Advertisement

Hydrogeology Journal

, Volume 27, Issue 1, pp 273–289 | Cite as

Application of influence diagrams for well contamination risk management: a case study in the Po plain, northern Italy

  • Enrico CameronEmail author
  • Giorgio Pilla
  • Fabio A. Stella
Report
  • 40 Downloads

Abstract

The aquifer of the Oltrepò Pavese plain (northern Italy) is affected by paleo-saltwater intrusions that pose a contamination risk to water wells. The report first briefly describes how the presence of saline water can be predicted using geophysical investigations (electrical resistivity tomography or electromagnetic surveys) and a machine-learning tool specifically developed for the investigated area. Then, a probabilistic graphical model for addressing the risk of well contamination is presented. The model, a so-called ‘influence diagram’, allows researchers to compute the conditional probability that groundwater is unsuitable for use taking into account the results of the geophysical surveys, the predictions of the machine-learning software, the related uncertainties and the prior probability of contamination in different sectors of the plain. The model, in addition, allows for calculation and comparison of the expected utility of alternative decisions (drilling or not drilling the well, or using another water source). The model is designed for use in ordinary decision situations and, although conceived for a specific area, provides an example that may be adapted to other cases. Some adaptations and generalizations of the model are also discussed.

Keywords

Groundwater development Wells Contamination Risk analysis Influence diagrams 

Application de diagrammes d’influence pour la gestion du risque de contamination de forages d’eau: étude de cas dans la Plaine du Pô, Italie du Nord

Resume

L’aquifère de la plaine de l’Oltrepo Pavese (Italie du Nord) est. touché par des intrusions d’eaux salées anciennes qui présentent un risque de contamination pour les forages d’eau. L’article décrit d’abord brièvement comment la présence d’eau salée peut être appréhendée en recourant à des investigations géophysiques (tomographies de résistivité électrique ou levés éléctromagnétiques) et à un outil d’apprentissage automatique développé spécifiquement pour la zone étudiée. Puis, un modèle graphique probabiliste, mis en œuvre pour répondre au risque de contamination des forages, est. présenté. Le modèle, nommé “diagramme d’influence”, permet aux chercheurs de calculer la probabilité conditionnelle que l’eau souterraine soit impropre aux usages, en prenant en compte les résultats des levés géophysiques, les prévisions du logiciel d’apprentissage automatique, les incertitudes liées et la probabilité a priori d’une contamination de différents secteurs de la plaine. De plus, le modèle permet le calcul et la comparaison de l’utilité attendue de décisions alternatives (réalisation ou non d’un forage, ou utilisation d’une autre source d’approvisionnement en eau). Le modèle est. destiné à une utilisation dans des situations de décisions habituelles et, bien que conçu pour un secteur spécifique, il fournit un exemple qui pourrait être adapté à d’autres cas. Certaines adaptations et généralisations de ce modèle sont également discutées.

Aplicación de diagramas de influencia para la gestión del riesgo de contaminación de un pozo: un estudio de caso en Planicie del Po, norte de Italia

Resumen

El acuífero de la llanura Oltrepò Pavese (norte de Italia) se ve afectado por intrusiones de agua paleo-salada que representan un riesgo de contaminación para los pozos de agua. El documento, en primer lugar, describe brevemente cómo se puede predecir la presencia de agua salina utilizando investigaciones geofísicas (tomografía de resistividad eléctrica o estudios electromagnéticos) y una herramienta de aprendizaje automático desarrollada específicamente para el área investigada. Luego, se presenta un modelo probabilístico gráfico para abordar el riesgo de contaminación del pozo. El modelo, denominado “diagrama de influencia”, permite a los investigadores calcular la probabilidad condicional de que las aguas subterráneas no sean adecuadas para su uso teniendo en cuenta los resultados de los estudios geofísicos, las predicciones del software de aprendizaje automático, las incertidumbres relacionadas y la probabilidad previa de contaminación en diferentes sectores de la llanura. El modelo, además, permite calcular y comparar la utilidad esperada de decisiones alternativas (perforar o no perforar el pozo, o usar otra fuente de agua). El modelo está diseñado para su uso en situaciones de decisión ordinaria y, aunque se concibió para un área específica, proporciona un ejemplo que puede adaptarse a otros casos. También se discuten algunas adaptaciones y generalizaciones del modelo.

