If I do a postmortem at work and conclude that, say, kernel X.Y's page cache makes deadlocks under unusually high load more likely, that doesn't mean that kernel X.Y caused the outage. Something had to provide the high load in the first place. But the kernel was a contributing factor - we could have survived the unusually high load otherwise.
30% higher than a base of 0.01%? 30% higher than a base of 10% probability?
That makes a difference.
Especially if other factors have much higher contributory rates. Like "letting underbrush accumulate increases chances of severe events by 400% YoY".
And it's not a postmortem if you only discuss a single factor. It's an opinion/position piece.
It's interesting you use this example because this has been covered by the RFS - the period during which hazard reduction burns can be safely done to clear the underbrush is shrinking because the fire season is longer and winter is shorter. They can only do those burns when it's both safe and not raining, so a change in climate can indirectly impact the severity.