Environmental Research Letters

LETTER • OPEN ACCESS

The evolution, propagation, and spread of flash drought in the CentralUnited States during 2012To cite this article: Jeffrey B Basara et al 2019 Environ. Res. Lett. 14 084025

 

View the article online for updates and enhancements.

This content was downloaded from IP address 128.104.109.141 on 07/08/2019 at 22:52

Environ. Res. Lett. 14 (2019) 084025 https://doi.org/10.1088/1748-9326/ab2cc0

LETTER

The evolution, propagation, and spread of flash drought in theCentral United States during 2012

Jeffrey BBasara1,2 , Jordan IChristian1 , RyannAWakefield1, JasonAOtkin3, EricHHunt4 andDavid PBrown5

1 School ofMeteorology, University ofOklahoma,Norman,OK,United States of America2 School Civil Engineering and Environmental Science, University ofOklahoma,Norman,OK,United States of America3 Cooperative Institute forMeteorological Satellite Studies, Space Science and Engineering Center, University ofWisconsin-Madison,

Madison,WI,United States of America4 Atmospheric andEnvironmental Research, Inc., Lexington,MA,United States of America5 Grazinglands Research Laboratory, USDAAgricultural Research Service, El Reno,OK,United States of America

E-mail: jbasara@ou.edu

Keywords: drought, flash drought, Central United States, land-atmosphere interactions, evaporative stress

Supplementarymaterial for this article is available online

AbstractDuring 2012, flash drought developed and subsequently expanded across large areas of the CentralUnited States (US)with severe impacts to overall water resources andwarm-season agriculturalproduction. Recent efforts have yielded amethodology to detect and quantifyflash droughtoccurrence and rate of intensification from climatological datasets via the standardized evaporativestress ratio (SESR). This study utilizes theNorthAmerican Regional Reanalysis and applied the SESRmethodology to quantify the spatial and temporal development and expansion offlash droughtconditions during 2012. Critical results include the identification of the flash drought epicenter andsubsequent spread offlash drought conditions radially outwardwith varying rates of intensification.Further, a comparison of the SESR analyses with surface-atmosphere couplingmetrics demonstratedthat a hostile environment developed across the region, which limited the formation of deepatmospheric convection, exacerbated evaporative stress, and perpetuated flash drought developmentand enhanced its radial spread across theCentral US.

1. Introduction

The United States (US) drought of 2012 was anextreme event of historic proportions. At its peak, itcovered most of the contiguous US and representedthe fourth largest drought by aerial extent since 1895(Knapp et al 2015). Within the Great Plains region ofthe Central US, precipitation deficits during the mostcritical portion of the growing season exceeded thoseof the most extreme drought years during the DustBowl (Schubert et al 2004), andwere the greatest notedin the modern-day historical record dating to 1895(Hoerling et al 2014). Severe impacts were observed onregional ecosystems (Mallya et al 2013, Knapp et al2015, Sippel et al 2016, Wolf et al 2016, Zhou et al2017) and agricultural productivity (Boyer et al 2013);in the heavily irrigated region of central Nebraska, the

drought led to increased water usage and subsequentgroundwater declines that took years to recover(Young et al 2018). In total, economic losses from thedrought were estimated to be in excess of $30 billionacross the entire nation (NCEI 2017). Impacts fromthe 2012 drought ultimately effected energy systems,created water shortages and navigational disruptions,and reduced employment in rural areas (Grigg 2014).

Drought in the Central US has been attributed tonumerous local to global factors including surface-atmosphere feedbacks, persistent synoptic patterns,and large-scale teleconnections (Basara et al 2013). Assuch, the overall evolution of the 2012 drought wasinherently complex. It was embedded within an exten-ded period of continuous drought in the US that span-ned late 2010 through mid 2015, a period that alsoincluded notable extreme drought events in the

OPEN ACCESS

RECEIVED

3May 2019

REVISED

20 June 2019

ACCEPTED FOR PUBLICATION

25 June 2019

PUBLISHED

31 July 2019

Original content from thisworkmay be used underthe terms of the CreativeCommonsAttribution 3.0licence.

Any further distribution ofthis workmustmaintainattribution to theauthor(s) and the title ofthework, journal citationandDOI.

