Hacker News new | comments | show | ask | jobs | submit login
Wind Map (hint.fm)
909 points by taylorbuley 2033 days ago | hide | past | web | 79 comments | favorite

Not actually a good representation for wind energy potential, because wind power is cubic in wind speed (v^2 for kinetic energy per unit volume, times v^1 for volume/(cross sectional area * unit time)). So energy potential is concentrated in the "long tail", infrequent high winds. (E.g. 30 mph is 8x more powerful than 15 mph, 27x more than 10 mph). The distribution looks much more "uneven" -- both in space (clustered) and time (fluctuating) -- than a linear-in-windspeed map shows.

Off topic: if anyone's looking for a slightly gigantic weather history dataset ("oodles of terabytes"), look at [1]. It has accumulated "a few thousand" weather stations, at 5 minute intervals, going back "a few" years (I forget the numbers).

[1] http://madis.noaa.gov/

It's easy enough to convert something like that to energy potential, but when determining whether a site has good potential for wind energy development we really do tend to think of the wind in terms of speed (in m/s) and not energy density.

Depending on the capture technology (even specific wind turbine models using a certain technology) and project arrangement the variation even at a single location is very high.

And this is a beautiful visualization of the flow.

If you are in the business of determining whether a site has good potential for wind energy development, answer me this if you can: what is the maximum realistic potential percentage of the U.S. energy consumption that windpower could satisfy in 10, 20, 30, 40, and 50 years given realistic conditions (funding, available land, etc.)?

That question is literally a massive research project and I couldn't begin to answer it here. The best place to start is probably the DOE 20 percent by 2030 report released in 2008.


The US has huge amounts of available wind energy, but constrained supply chain, development, and transmission resources. 20% by 2030 is technically possible but will require huge amounts of work.

Note for those unfamiliar with the field: there is a considerable difference between "X% of electricity use" and "X% of power use" (for the USA, about a 1:6 factor, if I read http://en.wikipedia.org/wiki/Energy_in_the_United_States correctly). That 20percentwind.org site, to me, has the former as a goal.

You have it right. 20 % of electricity use means, that on average power is the same 20 %. When it is blowing in large areas, more than 50 % of electric power is produced by wind turbines and when it's not blowing the percentage may be very small. In Portugal the maximum power share wind turbines have had is 75 %, though only 17 % on average is produced by wind.[1]

This creates two problems:

1) You need to have a lot of transmission capacity, which is used very inefficiently. With conventional powerplants you can forecast pretty easily, where and how much electricity is produced and consumed and you can size power grid accordingly (with backup capacities, of course). With wind power the production places and amounts are all the time changeing and you should have plenty of capacity, of which on average only small share is used.

2) To cover the consumption during non-windy times, you need to maintain backup power plants, which should be able to follow load rapidly. Water power is ideal for this, but as penetration level raises you need also other power plants, gas turbines for example (but who wants to maintain a turbine, which is used rarely...). Also batteries may be applied in the future, but in large scale that is nowadays pretty much science fiction.

[1] http://en.wikipedia.org/wiki/Wind_power#Penetration

Interesting point. Of course a wind-powered device cannot extract all of the power of the wind; the power goes more like v^2 times (delta)v, where the delta is basically how effectively the device can slow the wind. Of course, one might expect this term might be proportional to the original windspeed, so it still goes roughly as v^3. For a related analysis, see Betz' Law: https://en.wikipedia.org/wiki/Betz_law

Hey, thanks for that weather dataset link. I had tried to find something similar a couple of months ago and completely failed — somehow, I missed this site entirely.

This is beautiful and a brilliant, intuitive "100,000 feet up" picture of wind movements.

But it's not horribly useful if you are interested in wind conditions at a particular place. Trying to figure out if a spot is experiencing 15mph wind vs 10mph is hard, for instance.

Compare it to these maps: http://passageweather.com/ which are not pretty nor are they intuitive but they are useful for someone trying to use the map to get detailed, localized wind condition information.

It seems like there should be some kind of middle ground between this wind map and maps like passageweater.com, maps that are intuitive and pretty but also provide detailed, useful information.

