Recently, a couple very clever guys have been tinkering with NHDPlus data, generating static and interactive maps of the rivers and streams represented in this impressive geospatial data set created by U.S. EPA and USGS. I was particularly impressed with Mike Bostock's version, leading me to fool around with his package to generate the map above.
NHDPlus data is arranged into 18 regional drainage basins for the continental US, with the California basin being a combo of three sub-basins, which generally equate to the Klamath/North Coast, Sacramento/San Joaquin, and everything else along the Central/South coasts and inland desert regions. It wasn't until now that I noticed California is the only state that roughly follows the contours of its main watersheds - and the water geek in me likes to think California's founders took hydrology into account when establishing its borders.
Mike Bostock generated a package for reading the NHDPlus data to generate .json data and .png images for all 18 drainage basins in the continental US. He relied on the designated stream order to help assign line width - the higher the stream order the thicker the line. This can be a little misleading though, since stream order doesn't generally translate to flow and in some instances seasonal streams or ephemeral desert washes appear almost as significant as the Sacramento and San Joaquin Rivers - some of California's most significant water courses.
The beauty of this map is that it manages to effectively pack in all the thousands of rivers and streams in California into a single image, which I've never really seen before. The Delta is accurately represented as the hydrologic heart of the state and although numerous waterways can be found in California's southeastern deserts, most just fizzle out before reaching a river of much use to cities, farms, or fish.
For a little more abstract take on the map I let Illustrator do its magic - (sort of) showing the relative density of rivers and streams throughout California:
The innovative work of Nelson Minar apparently sparked Mike Bostock's interest in this data set, after creating a set of vector tiles for Polymaps, Leaflet, and D3. Jason Davies built on Nelson's efforts to create an modified version for D3 in the Albers projection. Although Mike's static version doesn't let you zoom or pan around, it does provide a visually stunning overview of the nation's catchments, which I think is a pretty great use of the data!
To deploy Mike's package on a Mac you need, or should have, several other packages, including Node.js, Homebrew, topojson, and canvas. If you're interested in this sort of thing it's worth figuring out though, since he's provided a great framework for converting shapefiles to png images or json data to create interactive maps with D3.