By: Eric Hunt, Julia Borman, Subodh Acharya, Jeffrey Amthor
The world has changed dramatically in the last century, but one constant has been the continued importance of agriculture to the U.S. economy. While on-farm jobs account for just about 1 percent of U.S. GDP and 1.3 percent of U.S. employment, farm output supports much larger food and textile sectors, which represent nearly 5 percent of U.S. GDP and 11 percent of the country’s employment.1 The value of U.S. crop production in 2018 was a robust $184 billion, with corn grown for grain topping the list at $52 billion.2
Despite increases in crop production efficiency due to advances in farming technology, the key driver of crop outcomes remains very much out of our control: weather. The year 2019 was the second wettest on record for the lower 48 states. Farmers were unable to plant roughly 19 million acres worth of crops, including 11.4 million acres for corn (13 percent of total U.S. corn acreage) and 4.5 million acres for soybeans (6 percent of total U.S. soy acreage).3 In 2020, the U.S. Department of Agriculture (USDA) estimated that a sudden and powerful derecho affected about 38 million cropped acres across the Midwest in a matter of hours.4
While recent years have featured stormy, wet weather that has damaged crops, we are considerably far removed from the last extended drought that posed major difficulties for U.S. farmers. This article will break down how flash droughts have previously wreaked havoc in the United States and use the AIR U.S. corn growth and yield model to simulate a contemporary flash drought event and project some of the disruptions that might ensue should one occur.
What are flash droughts?
Flash droughts are defined by a rapid onset and intensification of drought and are characterized by abnormally high temperatures, increased wind speeds, greater incoming solar radiation, and rapid depletion of soil moisture, leading to a marked decline in vegetation health.5 6 To understand the possible consequences of a flash drought today, it is worth considering past droughts for guidance.
Insurance sector
The insurance market for U.S. agriculture is dominated by the public-private partnership between approved insurance providers (private companies) and the USDA Risk Management Agency. Established in response to the 1930s dust bowl, the crop insurance program has evolved significantly, with large program overhauls in the mid-1990s introducing revenue-based products and expanding coverage.7
The U.S. Crop Insurance Program insures nearly 400 million crop acres and has averaged around $10 billion in premium annually across all crops over the past six years.8 Of that, corn accounts for about a third of the total program premium and about a quarter of the acreage insured (Figure 1). Most grain corn is protected under the product Revenue Protection, which sets a revenue guarantee for the producer that can be triggered by low production (yield), low prices during the harvest season, or a combination of the two.
U.S. MPCI (Multi-Peril Crop Insurance) and Corn specific Insured Acreage and Premium Totals. (Data from RMA Summary of Business)
Success of corn yields and revenue is vital to the agricultural economy in the United States and the crop insurance portfolio. In recent history, the 1993 flood, 2002 drought, and 2012 drought caused the greatest gross loss ratios (the ratio of claims paid to premium collected) in the program (Figure 2).
U.S. MPCI Program Gross Loss Ratios (Data from RMA Summary of Business)
What were some noteworthy 20th-century flash droughts?
The last drought that drastically reduced production simultaneously on corn, soybean, and wheat (CSW) was in 1988 when a flash drought developed over the central and eastern United States in late spring and reached its peak in August. The timing of this drought led to many locations experiencing significant soil moisture deficits during July and August, which is a critical time for all three crops' growth. Thus, during the period 1969-2018 (Figure 3), 1988 stands as the worst year off trend for CSW in many crop reporting districts in the north central United States. The AIR U.S. Multi-Peril Crop Insurance Model puts the 1988 event as a 1-in-27-year loss.
Crop reporting districts with the lowest yield (from trend) due to the 1988 drought over the period 1969-2018. Legend as follows: White (N, No crops with worst yield; Yellow (C, Corn); Light Green (S, Soybean); Dark Red (CS, Corn and Soybean); Orange (W, wheat); Brown (CW, Corn and Wheat). (Data from USDA National Agricultural Statistics Service).
