<%BANNER%>

UFIR



Surface radar-derived convective rainfall associations with Midwest U.S. land surface conditions in summers 1999 and 2000
CITATION PDF VIEWER
Full Citation
STANDARD VIEW MARC VIEW
Permanent Link: http://ufdc.ufl.edu/IR00001206/00001
 Material Information
Title: Surface radar-derived convective rainfall associations with Midwest U.S. land surface conditions in summers 1999 and 2000
Series Title: Matyas, C. J., and A. M. Carleton, 2010: Surface radar-derived convective rainfall associations with Midwest U.S. land surface conditions in summers 1999 and 2000. Theoretical and Applied Climatology 93, 315-330.
Physical Description: Journal Article
Creator: Matyas, Corene
Publisher: Theoretical and Applied Climatology
Publication Date: 2010
 Notes
Abstract: Previous research has suggested that spatial heterogeneities in soil moisture and/or vegetation cover promote the development of convective clouds. We examine the intensity of convective precipitation for the Midwest US Corn Belt in the summers of 1999 and 2000, which had contrasting synoptic circulation, atmospheric humidity, and soil moisture conditions. For days when synoptic scale atmospheric forcing is weak, we calculate a convective severity index (CSI) based on radar reflectivity composite values. Our results suggest that boundaries between soil types, and cropland and forest vegetation types in the western portion of the Corn Belt, enhance the development of convective precipitation. In the eastern part of the Corn Belt, less convection occurs, but we find a positive correlation between the intensity of convection and soil moisture conditions. Our results also demonstrate that the CSI is a simple yet effective technique for identifying where deep convection occurs relative to lighter precipitation.
Acquisition: Collected for University of Florida's Institutional Repository by the UFIR Self-Submittal tool. Submitted by Corene Matyas.
 Record Information
Source Institution: University of Florida Institutional Repository
Holding Location: University of Florida
Rights Management: All rights reserved by the submitter.
System ID: IR00001206:00001

Downloads

This item is only available as the following downloads:

9matyascarleton2010 ( PDF )


Full Text

PAGE 1

ORIGINALPAPERSurfaceradar-derivedconvectiverainfallassociations withMidwestUSlandsurfaceconditionsinsummer seasons1999and2000CoreneJ.Matyas&AndrewM.CarletonReceived:1August2008/Accepted:21April2009/Publishedonline:13May2009 # Springer-Verlag2009Abstract Previousresearchhassuggestedthatspatial heterogeneitiesinsoilmoistureand/orvegetationcover promotethedevelopmentofconvectiveclouds.We examinetheintensityofconvectiveprecipitationfor theMidwestUSCornBeltinthesummersof1999and 2000,whichhadcontrastingsynopticcirculation,atmospherichumidity,andsoilm oistureconditions.Fordays whensynopticscaleatmosphericforcingisweak,we calculateaconvectiveseverityindex(CSI)basedon radarreflectivitycompositevalues.Ourresultssuggest thatboundariesbetweensoiltypes,andcroplandand forestvegetationtypesinthewesternportionofthe CornBelt,enhancethedevelopmentofconvective precipitation.IntheeasternpartoftheCornBelt,less convectionoccurs,butwefindapositivecorrelation betweentheintensityofconvectionandsoilmoisture conditions.OurresultsalsodemonstratethattheCSIis asimpleyeteffectivetechniqueforidentifyingwhere deepconvectionoccursrelativetolighterprecipitation. 1Introduction Theextenttowhichlandsurfaceconditionsaffectthe developmentofconvectivecloudsandprecipitationisa subjectofcontemporaryclimateresearch,especiallyfor interiorcontinentallocationsduringthewarmseason(e.g., Carletonetal. 2008a b ).Modelingstudieshaveinvestigatedsoilmoistureinfluencesonprecipitationgeneration forbothmoistanddryenvironmentsoverarangeofspatial scalesandhaveshownthatthespatialheterogeneitiescan initiateconvectiverainfallwhenenvironmental(i.e.,free atmosphere)windsareweak(YanandAnthes 1988 ; OglesbyandErickson 1989 ;FastandMcCorcle 1991 ; ChenandAvissar 1994a b ;Giorgietal. 1996 ;Lynnetal. 1998 ).Otherstudieshaveshownthatvegetationtypeplays animportantroleinincreasingconvectiveprecipitation throughtheenhancementofevapotranspirationandalterationoftheboundary-layerwindcirculations(Blythetal. 1994 ;ClarkandArritt 1995 ;Bonan 2001 ;Carletonetal. 2001 ).Inparticular,manyofthesestudieshavedemonstratedthattheboundariesbetweendryandmoistsoils,and croplandandforestlandcovers,caninitiatemesoscale verticalaircirculationsthatincreaseconvectivecloud development.However,fewobservationalstudieshave investigatedtheresultingintensityoftheconvective precipitation. Spatio-temporaldifferencesinsoilmoisturecontent canaffecttheplanetaryboundarylayerdepthandother attributes,suchasitsstability,viathegenerationof “ non-classical ” (i.e.,non-seabreeze)mesoscalecirculations(NCMCs).Deepconvectionmaybeparticularly favoredoverandnearalternatingbandsofmoistand drysoils(YanandAnthes 1988 ;O ’ Neal 1996 ;Travis 1997 ;BrownandArnold 1998 ;Lynnetal. 1998 ;LeMone etal. 2007 ).ChenandAvissar( 1994a )explainedthatthe lowerBowenratiooverwettersoilsmayinducea differentialheatingratewithrespecttoadjacentdrysoils, givingrisetostrengthenedhorizontaltemperaturegradients.Giorgietal.( 1996 )statedthattheselocalscaleTheorApplClimatol(2010)99:315 – 330 DOI10.1007/s00704-009-0144-7 C.J.Matyas( ) DepartmentofGeography,UniversityofFlorida, 3141TurlingtonHall,P.O.Box117315,Gainesville,FL32611, USA e-mail:matyas@ufl.edu A.M.Carleton DepartmentofGeography,andEarthandEnvironmental SystemsInstitute,ThePennsylvaniaStateUniversity, UniversityPark,PA16802,USA

PAGE 2

latentandsensibleheatfluxesdifferentiatedbysurface typemaybeimportantinthedevelopmentorpersistence ofdroughtsandwetperiods.Atmosphericlatentheat increaseswithgreaterevaporationofmoisturefromwet soilsandprovidesadditional moistureforclouddevelopmentandprecipitation,aswellasgreaterconvective availablepotentialenergy.Alternately,whentheatmosphereissufficientlymoist,drysoilscancontributeto convectivecloudformationthroughupwardvertical motionduetoincreasedsurfacesensibleheat,orgreater Bowenratio(Rabinetal. 1990 ).Whentheatmospheric moisturecontentislow,areductioninsoilmoisturemay beassociatedwithincreasedsurfacetemperature,lower surfacepressure,andincreasedridgingaloft,whichtends toenhancethesurfacedryingasapositivefeedbackloop (OglesbyandErickson 1989 ). Severalstudiessuggestthatvegetationcoveralso enhancesconvectivedevelopment,especiallyoverotherwiseflatsurfaces,andmaypromoteprecipitation(e.g., Anthes 1984 ).ClarkandArritt( 1995 )foundthatamodeled vegetationcoverinitiatesconvectionduetoitshighersoil moisturecontentandreducesconductionofheatintothe soilduetoshadingofthesoilsurface.O ’ Neal( 1996 ) observedthatconvectivecloudsinthecentralUSAoccur morefrequentlyoverforestedregionsthanovercroplands. Blythetal.( 1994 )modeledtheboundarylayeroverforestcoveredsurfacesandfounda30%increaseinprecipitation amountscomparedwithbaresoil.Halfofthisincrease(i.e., 15%)wasfromthere-evaporationofprecipitation,or “ recycling ” ,whiletheotherhalfwasattributedtothe greaterroughnesslengthoftrees,whichleadstoenhanced verticalmotionofair.Carletonetal.( 2001 )investigated summertimeconvectiveclouddevelopmentintheUS Midwestassociatedwithcroplands,forests,andcropland/ forestboundariesandfoundevidenceofNCMCdevelopmentalongtheboundaries,givenrelativelylowwind speedsinthefreeatmosphere. Possiblybecauseoftheneedtoensurealackof topographicreliefoverwideareas,fewempiricalstudies havesoughttoestimatethefrequencyandintensityof convectionthatisenhancedbyquasi-linearheterogeneities insoilandvegetationtypes.Oncecomprisedofamixof forestsandgrasslands,thehumidlowlandsoftheMidwest USCornBelt(Fig. 1 )isonesuchregionofflatterrain.It comprisesanimportantrain-fedagriculturalregioninthe USAandisdominatedbycornandsoybeanproduction. Theheterogeneousmixtureofcroplandsandforests(Figs. 2 and 3 )causesmoisturegradientstoformneartheland surface.Moisturegradientsformwhereforestsborder croplandsasthecropsrequireandevapotranspirelarge amountsofwater,contrastedwiththesituationforthe remnantforests,especiallyinmidtolatesummer(Twineet al. 2004 ;Carletonetal. 2008b ).Accordingly,Adegokeand Carleton( 2000 ),amongothers,havefurtheridentifiedthe CornBeltasanidealregionforexamininginteractions betweenthelandsurfaceandtheatmosphere,becauseof theheterogeneousarrangementofsoil(HollingerandIsard 1994 )andvegetation(Bonan 2001 ).Theareaisalso subjecttointerannualprecipitationvariations,thedrought of1988andfloodof1993beingthetworecentextreme examples(Giorgietal. 1996 ).Severalresearchershavealso foundtheadjacentGreatPlainsregionofNorthAmericato beagoodlocationinwhichtoexaminelandsurface/ atmosphereinteractions(e.g.,Giorgietal. 1996 ;Baidya Royetal. 2003 ;KosterandSuarez 2004 ;Mahmoodand Hubbard 2004 ).However,theCornBeltismorehumidand Fig.1 ClimatedivisionscomprisingtheMidwestUSCornBelt whoseconvectiveactivitypatternsareexaminedinthecurrentstudy Fig.2 Exampleofheterogeneousarrangementofcroplandandforest vegetationintheMidwestUSCornBelt;inthiscase,south-eastIowa 316 C.J.Matyas,A.M.Carleton

