QGIS Plugin for Hydrological Network Classification

This plugin creates the channel network, joins and the watershed using SAGA algoritms. After that, a set of measures describing the networks are calculated. Finally, the set of measures is used by a fuzzy logic algorithm to define the type of watershed. The proccess follows the paper of Zhang & Guilbert (2013): Automatic drainage pattern recognition in river networks.

Description

The plugin for QGIS called “Clasificación de Redes Hídricas” tries to include some tools to try to classify, in an automated way, the type of channel networks inside a watershed. Some of the tools of the plugin, like channel and junction are linked to SAGA GIS tools. The tools used from SAGA are:

• Fill sinks by the algorithm of Wang & Liu
• Determine Channel network and Watersheds (includes junctions)

The rest of the algorithm are implemented in Python using the API of QGIS and tested under QGIS 3.x.

Inputs:

• A raster layer describing a Digital Elevation Model of the terrain

Outputs:

• The channel network (Vector. Linestring layer)
• Joins of the channel network (Vector. Point layer)
• The watershed corresponding to the channel network classified (Vector. Polygon layer with enrichment and classification fields)

Warning: The plugin is a experimental version and some steps could require a long time. In that time QGIS may seem blocked. Please be patient.

Github: The repository of the plugin is in: https://github.com/maurena/networkclassification

The plugin

The main window of the plugin is showed in the Figure. The plugin is divided in tabs that represents the steps that can be applied. The tabs are:

• Channel calculation: Determine the channels of a Digital Elevation Model (DEM) in raster. This process uses SAGA algorithms and create a Filled DEM, a Channel Network and the Junctions of this Channel Network.
• Enrichment: Determine additional attribute information for each watershed in order to be used in Classification
• Classification: Classify each watershed using the fuzzy functions proposed in Zhang & Guilbert (2013)

Detailed description of the tabs

The tab to determine channel network has three entries:

1. DEM Selection: A drop-down list that allows selecting between the loaded raster in the QGIS Project. The selected raster must be a Digital Elevation Model.
2. Minimum slope: In order to determine channel network, the DEM should be filled. Following the description of SAGA, this entry is “Minimum slope gradient to preserve from cell to cell; with a value of zero sinks are filled up to the spill elevation (which results in flat areas). Unit [Degree]”. The default value for the plugin is 0.01 degrees (SAGA considers a minimum value of 0.1)
3. Threshold: Represent the Strahler order to start a channel. The default value is 5 as is recommend by SAGA.

Tab for watersheds enrichment

This tab includes the attributes proposed in the paper of Zang & Guillbert (2013) and will include other as fractal dimension between others. In this tab, you must select a Channel layer and a Joint Layer in the first two drop-down lists. After that, you can select a watershed layer to calculate the attributes for each element in this layer or mark the check to apply the calculation for all channels by creating a convex hull that acts as watershed. Then you can select the attributes (Angularity, Length Ratio, Sinuosity, Elongation or Fractal Dimension) that will be calculated when “Calculate” button is pushed. The description of the attributes are:

• Angularity: Mean value, in degrees, of the angles in all junctions in the channel networks inside each watershed. If the angles in the junction is only one, this angle is used. In other case, more than two angles, the angle between the segment with highest Strahler order and the other segments is used. In case of more than a segment with highest Strahler order, the longest segment is used.
• Sinuosity: Calculate as the sums of the ratio between each segment of the stream and the distance from the spring to the outlet of the same stream.
• Length Ratio: Calculated as the ratio between the length of the main channel and the length of the tributaries. The length of the main channel is approximated by using the longest channel from all springs to the outlet.
• Elongation: Calculated using the oriented minimum rectangle (OMRE) of all segments in the network for each outlet. The elongation is the ratio between two sides of this OMRE.

Tab for classification

This last tab is used to determine the classification of the watershed. In order to do this it is necessary to choose the layer from the drop-down list of the window. After that, you can push the button "Apply Fuzzy Logic" to finish the calculation.

References

Ling Zhang & Eric Guilbert (2013) Automatic drainage pattern recognition in river networks, International Journal of Geographical Information Science, 27:12, 2319-2342, DOI: 10.1080/13658816.2013.802794

Bug tracker

Please, any bug you find send to the email of maurena.

20240325. First version of the plugin

No bugs detected until the cited date.