Switch shock time series to run off L1 blocks

[SeanMcOwen]

Where it says:

NOTE: ideally, this should be mapped either to relative timesteps or L1 time rather than fixed timesteps

so that the scenarios are invariant to number