PAGE 3

haslowerratesofevaporationincontrasttotheGreat Plains(Carletonetal. 2008a ). Thisstudyinvestigatesassociationsbetweenthefrequencyandintensityofconvectionandsoilandvegetation heterogeneitiesintheCornBeltregionattheclimate divisionscale.Weexaminespecificallythetwoconsecutive summerseasonsof1999and2000,usingcombinationsof radarreflectivityvalues,atmosphericsynopticdata,soil moisturedata,andlanduse/landcoverinformation.These twosummerspresentagoodopportunitytoexaminethe influenceofthelandsurfaceontheformationofconvective precipitationastheyfeaturedcontrastingsoilmoisture patterns.FromJunetoSeptember1999,dryconditions dominatedtheCornBeltregion.Climateforecastershad predictedthefollowingsummerof2000tobeevendrier, whichpromptedmanyfarmerstochangetheircrop schedules(Changnon 2002 ).However,thesomewhatdry soilconditionspresentinJune2000changedtonormaland evenmoistconditionsbytheendofSeptember2000,due tolargeamountsofprecipitation,especiallyinsouthern areas(Carletonetal. 2008a ). Becauseatmosphericdynamicsexertastrongcontrolon thedevelopmentofprecipitation(OglesbyandErickson 1989 ;FastandMcCorcle 1991 ;Blythetal. 1994 ;Clark andArritt 1995 ;O ’ Neal 1996 ),wefirstidentifydaysin summers1999and2000whenstrongsynopticscaleforcing wasabsent.Thisisundertakenusingtheapproachof Carletonetal.( 2008a ),whichinvolvesdeterminingthe categoryofmid-tropospheric(i.e.,500hPa)vectorwind speed( V (500))foreachdayoverthecentralpartofthe CornBelt( “ CentralCornBelt ” ).Wethenpresentthe compositeatmosphericconditionsprevalentwhenconvectionoccursonthissubsetofdays.Next,weperforma quantitativeanalysisofradarreflectivitydatafortheearly morningthroughearlyeveningperiodsondayshaving convection,todeterminethefrequencyandintensityof convectiveprecipitationwithintheCornBeltregion.For thispurpose,wedevelopaconvectiveseverityindex (CSI).TheCSIrepresentsanadvancementonthebinary andnon-spatialclassificatio nprocedureusedinCarleton etal.( 2008a b )byconsideringallcategoriesof convectiveintensityquantitatively.Ouranalysisalso improvesuponthatofCarletonetal.( 2008a b )inthat weanalyzesoil-typeboundariesinadditiontovegetationtype(forest,crop)boundarieswithintheCornBeltregion, fortheirinfluenceontheCSIvalues.Ourresultssuggest thatthemajorboundariesbetweendifferentsoiltypesand vegetationtypesenhancethedevelopmentofconvective precipitation,particularlyondaysofrelativelyweak backgroundwindspeed. 2Midweststudyregion TheCornBeltstudyregioncomprises32climate divisions(Fig. 1 ),includingallofIllinoisandIndiana, andadjacentareastothenorth,east,andwest.This geographicalareaincludestheCentralCornBeltsubregion(Carletonetal. 2008a b) ,aswellasadjacentareasof naturalforestedvegetation,andissimilartothelarger Midwestregionexaminedforlandsurface – atmosphere interactionsduringthe1988droughtbyCarletonetal. ( 1994 ).Ouruseofclimatedivisionsallowsadirect comparisonwiththeweeklyPalmerDroughtSeverity Index,whichweusetoinfersubregionalsoilmoisture conditionsduringthetwosummers,alongwithbiweekly measurementsofsoilmoistureforIllinois(Hollingerand Isard 1994 ).Weaggregatetheradarreflectivitydatatothe climatedivisionscaleinordertoexaminelandsurface – precipitationpatternsatlargermesoscales,assuggestedby theaforementionedstudies,andgivenageneralhypothesisthatconvectionwilldevelopoversubareasoftheCornBelthavingdifferencesinatmosphericconditions andsoilmoisturecontentbetweenthecontrastingsummersof1999and2000.Becausesoilmoisturecontentwas generallylowerduring1999,weexpecttheconvective precipitationoccurringtohavelowerrainrates,yielding lowerCSIvalues,comparedtosummer2000.Ineach year,theanalysisperiodspansJune15 – September30, whichistheseasonofgreatestconvectioninthisregion. Thefollowingtwosectionsdiscussthespatialdifferences Fig.3 CornBeltregionlanduse.Abruptboundariesoccurbetween agriculturalandforestedareaswhilemanyofthecroplandsaredotted withsmallforestedpatches RainfallassociationswithMidwestUSlandsurfaceconditions317

PAGE 4

insoilmoistureandatmosphericconditionsintheCorn Beltbetweenthetwosummers. 2.1Vegetationandsoilconditions Aspreviouslymentioned,themajorityofthestudy regionfeaturescropsasthedominantvegetationtype, butareasofforestsexistinmanyoftheclimate divisions.Toidentifywhichclimatedivisionshavethe greatestprobabilityofexperiencingenhancedconvective precipitationduetothepresenceofvegetation-type boundaries,wehaveplacedeachdivisionintooneof threegroups:(1)presenceofdenseforestsalongside croplands,(2)heterogeneousmixtureofforestsand croplands,and(3)lessthan10%forestcoverispresent (Table 1 ).Weexpectthehighestchanceofenhanced convectiveprecipitationtooccurwheredenseareasof forestsarelocatedalongsidecroplandstobothcreatea moisturegradientandwhereforestsmayenhanceupwards verticalmotion,suchaswithinclimatedivisionsIA3 (easternborder),WI7(middle),andIL6(westernborder). Climatedivisionsfeaturinglessthan10%forestcover, suchasIL5andIN3,mayhaveinsufficientforestcoverto producesignificantmoisturegradients.Anymoisture gradientsforminginclimatedivisionsthatfeature10 – 50%forestcover,butwhichhaveaspatialpatternwhere Table1 Characterizationofeachclimatedivisionaccordingtothespatialarrangementofvegetationtypesandsoiltypes Climate division Crop/forest boundary Mixedcrop/ forest Croplands dominate Soil-type boundary Alternatingsoil-type bands Dominantsoil type IA3XX IA6XXX IA9XX MO2X X WI7X X WI8XX WI9XX IL1XX IL2XX IL3XX IL4XX IL5XX IL6X X IL7XX IL8X X IL9X X IN1XX IN2XX IN3XX IN4X X IN5X X IN6XX IN7X X IN8X X IN9X X MI8X X MI9X X MI10XX OH1XX OH4XX OH8X X KY3X X Climatedivisionsinboldaremostlikelytoexperiencetheenhancementofconvectiveprecipitationduetothearrangementofvegetationandsoils 318 C.J.Matyas,A.M.Carleton

