I'm not a rocket surgeon, but it seems like in that situation they could have just stationed several people within radio range of each other in the tunnel so as to relay messages to the surface.

Depends on how the radios are set and used. Most modern radio system has a direct and trunk modes. But can't do both at the same time. Trunk modes allow channel control and stacking message by a centralised control. While direct mode is local only.

Radio repeaters are mandatory in large public buildings and special object (like tunnels) in a lot of countries to prevent this issue.

It's not new and the solutions is well known. Just having the will and resources to implement it varies. Not sure what the local DC law requires...

Tjin is right. The article linked by bws48 says it was a digtital trunked radio system. The simple minded way to think of it is that the signal goes as digital packets from one radio to a computer (usually through a repeater), then back to the other radio. While there are many advantages of this sort of system, one big disadvantage is that both radios need to be able to "see" the computer. If either one can't, perhaps by being down in a tunnel, then they can't talk to each other. I don't know how difficult it would be to have enough repeaters in the subway tunnel system to make it work in the tunnels, but I suspect it would be expensive. I'm not a radio expert so there are probably other work arounds to make it function in the subways.

There are however many advantages of digital trunked radio systems, which is why it is becoming the standard for public safety agencies. The biggest advantage is to allow many users to operate on a limited number of discreet frequencies. This is very advantageous in a big city, where there are many, many hundreds of fire, police, and EMS responders all working at one time. Also the dispatcher can very quickly set up a "channel" specific to an incident. All the reponders to that incident (fire, police, EMS) can talk to each other, but not have interference from nor interfere with other communications in the surrounding area. The signals can be easily encrypted, which is particularly important for law enforcement.

These systems are still relatively new for many agencies, so one of the issues is symply finding out where they do and don't work. To some extent you can model where the repeaters will cover, but there are almost always some unexpected holes in the coverage.

So in this case, it sounds like a few cheap walkie talkie's would have been better than the expensive digital-trunk radios.

IIRC the Metro system has comms with the trains. Maybe the regional (DC/VA/MD) FD's could develop a comm system that would leverage off the Metro comms while they are in the tunnel. The FD's could go into "Metro mode" and use that system to transition into their trunk. Probably more expensive than walkie-talkies or runners and it would require a bit of commonality -- can't have that...

Not necessarily. The digital trunk radios I've worked with can be switched into simple VHF simplex mode. In that case they work just like any ordinary VHF. The firemen would first have to realize that they weren't hitting a digital repeater, then have to switch over. It still might not work all that well in the tunnels. VHF is (more or less) line of sight. In a relatively straight tunnel it would probably work fine. If you were trying to talk to someone around a bend, it probably would not work very well, if at all. Even if you could talk using VHF simplex through the tunnel, you then would not be able to talk to the rest of the reponse team on the surface, who would presumably be working in digital trunk mode. (As I noted up thread trunk mode has many advantages for a big incident.) You might need to have a human "repeater" at the tunnel entrance, with two radios (one VHF simplex and one digital trunk) to relay information.
This is probably what is really required here.

Communication is always a huge issue in any incident. What little I know about radios has been learned the hard way on land SAR operations. Digital trunk systems have many advantages in normal (above ground) operations for a city. But special situations may require special solutions. I'm not sure the fire dept is totally at fault here. It sounds like there is a serious issue of no agency having safety oversight over Metro, with the ability to enforce changes. Given that emergencies have occured in the subways in the past, and will no doubt occur again, planning for robust emergency communication (and testing those plans) seems like a no brainer. METRO should have worked this issue before.

It would be very interesting to learn how other cities with extensive subway systems handle this problem. (New York? London?) I have to believe somebody has found solutions that work.

An Extech DV50 or something like it with a long fiberglass rod probably would do it. The DV50 has a telescoping 39" fiberglass pole.

The third rail doesn't constantly arc over to the cars, outer rails & every other conductor nearby so there's a limit to the voltage there.

... though I can't remember what AC frequencies the DV50 can handle and that might be a problem.

You're going to EDC a long fiberglass rod to go with your AC voltage detector?

Another small problem with the AC detector might be caused by the fact that the third rail is powered by DC.

But on a more serious note, the initial report has been released and the power to the third rail was not cut:

http://wtop.com/sprawl-crawl/2015/01/ntsb-releases-initial-findings-fatal-metro-smoke-incident/

(includes photo of the damage done by "the arcing incident")