Add hazard rate outputs

[patrickbarks]

A recent paper in MEE by Torbjørn Ergon et al argues for a greater focus on hazard rates rather than survival probabilities in population analyses.

I think it would be straightforward to add hazard rate outputs to Rage fns vitalRates, vitalRatePerturbations, and makeLifeTable.

If S is a stage-specific survival probability, the corresponding time-averaged hazard is

h = -log(S)

and the sensitivity of S with respect to h is

dS_dh = -exp(-h)