Automatic estimation of aquifer parameters using l

Authors: Ning Luo Walter A Illman

Publish Date: 2016/04/19

Volume: 24, Issue: 6, Pages: 1443-1461

PDF Link

Abstract

Analyses are presented of longterm hydrographs perturbed by variable pumping/injection events in a confined aquifer at a municipal watersupply well field in the Region of Waterloo Ontario Canada Such records are typically not considered for aquifer test analysis Here the waterlevel variations are fingerprinted to pumping/injection rate changes using the Theis model implemented in the WELLS code coupled with PEST Analyses of these records yield a set of transmissivity T and storativity S estimates between each monitoring and production borehole These individual estimates are found to poorly predict waterlevel variations at nearby monitoring boreholes not used in the calibration effort On the other hand the geometric means of the individual T and S estimates are similar to those obtained from previous pumping tests conducted at the same site and adequately predict waterlevel variations in other boreholes The analyses reveal that longterm municipal waterlevel records are amenable to analyses using a simple analytical solution to estimate aquifer parameters However uniform parameters estimated with analytical solutions should be considered as first rough estimates More accurate hydraulic parameters should be obtained by calibrating a threedimensional numerical model that rigorously captures the complexities of the site with these dataDes analyses d’hydrogrammes sur de longues durées perturbés par des événements variables de pompage ou d’injection dans un champ captant municipal d’un aquifère captif dans la région de Waterloo Ontario Canada sont présentées De tels enregistrements ne sont généralement pas considérés pour linterprétation des pompages d’essais Ici les variations de niveau deau sont calquées sur les variations de régime de pompage et d’injection en utilisant le modèle de Theis implémenté dans le code WELLS couplé à PEST Les interprétations de ces données aboutissent à un ensemble de transmissivité T et de coefficients d’emmagasinement S entre chaque piézomètre et forage de production Il s’avère que ces estimations ponctuelles prédisent mal les variations de niveau deau aux piézomètres voisins de forages non utilisés pour le calage Dun autre côté les moyennes géométriques des différentes évaluations de T et de S sont semblables à celles obtenues à partir de pompages dessai antérieurs effectués sur le même site et prédisent correctement les variations de niveau deau dans dautres forages Les analyses indiquent que les longues séries d’enregistrements municipaux de niveau deau se prêtent bien aux interprétations en utilisant une solution analytique simple pour estimer les paramètres des aquifères Cependant des paramètres uniformes estimés avec les solutions analytiques devraient être considérés en première approximation en tant qu’évaluations grossières Des paramètres hydrauliques plus précis devraient être obtenus en calibrant un modèle numérique tridimensionnel qui transcrirait rigoureusement les complexités du site à partir de ces donnéesSe presentan análisis de hidrogramas a largo plazo perturbados por eventos variables de bombeo / inyección en un acuífero confinado en un campo de pozos de abastecimiento de agua municipal en la Región de Waterloo Ontario Canadá Tales registros normalmente no se consideran para el análisis de los ensayos de bombeo En este caso las variaciones del nivel de agua son las huellas digitales de los cambios en los caudales de bombeo / inyección para utilizar el modelo de Theis implementado en el código WELLS junto con PEST Los análisis de estos registros producen un conjunto de estimaciones de transmisividad T y del coeficiente de almacenamiento S entre cada pozo de monitoreo y de producción Estas estimaciones individuales se encuentran para predecir pobremente las variaciones del nivel de agua en los pozos de monitoreo de las inmediaciones no usados en el intento de la calibración Por otra parte las medias geométricas de las estimaciones individuales de T y S son similares a las obtenidas de ensayos de bombeo previos realizados en el mismo sitio y predicen adecuadamente las variaciones de nivel de agua en otros pozos Los análisis revelan que los registros municipales de nivel de agua a largo plazo son susceptibles de análisis utilizando una solución analítica simple para estimar los parámetros del acuífero Sin embargo los parámetros uniformes estimados con las soluciones analíticas deben ser considerados como las primeras estimaciones aproximadas Parámetros hidráulicos más precisos deben ser obtenidos por la calibración de un modelo numérico tridimensional que captura rigurosamente las complejidades del sitio con estos datos本文分析了加拿大安大略省滑铁卢地区市政供水井场承压含水层变化着的抽水/注入事件扰动的长期水文曲线。