The only reason "Riverbank Subway Redux" would ever happen is if Greater Boston reached a self-consensus that Storrow Drive needs to be torn down between Kenmore and Copley on grounds that it's an induced demand trap and that severing the city from the Esplanade was a great mistake that needs rectifying.
Then and only then would this Blue extension be considered, because the terms of engagement for tearing down the parkway would likely have a required transit mitigation trade-in as part of the sell job. In that case:
-- Embankment Rd. between Beacon St. and MA 28/I-93 stays (but probably lane-dropped from 6 to 4 lanes), as that's a car artery that has natural (not induced) demand not replicated by the Mass Pike.
-- Soldiers Field Rd. stays between the Pike and the Kenmore exit because there's no direct Pike exit to Kenmore/BU/Comm Ave.
-- The eastbound carriageway of Storrow between Kenmore and Embankment along the Back St. retaining wall is given over to the Blue Charles-Kenmore extension ROW because Kenmore-Copley is the fully induced-demand portion of the parkway that the under-capacity Pike can easily absorb. The deep-cut EB ramps under Mass Ave. and the auto tunnel by the Copley exit also become Blue property.
-- Waive Pike intra-city Pike tolls (i.e. the westbound-only Allston exit goes free) as compensation so you can get between the current Allston tolls and I-93 without getting dinged in Allston. This is only fair, since you're consolidating what was a free parkway. Obviously all tolls starting Allston to points west remain stet.
-- The westbound carriageway of Storrow is retained as a speed-restricted 2-lane park access road with frequent at-grade crosswalks from each side street. Call it "Storrow Lane" or whatever, since you do need at least some sort of official vehicle access around the water. Have DCR wield the hammer at traffic restrictions so nobody ever thinks to use it as a shortcut, and close it off on weekends.
Now...with that settled, here's how you'd build Blue.
1. Red-Blue's study design has tail tracks spreading wide around the wedge-shaped Charles platform to avoid the pilings from the Red Line viaduct above. 2 ways to extend: (1) take the tail track that stubs out in front of the Charles Circle CVS (i.e. "inbound" to GC) and extend cut-and-cover under the Embankment Rd. offramp, then take the tail track way over on the other side that stubs out in front of the hotel in the ex- Charles St. jail and take a wider loop underneath Charles Circle to meet up. Or...(2) extend the CVS-side tail track as a 2-track tunnel under the offramp and convert Charles Under into 2 side platforms instead of a center wedge, relegating the opposite-side tail track to just MOW storage. You would choose (2) despite its extra disruptiveness at the station if the distended loop the hotel-side tail track makes in Option (1) creates some sort of schedule imbalance...but you'd hope that (1) works well enough because it's a lot cleaner and less intrusive to not have to modify Charles Under station at all. I doubt for only reaching Kenmore that the extra time chewed up by an outbound train making the wider sweep under Charles is consequential enough a schedule imbalancer to worry about, so (1) is probably A-OK.
2. Cut-and-cover shallow under Embankment Rd. eastbound to the auto tunnel portal. You're busting down Embankment to 4 lanes anyway because it now terminates at Beacon St. with significantly lower volumes, so space comes available for the tunnel by compacting the roadway layout.
3. Recycle the auto tunnel as a subway tunnel. This requires undercutting the concrete tunnel floor to increase vertical clearances for Blue Line cars, but since the several feet thick roadway base is not needed for tracks laying on top of a rock ballast base you can easily jackhammer up the floor to net a proper-clearances trackbed. Re-waterproof the tunnel. Plunk an intermediate station--"Esplanade"--where the tunnel goes wide for the Copley exit and repurpose most of the exit ramp incline as the concourse to the surface. This will be a heavy-use station being right by the Public Gardens and State Capitol, and being nearly 1700 ft. from the Arlington entrance on the other side of the Gardens.
