Scientists predicted a decade ago that Arctic ice loss would bring on worse western droughts. Arctic ice loss has been much faster than the researchers — and indeed all climate modelers — expected (see “CryoSat-2 Confirms Sea Ice Volume Has Collapsed”).
It just so happens that the western U.S. is in the grip of a brutal, record-breaking drought. Is this just an amazing coincidence — or were the scientists right and what would that mean for the future? I ask the authors.
Here is the latest drought monitor:
And that drought monitor predates the record-smashing heat wave now gripping the West.
Back in 2004, Lisa Sloan, professor of Earth sciences at UC Santa Cruz, and her graduate student Jacob Sewall published an article in Geophysical Research Letters, “Disappearing Arctic sea ice reduces available water in the American west” (subs. req’d).
As the news release at the time explained, they “used powerful computers running a global climate model developed by the National Center for Atmospheric Research (NCAR) to simulate the effects of reduced Arctic sea ice.” And “their most striking finding was a significant reduction in rain and snowfall in the American West”:
Where the sea ice is reduced, heat transfer from the ocean warms the atmosphere, resulting in a rising column of relatively warm air. The shift in storm tracks over North America was linked to the formation of these columns of warmer air over areas of reduced sea ice in the Greenland Sea and a few other locations, Sewall said.
I contacted Sloan to ask her if she thought there was a connection between the staggering loss of Arctic sea ice and the brutal drought gripping the West, as her research predicted. She wrote (back in late March):
Yes, sadly, I think we were correct in our findings, and it will only be worse with Arctic sea ice diminishing quickly. California is currently in a drought (as I watch every day — our reservoirs are at about 50% capacity right now, and I fear for the coming fire season, owning a house that backs up to greenspace and forest).
She directed me to her ex-student, now Assistant Professor of Geology at Kutztown University of Pennsylvania because he had done some additional work.
Sewall wrote me:
I am attaching a more definitive study (multiple fully dynamic models with greenhouse gas forcing) on the topic from 2005. The end result is about the same as the original 2004 study, just nailed down better.
Comparing current changes (2011 summer ice and 2011/2012 winter precipitation season) to the 2004 paper:
(1) Ice concentrations in August 2011 weren’t too far off from the ‘future’ in the 2004 paper. The “future” in the 2004 paper was 2050, so it seems we are moving faster than predictions (which has been seen in multiple studies of Arctic sea ice). That is likely due to the relatively conservative greenhouse gas scenarios that were used for the earlier IPCC assessments and associated simulations. Potentially the forthcoming AR5 will have more accurate/realistic/extreme responses in Arctic ice.
(2) Observed precipitation seems to be lower than in the 2004 simulations (50–70% of ‘normal’ in the Sierras vs ~85–90% of normal in the simulations) based on snowfall data from 2011/2012.
(3) The pattern of wetter conditions to the north of California is as predicted in the 2004 paper, Washington State reporting 107–126% of ‘normal’ precipitation, Southern Alaska reporting 106–148% of ‘normal’ precipitation for 2011/2012.
I think the hypothesis from 2004 and 2005 is being borne out by current changes. The only real difference is that reality is moving faster then we though/hoped it would almost a decade ago.
The “more definitive” study is “Precipitation Shifts over Western North America as a Result of Declining Arctic Sea Ice Cover: The Coupled System Response” (available here).
That study found that “as future reductions in Arctic sea ice cover take place, there will be a substantial impact on water resources in western North America.”
I asked NCAR’s Kevin Trenberth to comment on these findings and he was concerned that using an artificially high CO2 level to get the models to explore what happens when Arctic sea ice collapses might conflate the CO2 effect with the ice loss. He also added “Variations from year to year are quite large and depend hugely on ENSO in the west of N America. You can not say whether they have come true at this point.”
I asked Sewall for a reply to those comments and he wrote:
Re. the point that Kevin Trenberth raises below and your e-mail just now:
I am quite confident that the changes are due to the decline in Arctic sea ice. The 2004 study did not alter CO2. The study was done with a prescribed decrease in Arctic sea ice cover (and a corresponding increase in local sea surface temperatures to reflect “non-freezing” conditions). The climate response presented in the 2004 study is, thus, a clean response due only to the imposed decline in Arctic sea ice cover.
In the 2005 study, I then moved to look at fully coupled models where the decline in Arctic sea ice cover was the result of warming temperatures, which were, in turn, the result of elevated CO2. This is presumably the same response we are now seeing.
Those coupled simulations showed the same response in storm tracks and western precipitation that we had found earlier in the 2004 study.
Kevin is correct that if I had only looked at the coupled simulations, it would be very difficult to determine if the changes in rainfall were due to the CO2 or to the Arctic ice reduction.
However, because the 2004 study was a clean sensitivity study, I can confidently attribute the exact same changes seen in the simulations I viewed in 2005 as being a direct result of the declining Arctic sea ice (and,thus, an indirect result of elevated CO2).
Kevin is also correct that variations from year to year are large and that the impact of Arctic ice on storm tracks varies significantly with ENSO state.
(1) In unpublished work that a student of mine did, we found that under strong El Nino conditions, Arctic ice concentration had less impact on storm tracks and precipitation in the west. Under more neutral (weak El Nino or weak La Nina) conditions, Arctic sea ice had a larger impact on storm tracks and precipitation in the west.
(2) Both the 2004 paper and the 2005 paper present results as 50 year averages. This, to some extent, takes care of the annual variability issue and suggests that sum total changes on climatic time scales will, indeed, result in dryer conditions in the west.
(3) While neither study employed ensembles, the 2005 study looks at seven different models (all with slightly different parameterizations, resolutions etc. so the effect is similar to that of a seven member ensemble) and the response, in spite of differences between the models, shows declining sea ice and declining precipitation in the American west with increases in precipitation from Oregon on northward.
If indeed this research is being confirmed, it suggests that on average — allowing for yearly variations due to ENSO — the West is going to become hotter and drier faster than people had expected.
NOTE: Top figure (Arctic ice) by Andy Lee Robinson.