Deerfield Culvert Delineation - MassDOT

Point Delineation Documentaiton

This project serves as an example workflow to delineating point locations on a stream network using the NHDHRDV2 hydrography layers. Culvert and flow gage locations in the Deerfield watershed in MA & VT are delineated as part of the MassDOT Culvert Project (Figure 1). Additionally, watershed attributes are caluclated for each of the given locations. Any tools referenced are assumed for ArcMap 10.2. The documentaiton is broken into the following sections:
1. Source Layers
2. Layer Pre-processing
3. Location Editing
4. Watershed Delineation
5. Additional Analysis
6. Catchment Attributes

Figure 1: Deerfield watershed with culvert and gage locations

Source Layers

The intermediate hydrography processing layers from the NHDHRDV2 dataset are required for the delineation process. Not all of these layers are made public, but are available upon request. In addition to the source layers listed in the table below, a spatial points layer representing the locations to be delineated is required. If the points are not already mapped to the stream grid, this is done in the Location Editing section.

Table 1: Required source layers from NHDHRDV2 dataset

Layer Pre-processing

1. Define Processing Zone
The processing zone shapefile is created, encompasing the full extent of the potential contributing drainage area to the points being processed. This can be based on a HUC boundary layer, selected catchments, or some other processing area designation layer. Whatever the method, it is important to maintain network connectivity within the desired watershed. Splitting a watershed by excluding contributing drainage area will result in incorrect delineation.

2. Buffer Processing Zone
The processing zone shapefile is buffered (Analysis Tools > Proximity > Buffer) to ensure all features in the necessary processing range are captured. 1 - 2 km is a reasonable buffer length. This step is not necessary if the original shapefile certain to contain all required catchments and flowlines.

3. Generate Vector Layers
Features from each of the Drainage Line and Catchment layers that intersect the buffer layer are selected and exported as new layers.

4. Generate Raster Layers
The stream grid, dem, flow direction grid and flow accumulation grid are clipped to the buffer layer using the Extract by Mask tool (Spatial Analyst Tools > Extraction > Extract by Mask).

5. Create Snap Stream Grid
A stream grid used for snapping points (strSnap50) is created using the Reclassify tool (Spatial Analyst Tools > Reclass > Reclassify). In this case, cells in the flow accumulation layer with 50 or more contributing cells are reclassfied as a value of 1 to define the snapping grid. All other cells are reclassified as "NoData".

Location Editing

The delineation process requires location points to be snapped to the strSnap50 layer (Figure 2). Before this snapping is completed, the points must be manually inspected to ensure that they will snap to sensible locations on the stream grid. For visual inspection, the stream grid (strFinal) is displayed over the strSnap50 layer. The high resolution flowline vector layer (strHR) is also displayed for reference in the manual inspection process.

Figure 2: Example of snapping a point to the snap stream grid

As a general rule, relative positioning of the point on the strHR vector layer should be reflected in its position on the stream grid after snapping. Figure 3 shows an example of a correction to ensure the point is located on the correct branch of the stream grid with respect to it's location on the vector layer.

Figure 3: Adjustment of a point to the correct position on the stream grid

Differences between the stream and snap grids also require adjustment of points to ensure they fall on the correct channel. Figure 4 shows how a point may fall onto a seemingly accurate position on the snap stream grid (strSnap50), but actually needs to be corrected to fall into the main channel of the stream represented by strFinal.

Figure 4: Location adjustment to ensure point falls on the main stream channel

In some cases locations are on such small, undocumented streams that snapping should be turned off for that particular point (Figure 5). These points are noted and will be identified in the "SnapOn" column of the Batch Point Setup step of the next section.

Figure 5: Example of a point, located on an undocumented stream, that does not
get snapped to the stream grid

Watershed Delineation

The watershed delineation process takes advantage of the ArcHydro tools. Layers are conformed to the specifications of these tools. It is necessary to follow the steps in order to generate the watersheds polygon layer.

1. Adjoint Catchment Processing

Input
Drainage Line: DrainageLineFinal
Catchment: Catchment

Output
Adjoint Catchment: AdjointCatchment

2. Batch Point Setup

Five columns are added to the point location layer using the "Add Field" tool (Data Management Tools > Fields > Add Field). These columns are necessary for batch delineation and are described in the table below.

Table 2: Fields added to the point location layer to establish the "Batch Point" layer for delineation

3. Batch Watershed Delineation

Input
Batch Point: crossings_delineation
Flow Direction Grid: fdr
Stream Grid: strFinal
Snap Stream Grid: strSnap50
Catchment: Catchment
Adjoint Catchment: AdjointCatchment

Output
Watershed: Watershed
Watershed Point: WatershedPoint

Additional Analysis

This section describes the additional steps required to caluclate attributes specific to the MassDOT Culvert Project. These steps are not neccessary for creating watershed polygons.

1. Longest Flow Path

Input
Drainage Area: Watersheds
Flow Direction Grid: fdr

Output
Longest Flow Path: LongestFlowPath

2. Flow Path Parameters from 2D Line

Input
Longest Flow Path: LongestFlowPath
Raw DEM: dem

Output
Slope 1085 Point: Slp1085Point

Catchment Attributes

The steps in this section calculate the attributes for the watershed polygons. A small set of value rasters is used but can be added to if new attributes are desired.

1. Spatial Layer Processing

Once the spatial layers have been generated a post-processing script is run to add unique ID fields, calculate area/length, and export tables for use in later steps. The inputs are defined in the spatialPostProcessing.py script and it is executed in Arc Python.

Table 3: User-inputs to the spatialPostProcessing.py script

2. Calculate Zonal Statistics After the spatial layers are updated, the inputs to the zonalStatisticsProcessing.py script are defined and the script is executed in Arc Python.

Table 4: User-inputs to the zonalStatisticsProcessing.py script

3. Finalize Attributes The R script, finalizeZonalStatistics.R, edits the raw output from the "Zonal Statistics" tool for final output. The inputs are defined and the script is executed in R.

Table 5: User-inputs to the finalizeZonalStatistics.R script