I’ve not looked into the detail, but believe the land use change elements of the specification come from IPCC LULUCF inventory methodology.  Data from a review by Dawson & Smith (2006) suggests that conversion from tillage land to grassland can increase soil carbon by between 1.1 and 7 t CO2e/ha/yr which equates to 0.3 - 2 t/ha/yr carbon, but I must admit the 2t/ha/yr seems rather high. At the lower end this would therefore be about 6 t/ha over the 20 year period, which is similar to the 7 t/ha in the standard.