Dirty Data?

Are e-mail attachments bad for the environment? Data-storage expert Matthew Yeager thinks so. In an interview with Mother Jones, Yeager compares the greenhouse gases created from sending a 4.7 megabyte-sized email attachment to boiling a kettle of water 17.5 times. As we inch over the one zettabyte marker of stored data worldwide, that’s — well, a whole lot of kettles. Yeager argues that redundancy represents a huge share of stored data, something to consider before sending that cute bunny photo to 300 of your closest friends. [%comments]

Ronaldo Yamashita

A couple weeks ago I got the same question, did some back of the envelope calculations (got a power consumption of at least a city of 350,000 habitants).
But after all, I concluded that if we did not have attachments, we would send far more physical copies, requiring gas for trucks and airplanes. It would be far worse for the environment.
Also, some new technologies, like deduplication, will help to reduce environmental costs overtime.

David L

The good news is that as data transfer gets faster and high-speed connectivity becomes more ubiquitous, it becomes less important for us to store data locally. Cloud infrastructure and SaaS, right now, are sort of just trendy buzzwords, but in a future world connected by fiber optics and high-speed mobile data networks, remote data storage could portend some major synergies.

Ian Kemmish

I notice he doesn't actually produce evidence for that figure, but it seems very high. Let's apply the principles of "street math" you espoused a few weeks ago.

My iMac takes a fraction of a second on just one core to send that file. My ISP's servers presumably take even less. Lets say the email goes through ten computers on its way to the recipient. That's still (a lot) less than a second of CPU time, but less be generous and call it a second. At maybe 100 watts, that's 1/36,000 kWh. My modem is tied up for a minute or two uploading the data, but I haven't switched it on specially, so that's a fraction of a watt (busy vs. idle consumption) for a minute - still not much. 1/120,000 kWh? Adding a generous fudge factor for all those routers out there, lets say a total of maybe 1/20,000 kWh

Boiling a kettle is 3kW for maybe three minutes. I make that 3/20 kWh. The electricity it takes to boil one kettle would appear to be equal to the marginal electricity use of sending three thousand (give or take) such emails.

Of course, there's the amortised energy cost of actually building and shipping my iMac, but how much is that likely to be for just one email?


Ian Mason

These figures are ridiculous; I can see that some people failed their physics exams.

Let's examine the claim:

"sending an email attachment of 4.7 megabytes-the equivalent of about 4 photos
taken on a point-and-shoot digital camera-creates as much greenhouse gas as
boiling your tea kettle 17.5 times"

OK. A standard jug kettle contains 2 litres of water. Water has a specific
heat capacity of 4.2 kJ/KgK, comes out of the tap at circa 10 degrees C and boils
at 100 degrees C. So boiling one kettle full of water consumes:

4.2 kJ/KgK * 2 Kg * 90 K = 756,000 J (Joules of energy)

17.5 times that is 13,230,000 joules, or 13.230 megajoules (MJ). One kilowatt hour
of electricity (energy) is 3.6 Megajoules, so we're talking about a claimed
3.675 kilowatt-hours to send one email. That's more energy than required to run
a 3 bar electric fire for an hour.

So, the claim is that one 4.7 megabyte email takes 13.25 megajoules, that's
2.8 joules per byte!

To demonstrate how farcically wrong Yeager must be I'll point out that the muzzle
energy of a 5.56mm rifle bullet is 1780 joules, so by Yeager's claim a email
attachment of a mere 636 bytes embodies as much energy as a rifle bullet. Or
let's compare it to the iconic Hiroshima atom bomb, which released about 63
terajoules. Yeager's figures would make Hiroshima equivalent to 4.7 million
email attachments of 4.7 megabytes.


Albert Alexander

If you try to save the world by reducing clock cycles, you will get nowhere. Your picture of a cat pales in comparison to spam, bittorrent, porn, radio, TV, and cell phones. Or how about the "cost" of rendering a first-person shooter game at 60 fps for 20 hours a week?

If you're dying to make a difference, you could send all your email at night. During off-peak hours, power companies burn off megawatts, simply because its cheaper than turning off the factory. You're doing them a favor by consuming that power for a purpose.

Or hell, you could do 'em one better. Get a local fileserver to handle all your local friends' email. Spin up a big ol' flywheel all night and use it to power the server during the day.

Jason Morris

While I agree with the other comments that the statistics supporting this assumption are completely absurd.

Secondly (and far more importantly,) even if the numbers made sense, the fundamental argument is ridiculous:

Claiming that emails with large attachments hurt the environment is equivalent to claiming that 'building solar panels hurts the environment because of the resource intensive, high cost materials used during construction.'

Yes. Making and storing things (even data) requires resources and energy. But the message behind those annoying little reminds at the bottom of emails ("Think Green! Please consider the environment before printing this email!") are still far more important. The result of NOT sending that email with a large attachment are far more harmful than the environmental damage caused by the email.

With a blog title "Freakonomics" you'd think you would have heard of opportunity costs before.

Please go back to economics 101, you're hack-environmentalism is embarrassing.



Others have dealt with some of the absurdity of this. I'll tack on another point:

"Yeager compares the greenhouse gases created from sending a 4.7 megabyte-sized email attachment to boiling a kettle of water 17.5 times."

Really? The inputs are know to three figures of precision? That's surprising, given that he doesn't even talk about the volume of the kettle.

Also, what is meant by "created?" There's a big difference between marginal costs and average total costs.


I don't think it is the transmission of the file where the energy use is--it is the storage of data. Someone has to pay to house that file for the next 100 years because chances are it is not going to be deleted.

The bigger point is that externalities are usually difficult to see and measure.

Drill-Baby-Drill Drill Team

So that "Dirty Dancing " Download Peer Fileshare is also Dirty Data Dancing Needlessly.


***But after all, I concluded that if we did not have attachments, we would send far more physical copies, requiring gas for trucks and airplanes. ***

I disagree. The marginal cost of a big email attachment (storage, sending via email, archving) = zero. It shouldn't be.

If email attachments/computer server storage were somehow metered, people would "shockingly" be more economical of what gets sent/saved/created.


***While I agree with the other comments that the statistics supporting this assumption are completely absurd. ***

Simple solution.

Someone should ask Google what its storage/electricity costs would be if every gmail user used only 1GB of their 8GB gmail account.

Eric M. Jones

@4- Ian Mason.....Right on. Saved me the trouble.

It is always amazing what people will swallow without checking....Innumeracy.

Every year Americans eat enough ice cream to fill the Grand Canyon.....or maybe not.


Back in about 1990, a common program for accessing the Usenet newsgroups would print a message when you posted a message:

"This program posts news to thousands of machines throughout the entire civilized world. Your message will cost the net hundreds if not thousands of dollars to send everywhere. Please be sure you know what you are doing.

Are you absolutely sure that you want to do this? [ny] "

Given that there were millions of messages posted each day and the worldwide expenditure on Usenet servers and bandwidth was clearly not hundreds of millions (if not billions) of dollars per day, this was clearly an overestimate.

Dr J

The Freakonomists seem to be very weak in science, especially physics, and reminds me of some of the whoppers they made in their analysis of global warming - to echo a point made before where is their "street fighting" mathematics on this one?

Maneesh Reddy

Not so fast. He may have a point. Think about the email being stored on a server, not transmitted. It sits there forever. I have 5 year old emails in my inbox. It's probably stored in several different locations, just in case one of them has a problem. All of those servers burn lots of electricity to run, and especially to cool. Have you ever been to a data center? It's amazing how much energy gets used to store data. My 2 cents.


Nice work! Can't wait to email this blog to everyone I know.

Eileen M. Wyatt

Computer server storage on email accounts used to be metered and limited at a much more stringent level. Surely I'm not the only person who remembers the days when it was necessary to delete old emails with big attachments to have room for new mail?

Shouldn't an economist ask the email providers why it became economical to be less stringent?

Jon L.

He is right, for certain sizes of kettle.


The problem with email attachments isn't in the intial sending of them. Think of the viral spread of a 'funny picture' someone sends to their friends, and their frends send to their friends.

Soon that 1MB picture of humor just went to 200 people, who keep it on their email server and download it to their local pc. Now that file is kept spinning on all those hard disks. And spinning.

Think of all the searchable content through Google. They only search what is on spinning platters. Off-line CDRom/DVD storage cannot be included. So all those emails on all those servers are just spinning and spinning. Even with HDD storage density goin up, the information growth is faster.

That's the real energy cost.

Imprivements in the way email programs handle archiving data can go a long way to fixing this. Most are a linear file and once removed to 'archive' are not searchable anymore from the main program, so why archive? I keep my mail in folders to better group it, but upon archiving my email program dumps them all back in one big bin.

Maybe viral email forwards can reference a lnk to one copy 'in the cloud'.

Still, a huge portion of email boxes are stuffed with spam, spam, and more spam. Compare that energy consumption.


Eric M. Jones

I am often amused when the webmaster of an email list group spends the time to refutate their posters about the enormous volumes of information they have to store. I offer up Feynman's (pre-computer but always relevant) "There is Plenty of Room at the Bottom" talk:


Basically he teaches that if everything that humans have ever written were stored in the most compact means possible (text at least), then it would fit inside a grain of pepper...or so. Do the maths.

I have in front of me a thumb drive that holds 16 GB....which is more than sufficient to hold everything I have ever written, plus a half-dozen of my favorite movies in VHS resolution. The new IPods hold 14,000 tunes in 128 kbps resolution.

At 2kB per page, everything my webmaster has on the website (and all the archives) would easily fit into the thumbdrive in a 32 GB Swiss Army Knife (wow...."Victorinox Secure combines the Victorinox Flash USB memory stick with a biometric fingerprint sensor"). That's 16 MILLION pages of text. Not to worry.