# What’s the carbon footprint of an email?

E-mail is often regarded as cheap and more environmentally friendly than paper. But is this really the case? I suspect that loads more emails are sent nowadays than paper letters were sent before the Internet became popular.

What is the carbon footprint of an email, including drafting, sending, and reading it? And how much does adding a large attachment contribute?

• I suspect the actual formula would look like: CO2 produced = #recipients*(alpha + beta*attachmentsize). I'd love to find actual values for alpha and beta. Commented Sep 22, 2016 at 8:26
• The marginal footprint of a single email is highly variable, across several orders of magnitude. Is that what you're really looking for, or are you trying to establish whether email has a larger carbon footprint than letter-writing, perhaps on a per-capita basis?
– Nic
Commented Mar 2, 2020 at 16:01
• @Nic My goal was twofold; 1) make people aware that email has a carbon footprint and 2) get an estimate of how much email contributes to one's carbon footprint. Commented Mar 7, 2020 at 8:11

Here is an answer based on this 2010 Guardian article I found. Not sure how accurate it (still?) is.

• 4g CO2 for a 'normal' email.
• 50g CO2 for an email with a long attachment
• 0.3g CO2 for a spam email (that is filtered out by a spam-filter, or deleted quickly)

For comparison; this means that 10 emails a day, 8 without and 2 with long attachment, is on a yearly basis equivalent to burning 5.4 gallons (24.5 liter) of gasoline, or driving 119 miles (191.5 kilometer) with an average passenger car (source: 48kg filled in EPA calculator).

The article says that the footprint numbers are based on a book written by Mike Berners-Lee (the brother of Tim) called "How Bad Are Bananas? The Carbon Footprint of Everything" and assumes

a typical year of incoming mail for a business user – including sending, filtering and reading – creates a carbon footprint of around 135kg

and on research by McAfee that found that

62 trillion spam messages are sent every year, requiring the use of 33bn kilowatt hours (KWh) of electricity and causing around 20 million tonnes of CO2e per year.

The article also provides a rough estimate that the average email has 1/60th the footprint of an old-fashioned paper letter.

• As at 7 March 2020, The BBC are still quoting the same numbers.
– Fred
Commented Mar 7, 2020 at 12:10
• @Fred, indeed. The same numbers keep turning up in various places. I suspect no one redid that research. It's likely that the carbon footprint per email is lower now, but I think more emails are sent nowadays compared to 10 years ago, so the overall footprint of email may have gone up. Commented Mar 7, 2020 at 12:18

The guardian article mentioned by THelper to me is suspect.

It cites McCaffey that claims that 62 trillion spam messages take 33 billion kWh of power. That's 2000 messages per kWh.

Hmm. Let's suppose that the average message is a megabyte. Typical internet connection at the consumer level is aobut 5 Mbit/s So a megabyte takes about 1.3 seconds to send. Ignore the computer for now. The household router uses 8 watts. As you go up the internet food chain, both the power and speed go up. My bet is that the larger ones are more efficient, but lets pretend they aren't. Let's also say that the message takes 2 seconds. There are 3600 seconds in an hour, so a message takes 2/3600 * 8 watt hours. This time we will round the other way 1/2000 * 8 watt hours or 1/250 of a watt hour.

From me to gmail.com is 15 hops. Assume this is typical. So 30 hops gets my email to my destination. So 30 * 1/250 = 1/8 of a watt hour. So a kWh will handle 8000 messages.

Further: Lets take another look at that average message size. I currently have 30,000 messages in Gmail, using about 3 GB. So 10000 messages per GB or about 1/7 of a megabyte each. So our kWh could transport 80,000 messages.

Ok. Look at the electricity to make the message. I'm in front of my mac pro. It takes me 10 minutes to write 500 words to a friend. My computer with it's 3 monitors uses 200 watts. So 10/60 * 200/1000 = 1/30 kWh. Writing it is by FAR the big energy cost.

But: How many do I write? According to gmail, 2000 last year, so about 6 a day. (If I wrote it, I keep it.)Not counting spam, I receive about 60 a day. Writing still dominates. I use a fifth of a kWh to write. Reading most messages takes under 30 seconds. If we assume 1 minute each, then it works out that reading and writing are about equal. Both are monsters compared to moving the data around the net.

So since the average message is written once and read once, we have 2/5 of a kWh per message. So that works out to only 2.5 messages per kWh. Ouch.

Hmm.