4. From the west portal of the tunnel, take the current stone Back St. retaining wall and re-pour it as a concrete tunnel wall that then pokes a couple additional feet over the roof as a squat retaining wall. Dig down the Storrow EB roadpack at shallow depth such that you can frame the subway tunnel in a simple box against that Back St. retaining wall. The tunnel box would basically be two-thirds below the level of the current Storrow pavement, and one-third above the Storrow EB pavement but still a few feet below Back St. Top it off with gently sloping dirt and grass, sidewalks, and the side-street pedestrian and (limited) car access points from the Back Bay street grid to "Storrow Lane" on the WB carriageway. You would choose this kind of not-completely-subsurface box tunnel design (which is MUCH cheaper than cut-and-cover) because it creates a de facto seawall against Charles Basin flooding. The top of the subway tunnel poking slightly above current elevation of the Storrow pavement keeps the subway tunnel dry and free from the cost of having to install pricey active pumping, while the rebuild/reappropriation of the Back St. retaining wall and several-foot overtop of the subway roof provides much better flood protection for the whole of the Back Bay than the current porous stone retaining wall. This is an important 'get', and combining the wall + tunnel structures into one passive barrier brings bigtime economy-of-scale to the project. Think about how much doing that helps immediately rein in the rate-of-rise in flood insurance premiums for Back Bay brownstone owners, and it's easy to see where this project ends up paying back its up-front investment back into the economy way faster than most public works megaprojects of similar scope.
5. Use the EB deep cut underneath the Mass Ave. bridge to descend the tunnel to deeper depth for passing underneath the Muddy River, and use the wide area by those parallel ramps to put the next intermediate station: "Beacon/Mass Ave." (or whatever you want to geographically call it). Entrance is a not-too-far 1450 ft. from Hynes, but hits the #1/CT1 and is a much closer walk to the subway from MIT main campus than Central or Kendall despite the walk across the bridge. Just cap it off and pour dirt on top. It's 200 ft. in from the riverbank, so the dirt backfill underneath the Mass Ave. bridge on the former EB side is enough flood protection.
6. Start curving the tunnel so you're on a trajectory to trace a diagonal line from corner of Back St./Charlesgate E. to corner of Beacon St./Charlesgate W. Tunnel under the Muddy River by shivving a subterranean roof shield under the riverbed to prevent surface disruption. This starts the more expensive length of tunneling than the really shallow, semi-surface run + tunnel/cut recycling that got you from Embankment Rd. to Mass Ave.
7. Cut-and-cover deep under Beacon St., reaching depth 1 level below GL-Kenmore by Raleigh St. Settle any building mitigations on this block (Beacon is wide enough here that this really shouldn't be a big issue, but obviously any Old Boston structures are delicate for adjacent cut-and-cover work).
8. Put Kenmore Under following Beacon St. such that it's slicing diagonally under the Green level + Green loop at a tolerably narrow angle to limit the amount of required structural underpinning. One of the huge lessons from the failed Silver Line Phase III BRT tunnel build was how much structural underpinning was required at Chinatown: double-deck BRT platforms underneath a 110-year-old Orange station underneath an extremely narrow street intersection...and that design proved absolutely lethal to the project cost and engineering feasibility. You must minimize the footprint of those structural underpinnings to a tolerable minimum to have a realistic chance at handling these priciest parts of the project. Thankfully, if you keep to a simple 2-track island with generous-throughput egresses to the upper level (for a Sox game surge crowd) you've got a risk-minimizing trajectory for Kenmore Under that doesn't chew up too much $$$ having to underpin too broad a swath of the 1932 station upstairs. Put a side pump room down here to actively pump out both levels of the station in an emergency. The Blue extension doesn't incur much risk of the 'storm drain' effect because of the way you built it along the Charles Basin as de facto strengthened seawall, but the Green-D portal is still as big a risk as ever from the Muddy River and now incurs a much-elevated risk of any intrusion down the portal to Kenmore draining straight into new Kenmore Under first. If the sandbags or other fortified portal-blocking measures fail, you will need rigorous pumps at the station for the second line of defense now that you've added the a second level to the 'drain'.
9. Tail tracks stubbing out just beyond the footprint of Kenmore Upper upstairs. Would be easiest to keep it going straight under Beacon, stubbing out before the depth of the C/D tunnel makes for another underpinning job. Not a big deal any which way.