通常这些记录在含水层测试分析中未被考虑。在这里利用在WELLS 码实施的Theis模型耦合PEST在抽水和注入量变化中采集到水位变化痕迹。这些记录分析产生了每个监测井和生产井之间一组导水系数和储存系数值。发现这些单独的估算值预测校准中没有使用的监测井附近的水位变化效果很差。另一方面单独的导水系数和储存系数估算值的几何方法类似于同一地点先前进行的抽水试验获取的方法能够充分预测其他井的水位变化。分析结果显示采用简单解析方法长期市政水位记录能够用来分析以估算含水层参数。然而解析方法估算的统一参数应该作为基本的粗略估算值。通过校准依靠这些资料捕获此地复杂性的三维数值模型应该获取更准确的水力参数。São apresentadas análises de hidrogramas de longo termo influenciados por eventos de bombeamento e injeção em aquífero confinado no campo de poços de abastecimento público de água na região de Waterloo Ontário Canada Tais dados normalmente não são considerados para testes de aquíferos Neste as variações do nível de água são respostas às vazões de bombeamento/injeção pela aplicação do modelo de Theis a partir do codificador WELLS acoplado ao programa PEST As análises desses dados permitem estimar a transmissividade T e o armazenamento S individualmente para cada conjunto de poços de monitoramento e de produção Essas estimativas individuais são obtidas para prever aproximadamente as variações de nível de água em poços de monitoramento adjacentes e não são aplicadas nas calibrações de modelos Por outro lado as médias geométricas das estimativas de T e S individuais são semelhantes às obtidas nos ensaios de bombeamento preliminares conduzidos no mesmo local e que adequadamente preveem as variações de nível de água em outros poços As análises revelam que as séries históricas de dados de variação de nível de água em poços de abastecimento público permitem estimar parâmetros hidráulicos a partir de soluções analíticas simples No entanto os parâmetros estimados com essas soluções analíticas devem ser considerados como primeiras estimativas aproximadas Parâmetros hidráulicos mais precisos devem ser obtidos pela calibração de um modelo numérico tridimensional que rigorosamente reproduza as complexidades do local com esses dadosPlanning for the optimized use of groundwater resources is of paramount importance to water managers worldwide in the face of increased demands on groundwater resources the protection of groundwater resources from contamination and the increasing energy costs of community water systems The optimized design and management of groundwaterbased water systems requires the accurate estimation of transmissivity T and storativity S which are two important hydraulic parameters in predicting groundwater flow Traditionally the estimation of T and S is performed through the analysis of pumping tests in which drawdown data are analyzed using analytical or numerical models The Theis 1935 solution is the first analytical model developed for transient analysis of a pumping test in a confined aquifer but numerous type curve solutions have since been developed over the next few decades for different aquifer types and boundary conditions eg Hantush and Jacob 1955 Neuman 1974 Moench 1997 Mathias and Butler 2006 Mishra and Neuman 2011 The application of these analytical models is sometimes restricted in complex hydrogeological conditions where some features that significantly affect groundwater flows are not considered eg Mansour et al 2011 Although the use of analytical models may lead to good matches between observed drawdowns and type curves the estimated hydraulic properties may be scenariodependent—for example Wu et al 2005 demonstrated that the conventional analysis of aquifer tests yields biased T estimates that evolved with time and depended on the location of monitoring boreholes and the estimated S is mainly affected by the geology ie heterogeneity between the watersupply and monitoring boreholesAnother important issue is that planning and conducting these tests within a municipal well field can be expensive timeconsuming and impractical For example it is logistically infeasible to cease pumping/injection at a municipal well field to conduct a dedicated pumping test considering that municipal water supply cannot be interrupted and pressure influences induced by neighboring watersupply boreholes can affect results in uncertain ways Harp and Vesselinov 2011

Keywords: