The Most Detailed Map of US Waters That You've Ever Seen (2023) | Hacker News Reader

The Most Detailed Map of US Waters That You've Ever Seen (2023)

esri.com

210 pointssurprisetalk31 comments

The detail is impressive - it even has the little 12" wide, 6" deep creek that runs through my property and the neighbors seasonal puddle/not-even-a-pond that forms in his field after heavy rains.

It also has a lot of features that dont exist and is missing several that do.

Pretty strange. It misses a pond that is completely obvious on the lidar elevation map.

I can confirm the peer message. It is missing much more significant water features (ponds/streams/marshes) on land I am intimately familiar with.

I've been looking around creeks in my neighborhood and this layer is substantially less accurate than the base layer it's displaying on top of.. Not sure what's going on with this data.

Relatively less detail I'm sure but Raven Maps has a beautiful large laminated wall map of landforms and drainage in the continental US.

https://www.ravenmaps.com/48state-drainage-landform.html

That’s really beautiful.

It really is. I just had to get everything taken off my walls (and the house basically emptied) because of smoke damage remediation. A large framed map of Boston Harbor had its glass broken. I may just replace it with this which is probably more interesting as I put things back together.

Doesn't some of this change throughout the year? Like some rivers only run at certain times the year. If so when is this data from?

There are differences in representation for the mean annual flow of each waterway. Read the article.

Not bad. Areas like Tucson, AZ are deceptive however.

The Santa Cruz "river" typically doesn't have any water in it. :-)

The deception works well though. Some German POWs got a map of Arizona and tried to escape by rafting down the Gila River. Naturally all escapees were found in less than a week.

More detail isn't necessarily better. I own some property and this water map repeats the same seasonal streams that exist on the state's (very old) records.

they should validate with satellite data or something similar.

Where do they get their ground truth from? They show a lake near me which no longer exists, as of a few years ago when the dam was removed.

(OSM gets this right, FWIW.)

Lidar and some sort of algorithm probably. It seems to work ok but not perfectly and depending on your area the lidar point cloud could be almost 10 years old.

I’d be interested to know how they gathered all that data. Given the very detailed ponds and streams that may seem invisible to our daily lives, did they really send someone out there to survey all of it? Or is this via satellite imagery or something similar?

It's derived from LIDAR scans.

https://www.usgs.gov/ngp-standards-and-specifications/elevat...

I'm not sure it's their best product for elevation anymore, but USGS also publishes 3dep, which is a unified layer of the best available elevation data for the US.

It’s all those “ufos” with clearly marked FAA mandated navigation lights.

From high resolution Digital Elevation Models. They first compute the flow direction from every pixel; from flow direction, the compute flow accumulation at every pixel. They then filter based on some threshold for catchment area, and vectorize the flow paths.

The USGS NHD+ Users Guide includes a very detailed description of the technical process:

https://pubs.usgs.gov/publication/ofr20191096

Weird, just spot checking a few places I know relatively well I see a few issues that seem like they would show up from that method. I wonder if there are issue in build up environments? Unfortunately that link is down for me.

A collection of streams go uphill in this map where in reality I believe they flow into an underground pipe or something that drains into the lake in a different place. I wouldn't expect the underground section to be perfect, but the hill the streams ascend is ~50 feet.

In another place two ponds are not shown, maybe due to dense tree cover? I believe both have water year round, but one was fenced off after a kid drowned and the owners may have filled part of it in.

Hydrological modeling gets tricky when water flows underground, either naturally or unnaturally. A pixel surrounded entirely by higher elevation pixels is a "sink", and if you want to model the overland flow of water a preprocessing step is often to fill sinks and apply a slight gradient to them.

I believe that the NHD+ dataset uses an alternate strategy; it burns an artificial channel into a sink, connecting it to a lower-elevation outlet. The dataset includes attributes that identify these "underground flow" segments. I didn't examine ESRI's Living Atlas site closely enough to see if they've used a different symbology to represent these segments, but your comment makes it sound like they did not.

A pond often has no inlet and no outlet (as opposed to a lake), so it makes sense that some ponds may not show up in this dataset.

Thanks, that's interesting and makes sense that there are alternate strategies with different tradeoffs.

I checked the Lost River and some of the other areas on the Snake river plain. I can see how it would be hard to deal with water flowing underground. I noticed that the lava flows have effectively no streams, which I guess shouldn't be a surprise but it was.

While these particular maps may be from scans, the government really does send people into every little field for this type of thing. Between geological surveys and defense planning, I don't there is an acre of land anywhere in the US that hasn't welcomed a government surveyor at some point.

https://www.usgs.gov/special-topics/water-science-school/sci...

Wow the inner bay area is parch dry

Trying to scroll anywhere on the map crashes the website on iOS. Not sure what that’s about.

It is downloading a MASSIVE amount of data. Good stress test for browsers.

How can that be the case when e.g. chrome can handle opening multi-GB PDFs?

It appears to be streaming the data in and out to reduce ram load. On my desktop it was very laggy and had no detail for 2-4 mins and from then on I can zoom into any area and wait a few secs for the data to overlay. GPU usage is very high while rendering.

Interestingly though, Chrome can NOT handle a 200MB html file. Which is kind of interesting given HTML is the sole purpose of a browser.

Source: Was doing some interesting data extraction work and wanted to pull tables from a 200MB HTML file but chrome always crashed after a few seconds.

I’m not that surprised, I’ve always had the impression the whole HTML ecosystem is more amateurish than PDF.

“ Water plays a central role in all our lives.”

You don’t say.

Reply on news.ycombinator.com