© 2019TheAuthor(s). Published by IOPPublishing Ltd

Southern Great Plains (Hoerling et al 2013, Seager et al2013) and California (Griffin and Anchukaitis 2014,Seager et al 2015). In addition, La Nina conditionsduring the winter of 2011–2012, resulted in broadregions of the US experiencing drier-than-normalconditions (Rippey 2015). The resulting implicationsfor the regional atmospheric circulation, which inclu-ded a displaced stormtrack, limited moisture return,and broad subsidence, led to large negative standar-dized anomalies in precipitation, runoff, and soilmoisture and positive anomalies in temperature(Hoerling et al 2013). Persistent negative soil moistureanomalies were also observed (Otkin et al 2016), whichwere further strongly coupled to atmospheric pro-cesses during this period (Basara andChristian 2018).

The nature of the 2012 Central US drought’s evol-ution is of particular interest, specifically its rapid geo-graphic expansion and marked intensification fromnon-drought to extreme/exceptional drought condi-tions during the growing season (AghaKouchak 2014,McEvoy et al 2016, Otkin et al 2016). As noted byOtkinet al (2016) and displayed in figure 1, the vastmajority ofthe Central US domain (the region east of the RockyMountains and west of the Great Lakes) was drought-free or abnormally dry, D0, as late as mid-May 2012 asdefined by the US Drought Monitor (Svoboda et al2002). However, drought quickly enveloped the region,with large portions of the Central US deteriorating intoD3 (Extreme)orD4 (Exceptional)drought by July.Dur-ing this period of rapid intensification, drought inten-sities increased in magnitude by four or even five USDroughtMonitor categories in fewer than twomonths.

Such rapid intensification is consistent with thedefinition of flash drought by Otkin et al (2018).Flash droughts occur during periods when extreme pre-cipitation deficits combined with increased surface

temperature, wind speed, downwelling shortwave radia-tion, and/or vapor pressure deficits persist for severalweeks. This leads to the rapid depletion of soil moisture,increased evaporative stress, and the rapid intensificationof drought conditions and associated impacts. Christianet al (2019) developed a methodology to identify flashdrought using a standardized form of the ratio betweenevapotranspiration andpotential evapotranspiration: thestandardized evaporative stress ratio (SESR). This ratio isbased on previous satellite-based studies focused on eva-porative stress (Anderson et al 2007a, 2007b, Otkin et al2013) and is inversely proportional to the amount of eva-porative stress on the environment. The use of SESR isadvantageous for flash drought identification as it incor-porates numerous variables (Otkin et al 2014, Ford et al2015, Hobbins et al 2016, McEvoy et al 2016, Otkin et al2016,Otkin et al2019.

Because flash drought is a relatively new area ofstudy and yet has critical impacts, the objectives of thisstudy were to (1) quantify the rapid evolution andexpansion of the 2012 historic drought in the CentralUS, using the SESR, and (2) identify critical feedbacksvia land-atmosphere interactions that accentuated the2012 drought’s intensification, perpetuation, andpropagation.

2.Data andmethods

TheNorth American Regional Reanalysis (NARR)wasused to investigate the evolution of the 2012 flashdrought across the Central US, in particular toexamine land-atmosphere interactions that developand propagate of flash drought conditions. The NARRhas a spatial resolution of 32 km with 29 vertical levels(Mesinger et al 2006). Computed quantities include

Figure 1.Drought categories across theCentral US for 08May 2012 (i.e. prior toflash drought onset) and 09October 2012 (i.e. afterflash drought development) as indicated by theUSDroughtMonitor (Svoboda et al 2002).

2

Environ. Res. Lett. 14 (2019) 084025

the SESR, as well as the convective triggering potential(CTP) and low level humidity index (HI)methodologyoriginally developed by Findell and Eltahir (2003a).The domain for analysis was bounded by a north andsouth latitude of 49°N and 33.5°N, respectively, and awest and east longitude of −104°W and 87°W,respectively.

