The paper, which looks quite good, amounts to (stationary baseline + climate drift) explaining burning more than (stationary baseline) alone.
From the paper...
> Their respective uncertainty bands overlap for most of the time periods considered; however, starting in 2018, there is a complete separation between CMIP6-ALL and CMIP6-NAT scaling factor distributions suggesting a complete dominance of anthropogenic over natural forcing.
...or incorrect models or uncertainty bands. The paper looks good but like it presumed to find the desired results from the research outset.
Beware anything done with simulations that provides 3 significant digits (172% increase!). I don't doubt the influence of climate but I do doubt the certainty of these authors.
Going back to the GP, I still want to know how wet years, dry years, and winds are outside of reasonable expectations in SoCal. These are part of the (stationary baseline) the paper should include. From a quick pass on the paper, it's not clear to me that the paper captures hysteresis due to intermittent wet/dry periods. Or that they considered any additive non-stationarity, in (stationary baseline + climate drift), they wouldn't deem anthropogenic.
Not denying here. Just enjoying throwing rocks at a paper adjacent to a past life. I have published on uncertainty quantification for computational fluid dynamics that employed explicit slow/fast timescale decompositions.