資源描述:
《上海市水情長(zhǎng)期預(yù)測(cè)方法比較研究》由會(huì)員上傳分享,免費(fèi)在線閱讀,更多相關(guān)內(nèi)容在教育資源-天天文庫(kù)。
1、河海大學(xué)碩士學(xué)位論文上海市水情長(zhǎng)期預(yù)測(cè)方法比較研究姓名:陳其幸申請(qǐng)學(xué)位級(jí)別:碩士專業(yè):水文學(xué)及水資源指導(dǎo)教師:陳元芳;葛朝霞20050301摘要長(zhǎng)期水文預(yù)報(bào)是介于水文學(xué)、氣象學(xué)、氣候?qū)W等多學(xué)科之間的一門邊緣學(xué)科,由于影響因素的復(fù)雜性與目前科學(xué)水平的限制,長(zhǎng)期水文預(yù)報(bào)精度還不高。如何提高預(yù)報(bào)精度是水文學(xué)家關(guān)注的問(wèn)題。本文以上海市水情為例,開展長(zhǎng)期預(yù)測(cè)方法研究。重點(diǎn)是對(duì)比不同方法,以期找到更好的預(yù)報(bào)模型來(lái)提高預(yù)報(bào)精度。上海市年最高潮位預(yù)報(bào)是上海市水情預(yù)報(bào)最重要的部分,在以往的上海市年最高潮位預(yù)測(cè)中,發(fā)現(xiàn)結(jié)果都不理想。原因可能是年最高潮位受上游來(lái)水、天文潮、臺(tái)風(fēng)的共同影響,預(yù)報(bào)難度
2、大,致使用過(guò)去的預(yù)報(bào)方法,預(yù)測(cè)精度差。本次由于增加了物理成因法,引用7海溫、南方濤動(dòng)指數(shù)(SouthernOscillationindex),太陽(yáng)黑子數(shù)等資料,預(yù)報(bào)精度得到了提高,特別是年最高潮位預(yù)測(cè)中效果更為明顯。文中著重研究和應(yīng)用了多元逐步回歸、自然正交展開(EOF)、投影追蹤回歸(PPR)三種預(yù)報(bào)方法,并與過(guò)去研究的自回歸模型AR(P)、灰色模型GM(1,1),1r7限自回歸模型(ThresholdAutoregressiveModel)對(duì)比分析。另外,在定性預(yù)報(bào)中,對(duì)較新的加權(quán)馬爾柯夫模型與馬爾柯夫模型也作了研究與比較。關(guān)鍵詞:長(zhǎng)期預(yù)測(cè):多元逐步回歸;自然正交展開(E
3、OF);投影追蹤回歸丈PPR)一自回歸模型Aft(P);灰色模型GM(1,1);門限自回歸模型(ThresholdAutoregressiveModel);加權(quán)馬爾柯夫模型AbstractThelong-rangehydrologicalforecastisaboundaryscienceofhydrology,meteorology,climatologyandotherrelativesciences.Becauseofcomplexityofaffectingfactorsandrestrictionofcurrent-conditions,theresultsofth
4、elong-rangehydrologicalforecastarestillnotperfect.Howtoimprovetheforecastaccuracyisthekeyproblemconcernedaboutbyhydrologists.Inthispaper,thelong-rangehydrologicalforecastisstudiedandShanghaiwaterregimeisselectedasacasestudy.Emphasisisputonthecomparisonofdifferentmethodsinordertoimproveforec
5、astaccuracy.ThepredictionoftheannualhighesttidestageatHuang-puParkisveryimportant,butinthepastitisoflittleeffect.Why?Perhapsitisuplandwater,astronomicalbearingandtyphoonthatconducestodifficultiesofforecastingtheannualhighesttidestage.Inthepaper,because'seatemperature,SouthernOscillationinde
6、xandsunspotnumberarealsoinvestigated,forecastaccuracyisimproved,especiallyinpredictionoftheannualhighesttidestage.Thepaperlaysemphasisonresearchsofstepwisemultipleregression,empiricalorthogonalfunction,projectionpursuitregression,incomparisonwithautoregressionmodel,greymodel(1,1),andthresho
7、ldautoregressivemodel.Moreover,intheendweightedMarkovchainmodel,MarkovchainmodelandcomparisonofthemarestudiedthatrefertoqualitativeforecastKeywords:long-rangehydrologicalforecaststepwisemultipleregression;empiricalorthogonalfunctionprojectionpursuitregre