Abstract
There is a close relationship between groundwater level in a shallow aquifer and the surface ecological environment; hence, it is important to accurately simulate and predict the groundwater level in eco-environmental construction projects. The multiple linear regression (MLR) model is one of the most useful methods to predict groundwater level (depth); however, the predicted values by this model only reflect the mean distribution of the observations and cannot effectively fit the extreme distribution data (outliers). The study reported here builds a prediction model of groundwater-depth dynamics in a shallow aquifer using the quantile regression (QR) method on the basis of the observed data of groundwater depth and related factors. The proposed approach was applied to five sites in Tianjin city, north China, and the groundwater depth was calculated in different quantiles, from which the optimal quantile was screened out according to the box plot method and compared to the values predicted by the MLR model. The results showed that the related factors in the five sites did not follow the standard normal distribution and that there were outliers in the precipitation and last-month (initial state) groundwater-depth factors because the basic assumptions of the MLR model could not be achieved, thereby causing errors. Nevertheless, these conditions had no effect on the QR model, as it could more effectively describe the distribution of original data and had a higher precision in fitting the outliers.
Résumé
Il y a une relation étroite entre le niveau piézométrique dans un aquifère superficiel et l’environnement écologique de surface ; par conséquent, il est important de simuler précisément le niveau piézométrique et de le prévoir dans les projets de construction éco-environnementale. Le modèle de régression linéaire multiple (MLR) est une des méthodes les plus utiles pour prévoir le niveau piézométrique (profondeur) ; cependant, les valeurs prédites par ce modèle ne reflètent que la distribution moyenne des observations et ne peuvent pas s’ajuster efficacement aux valeurs extrêmes de la distribution (valeurs aberrantes du point de vue statistique). L’étude présentée ici construit un modèle de prédiction de la dynamique du niveau piézométrique dans un aquifère superficiel en utilisant la méthode de la régression quantile (QR), sur la base de données piézométriques observées et des facteurs associés. L’approche proposée a été appliquée à cinq sites de la ville de Tianjin, Chine du Nord, et le niveau piézométrique a été calculé en différents quantiles, à partir desquels le quantile optimal a été affiché au moyen de la méthode boîte à moustaches et comparé aux valeurs prédites par le modèle MLR. Les résultats montrent que les facteurs associés à la piézométrie dans les 5 sites ne suivent pas la distribution normale standard et qu’il y a des valeurs extrêmes dans les précipitations et dans les niveaux piézométriques des derniers mois (état initial), car les hypothèses de base du modèle MLR ne peuvent pas être atteintes, ce qui cause ainsi des erreurs. Néanmoins, ces conditions n’ont pas d’effet sur le modèle QR, car il pourrait décrire plus efficacement la distribution des données originelles et avoir une meilleure précision pour caler les valeurs extrêmes.
Resumen
Existe una estrecha relación entre el nivel del agua subterránea en un acuífero somero y el ambiente ecológico de superficie; por lo tanto, es importante simular y predecir con precisión el nivel de agua subterránea en proyectos de construcción eco-ambiental. El modelo de regresión lineal múltiple (MLR) es uno de los métodos más útiles para predecir el nivel de las aguas subterráneas (profundidad); Sin embargo, los valores pronosticados por este modelo sólo reflejan la distribución media de las observaciones y no pueden ajustarse de manera efectiva los datos extremos de la distribución (outliers). El presente estudio construye un modelo de predicción de la dinámica de la profundidad del agua subterránea en un acuífero somero utilizando el método de regresión por cuantiles (QR) sobre la base de los datos observados de la profundidad del agua subterránea y de los factores relacionados. El enfoque propuesto se aplicó a cinco sitios en la ciudad de Tianjin, norte de China, y la profundidad del agua subterránea se calculó en diferentes cuantiles, de la cual el cuantil óptimo se proyectó de acuerdo con el método de diagrama de caja y se compara con los valores predichos por el modelo de MLR. Los resultados mostraron que los factores relacionados en los cinco sitios no seguían una distribución estándar normal y que había valores extremos en los factores de la precipitación y del último mes de la profundidad del agua subterránea (estado inicial), ya que no podrían alcanzarse los supuestos básicos del modelo de MLR, lo cual provoca errores. Sin embargo, estas condiciones no tuvieron ningún efecto en el modelo de QR, ya que podría describir de manera más efectiva la distribución de los datos originales y tenía una mayor precisión en el ajuste de los valores extremos.