水井污染风险管理影响图表的应用:意大利北部Po平原的一个研究实例

摘要

(意大利)北部Oltrepò Pavese平原含水层受到古海水入侵的影响,古海水入侵给水井带来了污染风险。本文首先简要描述了利用地球物理调查(电阻率层析成像或者电磁勘查)及专门为调查区研发的机器学习工具怎样可以预测是否存在着海水。然后,论述了针对水井污染风险概率图解模型。所谓的“影响图解”模型可以使研究人员计算综合考虑地球物理勘查结果、机器学习软件的预测结果、相关不确定性及平原不同区域污染的先验概率而认为地下水不适合利用的条件概率。另外,模型还可以对供选择的决定(钻井或者不钻井,或者使用另一水源)的预期实用性进行计算和对比。模型为普通决策情况下使用而设计,尽管为特定区域而构思,但模型提供了可能适用于其它情况的例子。文章还论述了一些适应性和概括性。

Applicazione dei diagrammi di influenza per la gestione del rischio di contaminazione dei pozzi idrici: un caso di studio nella pianura alluvionale del Po, Italia settentrionale

Riassunto

L’acquifero della pianura alluvionale dell’Oltrepò Pavese è interessato da un particolare fenomeno d’intrusione di paleo-acque a elevata salinità che determinano un rischio di contaminazione dei pozzi idrici da realizzare. Nell’articolo si descrive brevemente come la presenza di acque salate possa essere prevista sia mediante indagini geofisiche (tomografia elettrica o indagini elettromagnetiche) sia attraverso un algoritmo di machine learning specificamente sviluppato per l’area di indagine. Viene presentato, successivamente, un modello grafico-probabilistico per la gestione del rischio di contaminazione dei pozzi. Il modello, un “diagramma di influenza”, permette di calcolare la probabilità condizionale che l’acqua di falda non sia idonea all’uso considerando i risultati delle indagini geofisiche, le previsioni dell’algoritmo di machine learning, le relative incertezze e la probabilità a priori di contaminazione nei diversi settori della pianura. Il modello, infine, permette di calcolare e confrontare l’utilità attesa di decisioni alternative quali realizzare o non realizzare un pozzo oppure usare un’altra fonte di approvvigionamento idrico. Il modello è progettato per essere impiegato in situazioni ordinarie e, anche se sviluppato per un’area specifica, rappresenta uno strumento che può essere adattato ad altri contesti idrogeologici con presenza di contaminazione. Nell’articolo sono discusse, anche, alcune varianti ed. estensioni del modello messo a punto.

Aplicação de diagramas de influência para gerenciamento de risco de contaminação em poços: estudo de caso na planície Po, norte da Itália

Resumo

O aquífero da planície de Oltrepò Pavese (norte da Itália) é afetado por intrusões de paleo-águas salinas que oferecem um risco de contaminação para poços de água. O artigo, primeiramente, descreve de forma breve como a presença de água salina pode ser estimada utilizando métodos de investigação geofísica (tais como tomografia de resistividade elétrica ou levantamentos eletromagnéticos) e uma ferramenta de aprendizado de máquina (inteligência artificial) desenvolvida especificamente para a área investigada. Com isso, o modelo gráfico de probabilidade para acessar os riscos de contaminação é apresentado. O modelo, chamado de “diagrama de influência”, permite que os pesquisadores calculem a probabilidade condicional de qual água subterrânea é imprópria para uso, levando em consideração os resultados dos levantamentos geofísicos, as estimativas do algoritmo de aprenziado de máquina, as incertezas relacionadas e a probabilidade de contaminação prévia de diferentes setores da planície. O modelo, em adição, permite que se calcule e compare a utilidade esperada de decisões alternativas (perfurar ou não perfurar um poço, ou utilizar outra fonte de água). O modelo é desenvolvido para uso em situações de tomadas de decisões comuns e apesar de ter sido concebido para uma área específica, apresenta um exemplo que talvez possa ser adaptado para outros casos. Algumas adaptações e generalizações do modelo também são discutidas.

