That source is primarily examining conventional feedlot approaches.
> An all grass cow-calf – to – finish operation was included as a minor component in the eastern and northwestern regions
They examined a single all grass finished operation. There are studies that have found grass finishing can cut emissions in cattle by as much as 80%.
> The modeled operations were not intended to be actual operations; they were developed to represent the practices found in each region.
> Environmental footprints for all individually simulated ranch and feedlot operations were integrated into full production systems within their respective study regions using two methods
They also don't appear to be measuring actual operations, but rather modeling operations based on surveys of farmers and ranchers about the characteristics of their operations and then extrapolating from there.
Models have their place for sure, but I wouldn't make any kind of declaration of certainty based on a single model-based study. You have to average the outputs of hundreds or thousands of models, and even then, you can't be sure you have the answer.
A better approach would be one that measures the actual output of each operation at various phases and averages across them. Difficult to do, but I've seen studies that attempted it.
Gladly, although, with the caveat that science works in the aggregate so any individual paper should be treated with skepticism and that I have no idea what quality of journals these are coming from because the open access literature is currently a royal fucking mess.
Here's another study that found that the soil type of the grassland and the dominant species of grass had a significant effect on whether intensively grazing it sequestered or released carbon from the soil overall:
And that one appears to be a literature review of other studies.
Related note, you have to be so careful looking through the studies, here's one that purports to compare the different systems. But when you read the abstract it uses "A deterministic model based on the metabolism and nutrient requirements of the beef population". In other words, they aren't actually measuring anything. They just have numbers in a database for "x lb of beef requires y inputs" and "x lb of beef gives z outputs" and they're crunching those numbers under assumptions about the productivity of each approach. Naturally, they find that conventional feedlots are the most environmentally friendly.
I mean, the paper uses data from the National Cattlemen's Beef Association™, which is a beef lobbying group. The third author of the paper is also from the NBCA, and the paper was funded by the Beef Checkoff™, which a beef marketing program. I have my reservations about the accuracy of this paper.