PAGE 5

croplandsandforestsareheterogeneouslymixed(e.g.,IL7 andIN9),mayfeaturemoisturegradientsthatdonotcover asufficientlylargeareatosignificantlyaffectthe developmentofconvectiveprecipitation. Asimilarclassificationofclimatedivisionsisperformed basedonthetypesofsoilandtheirspatialarrangement (Table 1 ).Alfisolsarethedominantsoiltypeinthestudy region,withmollisolscoveringmuchofthewesternarea (Fig. 4 ).Althoughbothsoilscanretainhighamountsof moisture,alfisolsthattypicallysupportforestedvegetation aregenerallymoremoistthanmollisols,whichtraditionally supportgrasslands.Inmanyareaswheremollisolsarethe dominantsoiltype,narrowbandsofalfisolsarealsopresent onscalesof20 – 30km.BrownandArnold( 1998 )and Travis( 1997 )suggestedthatthesealternatingbandsof alfisolsandmollisolshelppromotedeepconvectionin IllinoisandWisconsin.Therefore,weexpectheterogeneitiesinsoilmoisturetoincreaseconvectiveprecipitation developmentparticularlyinthenorthwesternportionofthe studyregion.Wehypothesizethatclimatedivisions featuringeitheraprominentboundarybetweentwosoil types(e.g.,IA3andIA6),orthathavealternatingbandsof differentsoiltypes(e.g.,IA6,IL6),haveahigherlikelihood ofinfluencingthedevelopmentofconvectiveprecipitation thanclimatedivisionscomprisedofasingledominantsoil type(e.g.,IL5andIN8;Table 1 ). WeutilizethePalmerDroughtSeverityIndex(PDSI; Alley 1984 )toindicatesoilmoistureataclimatedivision spatialscalefortheentirestudyregion.PDSIiscalculated weeklyfromstationprecipitationandtemperaturedata usingaformofthewaterbalanceequation.Differencesin soiltypemaynotinfluencethedevelopmentofconvective precipitationwhensoilswithinaclimatedivisionarewetter ordrierthannormalforthemajorityofthestudyperiod. Undertheseconditions,horizontalgradientsinsoilmoisturewillnotformasallsoilsintheregionmaybe uniformlywetordry.Whenprolongedperiodsofdry conditionsexistinaregion,evapotranspirationratesmay alsodecrease,whichinturnwoulddecreasetheamountof moistureavailableinthelowertropospheretoproduce convectiveprecipitation.ThroughoutJune1999,soilsare moistinthenorthwestpartofthestudyareabutdryinthe east(Fig. 5 a).MapsofweeklyPDSIvaluesshowthatthe westernhalfofthestudyareachangesfrommoisttodry duringthatsummer,whilethedryconditionsintheeastern regionintensifytoextremedrought(Fig. 5 b).Atthe beginningofsummer2000,mostoftheCornBelthas normaltomoderatelydrysoilmoisturelevelswithmoist conditionspersistinginthenorth(Fig. 5 c).Bytheendof thesummer,moistconditionsexistinthenorthernandthree ofthesoutherndivisions,whileclose-to-normalconditions prevailelsewhere(Fig. 5 d). 2.2Atmosphericconditions BecausestrongsynopticforcinglikelymasksorsuppressesNCMCs(FastandMcCorcle 1991 ),wecategorizethedayswithinthetwostudysummersaccordingto V (500)(O ’ Neal 1996 )intofiveclassesrangingfromweak tomoderatetomaximum.Theseclasses(Table 2 )are basedonthespatialrangeofwindspeedvaluesacrossthe CentralCornBeltassociatedw ithspecificcategoriesof V (300)forthesameregionclassifiedbyCarletonetal. ( 2008a ).Westratifyeachday ’ scombinedafternoonand eveningperiodcoincidingwiththetimeofmaximum surfaceheating(Tianetal. 2005 )intoconvectiveornonconvective,basedonaqualitativeevaluationofNext GenerationWeatherRadar(NEXRAD)compositeradar imagesobtainedonline( http://www4.ncdc.noaa.gov/cgiwin/wwcgi.dll?WWNEXRAD~Images2 ).Daysclassified asnon-convectiveeitherlackradarechoesorhavevery lowlevelreflectivityvaluesindicativeofprecipitationthat wouldnotreachtheground(i.e.,virga;Carletonetal. 2008a ).AsshowninCarletonetal.( 2008b ),theprimary determinantofwhetherconvectiondevelopsonagiven dayintheCentralCornBeltisthesignofverticalmotioninthefreeatmosphere:convection(no-convection)days typicallyhaveupward(downward)verticalmotionatthe 700-hPalevel. Thefrequencyofdaysexhibitingeachflowtypeandthe co-occurrenceofconvectiveactivityforbothsummers (Fig. 6 )issimilartotheresultsobtainedbyCarletonetal. ( 2008a ),whoutilized14fewerdaysforeachsummer ’ s Fig.4 Studyareaandsurroundingregionmapofsoiltypes illustratingprominentsoil-typeboundaries RainfallassociationswithMidwestUSlandsurfaceconditions319

PAGE 6

analysis(June15 – September15).Convectionismost enhancedontheprimaryflow-typedaysofweak,moderate, andmaximumin1999.Duringthesummerof2000, convectionoccursmostfrequentlyduringweak – moderate flowdays.Forthepresentanalysisofconvectiveintensity anditspossibleassociationswithlandsurfacesoiland vegetationconditions,wedonotstudythenon-convective days.Thesynopticatmosphericconditionsassociatedwith thenon-convectivedaytypesarepresentedinCarletonet al.( 2008a )andgenerallyhavereverseanomalies(e.g., positivevaluesofomega,negativevaluesofspecific humidity)tothoseondeepconvectiondays.Although Carletonetal.( 2008a )foundpossibleevidenceofland surfaceforcingevenonmoderate-maximumandmaximum days,whicharestronglycharacterizedbyatmospheric dynamics(e.g.,cyclonicvorticity,upper-airdivergence), thecurrentstudyconfinestheanalysistotheweakerflow days.Theconvectionthatformsonthestrongerflowdays tendstobemorewidespreadandorganizedandlastslonger thanthatformingduetoland -surfaceinteractionson weakerflowdays. Todeterminethesynopticatmosphericenvironments withinwhichdeepconvectionoccurredduringtheweaker synopticflowdays,wecompositetheNCEP/NCAR Fig.5 WeeklyPalmerDrought SeverityIndexmapsforthestart andendofeachstudysummer: a June12,1999; b Sept.30, 1999; c June17,2000; d Sept.23,2000 WeakWeak – moderateModerateModerate – maximumMaximum 5-114 – 1111 – 1710 – 1813 – 17 8.58.014.014.515.5 Table2 Rangeofwindspeeds at500hPaovertheCornBelt andaveragewindspeedforeach synopticclassificationinms 1 320 C.J.Matyas,A.M.Carleton

