>>and those LN could be responsible for an inordinant amount of volume
Theoretically yes, in practice no.
Routing complexity increases as usage increases without a commensurate increase in channel collateral, because increased usage leads to more channels using up the collateral for one direction of transmissions, which eliminates the transmissibility of one of the two routes in the channel. The network would thus naturally see rising collateral - either in existing channels or in new ones - to maintain routing efficiency, as usage increases.
Beyond that, more individuals joining the LN would naturally lead to more collateral being added to it, as individuals create new channels with new stashes of respective collateral.
For these reasons, low collateral reliably maps to low usage.
1.) I have a direct disincentive to include more collateral than would be absolutely required.
2.) Most of the time LN's only need to go one way.
3.) The only thing your original graph shows is that the price of bitcoin has fallen.
4.) I anecedotely know there has been more use in LN among all my crypto friends and in South/ central America where i have seen increased use in Cafe's and hotels.
5.) The total amount of BTC being used as collateral in LN has grown. (Even though this means absolutely nothing and your graph of 'TVL' Also means nothing when compared is USD).
6.) It's annoying to see someone spitting FUD on BTC/ LN with ETH in their name.
Theoretically yes, in practice no.
Routing complexity increases as usage increases without a commensurate increase in channel collateral, because increased usage leads to more channels using up the collateral for one direction of transmissions, which eliminates the transmissibility of one of the two routes in the channel. The network would thus naturally see rising collateral - either in existing channels or in new ones - to maintain routing efficiency, as usage increases.
Beyond that, more individuals joining the LN would naturally lead to more collateral being added to it, as individuals create new channels with new stashes of respective collateral.
For these reasons, low collateral reliably maps to low usage.