Improving American Whitewater’s National Whitewater Inventory

Nick Gottlieb and Seth Fitzsimmons, May 2019

Data is the heart of the National Whitewater Inventory. American Whitewater has been collecting and maintaining point data for put-ins, take-outs, gages, rapids, and other whitewater points of interest. Until now, however, our data have been limited to point-based descriptors, thanks to volunteer Stream Team members using GPS devices and satellite imagery. Thanks to a generous grant from the US Coast Guard, we’ve been able to embark on a project to use the USGS’s National Hydrography Dataset (NHD) to improve the National Whitewater Inventory’s map-based information.

The NHD represents the nation’s drainage networks and related features, including rivers, streams, canals, lakes, ponds, glaciers, coastlines, dams, and stream gages. The NHD, at 1:24,000 scale or better, is the most up-to-date and detailed hydrography dataset for the Nation.

Thus far, this phase of the project has involved preliminary data cleanup and snapping existing point data to the NHD Plus High Resolution (HR). We’re able to turn those points into lines, wending their way through canyons and valleys. Now that they’ve been correlated with NHD flowlines, a whole slew of analyses become possible:

  • Generation of partial flowlines corresponding to whitewater runs (check out the map below)
Elevation profile for the Upper Gauley in West Virginia (map).
  • Measurement of boatable miles within different management areas or conservation initiatives, e.g., whitewater river miles in a proposed wilderness area, a Congressional District, or a unit of the National Forest
  • Identification of watershed boundaries influencing particular stretches of river (supporting personally subjective gage correlation)
  • Generation of elevation profiles for whitewater runs, providing additional context over the traditional “feet per mile” metric

Data-wise, we’ve gone from this:

NWI put-in, take-out, and rapid locations in Northern California.

To this:

Generated reach segments.

Using this:

NHD flowlines.

What Does This Mean for the AW Paddler?

Over the coming months, we’ll be working on a few enhancements to the American Whitewater website that will allow display of and interaction with the new dataset. First and foremost, we’ll be replacing all the existing map interfaces with a more modern basemap (thanks, Mapbox!) and integrating our newly-enriched geospatial dataset into the map display. So, when you look at a river, you’ll see the actual segment of river highlighted — not just pins for put-in and take-outs — and you’ll see other georeferenced data (like rapids and stream gages) as well.

An example river detail page with the new map in place.
Rapid locations on a section of the Rogue River in Oregon.

In addition to detailed maps of each reach, we’ll be implementing a nationwide interactive map interface for browsing the National Whitewater Inventory. This interface will allow users to browse reaches based on their location, something currently limited to just browsing by state. We’re also planning to incorporate visual cues based on river metadata: Is the river running? What’s it rated? Is it going to rain?

You can explore the work in progress data on a map here. It will become more interactive as it’s integrated with the American Whitewater website.

Finally, we’ll be updating the Stream Team interface to allow creation of new reaches and modification of existing reaches using the NHD as a base. In other words, volunteers will be able to use the web interface to define both access points (as they always have) while simultaneously generating geospatial line data that represents the reach in question.

Demonstration of dragging point locations and having them snapped to NHD features.

Technical Details

To do so, we’ve followed this process (simplified somewhat here), pioneered by volunteer Joel McCune:

  1. Snap put-ins to NHD flowlines
  2. Starting at the put-in, navigate downstream until encountering the flowline segment that the corresponding reach take-out snapped to
  3. Stitch together flowline segments that were traversed
  4. Crop stitched linework to the snapped put-in and take-out locations

The code to perform this process is on GitHub; it consists primarily of SQL queries for navigating the NHD flow network in PostgreSQL and Make targets for producing various outputs.

As part of this process, we’ve been able to automatically generate lines for 5,346 reaches (of 5,573 with put-in and take-out locations, some of which are located outside the United States). Of these, 46 have been automatically flagged for review, typically because the put-in has been located downstream of the take-out (a common mistake for park and play spots).

The data are still not perfect by any stretch of one’s imagination. We’re looking for you, dear readers, to go through, flag, and help correct your local reaches to make it easier, safer, and more fun to get on your favorite river.

If you have mapping skills and would like to join our volunteer team please reach out to us by emailing evan@americanwhitewater.org.