added to the input watershed file.
Input files are: watershed delineations and DEM. This tool
does not use sampling stations, so that is not an input file.
Output file is: watershed delineation with five columns added
(one for each watershed slope property, in degrees) and
_YYYYMMDD_HH_MM_SS appended to the file name.
This tool requires the "Spatial Analyst" extension.
OK Cancel Environments... «Hide Help Tool Help
Figure 4-15. Watershed slope tool interface.
4.3.5. Dams, mines, NPDES, and CERCLA site preprocessing
Objective:
To determine how many dams, mines, NPDES, and CERCLA sites are located within each watershed.
Data use:
Watersheds with more of these sites may be more likely to have altered hydrology or water quality
and have increased human activity. Thus, they may not be good candidates for reference sites.
However, sites in these categories found in a given RMN watershed should be investigated because
their source databases may be out-of-date, they may be small or far from the sampling station, or
they may not have substantial water quality or flow impacts. Thus, watersheds should not be
excluded from "reference" status just because they contain these sites; each situation must be
further explored.
Required input files:
Part 1 input and output files (manual site identification precursor tool):
Input:
1. Watershed delineations with added characterizations thus far
Delineated watersheds
| MD_LU_BF_chan_slope_sin_20160930.11_36_06
£3
Watershed names
NAME
State
MD
DEM
| elev_cm_ma02a
d
£5
Output file name
| MD_LU_BF_CSS_wtrshd_slp_degrees
Output location
| ojects^egional_Monitoring_Networks\GISV^1id-AtlanticV-U+BF+CSS+wtrshd_slp |
£5
Delete intermediate files?
| Yes
44
-------
Output:
1. Watershed delineations with added characterizations thus far with eight fields necessary for
noting relevant sites in watersheds: a count field ("_ct") and a notes field ("_info") for each
of the four site types.
Part 2 input and output files (manual site identification):
Input:
1
2
3
4
5
6
Output:
1. RMN watershed delineation shapefile with the eight new fields populated (four site types,
each with a count field and a notes field)
Procedures:
NOTE: RMN watersheds are designated in areas where there are minimal human impacts. Thus, they
should have few dams, mines, NPDES, and CERLCA sites. The methods described here would be
more cumbersome for watersheds that are not selected to minimize the numbers of these sites.
1. Fill out the tool interface as each field directs (see Error! Reference source not found.)- See
the tool help for more information. Using this tool quickly adds eight standardized fields to
the input watershed file; it does not actually determine the number of each site type in each
watershed. That part must be done manually, following use of this tool. Recording how
many of each site type is in each watershed is deliberately not included in the tool because
this determination is best done using human discretion, and the absence of automatic
calculation forces the user to personally check and annotate sites.
2. Add shapefiles for dams, mines, NPDES, and CERCLA to the Table of Contents.
3. Use the "Select by Location" interface to select all dams, mines, NPDES, and CERCLA sites
that "are within a distance of the source layer feature" of the watershed file output by this
tool (see Error! Reference source not found.). Set the search distance to 500 ft. Including
this buffer around the watersheds will make sure that any sites whose coordinates are just
1USGS major dams shapefile:
https://catalog.data.gov/dataset/usgs-small-scale-dataset-maior-dams-of-the-united-states-200603-shapefile.
2USGS mines shapefile: https://mrdata.usgs.gov/metadata/mineplant.faq.html.
3EPA sites of interest: https://www.epa.gov/enviro/geospatial-data-download-service.
Sampling stations aligned to flowline raster:
RMN_primary_secondary_[STATEABBRV]_reproj_aligned.shp
Watershed delineations output from manual site identification precursor tool
Shapefile of dams from the USGS1 or a detailed, local file
Shapefile of mines from the USGS2 or a detailed, local file
Shapefile of NPDES sites3 from the EPA or a detailed, local file
Shapefile of CERCLA sites3 from the EPA or a detailed, local file
45
-------
outside the watershed but might actually be inside to due data inaccuracies can be
assessed.
