Question on PVPMC Sub Hourly Transposition Bias Poster

[shri]

Hello!

I noticed your PVPMC poster you presented on the bias between hourly and sub hourly datasets due to the Perez decomposition of the diffuse components. In your conclusion, you state that the diffuse component should not be corrected on the hourly simulation. Does this mean that the diffuse bias needs to be corrected sub hourly (but not hourly) while the direct bias would be corrected through the hourly simulation (but not sub hourly)?

Also if this bias exists, does that mean that the strategy of comparing hourly simulation results to decomposed and reconstructed one minute (or perhaps 5 or 15 min datasets) would be inaccurate due to the transposition bias from Perez and thus still under predict sub hourly clipping loss?