臺北市山坡地降雨量及地下水位之時間序列分析

Time Series Analysis of Rainfall and Groundwater Table for the Slopeland Watersheds of Taipei Area

蘇苗彬(Miau-Bin Su) ； 陳旺志(Wang-Jyh Chen)

《中華水土保持學報》 27卷2期 (1996/06) Pp. 127-138

https://doi.org/10.29417/JCSWC.199606_27(2).0005

摘要

有關地下水流的研究，國內多側重於對含水層的模擬，較少應用統計推論或時間序列的方法進行分析，尤其是在集水區中上游屬於水資源涵養的地區，由於缺乏實際觀測的資料，較少有相關之研究發表。有鑑於此，本研究針對臺北周圍山坡地土地利用型態差異較大及記錄較爲長期而完整之無名溪、青礜溪、松山區等三處集水區進行分析探討。首先將三處測站之雨量及地下水位資料進行整理及補遺，完成三站之日雨量、二小時雨量、月最高地下水位、日最高地下水位及二小時地下水位等不同期距之水文資料檔。再就此等資料檔分別以波譜密度分析、線性廻歸分析、二元交叉相關函數分析及ARIMA（p，d，q）模式之序率式時間序列分析等方法，分別探討各集水區其地下水位之週期性、長期趨勢性及雨量與地下水位之稽延時差關係，並探討以ARIMA（p，d，q）模式進行地下水位時間序列分析及預測之可行性。研究顯示各集水區月最高地下水位具有下降之長期趨勢，且以開發比率最大者爲甚。各集水區日最高地下水位之時間序列以ARIMA模式進行分析，結果顯示其對短期間內之地下水位預測應屬可行。各集水區之二小時雨量及地下水位經二元交叉相關函數分析其交叉相關係數均不高，然由其趨勢仍可初步判定雨量影響地下水位之稽延時差爲開發比率愈大者其稽延時差亦愈大。

關鍵字

地下水位 ； 降雨 ； 時間序列 ； 自我廻歸移動平均整合模式

並列摘要

Rainfall induces groundwater table fluctuation in slopeland area. In this study, time series analysis are chosen to study the effect of different land utilization to the relationship between rainfall and watertable. Rainfall and groundwater table data of these three watersheds, named Wu-Ming-Si, Ching-Chei-Si and Song-Sun-Chi, were collected and analyzed. The data file includes data of daily rainfall, 2 hours rainfall, monthly maximum groundwater table, daily maximum groundwater table and 2 hours groundwater table. Spectral density analysis of groundwater table were performed to check their periodicity of these three watersheds. Result of linear regression analysis of groundwater table show that, long term trend of groundwater table is droping and especially significant for the watershed having largest land use ratio. Cross correlation function analysis of rainfall and groundwater table were performed to calculate the lag time between rainfall and groundwater table of these three watersheds. The watershed having most paved area has a longer time lag which implies its infiltration had reduced. At the end, time series of groundwater table of these three watersheds were analyzed by stochastic method and their ARIMA (P, d, q) models were built.

並列關鍵字

Groundwater Table ； Rainfall ； Time Series ； ARIMA