Minnesota statewide altered watercourse project

This project was a concerted effort between the Minnesota Pollution Control Agency (MPCA) and Minnesota Geospatial Information Office (MnGeo) to create a statewide inventory of streams that have been hydrologically modified (e.g., channelized, ditched or impounded). These data were created in support of the MPCA’s water quality monitoring and assessment program and provide information about stream habitats that have been compromised through such alteration. The project visually reviewed approximately 105,000 miles of stream linework to classify streams into four major categories: 1) Natural, 2) Altered, 3) Impounded, and 4) No definable channel. A number of geographic information system (GIS) datasets were used in the process.

Original funding was provided by the Clean Water Land and Legacy Amendment.

Because altered stream channels (i.e., ditching and straightening of stream channels) have reduced habitat complexity, aquatic organisms that require diverse habitat may be negatively impacted. For this reason it is important to understand the magnitude of stream channel alteration when assessing streams and rivers for aquatic life uses. The extent of stream channel alteration in Minnesota is largely unknown because available geographic information system (GIS) datasets are outdated or incomplete (e.g., national hydrography dataset, public waters inventory, and Minnesota Department of Transportation).

Prior to this project, MPCA staff used multiple tactics to determine whether a stream was altered; the process was time-consuming, subjective and did not provide consistent results due to a lack of comprehensive data indicating where streams may have been modified and different interpretations of the available data. Therefore, in 2011 the MPCA entered into an inter-agency agreement with the Minnesota Geospatial Information Office (MnGeo) to develop a GIS based methodology for determining stream channel alteration, which was focused on visual interpretation of contemporary and historic aerial imagery (including LiDAR) as well as several other reference layers. The project developed a process whereby each stream segment (down to 150 meter scale) was assigned a status of natural, altered, impounded or no definable channel. The evidence used to designate each stream segment was identified and a code of high, moderate or low confidence was assigned to each delineation based on the evidence that was used to make the decision. This study provides the first comprehensive effort to distinguish altered stream miles on a statewide scale.

See the Results section below.

For more than 160 years streams and rivers in Minnesota have been modified, created and rerouted to accommodate various aspects of human development and entrepreneurship.

• Drainage has a long history in Minnesota, starting in the mid-1800s through present day
• Primary reasons for stream channel and landscape alterations
  • Increase agriculture capacity
  • Increase per acre land value
  • Improve highway and railway transportation
  • Public misconception of swamps and wetlands as “disease-breeding  areas”

Historic images

Note: All images are adapted from: Drache, Hiram M. 1992. Taming the Wilderness: The Northern Border Country 1910-1939. Interstate Publishers, Inc. Danvill, IL. pp. 376.

A floating dredge in 1914, picture likely taken west of Waskish, MN	Steam dredge in 1916 near Siggestad, MN.
Bay City dredge straddling the dug channel in 1914 near Waskish, MN	A completed Judicial ditch northwest of Baudette, MN from 1916
Steam dredge in 1916 near Grygla, MN

In 2008, the Land Management Information Center (now MnGeo) undertook a pilot project to develop a methodology to distinguish between natural and altered watercourses and to create a ‘natural/altered watercourse’ tool tied to the high-resolution National Hydrography Dataset (NHD). This project evaluated portions of three 8-digit HUC watersheds in different parts of the state in order to create a methodology that would work under the full range of hydrologic regimes found throughout Minnesota. During the pilot effort, areas were chosen in the North Fork Crow River (07010204), Snake River (St. Croix Basin - 07030004), and Redwood River (07020006) watersheds (Krumrie et al. 2008). The method development effort included a review of results by MPCA staff and quality control against data collected in the field by MPCA.

The statewide project built upon the efforts completed in the 2008 project and used this methodology along with new data processing tools and GIS layers (e.g., LiDAR and 2009-2011 aerial imagery) to complete a comprehensive visual analysis of all streams statewide.   For more information about the pilot project please download the final report and methodology document.

• Altered Watercourse Identification using GIS - A Pilot Project - Final Report
• Determining Altered Streams - Methodology

Watersheds involved in the 2008 Altered Watercourse Pilot Project (Snake, North Fork Crow and Redwood River).

Current GIS datasets that depict stream channel condition are often incomplete or inaccurate. This information is critical to understanding biological and chemical monitoring data, as channelized streams often have reduced habitat complexity, which reduces biological integrity. The four images below provide an example of one type of stream channel modification from an unnamed stream in South-Central Minnesota. The image on the bottom right provides an example of the inaccuracies with the National Hydrography Dataset and one reason for this project.

1938 image of an unnamed stream in South-Central Minnesota	1991 image of an unnamed stream in South-Central Minnesota
One meter resolution LiDAR DEM of an unnamed stream in South-Central Minnesota	Aerial Image from 2010, with National Hydrography Dataset (NHD) representing stream channel condition of an unnamed stream in South-Central Minnesota

• Reduction in both in-stream and overhanging cover
• Loss of natural riffle/run/pool sequences
• Creation of unstable substrates
• Loss of natural channel sinuosity
• Altered hydrology
• Overall loss of in-stream habitats

Rice Creek – An example of historic ditching and stream restoration in Minnesota

Although restoration activities have attempted to reestablish stream channels to much of their original course, a comparison of the historic channel and restored channel (based on this image extent) indicate that approximately 4 miles (57%) of the original stream channel has been lost due to stream channel modifications made sometime between 1910 and 1925.