The 1936 flash drought was unparalleled in recorded U.S. history
While the 1988 drought posed considerable difficulties, the severity of it pales in comparison with one just over 50 years prior. An exceptional flash drought that transpired during the spring and summer of 1936 led to extreme heat, large losses of human life, and significant reductions of crop production.9
An analysis of the Standardized Precipitation-Evapotranspiration Index (SPEI; Vicente-Serrano et al. (2010a)), which takes both precipitation and temperature into account, indicates that the flash drought originated over the southeastern United States in the spring of 1936 (Figure 4), spread northwest and merged with a secondary area of flash drought over the northern Plains. By early July, over 2 million square miles of the central and eastern North America were either in drought, or rapidly cascading toward drought. This immense area with an extremely dry land surface likely devolved into a feedback loop by potentially reinforcing an intense mid-to-upper level high pressure that further suppressed precipitation and allowed for prolonged stretches of extreme heat, first in the Midwest in July, and then further south in August.
Crop reporting districts where the median 1-month Standardized Precipitation Evapotranspiration Index (SPEI) was at or below -2.0 for the months of May, June, July, and August. An SPEI of -2.0 is considered very extreme drought. Data from https://spel.csic.es
The months of July and August 1936 are the hottest months on record in the United States, with numerous cities setting record highs for days on end. Thirteen U.S. states, ranging from Nebraska to New Jersey and North Dakota to Louisiana also set their record highs in those two months. At the spatial peak of the July heat wave, temperatures exceeded 100 degrees F everywhere from Montana's High Plains to Connecticut. This included highs above 104 degrees F and lows in the 80s in cities such as New York, Washington, D.C., Philadelphia, Detroit, Chicago, Minneapolis, and Omaha. By the middle of July, most of the Corn Belt had experienced at least 10 consecutive days of temperatures over 100 degrees F (Figure 5) and at least one day with temperatures over 110 degrees F. After a brief hiatus, the heat wave roared back in August over the south-central United States, bringing record high temperature, crop failures, and human suffering.
The consequences were severe. For a short period in July, the heat led to many northern U.S. cities sustaining death rates such that the Detroit Times reported that a “great city is dying of heat.” Newspapers in St. Paul and Toronto reported that hospitals were so overrun with heat stroke victims that they simply could not keep up. The relentless heat also sparked major forest fires across northern Minnesota, contributing to more lives lost. The crop failures were so profound in parts of the north central United States that over a third of corn and half of the spring wheat crop were not even harvested, contributing to farmers migrating out of the region.
Consecutive days over 100 degrees F in the summer of 1936.
Understanding impacts for the next flash drought
While the United States has been fortunate to have gone nearly a century without a repeat of this caliber of flash drought, warmer summers and increased precipitation variability due to climate change may make flash droughts, like the one in 1936, more likely in the coming decades; society should be aware of the potential impacts. As a contemporary example, previous research indicated that the 2010 flash drought and heat wave over Russia was 80 percent more likely in a warming climate.10 Therefore, public policy and economic sectors such as agriculture, health care, insurance, and utilities should contemplate a repeat scenario that, for the sake of this article, we are calling Drought X. While industry impacts might be far-ranging, agricultural output, in particular, could be significantly altered. To project the devastation that Drought X could trigger, we conducted a numerical experiment with the AIR U.S. corn growth and yield model.
The AIR U.S. corn growth and yield model simulates daily corn growth and development based on well-understood physiological responses of crops to rainfall, temperature, sunlight, and soil conditions. The model simulates the major effects of low and high temperatures on crop development, growth, and senescence. Temperature extremes can inhibit biomass accumulation and, in severe cases, may cause mortality. The model computes harvest index (the fraction of the corn plant that is kernels) based on developmental processes during the growing season. Final crop yield is the product of the harvest index and total crop biomass produced.
The model is parameterized for current (ca. 2019) crop genetics, farmer skill, and management practices to reflect current corn growth and yield potential. The yield prediction from the model was validated using the historic reported National Agricultural Statistical Service yields from 1974-2019 for all U.S. counties growing corn.
Drought X
Our Drought X scenario mimics the 1936 drought event by creating a scenario for rainfed corn in the U.S. Corn Belt region founded on 1988 observed temperature and precipitation reported by NOAA. Both precipitation and temperature adjustments are implemented at the agricultural district level.
Summer (June-August) total precipitation, which is critical to corn production, is adjusted downwards from 1988 levels. The smallest adjustments are in Iowa and western Illinois, which had near record low annual precipitation in 1988. Larger downward decreases are modeled in Nebraska, Minnesota, Wisconsin, and Indiana. Minimum and maximum temperatures are adjusted upwards to resemble the persistent and extreme heat waves experienced in the 1936 summer.