4. Check each of the four site shapefiles to see if any sites were selected (i.e., are within the
watersheds).
4ft Dam-mine-NPDES-CERCLA counts and properties
X
Delineated watersheds
I M D_LU_B F_C SS_wtrsh d_sl p_deg rees_20161006_11_44_15
"31
Watershed names
| NAME
Output file name
MD LU BF CSS WS
Output location
| iitoring_Networks \GIS V^lid-AtianticV-U -H3F +CSS +WS +dam -mine-NPDES -CERCLA | ^
Dam-mine-NPDES-CERCLA counts and
properties
Adds fields with standardized names to input watershed
delineation shapefile. This tool does not actually determine
how many of each site type there is in each watershed; that
must be done manually after running this tool.
After running this tool, add shapefiles of dams, mines.
NPDES, and CERCLA sites to your ArcMap table of contents.
Use the 'Select by Location* interface to select all dams,
mines, NPDES, and CERCLA sites that "are within a distance
of the source layer feature' of the watershed file output by this
tool. Set the search distance to 500 ft. to make sure that any
sites just outside the watersheds but still contributing to them
are captured. Then, investigate each facility that was selected.
Cancel
Environments... « Hide Help
Tool Help
Figure 4-16. Dams, mines, NPDES, and CERCLA site search support tool interface.
Select By Location X
Select features from one or more target layers based on their location in
relation to the features in the source layer.
Selection method:
select features from v J
Target layer(s):
~ MD_LU_BF_CSS_WS_sites_2Q 161102_10_19_19
0 mineplant
0 dams00xQ20
0 NPDES_Major
0 CERCLIS_NPL
1 1 Only show selectable layers in this list
Source layer:
MD LU BF CSS WS sites 20161102 10 19 19
Use selected features (0 features selected)
Spatial selection method for target layer feature{s):
| are within a distance of the source layer feature
H
Apply a search distance
| 500 | Feet
About select by location ^ Appl v gose
Figure 4-17. Selecting point source facilities within watersheds.
46
-------
5. Whether this characterization process is done depends on the results of Step 4.
a. If none are selected, use the "Field Calculator" to put Os in the fields "dam_ct",
"mine_ct", "NPDES_ct", and "CERCLA_ct". Then use the "Field Calculator" to put
"N/A"s in the fields "damjnfo", "minejnfo", "NPDESJnfo", and "CERCLAinfo".
It is important to note that there are no sites in these fields, so that it is clear
that this was evaluated. Leaving these fields blank could cause uncertainty for
subsequent users as to whether this step was completed. This concludes this
characterization procedure for these sampling stations.
b. If any dams, mines, NPDES, or CERLA points are selected, enter "Editor" mode
for the watershed file output by this tool. Proceed to Step 6.
6. Zoom to each selected site in turn (i.e., those within the watersheds). Increment the
appropriate count ("_ct") field and note pertinent information in the appropriate notes
("_info") field, such as distance from sampling station, size of lake behind dam, height of
dam, type of NPDES facility, whether the facility is active, etc. (see Figure 4-18). This
information is not meant to be an exhaustive profile of each site or its potential effects on
water quality. It simply indicates which sites might significantly alter the watershed and
should, therefore, be further investigated. If there are multiple sites of the same type within
the same watershed, notes on each of them should go in the appropriate "_info" field.
7. Once all sites within the watersheds have been examined and added to the proper
watershed attribute table fields, fill in any empty "_ct" cells with zero and any empty "_info"
fields with "N/A". This way, it is clear to future viewers that this step was done completely
and all relevant sites were evaluated.