Using the flash drought definition of Otkin et al(2018) and the methodology developed by Christianet al (2019), NARR grid points across the Central US in2012 were analyzed for flash drought occurrence viastandardized ESR (SESR) and changes in SESR. ESR isdefined as:

=ESRET

PET,

where ET is evapotranspiration and PET potentialevapotranspiration. Specifically within the Noah landsurface model in the NARR dataset, total surface ET iscomputed from soil evaporation, transpiration fromthe vegetation canopy, evaporation of dew/frost orcanopy-intercepted precipitation, and snow sublima-tion (Mesinger et al 2006)while PET is calculated usingthe modified Penman scheme from Mahrt and Ek(1984). SESR and the change in SESR are defined as:

s=

-SESR

ESR ESRijp

ijp ijp

ESR ijp

sD =

D - D

D( )SESR

SESR SESR,ijp z

ijp ijp

SESR ijp

where SESR ijp (referred to as SESR) and D( )SESRijp z

(referred to asΔSESR) are the z-scores of ESR and thechange in SESR, respectively, for a specific pentad (p)at a specific grid point (i, j). As described in Christianet al (2019), four restrictive criteria were employed toidentify flash drought occurrence at a given location.These include two criteria that emphasized the rapidintensification toward drought and two criteriafocused on vegetative impacts, summarized as:

• A negative trend of SESR for a minimum length offive pentads,

• Afinal SESR value below the 20th percentile,

• Pentad-to-pentad changes in SESR (ΔSESR) belowthe 40th percentile,

• Change in SESR over the entirety of the flashdrought below the 25th percentile.

This methodology applied to the NARR datasethas been shown to compare well with the satellite-based evaporative stress index (ESI), drought depic-tion from the USDM, and previously identified flashdrought cases (Christian et al 2019). For locationsidentified as having experienced flash drought, thedate whereby flash drought began was partitioned bymonth. In addition, the spatial coverage of flashdrought across the region was calculated from the

aggregate of grid points that experienced flashdrought.

Convection is a critical contributor to total pre-cipitation during the growing season across the Cen-tral US (Haberlie and Ashley 2019). As such, processesthat suppress convective initiation and propagationare critical to drought and flash drought development.Further, the Central US is a region known for strongland-atmosphere interactions (Koster et al 2004, Guoand Dirmeyer 2013, Koster et al 2016)which can yieldfeedback processes that enhance drought develop-ment (Hoerling et al 2013, Roundy et al 2013, Basaraand Christian 2018,Wakefield et al 2019). Gerken et al(2018) utilized the CTP and HI framework in the ana-lysis of 2017 flash drought conditions in the NorthernPlains of the US and noted that incorporating land-atmosphere interactions has broad utility to under-standing flash drought development. As such, toexamine feedback processes and convective suppres-sion related to flash drought development and expan-sion in 2012, the CTP and HI framework was utilized;the framework is described in Findell and Eltahir(2003a, 2003b)wherebyCTP is determined by:

(1) Locating the moist adiabat which intersects thetemperature profile 100 hPa above groundlevel (AGL).

(2) Integrating the area between this moist adiabatand the temperature profile from 100 hPa AGL to300 hPaAGL.

Further, HI is found by summing the dewpointdepressions at 50 and 150 hPaAGL as follows:

= - + -( ) ( )T THI Td Td .50 hPa AGL 150 hPa AGL

The framework quantifies pre-existing instabilitywithin the lower portions of the atmosphere, pre-existing moisture in the atmosphere before develop-ment of the convective boundary layer, and indirectlythe overall atmospheric moisture content in the lowertroposphere. Following the methods ofWakefield et al(2019), daily standardized anomalies of CTP and HIwere computed for the NARR dataset (1979 through2017). These standardized anomalies (hereafter, CTPzand HIz) allow for comparison of CTP and HI in bothtime and space where significant variations in meanCTP and HI values occur. When averaged over longertime periods, these values yield persistence ofmoisturewithin the boundary layer as well as covariability withvariables that impact evaporative stress (Wakefieldet al 2019).

3. Results

3.1. Flash drought expansionThe flash drought in 2012 began during April, with anepicenter located over north-central Kansas (figure 2)immediately adjacent to an area of long-term drought

3

Environ. Res. Lett. 14 (2019) 084025

Figure 2.Expansion offlash drought across theCentral United States during the growing season (April throughOctober) using SESRderived fromNARR. Red pixels indicate locations where flash drought began in a specifiedmonth. Black pixels indicate locationswhere flash drought began in a previousmonth. The last panel shows the cumulative andmonthly spatial coverage offlash droughtacross the domain.

Figure 3. 1month SPI fromNARRacross theCentral United States for the growing season (April throughOctober).