摘要
浅层地下水系统与地表生态环境关系密切,准确地模拟和预测地下水水位变化对于生态环境建设极为重要。多元线性回归模型是地下水位预测方法中最为常见的方法之一;但多元线性回归模型的预测值仅反映了实测值的均值分布情况,不能有效的拟合极端分布数据(异常值)。本文基于浅层地下水埋深及其影响因素实测资料的基础上,利用分位数构建浅层地下水埋深预测模型。将该方法应用于中国华北地区天津市,通过设置不同的分位数模拟计算出地下水埋深,并依据箱线图筛选出与原始数据分布最接近的预测值作为最优预测值,与多元线性回归模型计算结果进行对比。结果表明:五个站的影响因子分布不满足标准正态分布,其中降水量和上个月的地下水埋深(初始状态)均存在不同程度的异常值,故采用多元线性回归模型预测地下水埋深会存在较大偏差。而分位数模型不受异常值的干扰,能够更有效地描绘原始数据的分布情况,对异常值拟合具有更高的精度。
Resumo
Existe uma relação próxima entre o nível das águas subterrâneas em aquíferos rasos e o ambiente ecológico da superfície; assim, é importante simular precisamente e predizer os níveis das águas subterrâneas em projetos de construção ecoambientais. O modelo de regressão linear múltiplo (RLM) é um dos métodos mais úteis para predizer o nível das águas subterrâneas (profundidade); entretanto, os valores preditos por esse modelo refletem somente a distribuição média das observações e não pode ajustar efetivamente dados de distribuição extrema (valores discrepantes). O estudo aqui reportado construiu um modelo de predição da dinâmica dos níveis das águas subterrâneas em um aquífero raso usando o método da regressão quantílica (RQ) com base nos dados observados de nível das águas subterrâneas e fatores correlatos. A abordagem proposta foi aplicada em cinco locais na cidade de Tianjin, Norte da China, e a profundidade das águas subterrâneas calculada por diferentes quantis, dos quais quantis ótimos foram examinados quanto a sua adequação de acordo com o método dos diagramas de caixa e comparado aos valores preditos pelo modelo de RLM. Os resultados mostraram que os fatores relatados nos cinco locais não seguiram o padrão de uma distribuição normal e que houveram dados discrepantes na precipitação e fator de profundidade da água subterrânea do mês anterior (estado inicial) pois a hipótese básica do modelo de RLM não pode ser alcançada, consequentemente causando erros. Mesmo assim, essas condições não têm efeito no modelo de RQ, já que ele pôde descrever mais efetivamente a distribuição dos dados originais e ter uma maior precisão ao modelar os dados discrepantes.





Similar content being viewed by others
References
Banerjee P, Prasad RK, Singh VS (2009) Forecasting of groundwater level in hard rock region using artificial neural network. Environ Geol 58(6):1239–1246. doi:10.1007/s00254-008-1619-z
Borgoni R (2011) A quantile regression approach to evaluate factors influencing residential indoor radon concentration. Environ Model Assess 16(3):239–250. doi:10.1007/s10666-011-9249-3
Briollais L, Durrieu G (2014) Application of quantile regression to recent genetic and -omic studies. Hum Genet 133(8):951–966. doi:10.1007/s00439-014-1440-6
Dong ZG (2002) A review of predicting models for groundwater dynamic (in Chinese). West-China Explor Eng 04:36–39
Ebru Ç, Eban A (2011) Determinants of house prices in Istanbul: a quantile regression approach. Qual Quant 45(2):305–317. doi:10.1007/s11135-009-9296-x
Emamgholizadeh S, Moslemi K, Karami G (2014) Prediction the groundwater level of Bastam Plain (Iran) by artificial neural network (ANN) and adaptive neuro-fuzzy inference system (ANFIS). Water Resour Manag 28(15):5433–5446. doi:10.1007/s11269-014-0810-0
Hu QK, Li FW, Feng P (2013) Type-character based groundwater functional zoning in Tianjin (in Chinese). Water Resour Hydropower Eng 44(9):4–7
Izady A, Davary K, Alizadeh A, Moghaddam Nia A, Ziaei AN, Hasheminia SM (2013) Application of NN-ARX model to predict groundwater levels in the Neishaboor Plain, Iran. Water Resour Manag 27(14):4773–4794. doi:10.1007/s11269-013-0432-y
Jusseret S, Tam VT, Dassargues A (2009) Groundwater flow modelling in the central zone of Hanoi, Vietnam. Hydrogeol J 17(4):915–934. doi:10.1007/s10040-008-0423-x
Koenker R, Basset G (1978) Regression quantiles. Econometrica 46(1):33–50
Li X, Shu L, Liu L, Yin D, Wen J (2012) Sensitivity analysis of groundwater level in Jinci Spring Basin (China) based on artificial neural network modeling. Hydrogeol J 20(4):727–738. doi:10.1007/s10040-012-0843-5
