The Carbon Footprint of Digital Print

What is the carbon footprint of a digital book? We have to make some possibly heroic simplifying assumptions. The first point to note is that a digital book has a very, very low carbon footprint if no one reads/accesses it. This is a matter of some concern to librarians and archivists who may wish to simply preserve, or 'back up', large amounts of literature which will be little read. It can be held in computer memory for an infinitesimal energy cost. Well done the New York Public library and Oxford's Bodleian for using Google Book Search to archive books which will cost much more to move from the stacks than is spent on their digital archive. It is also very relevant that the ecological cost of printed books and magazines come up front: in the making of paper, the manufacturing of books/issues, significant numbers of which are 'returned' through the distribution channel and all of which may be bought but not read.

The carbon footprint begins to mount if the digital book is used. So let us assume that digital books are used in a service which has high throughput and which will deliver pages to customers at a price which will not be greater than the Amazon s3 service. Since Amazon is already delivering such a digital service with the Kindle, any seriously competitive digital publishing system will need to use a cost base with comparable or lower charges. The published tariffs of the s3 service tell us that a digital publisher should not really be paying more than 17c per GB for delivering content. A large digital distributor (eg Amazon itself) will obviously be paying a lot less than 17c, and the s3 scale goes down to 10c per GB for users who take up more than 150 TB a month. Now we can make a heroic guesstimate of the cost per digital book delivered. We need two more parameters:

1. How many books do we get for a Gigabyte of delivered content?
   
2. What percentage of the cost is attributable to electricity or to atmospheric pollution?

With a very rough back-of-the-envelope this looks like $100 per million pages delivered. If we say a typical book is read once page by page, for 200 pages, we have approximately 50 digital book-readings per $1 of server cost. Now how much of that $1 is attributable directly to energy costs, and how much to Amazon's margins the investment in computers, land etc? This envelope is getting rougher all the time, and I dont have the patience to plough through the details of Amazon's annual reports to find hard data on their energy expenditure, but if we say at most 50% of Amazon's web service costs are energy costs, it looks as though the $1 of server cost is translating to 50c in electricity, or about 1c per book-reading (10/20 book readings per kWh).

1 c or 1 penny is still a cost, but its not a big deal. Distribution costs do not disappear from the equation when we go digital, but they do almost vanish. Digital books cost almost nothing in the distribution chain and they have a much smaller environmental footprint.

What does a conventional book cost in energy? What is the carbon footprint of a typical book or magazine? According to David Reay quoted from the THES, a typical book costs 4.5 kWh or 3 kg of carbon dioxide:

What with production and transport, the average paperback has eaten its way through 4.5kWh of energy by the time it gets to a reader. In terms of climate impact, this is equivalent to about 3kg of carbon dioxide emissions for every glossy new textbook. So, for a print run of 10,000, there is a cost of 30 tonnes of carbon dioxide not mentioned on the dust jackets.

Digital books and magazines are at least two orders of magnitude more efficient than the print equivalents. These calculations may be back of the envelope, but they point to the urgent need to move to a more sustainable distribution system for the health of our planet and the long-term benefit of book and magazine publishing.