Model results indicate that, should a flash drought event comparable to 1936 occur, the loss of corn yield could be devastating in many areas. Applying the Drought X scenario to the AIR U.S. corn growth and yield model shows reductions from 1988 yields ranging between 7 percent and 80 percent with a median reduction of 43 percent. Figure 6 indicates the mean corn belt yield in 1988 was 110 bushels/acre, while under Drought X conditions that falls to 61 bushels/acre, a drop of about 45 percent.
Distribution of simulated yields for the study region under the Drought X scenario (blue) and 1988 observed weather (green). Dashed lines show average county yield. Under Drought X the average county yield is about 45 percent lower than 1988 yield.
The AIR U.S. corn model also indicates that the largest reductions would occur on the edges of the corn belt (Nebraska, Minnesota, and Indiana), with a notable decrease in northern Illinois as well (Figure 7).
(Left) County scale June to August rainfall percent reduction relative to 1988 in the Drought X weather dataset. Darker red indicates less rain in the Drought X weather pattern than in 1988 observed values; (Right) County scale modeled yield percent reduction under Drought X compared to 1988 modeled yields. Darker red indicates larger yield loss.
The implications of Drought X
Severe droughts that previously unfolded in the United States have rarely affected many different crops during the same event. However, Drought X’s impact would extend past corn and suppress yields of soybean in the Corn Belt, spring wheat and barley in the northern Plains, and cotton and peanuts in the southern states.
In the Drought X regions, insurance penetration is high. Under the current public-private partnership insurance setup in the United States, crop insurance market caps the losses that pass to insurance companies through the Standard Reinsurance Agreement, with the U.S. government paying the remainder of claims. Under the assumption of a low implied price volatility scenario, the U.S. Multi-Peril Crop Insurance model puts the Drought X event at a 1-in-29-year return period. This signals that the crop insurance program would function well under these circumstances, about as well as if 1988 outcomes occurred under the current program policies.
Nevertheless, there could be numerous impacts outside of the crop insurance sector. The extreme stress on pasture could contribute to a lack of forage for cattle and an increased corn price, which could lead to food rations for livestock. The combination of poor diet and extreme heat might lead to increased cattle mortality. We would also expect losses of hogs and chickens in buildings without means to mitigate the extreme heat. Ranchers and livestock producers would suffer financially, with increased likelihood of bankruptcy. The U.S. food supply chain might in turn suffer a significant reduction of meat and other raw materials that food processors rely on.
Suppose Drought X occurred in a season when the corn stocks-to-use ratio was already low and was exacerbated by a decline in corn production in another major corn-producing country (e.g., Brazil). In that case, there could be dangerous consequences for global food security.11 Such a destructive chain could play out as follows: As the global corn supply plummets because of Drought X, corn prices rise quickly, and countries institute export bans on corn to protect their own supply, especially in countries where grain storage capacity is minimal. This, in turn, causes further price increases for corn-based products, making many food items unaffordable in poor countries and for poor citizens in rich countries. The combination of expensive corn feed inputs and abnormally high mortality of cows, pigs, and chickens from the excessive heat leads to price increases for consumers of animal protein. The impacts trickle outside the United States to countries like Mexico that depend on U.S. corn imports.
Fortunately, due to advancements in technology in the last century, the impacts would probably take a different shape compared to 1936. Yet, such extreme heat would lead to increased demand for air conditioning, straining the aging electrical grid. As recently as August 2020, rolling blackouts were imposed during a massive heatwave in California.12 Even with the adoption of air conditioning, a long lasting and intense heat wave might cause spikes in mortality because not everyone has access to air conditioning. Those with air conditioning are conditioned to consistent temperatures and are less able to adapt to excessive heat. People with pre-existing health conditions may be particularly vulnerable in a blackout. Increased blackout periods, in turn, can cause significant disruption to businesses, increase the risk of mortality from heat stroke, food spoilage at supermarkets, restaurants, and schools, and may lead to civil unrest and riots.
How can the most severe impacts of Drought X be mitigated?
It is important to understand the extent to which a flash drought event could affect many sectors of the economy and civil society. Agriculture might experience losses unprecedented since at least the 1960s even with new technologies. There are also considerations regarding the supply chain for food, exports to other countries and national food security. Moreover, such a drought might affect the lives of those in these locations in drastic ways, paving the way for civil unrest if improperly handled.