Li F, Feng P, Zhang W, Zhang T (2013) An integrated groundwater management mode based on control indexes of groundwater quantity and level. Water Resour Manag 27(9):3273–3292. doi:10.1007/s11269-013-0346-8
Machiwal D, Mishra A, Jha MK, Sharma A, Sisodia SS (2012) Modeling short-term spatial and temporal variability of groundwater level using geostatistics and GIS. Nat Resour Res 21(1):117–136. doi:10.1007/s11053-011-9167-8
Mohanty S, Jha MK, Kumar A, Sudheer KP (2010) Artificial neural network modeling for groundwater level forecasting in a river island of eastern India. Water Resour Manag 24(9):1845–1865. doi:10.1007/s11269-009-9527-x
Olden JD, Jackson DA (2002) Illuminating the “black box”: a randomization approach for understanding variable contributions in artificial neural networks. Ecol Model 154:135–150
Ping JH, Li S (2006) Review and prospect of dynamic prediction model for groundwater (in Chinese). Water Resour Prot 04:11–15
Rakhshandehroo GR, Vaghefi M, Aghbolaghi MA (2012) Forecasting groundwater level in Shiraz Plain using artificial neural networks. Arab J Sci Eng 37(7):1871–1883. doi:10.1007/s13369-012-0291-5
Sahoo S, Jha MK (2013) Groundwater-level prediction using multiple linear regression and artificial neural network techniques: a comparative assessment. Hydrogeol J 21(8):1865–1887. doi:10.1007/s10040-013-1029-5
Sahoo S, Jha MK (2015) On the statistical forecasting of groundwater levels in unconfined aquifer systems. Environ Earth Sci 73(7):3119–3136. doi:10.1007/s12665-014-3608-8
Seeboonruang U (2015) An application of time-lag regression technique for assessment of groundwater fluctuations in a regulated river basin: a case study in northeastern Thailand. Environ Earth Sci 73(10):6511–6523. doi:10.1007/s12665-014-3872-7
Su Y, Wan YY (2009) The idea and application of quantile regression (in Chinese). Statistical Thinktank 10:58–61
Trichakis IC, Nikolos IK, Karatzas GP (2011) Artificial neural network (ANN) based modeling for karstic groundwater level simulation. Water Resour Manag 25(4):1143–1152. doi:10.1007/s11269-010-9628-6
Uddameri V (2007) Using statistical and artificial neural network models to forecast potentiometric levels at a deep well in South Texas. Environ Geol 51(6):885–895. doi:10.1007/s00254-006-0452-5
Wang SQ, Song XF (2008) Dynamic features of shallow groundwater in North China Plain (in Chinese). Acta Geograph Sin 63(5):438–445
Wang Y, Hung J, Kao H, Shih K (2011) Long-term relationship between political behavior and stock market return: new evidence from quantile regression. Qual Quant 45(6):1361–1367. doi:10.1007/s11135-010-9340-x
Wang KY, Wang WZ, Li QF (2014) Characteristics of changes of groundwater buried depth and influencing factors in Tianjin Plain area over past 21 years (in Chinese). Water Resour Pro 30(3):45–49. doi:10.3969/j.issn.1004-6933.2014.03.009
Zheng HM (2007) Groundwater three-dimensional numerical simulation of Tianjin (in Chinese). China University of Geosciences, Beijing
Acknowledgements
The authors would like to acknowledge the financial support for this work provided by the National Natural Science Foundation of China (grant No. 51579169), the Science Fund for Creative Research Groups of the National Natural Science Foundation of China (grant No. 51321065), and the Ministry of Water Resources Special Funds for Scientific Research on Public Causes (201401041). The authors declared that they have no financial conflicts of interest related to this work.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Li, F., Wei, W., Zhao, Y. et al. Groundwater depth prediction in a shallow aquifer in north China by a quantile regression model. Hydrogeol J 25, 191–202 (2017). https://doi.org/10.1007/s10040-016-1473-0
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10040-016-1473-0