I understand that High litter accumulation and N-fixation in dense bushland are obvious causes of low C:N ratio but why would the grassland have high soil organic carbon than bushlands?
The C:N ratios in grassland and savanna are within the range of expected values for terrestrial ecosystems. The bushland value is 'abnormal'. One explanation could be a lack of cellulose in the litter. Also, slow growth associated with low N-use efficiency leads to a low C:N ratio :)
See: The Logic of Carbon and Nitrogen Interactions in Terrestrial Ecosystems, M. Tateno and F. S. Chapin III, The American Naturalist, Vol. 149, No. 4 (Apr., 1997), pp. 723–744
Grassland have maximum carbon stock as compare to other ecosystems present on the earth. Pure grasslands have more C:N ratio as compare to mixed grasslands. That's why savanna has little bit lower C:N ratio as compare to pure grasslands. Because Savanna has scattered tree which affects grassland productivity.
However, some studies have found higher OC in bushlands vs grasslands and they justify that higher quality litter from leguminous woody plants vs grasses allows more decomposition of organic material. However, on my view I am also certain that although Acacia karoo is leguminous, its decomposition rate can be low mainly due to presence of secondary compounds that weaken the microbial activities.