May 8, 1947 Image of Middle Rice Creek near now Interstate 35 bridge (Rice Creek Parkway) and State Highway 3.

2008 image of Rice Creek following  Rice Creek Watershed District Stream Channel Restoration Project

1. Improve our understanding of disturbance within a watershed

• • Provide accurate information regarding the status of a sampling site
  • Assist in assigning beneficial uses under a Tiered Aquatic Life Use Framework
    • Provide another line of evidence for use designation
    • Provide boundary extent of designated uses

2. Improve our understanding of stream hydrology

• • Support stressor ID and TMDL planning

3. Determine the number of altered stream miles in the state
4. Identify areas where the National Hydrography Dataset (NHD) needs updating
5. Previous methods were time consuming and involved multiple reviews by numerous MPCA staff:

• • this  ‘standardized’ layer will:
    • Minimize review time
    • Maintain consistency
    • Be more accurate
    • Increase efficiency

Aerial photo from June 15, 2010, of a backhoe straightening a portion of the South Branch, Buffalo River, near Barnesville, MN; photo courtesy of Mike Hoppus (MN DNR)

This project utilized Geographic Information System (GIS) software, multiple data sources (e.g., LiDAR, 2001 – 2010 aerial imagery, historic air photographs) and a decision tree approach to visually describe channel conditions of all streams in Minnesota. Using a decision tree approach afforded us the capacity to track the type of delineation (i.e., natural, altered, no definable or impounded), the criteria used to make the delineation and the confidence of the decision (i.e., low, medium and high).

Examples of data sources

Reference layers - some of the data sources used in this project.

1938 Aerial image of an unnamed creek Southwest of Worthington, MN

2011 Aerial image of an unnamed creek Southwest of Worthington, MN

Altered watercourse decision tree

Altered watercourse decision tree

Larger version of flowchart: NHD Altered Watercourse Event Type Determination – Flowchart

Example of how different data sources were used to determine stream channel conditions

Aerial photograph from 1938, red arrows indicated natural bends (sinuosity) within the stream

1991 leaf-off satellite image. Note that the natural bends from the 1938 aerial photographs are no longer connected to the stream

One meter resolution Light Detection and Ranging (LiDAR) imagery (using a hillshade format), provides further indication that the natural stream channel has been modified; red arrows highlight the remnants of the natural bends. In this figure you are also able to see the soil piles created as a result of channelizations

Over 105,000 stream miles were visually reviewed multiple times as part of the altered watercourse project. Our review found that approximately 49.6% (41,204 miles) of streams in the State of Minnesota have been altered in some way (i.e., ditching, straightening or channelization) by humans. Because we used georeferenced historic air photos we were able to identify stream channel alterations dating back as early as the 1930’s. The map provided indicates stream channels that were designated as altered, natural and impounded. 

For more detailed explanations about how these data were developed, please use the following link to download the methodology document. The methodology document illustrates how the delineations were made and provides several examples of the different types of delineations.

• Altered Watercourse Determination Methodology

Altered watercourse layer map

Larger version of map: Altered Watercourse Layer – Map

Map of percent of modified streams by 8-digit HUC

Larger version of map: Percent of Modified Streams by 8-Digit HUC – map

Percentage of delineation types within each watershed

Larger version of chart: Delineation Percentage – chart

GIS on-line map

View larger map

This project also identified stream line-work that did not represent flowing waters (no definable channel). Some examples of these types of channels included:

• Swales through agricultural fields
• Shallow, wide, grass waterways
• Wetland and lakes, where no evidence of draining or impounding is present. 
• Streams that disappeared or are now subterranean; usually associated with large cities/suburban areas

An example of a swale through an agricultural field.

An example of a shallow, wide, grass waterway in south-western Minnesota.

Of the approximately 105,000 miles of stream line-work in the National Hydrography Dataset, this project identified approximately 22,000 miles of that does not represent an actual river or stream. Therefore, for MPCA purposes, we have removed the no definable channel miles from our analysis and we now estimate the flowing waters of Minnesota to be approximately 83,000 miles.

The following three images show an example of an unnamed creek near Lino Lakes, MN that has disappeared and is likely a subterranean stream connected to storm water ponds. The National Hydrography Dataset (NHD) stream line-work does not represent a stream and therefore the altered watercourse designation is no definable channel.

1947 image of an unnamed creek near Lino Lakes, MN

1991 image of an unnamed creek near Lino Lakes, MN

2010 image of ditch near Lino Lakes, MN

An example of a series of wetlands/small lakes near Circle Pines, MN, the National Hydrography Dataset (NHD) stream line-work does not represent a stream and therefore the altered watercourse designation is no definable channel.

• Altered Watercourse Determination Methodology. This document describes the tools, data and determination criteria used to interpret and create the altered watercourse determination and GIS layer. It is a revision of the pilot methodology developed in 2008. Revisions were made based on review and QA/QC of pilot results and availability of additional data sources.
• Statewide Altered Watercourse Project – Download the altered watercourse dataset and metadata
• Altered Watercourse Delineation Project Schedule: Red River Basin, Winter 2012; Lower Mississippi, Minnesota and Metro, Summer-Fall 2012; Arrowhead and Central Lakes, Fall 2012-Spring 2013 (see map below)