Maybe using colors instead of shades of gray would be clearer?

Also wouldn't mind seeing high pressure/low pressure systems, fronts, so you can see how they generate/impact wind.

actually, if you click on the map, you can zoom in, and you can see whether a particular spot is experiencing fierce winds or not. I haven't figured out how to zoom out yet though.

NINJA EDIT: There is an 'unzoom' button on the left!

I can't figure out how to zoom out either because I don't think they've implemented that functionality. You can also zoom in infinitely...which is strange.

Is a 5 mile an hour difference important?

Yes, absolutely, for many sports that rely on wind, such as wind- or kite-surfing.

To me? Yes.

You just hover over the spot of interest and then it tells you to one decimal place...

they just need to add mouseover details. Rough wind speeds are already indicated by line density, using color would just waste that variable so you couldn't use it for things like pressure. That would be a super cool visualization and again details could be provided by mouseover data.

It would also be cool if you could zoom in. There is definitely a lot of local variation in wind speed/direction that you have to filter out when looking at the entire US but would still be very interesting to see.

Beautiful visualization! It would be interesting to have an underlay showing the topography so one can correlate features in the wind pattern with mountain ranges, etc.

Great idea. You can clearly see how the wind patterns are distorted around the Sierra Nevadas and Rockies. While in the East wind patterns are much more smooth.

You could also add color to indicate the temperature of the air.

+1. Some of the linear disruptions in wind flow suggest the topography, it'd definitely be interesting to correlate changes in elevation w/ the wind patterns.

If you could get at the GIS data sets you would be golden. Anyone know of an easy way to integrate GIS data into web applications?

Would something like tilemill help? http://www.tilemill.com

That is the coolest thing I have seen in a very long time. Thank you!

This reminds me of the proof that there is always a point on the earth where there is no wind. This is a property of spherical surfaces: http://en.wikipedia.org/wiki/Hairy_ball_theorem

There is only a windless spot on earth if you take wind to mean air moving horizontal to the earth's surface. Otherwise imagine that we have wind that points north along the ground, with wind pointing up at the north pole and down at the south pole (and higher up wind comes back south).

That's a good point. I think the surface you are describing (that the wind would sweep out) would be a torus. The 2-torus is "combable."

And some people think mathematicians have no sense of humor!

I am used to seeing wind markings on aviation maps (e.g. http://www.aviationweather.gov/adds/winds/) but it's pretty cool how much clearer this map is. It makes me wish I could see a much larger portion of the world so I could get a better understanding of the currents that are entering/exiting the US.

That's amazing. It's so inspiring to see someone can have such a simple idea, execute, and create something not only incredibly intuitive and useful, but beautiful too.

It makes you wonder what other things are out there waiting to be done, and what you can do.

I'm not so sure it's exactly simple to implement. A quick glance at the code, it seems like they had to implement vector field support from scratch.

I haven't looked at the code and how they implemented vector field support, but couldn't a library like Sylvester have been used?


Paper.js looks like it could work as well:


been wanting to do this myself for like forever. it looks like they are using high level summary data from the NWS ... they also have this amazing thing called MADIS which focuses on micro climate research (and more) ... so in theory you could get a high resolution version of this for san francisco or nyc, and that'd be really great and possibly rather useful too. I think the Weather Underground is a subscriber of MADIS, and their "rapid update" feature has some of this, and also I should mention some of this is collected by ham radio people in their spare time too.

Never before has a piece of software made me wish for a hurricane to ravage my country simply out of curiosity.

This is beautiful, but I find it lacking because of the focus on the US mainland exclusively. It would be nice to see North America as a whole along with a good part of the Pacific and Atlantic, because these areas have a massive effect on wind patterns.

For example, the wind over the Chicago is travelling "backwards" for no apparent reason. This is most likely due to this being the "backward half" of a cyclonic system centred over Canada.

The other thing that struck me is the effect the Southwest desert has on wind patterns over all... just, wow.

