Call it highly unproven at this point, even accepting that she's talking about the tiny percent of beef in the US that's grass-fed. Typical grass-fed cattle aren't much better for greenhouse gas emissions than typical factory farmed. Niman has to resort to discussing techniques for reducing methane that are experimental or not widely adopted even among the small percent of cattle that's grass-fed.
She does have an interesting argument that one study shows a 19% increase in soil carbon sequestration from pasturing cattle instead of raising crops. My one-study rule applies though (one study's result isn't proof of anything, it might just show the need for lots of study to confirm the result). And I didn't know that turkeys could be grass-fed.
She left out the issue that most everyone does as well, that ranching provides an alternative land use to sprawl. She also left out the argument that lands that have been pastured for decades/centuries at the current rate of use are carbon neutral for that time period.
Finally, here in California the grass would have been originally grazed by elk and deer, while in the Midwest it would have been grazed by buffalo. It would be interesting to compare emissions on that basis.
Unrelated bonus blogging: Tim "Slob Hunter" Pawlenty may have gut-shot his reputation among the hunting community by abandoning a wounded deer in the field. This may well cause him more trouble in running for the Republican nomination than all the deaths from that bridge collapse in 2007.