I like to consider "growth" and "maturity" as different things. If the public capacity is growing, it doesn't necessarily mean that this new liquidity is very useful. For example, a new node could open hundreds of 1BTC channels and never create inbound liquidity and never spend anything on LN.

This node's actions would increase network capacity, but it would actually be a net drain on network resources. Other nodes would attempt to use these channels for routing, but since there's no inbound liquidity on this node, every payment will fail.

Collectively, nodes and wallets could waste years of time trying to use this doomed node as a router.

On the other hand, an efficient routing node may have dozens of unused channels. They might decide to close these channels since they aren't earning anything. This action would decrease public capacity, but it makes the network more efficient/mature. Further, since channel liquidity flows both ways, a tiny network with high transaction volume is more efficient than a large network with low tx volume.

LN is an evolving network that has cycles of growth and cutting (kinda like a body builder). This is what makes the network as a whole valuable, thousands of economic agents using their local knowledge of payment flows to invest in more efficient ways to pay one another. It's messy, opaque, and chaotic but this is what makes LN immune to death by 1000 forks.

A forked LN network wouldn't have the same exchange-rate-adjusted value as the original LN. There is dense information packed into the network topography. Information about who wants to pay who, etc.

A forked network wouldn't have the same payment needs as the original LN. Some nodes wouldn't respect the new fork, etc. Forking LN would essentially lose all this built up efficiency (compressed information about payment flows) expressed in the channel topography and the forked network would have to start fresh and evolve its own effective channel topography.