How should we build Melbourne Metro 2?

Diagram of proposed MM2 amongst Melbourne's network (via IV)

Recommendation 61 of Infrastructure Victoria's 30-year strategy is to "Prepare for Melbourne Metro Two and direct Geelong rail services". They're more positive about the project than last time, and it seems the potential for Geelong services to use it has been one of the key things to influence this. But they're less keen on the northern part of the project, recommending it be built in two stages and that the route from Southern Cross to Clifton Hill be reconsidered. Let's take a closer look at this. 

Firstly, some background. For those who haven't come across it, MM2 is the proposed successor to the Melbourne Metro 1 tunnel the government is currently building. Just as MM1 will link the existing Cranbourne/Dandenong and Sunbury lines on opposite sides of the city, adding capacity and taking pressure off the City Loop, MM2 would do basically the same thing for the Werribee and Mernda lines. As currently conceived, it would be built as one long bored tunnel, much the same as MM1 has been built.

This would free up space between Newport, Footscray and the city for Altona and Williamstown trains, and free up space from Clifton Hill through the City Loop for Hurstbridge trains - so not only would the lines that would plug into MM2 get to run more frequently, but the lines they used to share tracks with also get to run more frequently. 

The other big advantages of MM2 are added coverage, and improved directness. New stations would be built in Fishermans Bend and likely somewhere in the Fitzroy/Carlton area, where no stations currently exist, adding coverage. Trains from Werribee (and potentially future electrified trains from Geelong) would also get a quicker and more direct path from Newport to the city, compared with the fairly long and circuitous path via Footscray they take today. 

The Werribee and Clifton Hill lines will be over capacity by 2036 (via IV)

When it comes to the Newport-Southern Cross section, IV are very much on board with this vision. The modelling used by IV recognises that the Werribee line via Footscray, and the inner sections of the Mernda and Hurstbridge lines, will be over capacity by 2036 if nothing is done, and they recognise that Fishermans Bend will be such a big and dense development that a tram alone won't cope without heavy rail to support it. The potential to use it for Geelong services seems to have sealed the deal for them - it'd provide a nice clean run into Southern Cross for them, both in terms of being straighter and quicker but also less congested (one of the big issues approaching Southern Cross from the north). 

When it comes to the Southern Cross-Clifton Hill section, though, they're clearly much less keen. They note the Fitzroy/Carlton area is already very well-served by trams, so the uplift in PT access is proportionally less than in a PT desert like Fishermans Bend. So they want to cut costs in order to get better value for money, and suggest "an alternative alignment may be a more direct and less costly option, such as along the former Inner Circle railway line". 

The Inner Circle (purple) would be less direct than a tunnel (via OpenStreetMap)

This is so ridiculous I had to re-read it several times to make sure I hadn't misunderstood. The first problem is that any alignment that used the Inner Circle would be significantly less direct. If the point of using the Inner Circle is to avoid/minimise tunnelling then presumably the idea would be to use the existing Upfield line tracks between Southern Cross and Royal Park, like the original Inner Circle - which is way less direct than a tunnel. 

The second problem is, it's probably not feasible because those Upfield line tracks will also be very congested (which is one of the things their Recommendation 60 is designed to address). And if they're not intending to use the Upfield line, it would take substantial tunnelling to get from the CBD up to the Inner Circle - basically as much as just tunnelling directly to Clifton Hill would. 

The third problem is that this makes a connection with the Hurstbridge line at Clifton Hill station basically impossible. Even with a tunnel approaching from the south this is a substantial challenge - and the long-term desire to straighten out the notorious Rushall curves contributes to this - but if you're coming along the Inner Circle, how can you possibly connect at Clifton Hill and also head north to Merri? Maybe they don't plan to connect with Clifton Hill - but then why do all their diagrams show a connection? 

Like I said - completely ridiculous. How a thought-bubble like this got through the drafting process of a major government body I will never know. 

A full tunnel is pretty clearly the best option (via IV)

The other thing I think is worth mentioning is that trying to find a non-bored solution to the northern section, and splitting the project into two phases, are symptoms of Infrastructure Victoria being run by economists rather than infrastructure experts. The demand projects their modelling is based on does indicate that the southern section is required slightly earlier than the northern section, and to an economist this says "Well it'll be maximally efficient if we split it in two, then!" But to a project manager, splitting it in two means you lose a ton of efficiency in winding one phase down and winding the other up - way more efficiency than you'd ever gain from delaying the second half. 

This is particularly true if they're looking for a non-bored method of constructing the northern section - yes, bored tunnels can be more expensive than other construction methods, but bored tunnels also benefit massively from economies of scale. Long tunnels tend to cost less per-km than short ones - if we procure a TBM and get it to bore from Newport to Southern Cross, stopping and changing to a different construction method for the second half isn't likely to be cheaper than just continuing the TBM to Clifton Hill. Especially if it literally just does keep going straight away, rather than pausing for a few years between phases. 

As with Melton electrification, though, I'm pretty confident the people in the DOT will recognise this and word the government up. Remember that MM1 was also originally proposed to be built in two phases, and the government ultimately decided to just do it all in one go for efficiency reasons - I'd be pretty surprised if the same didn't happen here. In any case, let's hope they do follow the recommendation and commit to funding a business case ASAP, so we can sort through these questions and get cracking. 

More discussion on IV's 30-year strategy here