4

Environ. Res. Lett. 14 (2019) 084025

(which lingered from 2011) across southwest Kansas(figure 1). During May 2012, rapid drought develop-ment expanded north into Nebraska, South Dakota,and North Dakota, eastward across Missouri andIllinois, and southeastward into Oklahoma andArkansas. Conversely, localized above-normal preci-pitation (positive values of the one-month standar-dized precipitation index; SPI;McKee et al 1993, 1995)for the month ofMay across the northern Great Plainsand Midwest acted to inhibit flash drought develop-ment in southeast Nebraska and Iowa (figure 3). Over-all, the areal coverage identified to have experiencedflash drought across the central United States in May2012 spanned approximately 299 000 km (figure 2).

During June 2012, flash drought expanded inmul-tiple areas across the central Great Plains andMidwest.The first of these areas included southeast Kansas andsouthwest Oklahoma, located along the periphery of aregion that had already experienced flash drought dur-ing May. In addition, a broad region extending acrosssoutheast Nebraska, Iowa, Illinois, and the southernportion of Wisconsin experienced flash drought dur-ing the month of June. In total, these locations con-tributed approximately 361 000 km2 of new land areathat experienced rapid onset of drought, the highestamong allmonths of the 2012 growing season.

The remainder of the 2012 growing season sawmodest additional flash drought expansion in theCentral US. In July 2012, the primary regions exhibit-ing new flash drought included northern Iowa, south-ern Minnesota, western Wisconsin, central Missouri,and southern Arkansas, while during August 2012,portions of Oklahoma, Iowa, Minnesota, and NorthDakota also experienced flash drought. Flash droughtdevelopment during September was isolated to theMidwest across northwestern Illinois and several loca-tions in Wisconsin, as isolated areas or ‘gaps’ withinthe spatial coverage of flash drought developmentwerefinallyfilled.

The total land area across the Central US studydomain that experienced flash drought at any pointduring the 2012 growing season was approximately1008 640 km2, constituting over 40% of the total landarea in the domain. From a historical perspective,based on the Christian et al (2019) methodologyapplied to the entire 1979–2016 NARR period, the2012 flash drought ranks as the 2nd highest in totalspatial coverage within the study domain during thegrowing season. While some locations in the analysisand domain did not reach the flash drought criteria asdefined by Christian et al (2019), these locations didexperience drought onset.

Representative locations within the domain werechosen to illustrate the temporal evolution of flashdrought using SESR during different expansion peri-ods within the 2012 growing season (figure 4; locationsshown in figure 5). Near the epicenter of the 2012 flashdrought event, the overall change in SESR was excep-tionally dramatic with values declining 3.5 standard

deviations in only 30 d. Specifically, SESR decreased by2.4 standard deviations during the last 15 d of therapid drought intensification, which was the largestnegative change in SESR for that grid point at thatgiven time of year in the climatological record ofNARR (figure 4(a)). For regions to the southeast(north-central Oklahoma) and east (north-centralMissouri) of the flash drought epicenter in north-cen-tral Kansas, rapid drought development beganapproximately one week later (figures 4(b) and (c)).Conversely, evaporative stress increased rapidly incentral Nebraska in lateMay 2012, approximately fourweeks after rapid onset drought began in OklahomaandKansas (figure 4(d)).

In west-central Iowa, SESR began to decline inearlyMay 2012, but wasmoderated by precipitation inMay and June (figures 3 and 4(e)). In fact, June 2012precipitation in the regionwas at or slightly above nor-mal (figure 3). However, this precipitationwas not suf-ficient to prevent rapid intensification toward droughtdue to exceptionally hot and cloud-free conditions. Assuch, SESR sharply declined from late June to mid-July 2012 during a subsequent period of negative pre-cipitation anomalies.

As noted in figure 2, an extensive region of addi-tional flash drought development occurred in late July2012 for southeastern Minnesota and west-centralWisconsin. This can be seen in figure 4(f), where rapiddrought intensification lasted for approximately 40days and ended in late August 2012. Examination ofthe SPI reveals that the region received above-normaltotals through the early portion of the season whichinitially staved off drought development (figure 3).However, with the onset of negative precipitationanomalies, flash drought rapidly developed soonthereafter.

