Physical-based time series model applied on water table depths dynamics characteristics simulation

ABSTRACT Time series modelling applied to study water table depths monitoring data is an elegant way to model irregular and continuous data. When successive observations are dependent, future values may be predicted from past observations, and target parameters can be estimated. These may include expected values of groundwater levels, or probabilities that critical levels are exceeded at certain times or during certain periods. These target parameters are estimated with the purpose of obtaining characteristics of the development of a certain domain in time and such characteristics can, for instance, be extrapolated to future situations. In a system identification approach, is it possible to establish the dynamic relationship between water table perturbations and climatological events, vegetation, hydrogeological local conditions, management and groundwater abstraction. The aim of this work was demonstrate the use of a physical-based time series model to stablish the relationship between precipitation and water table depths from hydrogeological monitoring data. The results enabled to infer about water table dynamics even when it is affected by different climatological patterns, simulating mean, maximum and minimum states.

摘要：针对地下水位深度监测数据开展研究的时间序列建模（time series modelling），是处理不规则连续型数据的精巧方案。当连续观测值具有相关性时，可通过历史观测数据预测未来水位数值，并估算目标参数（target parameters）。此类目标参数可涵盖地下水位的期望值，以及特定时刻或时段内临界水位超标的概率。估算此类目标参数，旨在获取某一研究区域随时间演化的特征，例如可将这些特征外推至未来情景。借助系统辨识方法（system identification approach），可构建地下水位扰动与气候事件、植被状况、区域水文地质条件、人类管理活动及地下水抽取之间的动态关联。本研究旨在基于水文地质监测数据，采用基于物理机制的时间序列模型（physical-based time series model），构建降水与地下水位深度之间的关联关系。即便研究区域的地下水位受不同气候模式影响，本研究结果仍可用于推断地下水位动态变化，并可模拟地下水位的平均、最大与最小状态。