Irrigation might mitigate some of the impacts that a flash drought event would impose on crops. However, access to groundwater is not equitable across the Corn Belt. Since irrigation is typically not necessary for producing high yield crops over most of the region, many growers might not take on the cost of adding irrigation even if the required water distribution infrastructure is available, which is not the case for much of the Midwest. Finally, irrigation may only marginally offset the extreme temperatures accompanying a Drought X scenario, as two weeks of temperatures over 100 degrees F in mid-July across a wide swath of the Corn Belt would likely lead to very poor pollination of crops and therefore very low yield. Thus, a significant reduction in corn production would likely occur even under a scenario in which irrigation was added in places where it currently does not exist.
Investments in infrastructure are another mitigation strategy that could be employed. Recently we have seen that high heat events have severely stressed the electrical grid. Ensuring that an event like Drought X does not overly strain the electrical grid, but upgrading the present system might help mitigate livestock mortality, human mortality, food waste, and civil unrest that could result.
Authors
Eric Hunt, Ph.D., is a staff scientist, Atmospheric and Environmental Research (AER) at Verisk. Eric can be reached at ehunt@aer.com.
Julia Borman, Ph.D., is a senior scientist, research, AIR Worldwide at Verisk. Julia can be reached at JBorman@air-worldwide.com.
Subodh Acharya, Ph.D., is a scientist, research, AIR Worldwide at Verisk. Subodh can be reached at SAcharya@air-worldwide.com.
Jeff Amthor, Ph.D., is a assistant vice president, research, AIR Worldwide at Verisk. Jeff can be reached at JAmthor@air-worldwide.com.
1. “Ag and Food Sectors and the Economy,” United States Department of Agriculture, October 20, 2020, < https://www.ers.usda.gov/data-products/ag-and-food-statistics-charting-the-essentials/ag-and-food-sectors-and-theeconomy/ >, accessed on November 3, 2020.
2. “Crop Values 2019 Summary,” United States Department of Agriculture, February 2020, < https://www.nass.usda.gov/Publications/Todays_Reports/reports/cpvl0220.pdf >, accessed on November 3, 2020.
3. “2019 was the 2nd wettest year on record for the U.S.,” National Oceanic and Atmospheric Administration, January 21, 2020, < https://www.noaa.gov/news/2019-was-2nd-wettest-year-on-record-for-us >, accessed on November 3, 2020.
4. Tom Polansek, “Monday storm impacted an estimated 37.7 million acres of Midwest farmland,” Reuters, August 14, 2020, < https://www.reuters.com/article/us-usa-farms-storm/monday-storm-impacted-an-estimated-37-7-million-acres-of-midwest-farmland-idUSKCN25A2DP >, accessed on November 3, 2020.
5. “Flash drought,” American Meteorological Society Glossary of Meteorology, February 19, 2019, < https://glossary.ametsoc.org/wiki/Flash_drought >, accessed on November 3, 2020.
6. Jean Phillips, “Flash drought,” Space Science and Engineering Center, University of Wisconsin-Madison, June 22, 2018, < https://www.ssec.wisc.edu/news/articles/11039 >, accessed on November 3, 2020.
7. “History of the Crop Insurance Program,” United States Department of Agriculture, < https://rma.usda.gov/en/About-RMA/History-of-RMA >, accessed on November 3, 2020.
8. “Federal Crop Insurance Corp., Summar of Business Report for 2017 thru 2020,” United States Department of Agriculture, October 12, 2020, < https://www3.rma.usda.gov/apps/sob/current_week/sobrpt2017-2020.pdf >, accessed on November 3, 2020.
9. Eric Hunt et al., “The flash drought of 1936,” Journal of Applied and Service Climatology, September 19, 2020, < https://stateclimate.org/article/the-flash-drought-of-1936/ >, accessed on November 20, 2020.
10. Rahmstorf, S., D. Coumou, 2012: “Increase of extreme events in a warming world,” Proceedings of the National Academy of Sciences, 108 (44) 17905-17909; DOI: 10.1073/pnas.1101766108
11. Weston Anderson, “Climate Variability Poses a Correlated Risk to Global Food Production,” Columbia University, September 7, 2018, < https://academiccommons.columbia.edu/doi/10.7916/D8766Z04 >, accessed on November 3, 2020.
12. Debra Kahn, et al., “California has first rolling blackouts in 19 years and everyone faces blame,” Politico, August 18, 2020, < https://www.politico.com/states/california/story/2020/08/18/california-has-first-rolling-blackouts-in-19-years-and-everyone-faces-blame-1309757 >, accessed on November 3, 2020.
