Hi Scott. You are correct that there’s some oddities to the accounting for replacement cycles. However, extending the life span may only increase the uncertainties. The replacements in module b are tracked as a replacement of the same material with the same impacts. It’s hard to believe that future materials will have the same impacts, or even be the same materials, as the ones we use today. The further out we push our predictions, the less likely it is that our assumptions on material replacement will be accurate.
As for the end of life assumption - if a building owner is replacing a 50 year roofing material, but assuming that the building will only have 10 more years of life, they may opt to replace the longer lived material with a shorter cycle material, or they may find that extending the life of the roof may extend the life of the building. Regardless of what material they choose, that material will incur an embodied carbon cost, so while it might not be entirely accurate, it’s not unfair to apply the full carbon emission impacts to the selected material.
That isn’t to say that we can’t find guidance from looking out into the future at replacement cycles. We can use these studies with a bit of outside judgement to weigh the outcomes. For instance, we may choose a higher impact material only needs to be replaced once vs a lower impact material that needs to be replaced several times because there are additional benefits. You’re reducing raw material needs, you’re reducing the burden of repair for the building owner, you’re pushing the new carbon emissions for replacement out into the future (future carbon emissions should be weighed less than current emissions). The LCA results are a guide to help test assumptions, but shouldn’t be absolute.