PAGE 7

reanalysisdailydata(Kalnayetal. 1996 )forseveral atmosphericvariables,separatelybyflowstrengthcategory andforeachsummer.Specifically,wecompositeanomalies ofthe u windcomponent(west – east)at500hPa, v wind component(south – north)at700hPa,temperaturesand geopotentialheightsatthe500,700,and850hPapressure levels;vectorwindspeedsandspecifichumidityatthe surfaceandat850hPa;sealevelpressure,surfacetemperatures,outgoinglongwaveradiation(OLR),andmodelgeneratedprecipitationrates.Theanomaliesarecalculated relativetothemeansfor1968 – 1996.Severalofthese variablesareadditionaltothoseexaminedbyCarletonetal. ( 2008a ).Weutilizethe2.52.5NCEP/NCARreanalysis datasetasitiscomparableinscaletotheclimatedivision radaranalysisperformedinthecurrentstudy.Weutilize OLRandprecipitationrateanomalydatatodetermine wherecumulonimbuscloudsweremostfrequentineach summerandwouldhaveproducedhighrainratesforeach atmosphericflowcomposite.Weexpectnegative(positive) OLRandpositive(negative)precipitationrateanomalies withinclimatedivisionshavinghigh(low)CSIvalues. Additionally,wecompositespecifichumidityanomaliesat 1,000and850hPaasweexpectCSItobehigherwhere specifichumidityanomaliesatboth1,000and850hPaare positive,thanwheremoistureiseitherabovenormalatonly onelevel,orbelownormal. 3Radaranalysisandtheconvectiveseverityindex Wedeterminethelocationandintensityofconvective precipitationwithinthestudyareaforthetwosummer seasonsbyutilizinghourl ycompositeimageryofthe NEXRADnetworkavailableonlinefromtheNational ClimaticDataCenterforthedailyperiodsofgreatest surfaceheating:from1200UTC(7:00AMLST)to 2300UTC(6:00PMLST).Althoughthespatial resolutionofthiscomposite dradarreflectivitydatais coarserthanthenativelevelIIreflectivitydatafrom whichitisderived,itstillhasahighertemporal resolutionthanraingaugemeasurements,allowingone totrackconvectiveprecipitationbothvisuallyand numerically.Itshouldbenotedthatasmallpercentage ofthehourlydataaremissing(seeCarletonetal. 2008a ),butaccommodationsformissingdataaremadein thequantitativeanalysisoftheradarreflectivityreturns (seebelow).Weutilizeacomputerprogramtoaccomplish severallogisticalstepsrelatedtotheanalysisofthe NEXRADdataandpreparatorytoderivationoftheCSI, asthisprovidesanobjectivee stimationoftheconvection present,ratherthanalessprecisevisualinspectionofthe data.First,theprogramautomaticallyqueriesandobtains thehourlyreflectivityco mpositesfromtheNCDCwebsite.Next,foreachindividualimage,theprogramcounts thepixelscontainingagivenradarreflectivitylevelin eachimagebyrecognizingtheHTMLcolorcodeforeach pixel(Table 3 ).Eachcoloredpixelisallocatedtoits climatedivisionbyreferencetoanoverlaymap,andits intensityisrecordedaccordingtofourcategoriesof reflectivity:low(5 – 15dBZ),medium – low(20 – 30dBZ), medium – high(35 – 45dBZ),andhigh(50+dBZ).Theuse ofthesefourcategoriespermitsseparationofheavy convection(i.e.,highestcategory)fromreflectivity valuesthatmightbeproducedbygroundclutterorvirgaFig.6 Graphshowingthenumberofdayswithandwithout convectionthatoccurredunder eachsynopticflowtypeduring thestudyperiod RainfallassociationswithMidwestUSlandsurfaceconditions321

PAGE 8

(i.e.,lowestcategory).Thenumbersofpixelsper intensityclassarestoredbycategoryforeachclimate divisioninageographicinformationsystem(GIS). TheCSIprovidesanareanormalizationprocedurefor radarreflectivityacrossallfourintensitycategories consideredhere.TheCSIpermitsareasofheavy convectiveprecipitationto bedistinguishedfromareas receivinglittleorlightprecipitation,suchaswouldbe expectedtofallfromstratiformclouds(cf.Alfierietal. 2008 ).Eachintensitycategoryisassignedadifferent weightbasedonitsreflectivityrangetoindicateits importanceintheCSI:thus,low(1),medium – low(10), medium – high(50),andhigh(100).Theweightsallowthe finalCSInumbertodistinguishlighterrainfall(lowCSI values)coveringlargeareasfromveryheavyrainfall(high CSIvalues),whichtypica llyismorelocalized.This weightingisimportantinthecurrentstudyasconvection developinginassociationwit hsmaller-scalefeatures,such assoilandvegetation-typeboundaries,maynotcoveras largeanareaasconvectionthatismoredynamically forced.Theweightedcategoryvaluesarethensummed andmultipliedbythepercentoftheclimatedivision havingradarreturnsofanyintensitytoobtainaCSIthat canbecomparedamongclimatedivisionsofdifferent sizes.AlthoughweemploytheC SIattheclimatedivision scaleinthepresentstudy,themethodcouldbeutilizedat morelocalizedspatialscalesinfutureinvestigations. Toincludedayswhenonly11ofthe12hourlyradar imagesareavailable,wecalculatetheaverageofthehourly CSIvaluessothateachclimatedivisionhasoneCSIfor eachdaywithinthestudyperiod.Althoughtheaveraged CSIvaluesforthesedaysmaybeslightlylower,thiswould onlybethecaseifconvectionwaspresent.Becausefewer than10%ofthedaysinthestudyperiodaremissingoneof theradarimagesandfewerthanhalfofthesedaysoccurred duringweakorweak – moderatesynopticforcing,the averagingshouldnothavesignificantlybiasedtheresults. ThisaveragingalsoallowsCSIvaluestobestratifiedbythe daily V (500)categoriespreviouslyestablished.Finally,we averageCSIvaluesamongdayswiththesamebackground flowstrengthforbothyearstoobtainacompositeseasonal value.Thelatterprocedureallowsustospatiallycompare theseverityofconvectionamongclimatedivisionsandto helpidentifygroupsofclimatedivisionswherelandsurface conditionsmighthaveinfluencedconvectionduringweaker backgroundflowinthetwosummers. 4Results Asexpected,deepconvectionoccursmorefrequentlyand overalargerareaasbackgroundwindspeedsincrease (Fig. 7 ).Despitetheseasonalclimatedifferencesbetween thetwosummers,ondayswhensynopticscaleforcingis weakorweak – moderate,similaritiesareevidentwhen consideringtheregionshavinghighestCSIvalues.The northwestcorner(westernportion)ofthestudyregion featuresthehighestCSIvaluesduringweak(weak – moderate)flowdays.Overall,thelocationsofhighCSI occurwherevegetationandsoil-typeboundariesexistand lowertroposphericmoisturecontentishigh.Relativelylow CSIvaluesexistovertheeasternportionofthestudyregion inbothsummers.Wediscussresultsspecifictoseveral climatedivisionsinSection 5 Ondayswhensynopticforcingisweakest,the V (500)maximumislocatedfarthestfromthestudyregion (Fig. 8 a,b).Duringsummer2000,specifichumidity, cloudcoverage,andprecipitationrateanomaliesare higherthannormalinthenorthwesterncornerofthestudy region,particularlyneartheIllinois/Iowa/Wisconsinborders(Fig. 8 d).Giventhatthelowertroposphereisvery moist,itisnotsurprisingthatthehighestCSIvaluesoccur inthisregion,wherecropland/forestandmollisol/alfisol boundariesalsoexist.Duringweakflowdaysinsummer 1999,specifichumidityvaluesinthelowertroposphere arebelownormalovermostofthestudyregion,andOLR valuesindicatethatcloudcoverageislessthannormal (Fig. 8 c).Underthesedrieratmosphericconditions,the boundariesbetweencroplandsandforestsareexpectedto beineffectiveatenhancingrainfallratesindicativeofdeep convectionand,indeed,CSIvaluesarelowerthan observedduringweakflowdaysin2000. Forbothsummers,ondaysofweak – moderatesynoptic forcing,theregionofhighest V (500)windanomaliesis locatednorthwestofthestudyregion(Fig. 9 a,b).Specific Table3 Colorcodesandtheircorrespondingradarreflectivityvalues oftheNEXRAD “ composite ” datausedtoquantifytheseverityof convectionwithinthestudyregion ’ sclimatedivisions ColorHTMLcolorcodeReflectivityvalue(dBZ) Cyan#00FFFF5 Skyblue#6698FF10 Lightblue#0000FF15 Pastelgreen#00FF0020 Green3#4CC41725 Green4#34801730 Yellow#FFFF0035 Gold1#FDD01740 Orange#FF804045 Red#FF000050 Red2#E41B1755 Burgundy#80000060 Pink#FF00FF65 Darkpurple#80008070 322 C.J.Matyas,A.M.Carleton

