( ) Barnes Fengetal [13] ( 10km 10km ) [14] 1 1 ; Barnes ; [15] AUSPLI ;Chenetal [16]

Size: px

Start display at page:

Download "( ) Barnes Fengetal [13] ( 10km 10km ) [14] 1 1 ; Barnes ; [15] AUSPLI ;Chenetal [16]"

商岁姚
5 years ago
Views:

1 PLATEAU METEOROLOGY Vol.30 o.6 December [J] (6): * 3 3 ( ; ; ) : Barnes ( 10km 10km) :(1) ( ) mm d mm d % 4.5mm d % 93.12% (2) (3) : ; (4) : 7 12 ; (11 ) (8 ) : ; ; ; : (2011) :P468.0 :A 1 [9-12] [1] [2] [3] [4-8] : ; : : 973 (2010CB CB833503); ( ) : (1968-) xiongqf@cma.gov.cn * :. huangm@igsnrr.ac.cn

1616 30 ( ) Barnes Fengetal [13] 1951 2000 0.1 0.1 ( 10km 10km ) [14] 1 1 ; Barnes 1971 2005 0.1 0.1 ; [15] AUSPLI ;Chenetal [16] 1951 2005 753 7 2 0.5 2.

2 ( ) Barnes Fengetal [13] ( 10km 10km ) [14] 1 1 ; Barnes ; [15] AUSPLI ;Chenetal [16] :00~08:0024h [17] <0.5% [25] 1 [18-24] 1 ( ) ( :m) 1: 2: 3: 4: 5: 6: 7: 8: Fig.1 Distributionsofobservationstations(blackdot)and topography(colorareaunit:m)inchinaaswelasthe partitionofthesub-domains.1:ortheastchinaregion 2:Xinjiangregion3:EastregionoforthwestChina 4:orthChinaregion5:Middleandsouthregionof TibetanPlateau6:EastregionofSouthwestChina :Mid-andlower-reachesofYangtzeRiver8:SouthChina

3 6 : 1617 (λ0 φ 0) (λg ) φg 20~40km; : r g o = 50~100km 4km a [ 1 5km 2 ( cosφg +cosφ o )(λg -λo)] 2 + ( φ g -φ ) 2 槡 o 2.2 (2) [26-34] :a Jonesetal [28] 3 : Barnes [35] ; (IDW) [36] ;Cressman [37] Barnes [32] [38] ; [39] [40] ; Chenetal [16] (2) 20~40km10km 0.5% >70km 1% 0.1 [14] Barnes 2.3 [14] [14] : (1) 20 Barnes X e 5% 95% (2) X e>0.0 5% X e 0.0 (3) 20 X m X m 95% R g : 95% R g = Xe X m. (1) [28] X e M i=1

4 % : (1) : COR = 1 [X o (t)- 珡 X o ][X e (t)- 珡 X e (t)]. [X o (t)- 珡 2 X o (t)] [X e (t)- 珡 2 X e (t)] 5% (3) (2) : 20 MBE = 1 [X e (t)-x o (t)]. (4) (3) : MAE = 1 X 1 e (t)-x o (t). (5) (4) : 2 2 X e (t)-x o (t) RMAE = 100%. (6) X o (t) RMSE = [X e (t)-x o (t)] Barnes 槡 2. (7) (6) : 10 σ(o)= 1 [X o (t)- 珡 X o (t)] 槡 2. (8) (7) : % [X e (t)-x o (t)] 2 CRE = 100%. (9) [X o (t)- 珡 X o (t)] (8) : 3 σ(i)= 1 [X e (t)- 珡 X e (t)] 槡 2. (10) (9) : 2.4 [X e (t)- 珡 X e (t)] 2 RVAR = 100%. (11) [X o (t)- 珡 X o (t)] 2 ; 9 槡 X o X e 槡 (5) : 1