STOP here. Swallowing the D would be a waste of capacity because Brookline and Newton just don't chuck in enough ridership for 6-car HRT vehicles vs. 2-3 car LRV's. Going further would mean mounting a major Blue-specific flood protection project at the Fenway portal. And you will not be able to branch at Newton Highlands to convert the outer half of the Needham Line to rapid transit after escalating SW Corridor congestion forces a non-optional mode switch; those grade crossings on the Needham Branch won't work with HRT like they will just fine with LRT, and branching that far out of town will make headway management on Blue at-large too much more difficult. If you ever do figure out a westward trajectory in the next 100 years, the tail tracks are angled so you can shoot down to the D portal or turn off Beacon underneath the Pike to angle for some other trajectory into Allston. All bases covered, even though you have no idea today where or IF that will be.
You do amazing things simply providing Kenmore with HRT load relief and a fast 3-station shot from the Orange-Green-Red transfer stations downtown, because it frees up so much extra Green Line Central Subway capacity for tying together additional service patterns on the west end. For example, doing up the Urban Ring as real light rail pinging Lechmere-Kenmore on the Grand Junction with a short Comm Ave. subway extension to BU Bridge tying it in, and doing up the Allston-Harvard UR spur as another proper Green branch off that BU Bridge subway junction between the B and Ring mainline. Basically, Blue-Kenmore gives you the bandwidth to completely reimagine Green around various Ring-utilizing 'circuit' patterns and wider variety of branch frequencies by freeing up so much load on the Arlington-Kenmore trunk of the Central Subway. Those options, especially the UR-related ones, may collectively move more total riders than the Blue extension...but it was the Blue extension that served up the load-shifting that enables that whole universe of flexible new Green routing options.
-------------------------
Of course, ^^^^ALL^^^^ of this is contingent on Boston agreeing to a Storrow teardown, with transit trade-in as a mandatory condition of that parkway teardown. Under no other circumstance will you ever have the convergence of factors that pushes Blue past Charles near the top of the transit megaproject priority pile alongside completion of the Silver Line Phase III tunnel linking the Transitway with downtown/Back Bay (hopefully as light rail and not BRT this time), North-South Rail Link, the Grand Junction and/or Airport legs of the Urban Ring (again...hopefully as light rail), or certain long-coveted linear extensions of the HRT lines into the inner suburbs. The ROW acquisition isn't anywhere close to cost effective unless you can claim the Storrow EB carriageway and auto tunnel, do that shotgun marriage with beefed-up Charles Basin flood protection, and have the unique set of political deal-making that mandates "If we're taking away the road, it must be replaced by something that can move equivalent masses of people or we can't make Legislative sausage out of this." It's a very narrow and specific set of conditions that put this transit build on the table at all vs. never in any form...but if those narrow conditions re: Storrow trade-in are put in motion Blue-Kenmore will debut near the top of the priority pile out of political expediency.
It's not hard to imagine these conditions being met someday. It's just not going to be soon. We are not anywhere near ready as a region for making the Storrow Decision by the 2020's. We're still in the late stages of debate & acceptance about downsizing down less-impactful and clearer-cut induced demand roadways like the Forest Hills Casey Overpass (believe it or not, the opposition to tearing down that now torn-down bridge hasn't totally moved on), Bowker Overpass, McGrath Highway, or putting Morrissey Blvd. on an overdue lane diet...though all of those are tipping strongly towards downsize. Storrow's a much bigger debate than all of those put together. We might very well reach the comfort level someday of agreeing that it's expendable between Kenmore and Copley...but I doubt that reaches universal-enough consensus to act on before 2030. So whither Storrow and whither Blue as a trade-in is a fully legitimate debate...but not one that's going to be had by anyone in power seriously and out-in-the-open for at least another 15-20 years. And thus Blue past Charles is totally out-of-sight/out-of-mind as a possibility until then because the ROW and political opportunity don't exist until Storrow teardown hits critical mass.
In the meantime...lots and lots of work to get done just dragging Blue down the street from Bowdoin to Charles with the obvious Red-Blue build that should've been done 10 years ago and MUST be done to save Red and the Big Four transfer stations from choking on their own overload in years we can count on two hands.