PAGE 9

humidityvaluesandcloudcoveragearehigherthannormal overthewesternpartsofthestudyregion,althoughthe maximumanomalyvaluesarecenteredoversouthwestern (northwestern)Illinoisin1999(2000;Fig. 9 c,d).Thepeak inspecifichumidityvaluesduring2000occursnearthe borderbetweenIllinoisandWisconsin(Fig. 9 d),andCSI valuesindicatethatthisareaexperiencesmoredeep convectionduring2000thanduring1999.Bycontrast, climatedivisionsinnortheastIndianaandnorthwestOhio haveverylowCSIonweakandweak – moderateflowdays duringbothsummers. CSIvaluesarerelativelylowovertheeasternportionof theCornBeltonweakandweak – moderateflowdaysin bothsummers.Tomorecloselyinvestigatethepossibility thatsoilmoistureplaysabiggerroleindeterminingthe amountofconvectionthatwilloccurinthisportionofthe Fig.7 Convectiveseverityindex(CSI)valuesforeachclimatedivision.Highervalues indicatethatdeepconvection occursmorefrequentlyina givenclimatedivision. a Weak flow1999, b weakflow2000, c weak – moderateflow1999, d weak – moderateflow2000, e moderateflow1999, f moderate flow2000 RainfallassociationswithMidwestUSlandsurfaceconditions323

PAGE 10

CornBelt,wecomparetheCSIvaluestosoilmoisturedata atthe0 – 10-cmdepthavailableeachdayfromthemodeled NCEP/NCARreanalysisdata.Afterestimatingtheaverage soilmoisturevalueforeachclimatedivision,werankthe soilmoistureandCSIvaluesforeachclimatedivisionfrom highesttolowestandcalculateSpearmanrankcorrelation coefficientsforeachrankedset.Weperformtheanalysisfor weakandweak – moderateflowdayscombinedforboth summerstoincreasethesamplesize.Manyoftheclimate divisionsinIndiana,aswellaseasternIllinoisandwestern Ohio,exhibitstatisticallysignificantpositivecorrelations, meaningthatCSIishighwhensoilmoistureisalsohigh (Table 4 ).Resultsarenotstatisticallysignificantoverthe westernportionofthestudyregion,whereCSIvaluesare higherbecausedeepconvectionismorefrequentand coversalargerareathantheeasternportionoftheregion. Higherhumidityinthelowertroposphereoverthewestern CornBeltmayplayalargerrolethansoilmoisturein aidingthedevelopmentofthedeepconvectionthere. Toconfirmthepossibilitythatrelativelydrysoil conditionsexistedalongsiderelativelymoistsoilconditions inpartsoftheCornBeltduringbothsummers,whichmight haveinfluencedthedevelopmentofconvectiveprecipitation,weobtainmeasurementsofsoilmoistureat0 – 10cm depthfrom17stationsintheIllinoisClimateNetwork. Neutronprobemeasurementsareavailablebiweeklyforthe Fig.8 Thegeographicalextents ofpeakanomalyvaluescompositedfordayswithweak synopticscaleforcing.Pressure, temperature,andmidtroposphericwindsareshownfor 1999( a )and2000( b ).Lower troposphericmoisture,winds, cloudcoverage,andprecipitationratesarefor1999( c )and 2000( d ). Symbolsutilizedinthe figures arelabeledin e 324 C.J.Matyas,A.M.Carleton

PAGE 11

summermonthsatseveraldepths(HollingerandIsard 1994 ).Comparingthesoilmoisturemeasurementsfor stationswithineachclimatedivisiontothePDSIvalues fortheweekthatthesoilmoisturedatawerecollected allowsustodeterminetherangeofsoilmoistureconditions thatexistforeachcategoryofthePDSI(severedrought( 3 to 3.9),moderatedrought( 2to 2.9),etc.).Wefindthat soilmoisturecanvaryonagivendayamonglocationsin thesameclimatedivisionandamongstationsinneighboringclimatedivisions,confirmingthatdifferencesinsoil moistureathighspatialresolutionsexistwithintheCorn Beltregion.Becauserelativelymoistanddrysoilscancoexistwithinasingleclimatedivision,alargerangeofsoil moistureconditionscanalsooccurforeachcategoryofthe PDSI(Table 5 ).EvenwhenPDSIvaluesindicatedry conditions,aclimatedivisionmaycontainbandsofmoist soilsthatmighthelptoinitiateNCMCspromoting convectiveprecipitation.Thesoilmoisturemeasurements alsoconfirmthatsoilmoisturewaslowerin1999thanin 2000foreachPDSIcategoryinIllinois. Comparingtheradar-deri vedCSIvalueswiththe compositemeanprecipitationmapspresentedinCarleton etal.( 2008a ,theirFig. 5 a,c),yieldsstrongsimilaritiesin thelocationsofdeepconvectionandhighprecipitation amounts,particularlyfor1999.Forweakflowdaysduring summer1999,meanprecipitationislowestinsoutheastern Iowa,northeasternMissouri,southernIllinois,Indiana, westernOhio,andsouthwesternMichigan.Carletonetal. ( 2008a )showthattheseregionsrecorded0 – 6.7mmof precipitation,andtheycorrespondtoCSIvaluesof9orless Fig.9 AsinFig. 8 fordays experiencingweak – moderate synopticforcing( a – d ) Table4 Spearmanrankcorrelationcoefficientsforrankedsoil moistureandCSIvaluesgroupedbysummerfordayshavingweak andweak – moderatesynopticforcing Weakandweak – moderate1999Weakandweak – moderate2000 Climatedivision rsClimatedivision rsIL8 0.550 IL2 0.412 IL9 0.442 IL5 0.518 IN2 0.386 IL7 0.432 IN4 0.507 IN1 0.445 IN5 0.584 IN4 0.455 IN7 0.448 IN5 0.600 IN8 0.583 IN8 0.429 IN9 0.530 IN9 0.411 MI10 0.417 OH1 0.471 KY3 0.554 OH4 0.381 KY3 0.528 Valuesinboldarestatisticallysignificantat =0.05,italicsindicate significanceat =0.1 RainfallassociationswithMidwestUSlandsurfaceconditions325

