上海国土资源
上海國土資源
상해국토자원
SHANGHAI LAND&RESOURCES
2012年
2期
74-78
,共5页
谢华亮%戴志军%彭伟%张小玲
謝華亮%戴誌軍%彭偉%張小玲
사화량%대지군%팽위%장소령
杭州湾北岸%岸线变化%径向基%神经网络模型
杭州灣北岸%岸線變化%徑嚮基%神經網絡模型
항주만북안%안선변화%경향기%신경망락모형
the northern bank of Hangzhou bay%shoreline changes%radial basis function%neural network
基于杭州湾北岸龙泉—南竹港弧形岸段实测岸滩断面与长江入海控制站大通站年输沙量资料,对杭州湾北岸岸线的变化及其趋势进行探讨。研究结果表明:受长江入海泥沙减少以及热带气旋和人类活动作用等影响,近10年来杭州湾北岸岸滩处于侵蚀状态。灰色关联分析进一步揭示,杭州湾北岸不同等深线的进退对长江入海泥沙减少的响应在时间尺度上有一定的滞后。同时,不同等深线的进退亦展现较强的相关性特征。鉴于此,利用大通年输沙量和不同等深线进退的耦合关系,进一步构建了基于长江入海泥沙和杭州湾北岸等深线变化的径向基神经网络岸线预报模型,其中模型输入向量为当年大通站年输沙量和杭州湾北岸-3m、-5m、-8m等深线距离大堤的位置,输出向量为次年0m岸线距离大堤的位置。经检验,构建的径向基神经网络岸线预报模型误差小于20%,可用以预报杭州湾北岸岸线的动态变化。
基于杭州灣北岸龍泉—南竹港弧形岸段實測岸灘斷麵與長江入海控製站大通站年輸沙量資料,對杭州灣北岸岸線的變化及其趨勢進行探討。研究結果錶明:受長江入海泥沙減少以及熱帶氣鏇和人類活動作用等影響,近10年來杭州灣北岸岸灘處于侵蝕狀態。灰色關聯分析進一步揭示,杭州灣北岸不同等深線的進退對長江入海泥沙減少的響應在時間呎度上有一定的滯後。同時,不同等深線的進退亦展現較彊的相關性特徵。鑒于此,利用大通年輸沙量和不同等深線進退的耦閤關繫,進一步構建瞭基于長江入海泥沙和杭州灣北岸等深線變化的徑嚮基神經網絡岸線預報模型,其中模型輸入嚮量為噹年大通站年輸沙量和杭州灣北岸-3m、-5m、-8m等深線距離大隄的位置,輸齣嚮量為次年0m岸線距離大隄的位置。經檢驗,構建的徑嚮基神經網絡岸線預報模型誤差小于20%,可用以預報杭州灣北岸岸線的動態變化。
기우항주만북안룡천—남죽항호형안단실측안탄단면여장강입해공제참대통참년수사량자료,대항주만북안안선적변화급기추세진행탐토。연구결과표명:수장강입해니사감소이급열대기선화인류활동작용등영향,근10년래항주만북안안탄처우침식상태。회색관련분석진일보게시,항주만북안불동등심선적진퇴대장강입해니사감소적향응재시간척도상유일정적체후。동시,불동등심선적진퇴역전현교강적상관성특정。감우차,이용대통년수사량화불동등심선진퇴적우합관계,진일보구건료기우장강입해니사화항주만북안등심선변화적경향기신경망락안선예보모형,기중모형수입향량위당년대통참년수사량화항주만북안-3m、-5m、-8m등심선거리대제적위치,수출향량위차년0m안선거리대제적위치。경검험,구건적경향기신경망락안선예보모형오차소우20%,가용이예보항주만북안안선적동태변화。
Based on observed data of the annual mean sediment discharge at Datong station and measured profile data from the Longquan–Nanzhu coast,we examine shoreline change within the northern Hangzhou Bay(NHB),and perform a numerical simulation of shoreline change.The results show coastal erosion within Longquan–Nanzhu harbor in recent decades due to a sharp decrease in sediment discharge from the Yangtze River,the effects of typhoons,and anthropogenic influences.The results of grey relational analysis indicate that temporal variations in the locations of isobaths within the NHB lag behind the reduction in sediment supply from the Changjiang Estuary.A strong positive relation is found among the changes in the locations of various isobaths.Based on the coupling relations between the annual sediment discharge at Datong and changes in the locations of isobaths,a radial basis functional neural network mode(RBF) is established,using input vectors of the annual sediment discharge at Datong,and the locations of isobaths(–3m,–5m,and –8m).The output vector is the shoreline location(0m isobath) in the following year.The RBF network mode has an error of less than 20%,indicating it is useful for predicting the nature of future shoreline changes.