5 6 : mm d % [16] (7.22 mm d -1 ) (7.94mm d -1 ) 93.12% ; mm d -1 ; mm d mm d mm d mm d (3)~(11) 9 mm d mm d mm d % [16] %; [41] % 45.28% ; mm d -1 [16] IDW2 1 Table1 Averagestatisticsofalobservationstationsincros-validation (3~5 ) (6~8 ) (9~11 ) (12~2 ) /mm /mm /% /mm /mm /% /mm /%

6 (a) :0~50mm(b) :50~200mm Fig.2 Thedistributionsofprobabilitydensityfunction(PDF)fortheobservedandinterpolated precipitationincross-validationatalstationsofchinaintheobservationperiod.(a)daily precipitationfor0~50mm(b)dailyprecipitationfor50~200mm ; [42] ±1 mm 0mm 1mm (0.30) (0.18) (0.17) ( 2) (0.16);1mm <2 mm 3 2~4mm (0.10) -0.2~0.2mm ; 2~30mm (0.09) -4mm (0.06) ; >30mm >50mm % ~80% Barnes ( 4) 4a 0~5 mm d -1 2~30mm 3.3 ( ) <0.5 mm d -1 ; 0.5~1 mm d -1 ; ( 3) 1~1.5mm d -1 ;

6 : 1621 2mm d -1 ; 2~4mm d -1 ; 4~8mm d -1 ; 3 8mm d -1 16 mm d -1 Fig.3 PDFdistributionsofinterpolationerrorin cross-validationatalstationsineachseason andwholeyear 3 ~4 ; [43] 1.

7 6 : mm d -1 ; 2~4mm d -1 ; 4~8mm d -1 ; 3 8mm d mm d -1 Fig.3 PDFdistributionsofinterpolationerrorin cross-validationatalstationsineachseason andwholeyear 3 ~4 ; [43] 1.5~3mm d -1 ; 3-5 mm d -1 4c < ( 4b) 50% 105 E 4 (a) ( :mm d -1 )(b) ( :mm d -1 )(c) ( :%)(d) Fig.4 Thedistributionoferrorstatisticofdailygriddedprecipitationthroughcross-validation. (a)meanabsoluteerror(unit:mm d -1 )(b)rootmeansquareerror(unit:mm d -1 ) (c)meanrelativeerror(unit:%)(d)correlationcoeficient

8 (RMSE) 5 0mm 1 7 > 3.41mm 8.00mm; 8 100% 4d E >0.8 <0.6 [25] : ; 0.6~0.8 ( 1) ; mm 0.55 mm 8 5 (MBE) (MAE) 5 (MBE) (MAE) (RMSE) (a) (b) (c) (d) (e) (f) (g) (h) (i) Fig.5 Themonthlychangeofmeanbiaserror(MBE)meanabsoluteerror(MAE)androotmeansquareerror(RMSE) inprecipitationcross-validationinchinaandsub-regions.(a)china(b)ortheastchinaregion(c)xinjiangregion (d)eastregionoforthwestchina(e)orthchinaregion(f)middleandsouthregionoftibetanplateau (g)eastregionofsouthwestchina(h)mid-andlower-reachesofyangtzeriver(i)southchina

9 6 : 1623 [44] ; 6~7 7 : (1) mm % 4.5 mm d % (7.22mm d -1 ) 12 (7.94 mm d -1 ) 93.12% (2) ; ; ; 2mm >30mm d mm d -1 2~30 mm (3) : ; 8 (0.80) (4) : ; ; 11 8 ; ; [43] 1.1mm d ; [21] mm d -1 ; 4.5 mm d % [43] 42.09% 6 Fig.6 Themonthlychangeofcorrelationcoeficientsof precipitationincross-validationinchinaandsub-regions 2400 [43] 10000