PAGE 12

inthepresentstudy.Slightlymoreprecipitationoccursin southeasternMichiganandnorthwesternIllinois(4.5 – 11.2mm),andtheCSIvaluesfortheseareastotal15 – 19. AsCarletonetal.( 2008a )combinedthemeanprecipitation forweakandweak – moderatedaysinsummer2000,direct comparisonsbetweenthetwostudiesarenoteasilymade. However,Carletonetal.( 2008a )showthatprecipitationon weakandweak – moderatedaysaveragedashighas 12.3mmineasternIowa,northeasternMissouri,and northernIllinois.TheseregionshadCSIvaluesgreater than25onweak – moderatedaysand10 – 24onweakflow days.ThehighaverageprecipitationamountsinIndiana andsoutheasternIllinoisindicatedbyCarletonetal. ( 2008a )forweakandweak – moderatedaysdonotagree withtheCSIvaluesthatwederiveinthepresentstudy.A closerinspectionoftheradarreflectivitydatabetween1200 and2300UTCforeachdayshowsthatclimatedivisions IL7,IN7,andIN8didnotexperiencereflectivityvaluesin excessofthoseoccurringinsurroundingclimatedivisions onanyweakorweak – moderatedayin2000.Because Carletonetal.( 2008a )utilizedailytotalprecipitationfrom cooperativestationraingauges,thehigherprecipitation amountscouldhaveoccurredpriortoorafterthe1200 – 2300UTCtimesthatthepresentstudyusestocalculatethe averagedCSIvalues. Synopticscaleatmosphericforcingofdeepconvection becomesmostevidentduringdaysclassifiedasmoderate flow,whenthe V (500)maximumwindsarelocated adjacenttothestudyregion(Fig. 10 a,b;cf.Carletonet al. 2008a ).Thus,itisnotsurprisingthatCSIvaluesalsoare highestforthesedays(Fig. 7 e,f).Relativelyfastsoutherly windsat850hPasuggestthatalow-leveljet(LLJ)is importantforthehighamountsoflowertropospheric moistureandinstabilityacrossthestudyregion(Bonner 1968 ;AndersonandArritt 2001 ;ZhuandLiang 2007 ), particularlyduring2000(Fig. 10 d).Windspeedsat 850hPaaretwiceasfastduringmoderateflowdaysof 2000thanduringmoderateflowdaysof1999.Alsoon moderateflowdaysinsummer2000contrastedwith1999, theentirestudyregionhasaboveaveragecloudcoverage, indicatedbybelownormalOLRvalues,andtheCSI confirmsthatwidespreaddeepconvectiondevelopsacross mostofthestudyregion,probablyasaresultofhigh moisturevaluesatboth1,000and850hPa(Fig. 10 d). Becausedayswhenmoderatelystrongsynopticscale atmosphericforcingexistsdonotprovideanadequate opportunitytodiscernwhe therthelandsurfacealso contributestothedevelopmentofconvectiveprecipitation, wedonotfurtherdiscusstheresultsforthemoderateflow days. 5Focusonspecificclimatedivisions InSection 2.1 ,weidentifiedtheclimatedivisionswhere land-surfaceforcingismostlikelytocontributetothe developmentofconvectiveprecipitationasthesedivisions featurebothadistinctboundarybetweencroplandsand forests,andboundariesbetweentwodifferentsoiltypes. Webasedourselectionofcandidateclimatedivisionson thespatialarrangementofvegetationandsoiltypespresent (Table 1 ).Thissectiondiscussestheresultsofouranalysis forseveralofthosedivisions. Oneclimatedivisionlikelytoexperiencelandsurface forcingonthedevelopmentofconvectiveprecipitationis IA3(Table 1 ).Adistinctboundarybetweenmollisolsand alfisolsrunsthroughthemiddleofthisdivision,andaforestedareaislocatedalongitseasternedgewhiletherest ofthedivisionisdominatedbycroplands.TheCSIvalues supportourhypothesisthatland-surfaceforcingofconvectiveprecipitationishighlylikelyinthisclimatedivision (Fig. 7 ).DivisionIA3hasthehighestCSIvalueamongall divisionsinthestudyregionunderweakflowconditionsin 2000,andhighCSIvaluesarealsopresentduringweak – moderateflowdaysin2000.Abovenormalspecific humidityinthelowertroposphere(Fig. 8 d)mayhave helpedpromotedeeperconvectionoverthisclimate divisionduringthesedaysin2000.However,theresults PDSIcategoryNumberofsoilmoistureobservationsAverageStandarddeviationRange 1999Data 2to2.9629.834.9922.23 – 34.98 1.9to1.912225.669.767.47 – 44.15 2to 2.91613.577.688.53 – 39.93 2000Data 3to3.9338.335.3133.64 – 44.09 2to2.9837.376.8227.41 – 46.02 1.9to1.912428.0715.2111.76 – 47.07 2to 2.9729.418.1618.94 – 33.46 3to 3.9231.396.5526.75 – 36.02 Table5 Average,standarddeviation,andrangesofsoil moisturewaterequivalentvalues (millimeter)inIllinoisforeach PDSIcategory 326 C.J.Matyas,A.M.Carleton

PAGE 13

alsoshowthatthesoilmoisturecontentasinferredfromthe PDSIisalsoanimportantconsideration.During1999, convectionislessintenseoverthisclimatedivisionthanin 2000.PDSIvaluesshowthatthisareaisunusuallyto extremelymoistthroughoutthesummerof1999,andthese uniformlymoistconditionsdonotallowstronggradientsin soilmoisturetodevelop.Thus,CSIvaluesforweakand weak – moderateflowdaysduring1999aresimilartoone anotherbutlowerthanthevaluesin2000duetothe uniformlymoistsoilconditions. Asimilarexplanationmayalsoapplytoclimatedivision WI7.Thedominantland-surfacefeaturesherearea boundarybetweenforestsinthenorthandcroplandsin thesouth,anddifferencesinsoiltypebetweenthenorth thatisdominatedbyalfisolsandthesouthwheremollisols alternatewithalfisols.LikeIA3,WI7experiencesnormal moistureconditionsformostof2000butisextremelymoist in1999.CSIvaluesarehigherin2000duringbothweak andweak – moderatedaysthanin1999(Fig. 7 ),likelydue toboththenear-normalsoilmoistureconditionsthatallow gradientsinsoilmoisturetodevelop,andtheanomalously highamountsofmoisturepresentinthelowertroposphere inthispartofthestudyregion(Fig. 8 d). Aclimatedivisionadjacent toIA3thatcontainssoiltypeboundariesbutexperiencesnormalsoilmoisture conditionsforthemajorityofbothsummersisIA6. Figure 2 ,takeninsummer2000,typifiesthelandscapein division(IA6).Aboundarybe tweenmollisolsandalfisols (easternportion)andalternatingbandsofthesesoilsboth existwithinIA6(Fig. 4 ).These20-km-widebandsmay helpsupplythelow-levelmoisturenecessarytoinitiate weakconvection,asBrownandArnold( 1998 )suggested forsimilarboundariesinIllinois.TheaveragedCSIvalues arethesameduringweakflowdaysin1999and2000and areslightlyhigherduringweak – moderateflowdaysin bothsummerslikelyduetohigherspecifichumidity valuesinthelowertroposph ere.ThesimilaritiesinCSI valuesforbothsummerssuggestthatgradientsinsoil moisturemayindeedformwithinthisclimatedivisionand helppromotethedevelopmentofconvectiveprecipitation. Asmorethan90%ofthisdivisioniscoveredby croplands,wedonotbelievethatdifferencesinvegetation typeplayedasignificantroleinenhancingconvective precipitation. Anotherclimatedivisionthatweidentifyasalikely candidateforeithersoilmoistureand/orvegetation-type Fig.10 AsinFig. 8 fordays experiencingmoderatesynoptic scaleforcing( a – d ) RainfallassociationswithMidwestUSlandsurfaceconditions327