3.2. Surface-atmosphere interactions and feedbacksPositive HI values represent a drier-than-normallower atmosphere and reduced relative humidity inthe boundary layer during the day while CTP ismeasured within a layer that slightly overlaps that ofHI but is generally at higher altitude. As such, positivevalues of CTP further indicate a drier atmosphericcolumnaloft. Turbulence in the boundary layer duringthe day mixes this drier air aloft toward the surfacewhich reduces the relative humidity and enhancesevaporative stress. Thus, a combination of positiveCTP and HI z-scores (hereafter referred to as CTPzand HIz to distinguish this framework from thatpresented by Findell and Eltahir 2003a, 2003b) repre-sents an abnormally dry air mass conducive forenhanced evaporative stress in the surface layer andhostile to precipitation generation within the bound-ary layer and lower troposphere.

Persistent, positive HIz was established through-out the Great Plains and Corn Belt beginning inApril 2012, while CTPz was more variable (figure 5).

5

Environ. Res. Lett. 14 (2019) 084025

Positive CTPz over one region may be measured asnegative downstream, where the air mass originally incontact with the drought affected land surfaceupstream may become elevated over the downstreamland surface and can serve as a strong inversion or ele-vated mixed layer in the region of the atmosphericprofile where CTP is computed (Wakefield et al 2019).This is especially likely when coupled with positiveHIz. An elevated mixed layer can then suppress con-vective precipitation, aiding in the reduction of rain-fall. The first Central US locations to experience flashdrought (figure 2), and in particular the flash droughtepicenter in Kansas, were characterized by positiveCTPz and HIz during April and May 2012. Remnant,long-term drought in western Texas and eastern NewMexico remained present through the winter monthsof 2011–2012 and into the spring of 2012. As such,much of the region upstream, over, and downstreamof the flash drought epicenter was characterized bymean positive CTPz and HIz for both April and May2012, indicating that the entire lower atmospheric col-umnwas anomalously dry.

The presence of mean positive CTPz and HIz con-ditions was due, in part, to the circulation around alower-tropospheric ridge centered over the Gulf ofMexico (GOM) which led to continuous advection ofdrought modified air masses from Texas and NewMexico into Oklahoma, Missouri and Kansas andeven further downstream into Nebraska, Iowa andMinnesota. As a result, expansion of drought wasobserved along lower tropospheric streamlines during

May 2012, particularly from Kansas through Illinoiswhere droughtmodified airmass was transported overadjacent regions where drought intensified. In June2012, the northward shift of the GOM ridge center ledto a stronger southerly component to lower tropo-spheric winds and enhanced advection of positive HIzinto Nebraska, Iowa and Minnesota. The CTPz valueswere negative across much of the same region duringthis same period, but when coupled with positive HIzsuggests that the lower atmosphere was dry andstrongly capped making conditions unfavorable forsurface-based convection. In addition, local HIz max-ima were observed in regions where flash drought hadalready occurred, particularly north-central Kansas.As such, the southwest/northeast axis of flash droughtexpansion during June 2012 was aligned with lower-tropospheric southwesterly winds that incorporated asubstantial fetch over areaswith positiveHIz.

The advection of the hostile air mass into Minne-sota occurred during May and June. CTPz and HIzwere both positive throughout much of the period,such that local evaporative demand remained elevatedwithin the surface layer and reinforced by boundary-layer feedbacks. Further, the trajectory of lower-tro-pospheric winds (e.g. 850 mb) during August 2012over eastern Kansas, Missouri and Oklahoma allowedfor persistence of drought conditions, while northernlocations were influenced predominantly by westerly/northwesterly winds during August and September2012. Such conditions (figure 5) prevented droughtrecovery through continued advection of dry, hostile

Figure 4. SESR (red line) from theNARRdataset for six locations across theGreat Plains andMidwest. The shaded tan region on eachpanel represents the temporal period for flash drought from theflash drought identificationmethodology.

6

Environ. Res. Lett. 14 (2019) 084025

air masses from drought affected regions upstream.Locations with the earliest onset of flash drought hadpersistent, positive HIz with greater magnitudes inJuly and August 2012 than those with later flashdrought onset. The range of CTP and HI z-scores wasalso larger followingflash drought onset.