Very cool. It's interesting to see how elevation affects the wind here (http://www.theodora.com/maps/new9/usa_elevation_map.gif), but I'd wish there was a worldwide version. (I'm aware that you need a data source)

Should you be interested, here's data from the Netherlands: http://www.knmi.nl/samenw/hydra/cgi-bin/register.cgi

They are using a bad dataset. Some of the discontinuities you see are administrative boundaries between National Weather Service forecast offices.

http://nixweb.com/dust/ is my version, for the SF Bay Area only. It was 15 years ago so pardon my Java applet.

Where is the wind over the water? Who cares about what the is on land? As a sailor, I find this map frustrating and upsetting for showing me beautiful wind everywhere I can't sail a boat.

Beautifully done. Just discovered you can zoom in by double clicking.

very pretty. related: http://windhistory.com

I got lost for quite awhile looking at all the wind and how it moves around the area I'm pretty sure is my mountainous home. Kudos to the maker(s).

This appears to be an animated form of line integral convolution. Cabral and Leedom paper's from SIGGRAPH '93, Imaging Vector Fields Using Line Integral Convolution (PDF):


This is great, Mike - thanks for sharing the doc.

They should add Canada, Mexico and the ocean if possible to see where these in and out flows are coming from.

or tie it in with the ocean viz from nasa from the other day! unfortunately most of the data i could find for canada was only in the south, and the mexico data was similarly in only the more densely populated areas.

My favorite wind potential graph, NPR's "Visualizing the U.S. Electric Grid"


Makes offshore look pretty good.

This visualization is incredible but it made my browser freeze for almost a minute :(

Really a stunning presentation of this data. A huge "whoa." moment when it loaded.


On a side node: Anyone else reminded of Deluxe Paint palette cycling animations by this?

Yes, of course! The master of this technique is here: http://www.effectgames.com/demos/canvascycle/


This would be even cooler if you could display wind speeds at different elevations above the ground. I am guessing that at a certain elevation you have higher winds that kites might be able to capture energy from.

It's also neat because you can tell where the mountains are. They're the sections where the wind was pretty slow, and then it speeds up and then slows down again. You can see it over the Rocky Mountains.

I can zoom really far in by double-clicking, but I cant zoom back out.

I have to refresh to zoom out.


It would be cool to see the different speed ranges reflected by [color of my choice] as well!

There's an UNZOOM button on the left.

What do you want? Me to actually READ the websites!?

This would be even better if it could display change over time. To support, for example, animate the change in wind direction for the past week.

some previous visual inspiration possibly for project: (based on lidar data) http://www.datatelling.com/2011/07/29/wind-flow/


as well as the big nasa ocean viz from the past few days

predates both of those by quite a bit: http://nixweb.com/dust/

Odd that they didn't label any of the cities in the places that seem to have the greatest generation potential.

This will be a great tool for paragliders. I will post this on paraglidingforums Thanks for the link.

A nice addition might be the current locations of wind farms (if that data is accessible someplace).

Have already been working on an app for this (just for lulz); haven't had time to finish it yet. I pull from public DOE data regarding generation locations.

Looks beautiful but barely usable. Winds have different speed and direction at different altitudes.

This. The site should clearly state whether they're showing 1000ft prevailing winds, jetstream-level currents (around 30,000 feet) or more surface-level effects. All the people talking about power generation from this data should be aware that the only wind that matters there is below 1,000 ft (in many cases, far below 1,000 ft.)

"Surface wind data from the National Digital Forecast Database."

its been said by others but this is an incredible visualization of a massive amount of data in a beautiful and intuitive medium. it'll be interesting to have this for other large datasets eg. real time web usage...

So... I want this for all of the world, extrapolated over the oceans.

Looks a lot like a chunk of hair. Very neat implementation!

I wish they would make this open source. It is beautiful.

Wow that's cool. Where do you get the wind data?

I love this i think it is really cool. Where is Alaska and the rest of the united states?

I want color!

I want this as a screen saver!

I was thinking a desktop background version.

Wind power, baby.

beautiful, but not sure how useful it is?

great stuff, but my cputemp is in the clouds

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | DMCA | Apply to YC | Contact