10 [21] 8000 Climae200424(7): [14] (6): [17] [40] (3): [18] [J] [19] [1] GheliALalaureteF.Verifyingprecipitationforecastsusingup- scaledobservations[j].ecmwf ewsleter200087:9-17. [2].GRAPES Meso [J].J Geophys Res D04110doi: / [J] (2): JD [3] HongYix H AHutchinson M Fetal.Spatialinterpola- tionofmonthly meanclimatedataforchina[j].interjcli- mat200525(10): [4] MitchelK ELohmannDHouserPRetal.Themulti-in- stitutionnorth Americanlanddataassimilationsystem (L- DAS):UtilizingmultipleGCIPproductsandpartnersinacon- tinentaldistributedhydrologicalmodelingsystem[j].jgeo- physres d07s90doi: /2003jd [5] HufmanGJAdlerRFMorrissey M Metal.Globalpre- cipitationatone-degreedailyresolutionfrom multisateliteob- servations[j].jhydromet20012(1): cipitationforthewestern UnitedStates[J].JApplMeteor (9): [6] EischeidJKPasterisPADiazH Fetal.Creatingaserial- lycompletenationaldailytimeseriesoftemperatureandpre- [7] SilvaV BSKouskyV EShiWetal.Animprovedgridded historicaldailyprecipitationanalysisforbrazil[j].jhydrom- et20078(4): [8] BrantLDaveA.DailyprecipitationgridsforsouthAmerica [J].BulAmerMeteorSoc200586(11): [9].ApplMeteor200140: [J] (1): [10].X [J] (1): [J] (2): [32]BarnesSL.ApplicationsofBarnesobjectiveanalysisscheme [11]. Z- PartΙ:Efectsofundersamplingwavepositionandstation R [J] (4): randomness[j].jatmosoceantech199411: [12].BP [33]. [J] (4): logicaldatainchina :anewdataset[j].interj 0.1 [J] (1): [15]. AUSPLI [J] (2):7-15. [16]ChenDLOuT HGongLBetal.Spatialinterpolationof dailyprecipitationinchina: [j].advatmossci. [J].. MODIS (2):. [J] (6): [13]FengSHuQQian W H.Qualitycontrolofdailymeteoro- [20]DalyC.Guidelinesforassessingthesuitabilityofspatialcli- matedatasets[j].interjclimat200626: niquesforgauge-basedanalysesofglobaldaily precipitation [22] [J] ( ): [23]. [J] (5): [24] [J] (3): [25]. [Z]. : 2006:1-72. [26]AhrensB.Distanceinspatialinterpolationofdailyraingauge data[j].hydrologyearthsystemsci200610: [27]. [J] (1): [28]JonesD AWang WFawcetR.High-qualityspatialclimate data-setsforaustralia[j].aust MeteorOceanJ200958: [29]. [J] (2): [21]Chen M YShiWXiePPetal.Assessingobjectivetech- [30]ShenSSPDzikowskiPLiGLetal.Interpolationof dailytemperatureandprecipitationclimatedataontoAl- bertapolygonsofecodistrictandsoillandscapeofcanada[j].j [31]. GIS [J] (3):