PAGE 14

influencesonconvectiveprecipitationisIL6.Thisdivisionfeaturesadenselyforestedareaonitswesternborder andmixtureofforestsandcroplandsthroughoutmostof thedivision.Alfisols(mollisols)dominatethewestern (eastern)halfofthedivision,andareaswheresoiltypes alternateexistinthemiddleofthedivision.Soilmoisture conditionsarenormalduringmostof1999,whereasthey rangefromunusuallydrytoverymoistduring2000.CSI valuesareslightlyhigherin1999than2000duringweak flowdays,andareninepointshigherduringweak – moderatedaysin1999whenlowertroposphericspecific humidityisanomalouslyhighinthisarea.Whensoil moistureconditionsareuniformlybelownormal,gradients insoilmoisturemaynotexistintheregion.Prolonged dryconditionsalsomeanthattranspirationisreduced,so thatmoisturegradientsfromdifferentvegetationtypes alsoarelesslikelytooccur.Boththereducedtranspiration fromplantsandtheinabilityofsoilmoisturegradientsto formwhenconditionsarewetterthannormal,aswellas whentheyaredrierthannormal,mayhelpexplainwhy deeperconvectionformedovertheIL6divisionin1999 asopposedto2000. 6Discussion Thecurrentstudyfindsthatdifferencesinsoilmoisture acrossasoil-typeboundaryappeartoenhancedeep convection.Weidentifyseverallocationalexamplesin whichconvectionpreferentiallydevelopsoverregions wheremollisolsarethedominantsoiltype,yetwherethey aremixedwithsmaller-scalebandsofalfisols(e.g.,IA6, IL1,IL6).Intheirmodelingwork,YanandAnthes( 1988 ) andLynnetal.( 1998 )foundthatseabreeze-likefronts developedalongtheboundariesbetweenwetanddrysoil patches.Eventhoughthecirculationwasstrongestoverdry land,theadditionofmoisturealongtheboundaryofdry landwithmoistlandalloweddeepconvectiontoform, producinghighrainrates.Bothsetsofsimulationsfound thatconvectiverainfallwasheaviestnearalternating patchesofdryandmoistlandifthesearewiderthan 95km.However,theobservationalanalysesofO ’ Neal ( 1996 )fortheMidwestsuggestthatpatchesofdifferentsoil andvegetationcomplexescanaffectconvectivecloud developmentassmallasonly10kmwide.Thebandsof alfisolswithinthemollisolsofeasternIowaandwestern Illinoisareapproximately15 – 30kmwide(Fig. 4 ).Our findingthatheterogeneitiesinsoiltypescoincidewiththe greaterCSIvaluesintheseportionsofthestudyregion supportstheworkofpreviousresearchers,includingthatof BrownandArnold( 1998 )forIllinois. Additionally,therearelocationalexamplesinthetwo studysummerswhereconvectiveprecipitationisenhanced overclimatedivisionsthatfeatureamarkedboundary betweenforestsandcroplands(e.g.,IA3,WI7,IL6),rather thanoverregionswherecroplandsaloneoccur(e.g.,IN3). ThisfindingalsosupportstheworkofO ’ Neal( 1996 ),who observedthatforestedareashavemoreconvectivecloud coveragethancroplands.Physicalandphysiologicaldifferencesinvegetationtypesinfluencethetransferofmoisture, heat,andmomentumfromthelandsurfaceintothe atmosphere.Moisturegradientsaremaximizedduringthe earlysummerwhencropsareimmatureandtranspiring littlemoistureintotheatmosphere,therebycreatinga strongsurfacemoisturegradient,whileadjacenttreeshave highertranspirationrates.Whencropsmatureinmidtolate summer,theyrapidlydepletesoilmoisture,whichisthen transferredtotheatmospherethroughevapotranspiration (cf.Bonan 2001 ;Carletonetal. 2008b ).Throughtheir deeperroots,forestsalsodepletesoilmoisture,butahigh stomatalresistancemeansthattranspirationratesare comparativelylowerinthemidtolatesummerfromtrees thanfromcrops.Blythetal.( 1994 )foundthatforestsmay enhancerainfallduetothere-evaporationofintercepted rainfall,whileCarletonetal.( 2008b )addthatdifferencesin evapotranspirationratesofforestsascomparedtoimmature (mature)cropsthatproducelow(high)evapotranspiration ratescanalsoenhancerainfall.Thechangeinroughness lengthbetweencroplandsandadjacentforestsaffectsboth themomentumandconvectivefluxes.Ourpresentfindings supporttheseearlierstudies:asinferredfromthespatial associationsofland-surfaceconditionswiththeCSIvalues, convectiveprecipitationisenhancedbythepresenceofa heterogeneousland-surfacecoverwithinagivenclimate division. Thecurrentstudyalsofindsthattheamountofmoisture presentinthelowertroposphereseemstohavethelargest influenceontherangeofCSIvaluesmeasured,atleast,for thetwocontrastingsummersexaminedhere.Whenspecific humidityvaluesinthelowertroposphereareabovenormal, higherCSIvaluesoccurfortheclimatedivisionsthatare co-locatedwiththeanomalouslyhighmoisture.Whenthe atmosphereismoist,surfacemoisturefromsaturatedsoils andforestvegetationraisesthelatentheatflux;Segaletal. ( 1995 )foundthatthisincreasedlatentheatfluxrapidly destabilizestheatmosphereandleadstodeepconvection. Thus,conditionsatboththelandsurfaceandinthelower tropospherecancombinetoproduceheavyrainfall. 7Conclusionsandfutureresearch Thisstudy ’ sgoalwastoidentifywhetherland-surfaceboundariessuchasthealternatingmollisolsandalfisolsin easternIowaandnorthwesternIllinois,andheterogeneities invegetationtypesuchascropforestboundariesin 328 C.J.Matyas,A.M.Carleton

PAGE 15

southwesternWisconsin,couldtogetherorseparately enhanceconvectiveprecipitationwithintheUSCornBelt duringtworecentsummershavingdissimilaratmospheric andsurfacemoistureconditions.Becauseweakerflowdays areanindicatorofminimalsynopticscaleforcing,they presentthebestopportunitiestoascertainwhetherlandsurfacemesoscaleheterogeneitiesenhancethedevelopment ofdeepconvectionwithinthestudyregion.Wefirst identifieddaysduringthesummersof1999and2000 whensynopticscaleatmosphericforcingwasweakto moderate.Weexaminedthecompositeconditionsinthe midandlowertroposphereoneachofthesesetsofdaysto identifypropertiesthatcouldaidorinhibitthedevelopment ofconvectiveprecipitation,suchasthepresenceof anomalousatmosphericmoistureoraLLJ.UsingNEXRADradarreflectivityreturns,wethendevelopedanindex todeterminequantitativelytheseverityofconvective precipitationataclimatedivisionscale.Weaveragedthe CSIvaluestemporallytouncoverwherethedeepest convectionoccurredmostfrequentlyforagivenbackgroundwindspeedcategory. Ourresultsindicatethatunderthetwoweakestsynoptic flowregimesinthetwosummersstudied(weak,weak – moderate),theboundariesbetweenrelativelymoistanddry soilsandbetweencroplandandforestlandcoversboth appeartoenhancethedevelopmentofconvectiveprecipitationinthewesternportionoftheCornBelt.Wefindthat climatedivisionsIA3,WI7,IA6,andIL6havethestrongest evidenceoflandsurfaceforcingonthedevelopmentof convectiveprecipitationduetothepresenceofsoil-type and/orvegetation-typeboundaries.Lowertropospheric moistureisakeyfactorindeterminingwhichlandsurfaces enhanceconvectiononagivenday.Whenhumidityvalues at1,000and850hPaareanomalouslyhigh,convection formsovermoistland,particularlywhereforestsgrowatop alfisol-typesoils.Becausealfisolsbetterretainmoisture thanmollisols,theylikelyprovidemorelatentheattothe atmosphere.However,whereandwhensoilmoisturevalues areanomalouslymoistordry,gradientsinsoilmoisture cannotformandthusdonotenhancethedevelopmentof convectiveprecipitation.Lessconvectionoccursoverallin theeasternCornBeltinbothsummers.Ondayswhenthe CornBeltexperiencesatmosphericforcingofatleast moderatestrength,convectivedevelopmentisenhanced byaLLJ(seealsoCarletonetal. 2008a ). Theseobservationalfindingssupportthoseofprevious researchers,especiallymodel-generatedresults.Additional examinationoftheroleofland-surfaceheterogeneitiesin enhancingconvectiveprecipitationwithintheUSCornBelt foragreaternumberofsummerseasonsclearlyis warranted.Futureresearchshouldinvestigatecropland/ forestandalfisol/mollisolboundariesinIA3,WI7,IA6,and IL6suggestedbythisstudyasbeingimportantforthe developmentofconvectionathigherspatialresolutions. Theanalysisundertakeninthecurrentstudyutilizeda climatedivisionspatialscaletoaccommodatethePDSI valuesalongwiththeCSIandlandsurfaceconditions. Becausethetimingofconvectivedevelopmentisatleast partlydependentontheunderlyingsurface,thetimeofday atwhichconvectivedevelopmentbeginsorreachespeak intensityshouldalsobeexamined.Allocatingthedevelopmentofconvectiveprecipitationtoaparticularland-surface boundaryandtimeofdaywillaidprecipitationforecastsin theCornBeltandprovideamorepreciseassessmentofthe amountofprecipitationthatheterogeneitiesintheland surfacecaninduceundervaryingatmosphericconditions. AfuturestudyfortheCornBeltregionshouldalso examinethelead-lagassociationbetweensoilmoistureand rainfall.FindellandEltahir( 2003 )identifiedtheroleofsoil moistureonthedevelopmentofconvectionoverIllinois andfoundthatwetsoilscanhelptriggerdeepconvection. Thisenhancedconvection,inturn,moistensthesoils.Fora largernumberofsummerseasonsthanexaminedhere, rainfalldatashouldbeexaminedfordaysaftersoilmoisture valuesbecomehightodeterminehowoftenrainfalldevelops inresponsetowetsoilconditions.Thisinformationcouldbe combinedwithresultsgeneratedbymodeledsoilmoisture fieldstofurtherinvestigatethelead-lagassociationbetween convectiveprecipitationandsoilmoisture. Finally,theCSIdevelopedinthisstudytohelpseparate deepconvectionfromlighterrainfallcouldusefullybe appliedtothequantitativeanalysisofradarreflectivitydata inGISframeworksforadditionalclimateresearchprojects. TheCSIindexprovidedimportantinsightsintowhich climatedivisionscomprisingtheMidwestUSCornBelt experiencedthemostfrequentandintenseconvective precipitationforthetwocont rastingsummerseasons. BecausetheCSIcanbecalculatedforradarreflectivity valuesatvariousspatialscales,employingthistechnique withlevelIIbasereflectivityradardatashouldyieldvalues ataspatialresolutionof2km2andatemporalresolutionof 5or6min.Aggregatingradar-derivedconvectiveactivity overlongertemporalscalessuchasweeksormonthswould facilitatehigher-resolutioncomparisonsofconvectiverainfallratesformultipleseasonsandacrossmultipleyears.Acknowledgments ThisworkwassupportedbytheNational ScienceFoundationATM98-76753.ThefirstauthorthanksAndy Sherwoodforhisassistanceinwritingthecomputerscriptsutilizedfor thisproject.Wealsothanktwoanonymousreviewersfortheirhelpful commentsthatimprovedthismanuscript.ReferencesAdegokeJO,CarletonAM(2000)Warmseasonlandsurface – climateinteractionsintheUnitedStatesMidwestfrom RainfallassociationswithMidwestUSlandsurfaceconditions329