8. Save edits and exit "Editor" mode. This concludes this characterization procedure for these
sampling stations.
dam_ct
damjnfo
mine_ct
minejnfo
npdes_ct
npdes_
CERCLA_ct
~
0
N/A
0
N/A
0
N/A
0
N/A
0
N/A
0
N/A
0
N/A
0
N/A
0
N/A
0
N/A
0
N/A
0
N/A
0
N/A
0
N/A
0
N/A
1
Aerospace cor
0
N/A
0
N/A
0
N/A
0
N/A
0
N/A
0
N/A
0
N/A
0
N/A
0
N/A
0
N/A
0
N/A
1
Aerospace cor
0
N/A
0
N/A
0
N/A
0
N/A
1
Dam 0.75 kmA2 5.5 km from sampling station on mainstem
1
Soil products mine 10.8 km from sampling station on tributary
0
N/A
0
N/A
0
N/A
0
N/A
0
N/A
0
N/A
1
Dam 0.32 1071*21.5 km from sampling station on mainstem
0
N/A
0
N/A
0
N/A
Figure 4-18. Example site counts and descriptions. Watersheds without any of a site
type have a zero for the count and N/A for the notes.
47
-------
5. Technical support, updates, and known issues
5.1. Technical support
If you have questions about this guide or are having technical problems, e-mail Britta Bierwagen at
bierwagen.britta@epa.gov. Please provide any relevant screenshots or outputs from ArcMap, as well as
your input files.
5.2. Updates
Although tools may be added, updated, or revised over time (e.g., to add new features, fix bugs, or make
compatible with new versions of ArcMap), no updates to this guide are currently anticipated. E-mail
Britta Bierwagen to inquire about tool changes.
5.3. Known issues
There are currently no known issues with these tools that have not been described elsewhere in this
guide. Please send any issues or suggestions you have. They will be compiled in the event that an
updated guide is released.
48
-------
6. Technical details of scripts
Below are technical outlines of the watershed characterization scripts for users who want more
information on how these scripts work. The scripts were written May to October 2016 and tested in
ArcMap 10.3.1. They were retested with ArcMap 10.4 and 10.5 in December 2017.
6.1. Land use composition
NOTE: requires the Spatial Analyst extension
1. The tool creates a subfolder within the output folder in which the processing happens (the
"intermediate folder"). All intermediate files are created (some of which are deleted) in the
intermediate folder. The final output delineation shapefile is the only file that is placed in
the output directory the user specified. The intermediate folder name includes the final file
name and a timestamp.
2. The tool checks the watershed delineation names for certain unusable characters. If any are
found, the script returns an error and terminates.
3. The tool reprojects the input shapefiles to the same projection as the land use raster. That
process will generally convert all X-Y units into meters.
4. The tool creates buffer polygons of two specified radial distances around each sampling
station (1,000 and 5,000 m by default). If the user wants to get land use for more than two
buffers, the user will have to run the tool multiple times, changing the script to different
buffer distances each time.
5. The tool clips the station buffers to the boundaries of the corresponding watershed. The
tool matches the names of the watersheds to the names of the buffers. Because the names
of the buffers come from the name field selected for the sampling stations, the watershed
names must have matching sampling station names (> buffer names).
6. The tool clips the input land use raster to the watershed delineations, the outer buffers
clipped to watersheds, and the inner buffers clipped to watersheds. Each of these is a new
raster, which is saved in the intermediate folder of the selected directory. These rasters
have two fields added: the clipping extent (none or the inner or outer buffer distances) and
the watershed name/station ID. These are used for transferring land use fractions to the
output shapefile later in the script.
7. The tool adds a field to each clipped land use raster, which will store the fraction of each
category of each land use. Then the tool calculates the fraction of each land use for each
clipped raster based on the pixel count in that category and in the whole raster.
8. The tool copies the watershed delineation shapefile, assigns it the user's selected name, and
adds one field for each possible land use-clipping extent combination (e.g., LU31_whole,
LU42_5000). Note that this step adds all possible combinations based on the NLCD 2011
land use categories, so it may include land uses that are not encountered in the study area
or within 1,000 or 5,000 m of any sampling stations. Then the tool iterates through each
49
-------
clipped land use raster file and each land use in that raster to extract the land use fraction
and copies it to the correct column of the output shapefile. Finally, it transfers the
delineations with the land use fraction fields to the selected output folder.
