We may need to revisit the assumptions around compounding vs linear module degradation in the PV models. Tim Silverman and Dirk Jordan have suggested that even though compounding is the actual physical mechanism, linear may be a better representation of the findings in the data.