As a hydrologist, I am quite familiar with using past observations to assess risk, whether it be flood frequency analysis or runoff prediction, we use past data to provide "the most likely outcome" and define confidence limits about that estimate. It's what we do. But increasingly, we see trends in the data - it is getting warmer, especially in the oceans, we see longer, more damaging storms - possibly an effect of the weakening jet streams, etc. I note that this past winter's lingering incursion of the "polar vortex" was largely due to that weakening polar jet stream. It is deviations from normal tropical and polar jet streams that caused the distribution of observed precipitation during prior el Ninos. Hence, recent climatological trends suggest that our models are becoming less reliable and I would argue that we (hydrologists/climatologists) pepper our predictions of outcomes with a disclaimer like: To the extent that our models based on past observations are analogous to current observations...." I would also argue that ongoing changes in ocean currents and atmospheric wind patterns are underreported - those severe storms across mid-America this spring that lingered and lingered (including I suspect, those flooded highway photos above) are related to these trends and very likely are expressions of global climate change (just my opinion).