6.2. Base flow
NOTE: requires the Spatial Analyst extension
1. The tool checks the watershed delineation names for certain unusable characters. If any are
found, the script returns an error and ends.
2. The tool extracts the base flow value at each sampling station. By default, those values are
put in a field in the sampling station shapefile called "RASTERVALU".
3. The tool copies the watershed delineation shapefile, creates a new field in the watershed
delineation shapefile called "BASE_FLOW", and transfers the base flow from the sampling
station to the new field in the delineation shapefile. This involves a nested "for" loop in
which the tool matches each watershed with its sampling station and transfers the base
flow values to the output watershed file.
6.3. Channel slope and sinuosity, pre-stream trace
NOTE: requires the Spatial Analyst extension
1. The tool creates a file geodatabase with a name that includes the state abbreviation and the
time of creation.
2. The tool creates a feature class in that geodatabase with a name that includes the state
abbreviation and the time of creation.
3. The tool projects the feature class to the projection of the flowline raster and adds a field in
which the user will later note whether the stream trace segment is upstream or
downstream of the sampling station.
4. The tool creates a 2,000 m buffer around each sampling station in the input shapefile.
2,000 m was chosen because it is several times larger than the usual trace length.
5. The tool clips the flowline raster to that 2,000 m buffer.
6. The tool converts the flowline raster within that 2,000 m buffer into polylines and saves it in
the geodatabase as flow_polylines_[STATEABBREV]_[YYYYMMDD_HH_MM_SS].
6.4. Channel slope and sinuosity, post-stream trace
NOTE: requires the 3D Analyst extension and ArcGIS Desktop Advanced license
1. The tool reprojects the manual stream traces and sampling points to the projection of the
DEM. That way, the raster is not being reprojected, which can cause accuracy problems. The
reprojected traces and sampling stations are used for the remaining steps. The final
delineation shapefile from this tool is in the projection of the input traces.
50
-------
2. The tool checks whether any of the stations do not have enough total trace length around
them (upstream + downstream). "Enough" is twice the user input trace length, which is the
desired trace length in each direction. For example, if the user inputs a trace length of
1,000 m, each station must have a total of 2,000 m of stream trace (any combination of
upstream and downstream). The tool counts how many stations do not have enough stream
trace length and tells the user how many combined traces will be shorter than the specified
amount. This is shown in the geoprocessing results window.
3. The tool checks whether any upstream and downstream traces are shorter than the desired
user-input trace length. If any are too short, the tool reports how many are too short.
Regardless of how many are too short, the tool will proceed. This step merely lets the user
know if any traces are shorter than the input length and, therefore, will need to have extra
trace length used in the other direction.
4. As the first step to trimming the traces to the desired length, the tool converts the traces
into m-routes so that distance along them can be measured. The larger the m-value, the
further from the sampling station.
5. The tool creates points at the specified distance from the sampling stations along each
m-route trace. Whenever possible, the tool creates points at the trace length input by the
user. For any traces that are shorter than that length, the tool creates a point at the end of
the trace and adds the corresponding distance to the other trace at that station so that the
total trace length compensates for insufficient trace length in one direction.
6. The tool splits the m-route stream traces where the aforementioned points were created.
This effectively trims the stream traces to the desired distance. For any trace shorter than
the desired trace length, the whole trace is included (i.e., not trimmed), while the other
trace for that station is trimmed to a longer length to compensate for the shorter trace
length.
7. The tool then combines the trimmed upstream and downstream traces (i.e., those within
the specified distance of the stations, or the "inner" traces) into a single trace for each
station. It does this by adding a field to each trace which is filled with the concatenated X
and Y coordinates of the starting vertex. Only traces at the same station will have the same
value. Because "inner" traces around each sampling station share the same starting X and V
coordinates (if they were snapped to the station, as required), inner traces starting at the
same sampling station can then be dissolved into a single trace on the basis of their starting
coordinates. There is now one trace per sampling station.
