Regarding the finances - projects funded out of peoples' pockets and small funding sources won't be able to build the sort of serious five to six nines reliability networks with significant N+1 redundancy that will survive a major disaster. The amount of money that has been thrown around via the US DHS to local agencies in the past 14 years is immense. Similarly, large ISPs and major telecoms and their infrastructure providers (major IX points, datacenters, colos) have put a metric shitload of money into being able to survive a major disaster. Look at the engineering of the NAP of the Americas in Miami, for instance. In a major disaster I know of at least a dozen ISPs that might lose 75% of their network, towards the edges, as sites run out of battery and generator and fall of the network, but the remaining 25% of the network including their core will remain online for multi-week periods with fun things like 50,000 liter diesel tanks and N+1 2 megawatt generators. The telecom industry has a long tradition of seriously overbuilding stuff to survive disasters.
In a serious disaster, small transportable first responder command posts and similar need access to services that can only be delivered over reasonably-close-to-broadband IP services: GIS mapping, sending/receiving photos and schematics, logistics manifests, VoIP with 0.0% packet loss, medical data and a myriad of other things that you can't do by half duplex voice alone.