PAGE 16

mesoscaleobservations.In:KnivetonSJMADR(ed)Linking climatechangetolandsurfacechange.Kluwer,TheNetherlands,pp83 – 97 AlfieriL,ClapsP,D ’ OdoricoP,LajoF,OverTM(2008)Ananalysis ofthesoilmoisturefeedbackonconvectiveandstratiform precipitation.JHydrometeorol9:280 – 291 AlleyWM(1984)ThePalmerDroughtSeverityIndex:limitationsand assumptions.JClimApplMeteorol23:1100 – 1109 AndersonCJ,ArrittRW(2001)RepresentationofsummertimelowleveljetsinthecentralUnitedStatesbytheNCEP-NCAR reanalysis.JClim14:234 – 247 AnthesRA(1984)Enhancementofconvectiveprecipitationby mesoscalevariationsinvegetativecoveringinsemiaridregions. JClimApplMeteorol23:541 – 554 BaidyaRoyS,HurttGC,WeaverCP,PacalaSW(2003)Impactof historicallandcoverchangeontheJulyclimateoftheUnited States.JGeophysResAtmos108(D24):4793 BlythEM,DolmanAJ,NoilhanJ(1994)Theeffectofforeston mesoscalerainfall — anexamplefromHAPEX-MOBILHY.J ApplMeteorol33:445 – 454 BonanGB(2001)Observationalevidenceforreductionofdaily maximumtemperaturebycroplandsintheMidwestUnited States.JClim14:2430 – 2442 BonnerWD(1968)Climatologyofthelowleveljet.MonWeather Rev96:833 – 850 BrownME,ArnoldDL(1998)Land-surface – atmosphereinteractions associatedwithdeepconvectioninIllinois.IntJClimatol 18:1637 – 1653 CarletonAM,AdegokeJ,AllardJ,ArnoldDL,TravisDJ(2001) Summerseasonlandcover — convectivecloudassociationsfor theMidwestU.S. “ Cornbelt ” .GeophysResLett28:1679 – 1682 CarletonAM,ArnoldDL,TravisDJ,CurranS,AdegokeJO(2008a) Synopticcirculationandlandsurfaceinfluencesonconvectionin theMidwestU.S. “ CornBelt ” ,summers1999and2000.PartI: Compositesynopticenvironments.JClim21:3389 – 3415 CarletonAM,TravisD,ArnoldD,BrinegarR,JelinskiDE,Easterling DR(1994)Climatic-scalevegetation – cloudinteractionsduring droughtusingsatellitedata.IntJClimatol14:593 – 623 CarletonAM,TravisDJ,AdegokeJ,ArnoldDL,CurranS(2008b) Synopticcirculationandlandsurfaceinfluencesonconvectionin theMidwestU.S.CornBelt,summers1999and2000.PartII: Roleofvegetationboundaries.JClim21:3617 – 3641 ChangnonSA(2002)ImpactsoftheMidwesterndroughtforecastsof 2000.JApplMeteorol41:1042 – 1052 ChenF,AvissarR(1994a)Impactofland-surfacemoisturevariability onlocalshallowconvectivecumulusandprecipitationinlargescalemodels.JApplMeteorol33:1382 – 1401 ChenF,AvissarR(1994b)Theimpactofland-surfacewetness heterogeneityonmesoscaleheatfluxes.JApplMeteorol 33:1323 – 1340 ClarkCA,ArrittRW(1995)Numericalsimulationsoftheeffectof soil-moistureandvegetationcoveronthedevelopmentofdeep convection.JApplMeteorol34:2029 – 2045 FastJD,McCorcleMD(1991)Theeffectofheterogeneoussoil-moistureonasummerbarocliniccirculationintheCentral United-States.MonWeatherRev119:2140 – 2167 FindellKL,EltahirEAB(2003)Atmosphericcontrolsonsoil moisture-boundarylayerinteractions.PartI:Frameworkdevelopment.JHydrometeorol4:552 – 569 GiorgiF,MearnsLO,ShieldsC,MayerL(1996)Aregionalmodel studyoftheimportanceoflocalversusremotecontrolsofthe 1988droughtandthe1993floodovertheCentralUnitedStates. JClim9:1150 – 1162 HollingerSE,IsardSA(1994)Asoil-moistureclimatologyofIllinois. JClim7:822 – 833 KalnayE,KanamitsuM,KistlerR,CollinsW,DeavenD,GandinL, IredellM,SahaS,WhiteG,WoollenJ,ZhuY,ChelliahM, EbisuzakiW,HigginsW,JanowiakJ,MoKC,RopelewskiC, WangJ,LeetmaaA,ReynoldsR,JenneR,JosephD(1996)The NCEP/NCAR40-yearreanalysisproject.BullAmMeteorolSoc 77:437 – 471 KosterRD,SuarezMJ(2004)Suggestionsintheobservationalrecord ofland – atmospherefeedbackoperatingatseasonaltimescales.J Hydrometeorol5:567 – 572 LeMoneMA,ChenF,AlfieriJG,TewariM,GeertsB,MiaoQ, GrossmanRL,CoulterRL(2007)Influenceoflandcoverand soilmoistureonthehorizontaldistributionofsensibleandlatent heatfluxesinsoutheastKansasduringIHOP_2002andCASES97.JHydrometeorol8:68 – 87 LynnBH,TaoWK,WetzelPJ(1998)Astudyoflandscape-generated deepmoistconvection.MonWeatherRev126:928 – 942 MahmoodR,HubbardKG(2004)Ananalysisofsimulatedlong-term soilmoisturedataforthreelandusesundercontrastinghydroclimaticconditionsintheNorthernGreatPlains.JHydrometeorol5:160 – 179 O ’ NealM(1996)Interactionsbetweenlandcoverandconvective cloudcoverovermidwesternNorthAmericadetectedfrom GOESsatellitedata.IntJRemoteSens17:1149 – 1181 OglesbyRJ,EricksonDJ(1989)Soil-moistureandthepersistenceof North-Americandrought.JClim2:1362 – 1380 RabinRM,StadlerS,WetzelPJ,StensrudDJ,GregoryM(1990) Observedeffectsoflandscapevariabilityonconvectiveclouds. BullAmMeteorolSoc71:272 – 280 SegalM,ArrittRW,ClarkC,RabinR,BrownJ(1995)Scaling evaluationoftheeffectofsurfacecharacteristicsonpotentialfor deepconvectionoveruniformterrain.MonWeatherRev 123:383 – 400 TianBJ,HeldIM,LauNC,SodenBJ(2005)Diurnalcycleof summertimedeepconvectionoverNorthAmerica:asatellite perspective.JGeophysResAtmos110:12 TravisD(1997)AninvestigationofWisconsin ’ santhropogenicallygeneratedconvergenceboundaryandpossibleinfluenceson climate.WisGeogr12:34 – 46 TwineTE,KucharikCJ,FoleyJA(2004)Effectsoflandcoverchange ontheenergyandwaterbalanceoftheMississippiRiverbasin.J Hydrometeorol5:640 – 655 YanH,AnthesRA(1988)Theeffectofvariationsinsurface moistureonmesoscalecirculations.MonWeatherRev116:192 – 208 ZhuJH,LiangXZ(2007)RegionalclimatemodelsimulationsofUS precipitationandsurfaceairtemperatureduring1982 – 2002: interannualvariation.JClim20:218 – 232 330 C.J.Matyas,A.M.Carleton