圖書典藏及影音數位平台

本計畫就105年水利署「地下水觀測與地質資料檢討分析水文地質參數-以屏東平原為例」計畫成果，進一步驗證及研發，期能發展訂定可適用於無複井觀測井水文地質參數之推估方法，以供未來地下水資源利用及管理之參據。

Groundwater resources play a vital role on regional water supply. To achieve the goal of sustainable management for groundwater resources, an accurate field of specific yield and storage coefficient is crucial. The investigations of specific yield and storage coefficient require multiple-well systems for pumping tests. In Taiwan, over 700 observation wells have been installed. Only one fifth of wells are designed as multiple-well systems and others are designed as single-well systems. This project proposes a collection of identification approaches based on the observations of hydrology and meteorology to identify the specific yield and storage coefficient. Hydrology and meteorology signals such as groundwater levels, precipitation, ocean and earth tide, barometric pressure and river stages are collected and signal process and data mining based approaches are applied to simulate the quantitative relationships between each signal. A linear signal model (LSM) which is applied to simulate the underground hydrology basing on groundwater levels, precipitation and river stages can estimate the specific yield of wells in unconfined aquifers. A harmonic analysis is also used to calculate the amplitude and phase lag of M_2 tide and the result can be used to determine the transmissivity of aquifer. An elliptical loop and histogram approach which is inspired from the elliptical loop graphic approach is used to determine the barometric efficiency (BE) and the BE can be further used to identify the storage coefficient in confined aquifers. A river tomography test is also used to identify the spatial distribution of hydraulic diffusivity on the first aquifer. A procedure of numerical experiment is also proposed to verify the accuracy of parameter identification. Fifty sets of parameters which are generated randomly are assigned in MODFLOW simulations and the associated simulated levels are used in the identification process of LSM. The errors and correlations between identified results and the assigned parameters are calculated to verify whether the identified ability of the proposed method is reliable. In the numerical experiment, the fact that the error is low and the correlation is high give a strong support for the proposed process of LSM.