A team of atmospheric scientists, led by the National Oceanic and Atmospheric Association, issued a report this week that presented initial results of an examination into the extreme 2012 US drought. Its core finding was the drought likely resulted mostly from natural variability. Any climate change signal is relatively small but likely made conditions across the Midwest US a little dryer and a little warmer than they otherwise would have been absent climate change.
The 2012 drought did not grow out of the 2010-2011 Southern drought that impacted Texas and Oklahoma, as many, including myself, theorized as the drought developed. Instead, a stubborn ridge of high pressure took hold over the Plains, which cut off the vital Gulf of Mexico water supply upon which the region depends for agriculture.
This sentence, in the Executive Summary, is key: “The interpretation is of an event resulting largely from internal atmospheric variability having limited long lead predictability.” Many people think severe weather events should be easy to forecast, but the opposite is true. The rarer the event, the more difficult it is to accurately forecast with any kind of time difference. Additionally, the connection to low-frequency climate oscillations (i.e., La Niña: “the 2012 drought occurred in concert with an appreciably warmer ocean in most basins than was the case for any prior historical drought”) were minimal in the 2012 drought, contrary to what I have theorized. That’s the beauty of science, of course. You can be incorrect about something and demonstrate as such when data are analyzed.
Recently, some folks have characterized this event as a “flash drought”, owing to the sudden onset of such an event, as the first graphic below shows. The term obviously borrows from the better known “flash flood” concept. Unlike a flood however, droughts have longer-term impacts on human and ecosystems. Costs are still only estimated at this time (because the drought is ongoing) at $12 billion. While significant, the 1980 drought event that caused 56 billion (2012$) and the 1988 drought that caused 78 billion (2012$) of damages eclipsed the 2012 event (so far). The $12 billion figure is likely to grow as the drought impacts water supply reductions and livestock. The 2012 crop yield deficit was the greatest since 1866.
Figure 1 – U.S. Drought Monitor maps showing the evolution of the 2012 “flash drought” across the US Great Plains. Little evidence existed in November 2011 or even May 2012 that the drought would achieve the extent and intensity that it did.
The drought was the worst on record for WY, CO, NE, KS, MO, and IA, as the following graphic shows. The region experienced a 53% rainfall deficit (39.3mm vs. 73.5mm) in 2012. 1934 held the previous record of -28.4mm deficit. The 2012 deficit corresponds to a 2.7 standardized deficit, which approaches a 1-in-100 event. This relates well to the precipitation time series in the graph below.
Figure 2 – Precipitation and temperature departures from normal for the six states impacted by the 2012 drought. Note the extreme minimum in precipitation on the right side of the top graph. 2012 temperatures as a whole were not as extreme as those recorded twice during the 1930s, but July 2012 still ranks as the warmest month on record for the six states as well as the entire US.
The analysis also suggests that we should not expect similar 2013 precipitation anomalies on the basis of 2012 anomalies alone (based on the report’s Figures 10 and 11). Put another way, just because 2012 was drier than normal, 2013 shouldn’t automatically be drier also. Dry epochs occurred in this region before: in the 1930s and 1950s. Subsequent dry years occurred then due to longer-term changes in natural variability as well as land use practices. The currently is no indication that the 2010s will similarly be a dry epoch. As with the 2012 drought, such a prediction remains beyond current skill.
The diagnosed linkage to low-frequency forcing is interesting. Warm tropical sea-surface temperatures (SSTs) in the Indo-West Pacific Oceans and cold east Pacific conditions tend to dry the mid-latitudes in the winter/spring season and not the summer season. As the first graphic demonstrates, the 2012 drought flashed in the summer and not the winter. So despite primed conditions for drying in winter 2011-12, the Great Plains drought occurred for different reasons.
Of further interest to the future is the following graphs. The researchers generated a 20-member NCAR CAM-4 ensemble with monthly varying SSTs, sea ice, and specified external radiative forcings consisting of greenhouse gases (e.g. CO2, CH4, NO2, O3, CFCs), aerosols, solar, and volcanic aerosols via observations through 2005 and then an emission scenario thereafter (RCP6.0, a moderate emissions scenario pathway developed for the upcoming IPCC’s AR5).
Figure 3 – Model results of the 1996-2012 precipitation minus the 1979-1995 precipitation.
The NCAR CAM4 model might be representing the actual climate well for this time period. Left unsaid in the report is any analysis of the model’s future projections. Other model studies suggest that the central US could experience 2012-type temperature and precipitation conditions more regularly by the end of the 21st century.
Figure 4 – Model probability density functions of precipitation deficits for the six study states.
This figure suggests that the latter half of the time period (1996-2012) modeled had a higher probability of being drier than did the former half (1979-1995). The report did not present a potential cause for this shift in probability. If this probability does not revert back to the 1979-1995 distribution, dry conditions could become a more regular feature of future years.
Figure 5 – Model probability density functions of precipitation surpluses for the six study states.
This figure is not the logical companion to the previous figure. The probability of being wetter and drier could increase if the overall probability density function existed in a certain way. This is not the case however. Instead, the probability of the six states experiencing wetter conditions in the second half of the period studied decreased with respect to the first half.
This report is useful in diagnosing what happened prior to and during the 2012 US drought and in trying to ascertain how predictable such an event might have been. There is considerable interest in accurately predicting this type of event well in advance so as to prepare those who might be affected. This capability remains beyond us for now since this event was primarily driven by natural variability enhanced slightly by underlying change. With climate model projection studies indicating a much warmer and somewhat drier future for this region, stakeholders will likely have to adapt farming and ranching practices. Similarly, municipalities will have to prepare for extremely dry years in their infrastructure planning and practices. Of course, future change could be reduced as a result of our efforts to mitigate anthropogenic forcing. The scale of that endeavor is much larger than most people are aware and thus not likely to take place any time soon. Climate and energy policies need significant revamping at all levels.