Way to go back and forth between one staff and two?

Is there a way to have a single staff split into two staffs, and then later switch back again? (I want this as a workaround for the inability to specify different dynamics for different voices on the same staff. For example, I'm working from a pocket score of Mozart's Serenade #10 KV361; each staff has two voices on it, e.g., Oboe 1 & 2 are on the same staff, Clarinet 1 & 2 are on a common staff, etc., (except for the double bass). For ease of checking my input against the hard copy, and because to see the entire score (vertically) with each part on its own staff, I have to zoom so far out that it becomes hard to read (that's the way I started), I'd prefer to enter it into MuseScore as I have it printed: one staff per instrument, two voices per staff. However, I've quickly (measure 2) run into the problem of not being able to independently control the volume of each voice (e.g., Clar. 1 is supposed to play its first phrase piano, and Clar. 2 is supposed to come in at forte, playing something entirely different, mid-way through Clar. 1's phrase; I've verified by muting all the other parts that the Clar. 1 phrase does switch to forte at the _forte_mark attached to Clar. 2's entry note). I've read the "workaround" of manually adjusting the velocity of individual voices, but it's unclear from what I've read if this manual change persists until a new dynamic or a new customization is entered, or if it only effects individual notes, so that it needs to be applied to every note in the "deviant" phrase. In either case, this strikes me as uninvitingly laborious, so I've thought of another workaround, at least in theory: temporarily split the one staff into two staves--since dynamics' effects are "by the staff"--for the duration of the two voices having different dynamics, then re-consolidate down to one staff while they share dynamics, and go back and forth as necessary. As, most of the time, not all staves require different voice dynamics at once, this would get around the different voices different dynamics problem, while still keeping most of the score "condensed." So, is this presently doable?)