An article on The Guardian mentions a flaw that has been raised with the book's methodology: that the meat used for pet food is not grown for the purpose and would still exist even if pets didn't.
"Far from being unsustainable, pet-food manufacturing uses material from animals which are inspected by vets as fit for human consumption but which are surplus to the requirements of the human food industry," says Michael Bellingham, the chief executive of the Pet Food Manufacturers' Association. "These byproducts must meet the very high safety and quality criteria laid down in European legislation. Without us adding value annually to around 630,000 tonnes of animal byproducts in the UK, it might otherwise have to be disposed of via landfill or incineration. Not very green. Furthermore, a recent report by the Waste and Resources Action Programme [more commonly known as Wrap] is rightly damning of the enormous amounts of food – around 30% – that goes to waste each year. Compare that with the 1% of pet food they found went to disposal."
I haven't yet been able to find whether the authors included this in their calculations or whether they used the same footprint per kilogram of meat as was used for human consumption.
If the claim about meat-based pet food being primarily made up of off-cuts that would be otherwise burned is true, and the authors didn't include this in their calculations, the carbon footprint of meat-eating pets would be significantly smaller than the book claims.
Phys.org also expresses some skepticism:
Clark Williams-Derry, chief researcher at the Sightline Institute, a nonprofit sustainability think-tank in Seattle, scoffed at the study, which is how scientists express disdain.
"When I saw the study I ran some quick numbers," Williams-Derry said. "The average dog has to eat at least twice as much as the average person for this to be right. People are just heavier than dogs so, I just had to scratch my head at that.
The Car Connection points out that the book authors have used a yearly travel distance that is around half the U.S national average:
then by assuming that the vehicle will only be driven 10,000 km (about 6,200 miles) annually. According to the Department of Energy, the true U.S. average is more than twice that, at 13,700 miles.
Amongst other errors such as severely underestimating the value for the carbon footprint of the car's manufacture. They also suggest that the authors have not included "indirect impact like roads, bridges, and fuel infrastructure."
An article on Grist compounds up all these errors and adds a few more to calculate that the book authors are off by a factor of 18.
combining the underestimates of SUV impacts with the overestimates of dog food impacts — the anti-doggites are off by a factor of at least 18, and probably more.