8. The tool duplicates the sampling point shapefile, adds a field for storing their X and V
coordinates in the same format as the stream traces' starting X and V coordinates, and then
populates that field with the concatenated X and Vcoordinates.
9. The tool then uses the concatenated X and V coordinate field in the sampling station copy to
join the station attributes to the traces. That way, each trace can be identified by its
waterbody or station ID.
51
-------
10. The tool calculates trace sinuosity using planar distance. The formula is channel
length/valley length. Channel length is the planar length of the trace. Valley length is the
straight line distance between the trace end points, which is calculated using the standard
two dimensional distance formula (sqrt[(Xl-X2)2 + (VI—V2)2]) because the distances involved
are fairly small. Sinuosity and precursor values are stored in the trace feature class.
11. The tool calculates channel slope. Channel slope = change in elevation 4- trace length. To do
this, the traces are converted to z-enabled lines using the supplied DEM, and the starting
and ending elevations are stored in new fields of the trace. Note that the starting and
ending elevations are based on bilinear interpolation, so they won't exactly match the
values of the DEM pixels underneath them but rather should be an interpolation of nearby
pixels. With the starting and ending elevations in the DEM's vertical units in hand, the script
calculates the starting and ending elevations in meters and uses those values for slope
because the length of the traces is provided in meters.
12. The tool copies the input delineation shapefile into the user-specified directory, joins the
stream trace table to it, and copies the trace length, the sinuosity, and slope to the
delineation shapefile. The trace length is copied to make it easy for users to check whether
all traces were the correct length (i.e., none should be longer or shorter than twice the input
trace length).
13. The tool deletes all the intermediate files that were created inside the geodatabase. If the
user wants to keep those files for error checking, debugging, or other uses, they can be
disabled by a comment code.
6.5. Watershed slope
NOTE: requires the Spatial Analyst extension
1. The tool reprojects the watersheds to the projection of the DEM. That way, the raster is not
being reprojected, which can lead to accuracy problems. The reprojected watersheds are
used for slope calculations in each watershed but the final output shapefile from this tool is
in the projection of the input watershed shapefile.
2. The tool calculates slope (in degrees) for the entire DEM, creating a slope raster for the
study area.
3. The tool copies the watershed delineation shapefile to the output folder and adds five
empty fields (maximum, minimum, mean, range, and standard deviation of slope).
4. The tool creates layers out of the reprojected watersheds and the final output watersheds.
It creates layers from them so they can be individually processed in the next step.
5. The tool iterates through each feature in the reprojected watersheds layer and uses Zonal
Statistics as Table to get watershed slope parameters for each watershed individually. These
are stored in a one-row table, which is joined to the final output delineation layer. The slope
values are copied to the output shapefile layer, then the one-row table is deleted. The tool
iterates through the watersheds individually rather than doing Zonal Statistics on all of them
52
-------
at once because there are complications with using Zonal Statistics on nested watersheds.
Processing each watershed individually avoids the problems with nested watersheds.
6. The tool optionally deletes the intermediate files. Because the deletions occur at the end of
the script in sequential lines, they can easily be prevented by placing a comment code (i.e.,
"#") at the beginning of each line.
6.6. Dams, mines, NPDES, and CERCLA site preprocessing
1. The tool copies the watershed delineation shapefile from the input location to the output
location and gives it the user-specified name.
2. The tool adds eight fields: four to store the counts of the dams, mines, NPDES, and CERCLA
sites, and four to store notes about each of them. The former fields are of type SHORT
integer and the latter are of type TEXT (250 characters maximum).
53
-------
SEPA
United States
Environmental Protection
Agency
PRESORTED STANDARD
POSTAGE & FEES PAID
EPA
PERMIT NO. G-35
Office of Research and Development (8101R)
Washington, DC 20460
Official Business
Penalty for Private Use
$300
------- |