Hi Bai:
ACSNOM is accumulated melt from the snowpack. You can see the definitions of all of the model output variables here:
Of course, this is just melt from the snowpack, so does not necessarily tell you where that water went (i.e., to infiltration then to ET or recharge vs. direct runoff). Unfortunately I don't know of an easy way to partition this water. With code changes you could potentially zero out this flux right after it leaves the snowpack, then do the same run with and without this flux and compare streamflow, but there would be other impacts (e.g., ET changes) than just on runoff.
Hopefully someone in the community has a better idea...
Thanks!
Aubrey