11 6 : 1625 [35]ThiessenA H.Precipitationaverageforlargeareas[J].Mon WeaRev191139: [36]FrankeR.Scatereddatainterpolation:Testofsomemethods [J].MathematicComputa198233: [37]CressmanG P.Anoperationalobjectiveanalysissystem[J]. Mon WeaRev195987: [38]Hutchinson M F.Interpolating meanrainfalusingthinplate smoothingsplines[j].interjgeograpinformsystems1995 9(4): onfordailyhydrologicmodelingin mountainousregions[j]. WaterResBul199430: ousterrain[j].japplmeteor199433(2): [41]XiePPYatagaiAChen M Yetal.Agauge-basedanalysis ofdailyprecipitationovereastasia[j].jhydrometeorology 20078(3): [42]. [J] (2): [43]Hofstra Haylock Mew Metal.Comparisonofsix methodsfortheinterpolationofdailyeuropeanclimatedata [J].J Geophys Res D21110doi: / 2008JD [44]. [J] (3): Cross-ValidationErrorAnalysisofDailyGriddedPrecipitation BasedonChinaMeteorologicalObservation XIOG Qiu-fen 1 HUAG Mei 2 XIOG Min-quan 3 HUJiang-lin 3 (1.China Meteorological Administration Training CenterBeijing100081China;2.Instituteof GeographicSciencesand atural Resources ResearchChinese Academyof ScienceBeijing100101China;3.ational MeteorologicalCenterBeijing100081China) matedmeansquaredeviationσ(i)therelativevarianceratio(rvar)andtheprobabilitydensityfunction [39]GarenDCJohnsonGLHansonCL.Meanarealprecipitati- [34]. [40]DalyCeilsonR PPhilipsD L.Astatistical-topographic [J] (5): modelformappingclimatologicalprecipitationovermountain- Abstract:Basedondailyprecipitationof2403gaugeobservationsinChinafortheperiodsfrom2006 to2008agriddeddatasetwith (about10km 10km)isestablishedusingthe methodsby combiningthebarnesinterpolationandobservationprecipitationfrequency.thentheprecisionofgridded dataisassessedbyexaminingcross-validationstatisticsofthecorrelationcoeficient(cor)themeanbias error(mbe)themeanabsoluteerror(mae)themeanrelativeerror(rmae)therootmeansquareer- ror(rmse)compoundrelativeerror(cre)observationmeansquaredeviationσ(o)interpolationesti- (PDF)distributionthetemporalandspatialcharacteristicsofthestatisticsareinvestigatedandtheannual cycleofinterpolationerrorforthegriddeddataisdiscussed.theresultsshowthatdailygriddedprecipita- tionhasahigherprecisionwiththecorrelationcoeficientat0.81themeanbiaserroris-0.02mm d -1 themeanabsoluteerroris1.3mm d -1 themeanrelativeerroris58.67%therootmeansquareerroris 4.5mm d -1 andcompoundrelativeerroris41.17%andtherelativevarianceratiois93.12%.moreo- verthepdfdistributionofinterpolationprecipitationisclosertoobservationaswelastheinterpolation erroristhenormaldistribution.themeanabsoluteerrorsandtherootmeansquareerrorsforgriddedpre- cipitationincreasefromnorthwestchinatosoutheastchina.themeanrelativeerrorsaresmalintheeast ofchinaespecialyintheflatlandbutlargeinthewestofchina.inthemeanwhilethecorrelationcoefi- cientsshowhighintheeastpartofchinaandrelativelylowinthewestpartofchina.inadditionthean- nualcycleofmonthlymeanabsoluteerrorsandtherootmeansquareerrorsforgriddedprecipitationarein- creasingwiththeaugmentofmonthlytotalprecipitationthemaximumerrorsinjulywhiletheminimum indecemberingeneral.buttheprecipitationcharacteristicssuchasconvectionactivitieshaveaheavy impactonthedistributionoferrorstatistics.thecorrelationcoeficientreachingitspeakin ovember andvaleyinaugustassociateswiththeprecipitationcharacteristicsthatthehighcorrelationsarerelated tolarge-scaleandsystematicprecipitationwhilethelowcorrelationsarerelatedtolocaltheconvectivepre- cipitation. Keywords:Precipitation;Spatialinterpolation;Cross-validation;Errorstatistic

Microsoft Word - 01文本-A3-0730.doc

Microsoft Word - 01文本-A3-0730.doc 苏州市规划设计研究院有限责任公司技术文件证书编号 :[ 建 ] 城规编 ( 甲 ) 第 (081046) 证书等级 : 甲级苏州市滨湖新城 ( 吴中片 ) 公交场站专项规划审定审核项目负责人设计人员钮卫东施进华张海军吴素丽殷琪慜 2012 年 8 月目