Page 85 - 《水资源与水工程学报》2024年第2期
P. 85
! 2# ¢£G, :: mn¤¥D¦§@ VMD-LSTM¨<=>%©W=ª./ 8 1
Ä»'6¢B¾c^Óצ VMD- 2004 [J].JournalofClimate,2009,22(10):2773
LSTM)¿À, Ó~¦)ñ)p¶ 2792.
[|z, 5³j·1ÈÑ¿À¦ñÔ8¿À [5]HUANGYi,CAIJialiang,YINHe,etal.Correlationof
precipitationtotemperaturevariationintheHuangheRiver
ñ, º'./¦ÓJ»#$»<¿À7=:; j
(YellowRiver)Basinduring1957-2006[J].Journalof
µ, ѳkÌÍsª¥n¾¿ÀrY)¦
Hydrology ,2009,372(14):18.
u+Õ¼, Ð5, Ú"ÔÄ2¿À, Ñ¿Àí¹
[6]DONGQuan,CHENXing,CHENTiexi.Characteristics
265¦Ä2/e826ÏЦ%Þ
¨
andchangesofextremeprecipitationintheYellow-Hua
)pº¶, ¿À¿Çëì¦óÊÔÔ¥nÞ¦
iheandYangtze-HuaiheRiversBasins ,China[J].Jour
¥+D。Õ¶kO,Æ7=¥n¦%Þ
7 nalofClimate,2011,24(14):37813795.
8Ñ9¨¤¥:Ë¿À¦?@=, Ë¿À¦? [7]KREIBICH H,VANLOONAF,SCHROTERK,etal.
@éê¶[,-Ñ)¿À¦y@。 Thechallengeofunprecedentedfloodsanddroughtsinrisk
management [J].Nature,2022,608(7921):8086.
5 I N
[8]NEARINGGS,KRATZERTF,SAMPSONAK,etal.
'6,Æ26*+'<Ü9m[B¦<ô Whatroledoeshydrologicalscienceplayintheageofma
chinelearning ?[J].WaterResourcesResearch,2021,
jÞ¦Ú;, MNÌû VMD:G¦^
57(3):e2020WR028091.
¿ÇË8o;¦7=j, ×{yÓÔÚ;
[9] 6,
Ö. µÄ26·(¼[J]. 2]23*
|z¦ VMD^ LSTM)¿ÀY,&c^&
=¶@,2005(1):5660.
x。
[10]\Çð, [ ë, ¨ ì, :. ÓÔµ#¿À^ LSTM
( 1)VMD-LSTM )¿ÀÚ";¦ LSTM
)¦éãTÚã¿Ç[J/OL]. 2QRfü:112
¿Àkö÷O±¿Çc, Óý³×。 [20230323].http://kns.cnki.net/kcms/detail/32.
(2) ]@ VMDË LSTM¿À¦7=、 7G¶· 1356.TV.20230316.1547.008.html.
¿G, ×ë)Ú;|z¸f¿À+6Qa [11]Ñ [, CãÆ, Ñ×Ò, :. Ö)78N>µ%D¦Ãc
¿½¦OÑ, û^¿ÇË8o;¦7=¡, k +626¿À[J]. 2*+¶@,2022,33(6):904913.
ö÷O& LSTM¿ÀËÚãµ*¦·¸8ߺ [12]ØÇÊ, Ìpê, Ë9o, :. bâ»úÚãË78N>
´¿。 ¦¦§[J]. 2QRfü,2023,39(1):119126.
[13]; Ù, z¢m, 2 /, :. ÓÔ,×?¿Gü
( 3) q VMD^ LSTM¿À³j)~¦Ë
b¿Èµ¶ab¦ÿÚã·[J]. 2]+,2023,54
", ®s~ D( 7=、 7G©G×H7=
1
(2):172183+198.
¡) ¹cÿ&«ÿó¦)~。
[14]UÐ, ùúû, LÖ. ÓÔ¿G¦Ú
( 4) ^ùþüÆ«Ú", !"ÚãËÔ LSTM¿
ã·¿À[J]. 2QR^245+,2023,34(1):
À¦?@»¹9N, Úã¿ÇmY)³kº¾¦ 8490.
s3。 [15]CHENShu,RENMiaomiao,SUNWei.Combiningtwo
stagedecompositionbasedmachinelearningmethodsfor
LM: annualrunoffforecasting [J]. JournalofHydrology,
[1]KIKSTRAJS,NICHOLLSZRJ,SMITHCJ,etal.The 2021,603:126945.
IPCCSixthAssessmentReportWGIIIclimateassessmentof [16]z<©, 0,B, Ó ë, :. <R¥nÖ)¦Ãc+6
mitigationpathways :fromemissionstoglobaltemperatures Úã·¿À[J]. 2¿«3+,2023,42(5):4352.
[J].GeoscientificModelDevelopment,2022,15(24): [17]BENGIOY,SIMARDP,FRASCONIP.Learninglong
90759109. termdependencieswithgradientdescentisdifficult [J].
[2]à ?, éL9, W"?, :. ÓÔ<ôõ)¿À¦Úã· IEEETransactionsonNeuralNetworks,1994,5(2):
./[J]. 2]+,2020,51(1):112125. 157166.
[3]WALLINGDE,FANGD.Recenttrendsinthesuspended [18]GRAVESA.Longshorttermmemory[J].SupervisedSe
sedimentloadsoftheworld ’srivers[J].GlobalandPlane quenceLabellingwithRecurrentNeuralNetworks ,2012,
taryChange,2003,39(12):111126. 385:3745.
[4] DAIAiguo,QIAN Taotao,TRENBERTH K E,etal. [19]-Úî, ³¢á, ^UT, :. ÓÔ LSTMabú_EÛ
Changesincontinentalfreshwaterdischargefrom1948to ãTÚã¿Ç./[J]. VmZc:++( ,-*