The National Stream Internet Project: An analytical infrastructure for data on stream networks

Accurate, high resolution information does not exist for consistent status and trend assessments of water quality and aquatic biotas throughout the >3,000,000 kilometers of rivers and streams in the U.S. Without that information, prioritization of limited resources for conservation and management proceeds inefficiently. In recent decades, however, massive amounts of water quality, biological surveys, and habitat condition data have been collected by state, federal, tribal, and private organizations.

Examples of big aquatic databases



Examples of big aquatic databases that are common in many parts of the country. Stream temperature measurements (a), stream discharge measurements (b), fish sampling locations (c), and water chemistry measurements (d).


In many cases, high-quality information could be developed from those data if a nationally consistent analytical infrastructure for data on stream networks existed. The National Stream Internet (NSI) Project was funded by the Landscape Conservation Cooperative Network as a national initiative grant to develop that infrastructure.

The project has three simple objectives:

  1. develop and refine free statistical software for the analysis of data on stream networks,
  2. ensure compatibility between the stream software and the National Hydrologic Dataset, and
  3. conduct a workshop with researchers and aquatic program leaders to brainstorm about national priorities and the opportunities that big data and stream statistics now provide for developing better information about aquatic resources.

The national workshop was held in Boise in late April of 2015 and brought together a diverse group of academic researchers, stream statisticians, LCC staff, and aquatic program leaders from USFS, USGS, NOAA, BLM, USFWS, EPA, and the National Fish Habitat Partnership. A robust and wide ranging discussion ensued over a two day period in which many opportunities and some obstacles were identified.

Stream internet graphic











Elements comprising the National Stream Internet that could be used to develop valuable new information at relatively low cost anywhere in the country.


Prior to the NSI workshop, the Boise Aquatic Sciences Lab held the 3rd annual training workshop for stream network statistics, which has been conducted since 2013 to develop a grassroots level user-base. The statistics workshop again proved popular and 100 participants either tuned in for the webinar short course or traveled to Boise to work with course instructors and apply the stream models to their datasets.

Students study stream network statistics at 3rd annual training workshop in Boise early 2015.
Students study stream network statistics at 3rd annual training workshop in Boise.


The spatial network models can be applied to databases characterized by clustered locations, which provides a strong incentive to develop comprehensive, interagency databases. Moreover, the spatial models outperform traditional statistical techniques and enable predictions at ungaged/unmonitored sites, which facilitates development of high-resolution status maps for full river networks. Two recent applications of NSI technologies within the northwestern U.S. are the NorWeST project  and Climate Shield project.

When technical aspects of the NSI project are completed near the end of 2015, a website will be launched that describes the project in detail and an analytical infrastructure will exist for using the spatial statistical network models with many types of stream datasets. As NSI applications like NorWeST or Climate Shield become more common throughout other parts of the country, managers and conservationists will be empowered with better information for decision making and resource investments.

For more details about the NSI project, contact Dan Isaak; 208.484.6554.

Additional resources