4.Discussion and conclusions

Not all drought is the same, as fundamentally, variousphysical processes can inhibit precipitation thus lead-ing to drought at multiple spatial and temporal scales.Overall, the 2012 Central US drought can be consid-ered an historic and extreme event. The rapid intensi-fication and expansion of drought in the Central USoccurred at shorter timescales than other locations

impacted by the same broad, long-term drought (e.g.Southwest US). This rapid development was consis-tentwith the characteristics of aflash drought, one thatbegan with an epicenter in the central Great Plainsfollowed by quasi-radial expansion throughout thelarger Central US domain.While the 2012 drought as awhole has been well established as an historic event,the flash drought component observed in the CentralUS was also historic, with over 1250 000 km2 meetingthe criteria for flash drought. This would rank near thetop of total flash drought coverage in a year across theCentral US domain during the satellite record.

Two main results were revealed from the analysisof flash drought development and extent. First, theCentral US domain had near-normal or above-normalSESR values just before the rapid development ofdrought. Thus, antecedent surface conditions did not

Figure 5.Monthly averagedCTP andHI standardized anomalies (CTPz andHIz) for April through September of 2012. Red pixels areabove normal while blue are belownormal. Arrows represent themeanmonthly 850 mbwind speed and direction. Black dots indicatethe location of eachflash drought case from figure 4.

7

Environ. Res. Lett. 14 (2019) 084025

serve to inhibit the development of flash drought. Sec-ond, the time series and spatial analysis of SESRdemonstrated that flash drought did not develop uni-formly across the region. Rather, the majority of flashdrought in the Central US occurred along areas adja-cent to prior flash drought development earlier in thestudy period. Therefore, the Central US flash droughtshould not be considered a single, linear event, butrather as an evolving phenomenon that propagatedand spread through space and time.

Additionally, the timing of flash drought onset wasclosely tied to the advection of a hostile air mass fromregions already impacted by drought. CTPz and HIzprovide information concerning the atmosphericcomponent of land-atmosphere feedbacks and, in thecase of the 2012 drought, they demonstrated anoma-lies of moisture in the lower atmosphere along withthe subsequent persistence and movement of the dryair mass. Thus, land-atmosphere feedbacks throughthe boundary-layer contributed to anomalously dryair (elevated HIz) over drought impacted areas in theCentral US. As such, while large-scale dynamics yiel-ded overall environmental conditions conducive fordrought development, the downstream advection ofhostile air from adjacent, drought impacted areas (1)increased the atmospheric demand, (2) elevated theevaporative stress, all while (3) further suppressingdeep convection critical to overall precipitation andwater availability of the region. In the end, the timingof increased evaporative stress and positive CTPz andHIz values not only drove flash drought developmenton a local scale, but accentuated the propagation offlash drought to adjacent areas, eventually spreadingacross themajority of the Central US.

The 2012 drought in the Central USwas a functionof many critical land-atmosphere components. Whilethis study focused on rapid drought intensificationthrough time and space, along with the impacts oflocal and non-local feedbacks associated with land-atmosphere interactions, it is important to note thatlarge-scale drivers of drought were also extremelyimportant (Hoerling et al 2013, Rippey 2015, Otkinet al 2016). In fact, it was the combination and interac-tion of various physical processes at multiple time-scales that led to the cascading evolution of flashdrought, sustained drought, and drought impacts dur-ing the 2012 growing season.

It is also critical to note that the methodologydeveloped by Christian et al (2019) provides an inno-vative framework for examining flash drought givenits incorporation of multiple environmental condi-tions within a standardized approach. As such, itshould be utilized in drought monitoring as well asfuture studies that examine additional past droughtevents in the Central US (e.g. 1988, 2006, etc) alongwith historical drought development across the globeto better quantify the role and mechanisms of flashdrought and land-atmosphere feedbacks in cascadingevents.

Acknowledgments

This work was supported, in part, by the NOAAClimate Program Office’s Sectoral ApplicationsResearch Program (SARP) grant NA130AR4310122,the Agriculture and Food Research Initiative Compe-titive Grant No. 2013-69002 from the USDA NationalInstitute of Food and Agriculture, the USDA NationalInstitute of Food and Agricultural (NIFA) Grant No.2016-68002-24967, the USDA-ARS Southern GreatPlains Climate Hub Cooperative Agreement 58-3070-8-012, and the NASA SMD Earth Science Divisionthrough theEarth ScienceApplication:WaterResourcesProgram. The data from the NARR used in this study isavailable at https://esrl.noaa.gov/psd/data/gridded/data.narr.html. The authors would also like to acknowl-edge the anonymous reviewers who provided criticalfeedback that improved thismanuscript.

ORCID iDs

Jeffrey B Basara https://orcid.org/0000-0002-2096-6844Jordan I Christian https://orcid.org/0000-0003-2740-8201

References

AghaKouchakA 2014A baseline probabilistic drought forecastingframework using standardized soilmoisture index:application to the 2012United States droughtHydrol. EarthSyst. Sci. 18 2485–92

AndersonMC,Norman JM,Mecikalski J R,Otkin J A andKustasWP2007aA climatological study ofevapotranspiration andmoisture stress across the continentalUnited States based on thermal remote sensing: I.Modelformulation J. Geophys. Res. 112 921

AndersonMC,Norman JM,Mecikalski J R,Otkin J A andKustasWP2007bA climatological study ofevapotranspiration andmoisture stress across the continentalUnited States based on thermal remote sensing: II. Surfacemoisture climatology J. Geophys. Res. 112 1100

Basara J B andChristian J I 2018 Seasonal and interannual variabilityof land–atmosphere coupling across the SouthernGreatPlains ofNorthAmerica using theNorth American regionalreanalysis Int. J. Climatol. 38 964–78

Basara J B,Maybourn JN, PeiranoCM,Tate J E, BrownP J,Hoey JD and Smith BR 2013Drought and associated impactsin the great plains of theUnited States—a review Int. J. Geosci.4 72–81

Boyer J S et al 2013TheUS drought of 2012 in perspective: a call toactionGlob. Food Secur. 2 139–43

Christian J I, Basara J B,Otkin JA,Hunt ED,WakefieldRA,FlanaganPX andXiaoX2019Amethodology forflashdrought identification: applicationofflash drought frequencyacross theUnitedStates J.Hydrometeorol.20833–46

Findell K L and Eltahir E AB 2003aAtmospheric controls on soilmoisture–boundary layer interactions: I. Frameworkdevelopment J. Hydrometeorol. 4 552–69

Findell K L and Eltahir E AB 2003bAtmospheric controls on soilmoisture–boundary layer interactions: II. Feedbacks withinthe continental United States J. Hydrometeorol. 4 570–83

FordTW,McRobertsDB,Quiring SMandHall R E 2015On theutility of in situ soilmoisture observations for flash droughtearly warning inOklahoma,USAGeophys. Res. Lett. 429790–8

8

Environ. Res. Lett. 14 (2019) 084025

GerkenT, BromleyGT, Ruddell B L,Williams S and Stoy PC2018Convective suppression before and during theUnited StatesNorthern great plains flash drought of 2017Hydrol. EarthSyst. Sci. 22 4155–63

GriffinDandAnchukaitis K J 2014Howunusual is the 2012–2014California drought?Geophys. Res. Lett. 41 9017–23

GriggNS2014The 2011–2012drought in theUnited States: newlessons froma record event Int. J.Water Resour.Dev. 30 183–99

GuoZ andDirmeyer PA 2013 Interannual variability of land–atmosphere coupling strength J. Hydrometeorol. 14 1636–46

Haberlie AMandAshleyWS 2019A radar-based climatology ofmesoscale convective systems in theUnited States J. Clim. 321591–606

HobbinsMT,WoodA,McEvoyD J,Huntington J L,MortonC,AndersonMandHainC2016The evaporative demanddrought index: I. Linking drought evolution to variations inevaporative demand J. Hydrometeorol. 17 1745–61

HoerlingM, Eischeid J, KumarA, LeungR,Mariotti A,MoK,Schubert S and Seager R 2014Causes and predictability of the2012 great plains droughtBull. Am.Meteorol. Soc. 95 269–82

HoerlingM et al 2013Anatomy of an extreme event J. Clim. 262811–32

KnappAK,Carroll C JW,Denton EM, La Pierre K J, Collins S L andSmithMD2015Differential sensitivity to regional-scaledrought in six central US grasslandsOecologia 177 949–57

Koster RD,Chang Y,WangH and Schubert SD 2016 Impacts oflocal soilmoisture anomalies on the atmospheric circulationand on remote surfacemeteorological fields during borealsummer: a comprehensive analysis overNorthAmericaJ. Clim. 29 7345–64

Koster RD et al 2004Regions of strong coupling between soilmoisture and precipitation Science 305 1138–40

Mahrt L and EkM1984The influence of atmospheric stability onpotential evaporation J. Clim. Appl.Meteorol. 23 222–34

MallyaG, Zhao L, SongXC,Niyogi D andGovindaraju R S 20132012midwest drought in theUnited States J. Hydrol. Eng. 18737–45

McEvoyD J,Huntington J L,HobbinsMT,WoodA,MortonC,AndersonMandHainC2016The evaporative demanddrought index: II. CONUS-wide assessment against commondrought indicators J. Hydrometeorol. 17 1763–79

McKee TB,DoeskenN J andKleist J 1993 The relationship ofdrought frequency and duration to time scales 8th Conf. onApplied Climatology (Anaheim, CA) pp 179–84

McKee TB,DoeskenN J andKleist J 1995Droughtmonitoringwithmultiple time scales 9th Conf. on Applied Climatology (Dallas,TX) pp 233–6

Mesinger F et al 2006North American regional reanalysisBull. Am.Meteorol. Soc. 87 343–60

National Centers for Environmental Information 2017 Billion-dollar weather and climate disasters: Overview, NOAANCEI(https://ncdc.noaa.gov/billions/)

Otkin J A, AndersonMC,HainC,Mladenova I E, Basara J B andSvobodaM2013 Examining rapid onset drought

development using the thermal infrared–based evaporativestress index J. Hydrometeorol. 14 1057–74

Otkin J A, AndersonMC,HainC and SvobodaM2014 Examiningthe relationship between drought development and rapidchanges in the evaporative stress index J. Hydrometeorol. 15938–56

Otkin J A, AndersonMC,HainC, SvobodaM, JohnsonD,Mueller R, Tadesse T,WardlowB andBrown J 2016Assessingthe evolution of soilmoisture and vegetation conditionsduring the 2012United States flash droughtAgric. For.Meteorol. 218–219 230–42

Otkin J A, SvobodaM,Hunt ED, Ford TW,AndersonMC,HainC andBasara J B 2018 Flash droughts: a review andassessment of the challenges imposed by rapid-onsetdroughts in theUnited StatesBull. Am.Meteorol. Soc. 99911–9

Otkin J A, ZhongY,Hunt ED, Basara J, SvobodaM,AndersonMCandHainC 2019Assessing the evolution ofsoilmoisture and vegetation conditions during aflashdrought—flash recovery sequence over the south-central USJ. Hydrometeorol. 20 549–62

Rippey BR 2015TheUS drought of 2012Weather Clim. Extremes 1057–64

Roundy J K, FergusonCR andWoodE F 2013Temporal variabilityof land–atmosphere coupling and its implications fordrought over the SoutheastUnited States J. Hydrometeorol. 14622–35

Schubert SD, SuarezM J, Pegion P J, Koster RD andBacmeister J T2004On the cause of the 1930s dust bowl Science 303 1855

Seager R,Goddard L,Nakamura J,HendersonN and LeeDE 2013Dynamical causes of the 2010/11Texas–NorthernMexicodrought J. Hydrometeorol. 15 39–68

Seager R et al 2015Causes of the 2011–14California drought J. Clim.28 6997–7024

Sippel S, Zscheischler J andReichsteinM2016 Ecosystem impacts ofclimate extremes crucially depend on the timing Proc. NatlAcad. Sci. 113 5768

SvobodaM et al 2002The droughtmonitorBull. Am.Meteorol. Soc.83 1181–90

Wakefield RA, Basara J B, Furtado J C, IllstonBG,FergusonCR andKlein PM2019Amodified framework forquantifying land-atmosphere covariability duringhydrometeorological and soil wetness extremes inOklahomaJ. Appl.Meteorol. Climatol. 58 1465–83

Wolf S et al 2016Warm spring reduced carbon cycle impact of the2012US summer drought Proc. Natl Acad. Sci. 113 5880

YoungAR, BurbachME,Howard LM,WaszgisMM,Lackey SO and Joeckel RM2018Nebraska StatewideGroundwater-LevelMonitoring Report 2018University ofNebraska-Lincoln 24

ZhouY et al 2017Quantifying agricultural drought in tallgrassprairie region in theUS SouthernGreat Plains throughanalysis of awater-related vegetation index fromMODISimagesAgric. For.Meteorol. 246 111–22

9

Environ. Res. Lett. 14 (2019) 084025