Notes

Acknowledgements

The authors wish to thank Dr. Patrizio Torrese for his help in defining the expert probabilities mentioned in section ‘Predicting contamination in the Oltrepò Pavese plain’ and Dr. Jean-Michel Lemieux, Dr. Serge Brouyère and two anonymous reviewers for their suggestions for improving the article.

References

  1. Agip (1994) Acque dolci sotterranee. Inventario dei dati raccolti dall’Agip durante la ricerca di idrocarburi in Italia (dal 1971 al 1990) [Underground freshwater. Inventory of the data collected by AGIP during hydrocarbon exploration in Italy (from 1971 to 1990)]. AGIP S.p.A. Direzione Generale Servizi Centrali per l’Esplorazione, Milano, ItalyGoogle Scholar
  2. Apostolakis G, Bell DC (1995) Modeling the public as a stakeholder in environmental risk decision making. IEEE Trans Syst Man Cybernet 2:1021–1026Google Scholar
  3. Bersan M (2012) Studio dei meccanismi di risalita di acque profonde a elevata salinità nell’acquifero dell’Oltrepò Pavese [A study on the uprising of deep highly saline water into the Oltrepò Pavese aquifer]. PhD Thesis, University of Pavia, Pavia, ItalyGoogle Scholar
  4. Bersan M, Pilla G, Dolza G, Torrese P, Ciancetti G (2010) The uprising of deep saline waters into the Oltrepò Pavese (northern Italy) aquifer: early results. Italian J Eng Geol Environ 1:7–22.  https://doi.org/10.4408/IJEGE.2010-01.O-01 Google Scholar
  5. Bielza C, Gómez M, Shenoy PP (2011) A review of representation issues and modeling challenges with influence diagrams. Omega 39(3):227–241CrossRefGoogle Scholar
  6. Boni A (1967) Note illustrative della carta Geologica d’Italia F. 59 Pavia [Illustrative notes of the Italian geological map F sheet 59 Pavia]. Stabilimento L. Salomone, Calibria, ItalyGoogle Scholar
  7. Braga G, Cerro A (1988) Le strutture sepolte della pianura pavese e le relative influenze sulle risorse idriche sotterranee [The buried structures of the Pavia alluvial plain and their influences on the ground water resources (south western Lombardy, Italy)]. Atti Ticinensi Sci Terra 31:421–433Google Scholar
  8. Brambilla G (1992) Prime considerazioni cronologico-ambientali sulle filliti del Miocene superiore di Portalbera (Pavia-Italia settentrionale) [First chronoenvironmental considerations about the Upper Miocenic phyllites of Portalbera (Pavia-Northern Italy)]. In: Nuove Ricerche Archeologiche in Provincia di Pavia, Proceeding of the Convegno di Casteggio. Civico Museo Archeologico di Casteggio e dell’Oltrepò Pavese, Casteggio, Italy, pp 109–113Google Scholar
  9. Cameron E, Pilla G, Stella FA (2018) Application of statistical classification methods for predicting the acceptability of well-water quality. Hydrogeol J 26(4):1099–1115.  https://doi.org/10.1007/s10040-018-1727-0 CrossRefGoogle Scholar
  10. Conti A, Sacchi E, Chiarle M, Martinelli G, Zuppi GM (2000) Geochemistry of the formation water of the Po plain (northern Italy): an overview. Appl Geochem 15:51–65.  https://doi.org/10.1016/S0883-2927(99)00016-5 CrossRefGoogle Scholar
  11. Cox LA (2015) Breakthrough in decision science and risk analysis. Wiley, Chichester, UKGoogle Scholar
  12. Dash CJ, Sarangi A, Singh DK (2010) Spatial variability of groundwater depth and quality parameters in the National Capital Territory of Delhi. Environ Manag 45(3):640–650.  https://doi.org/10.1007/s00267-010-9436-z CrossRefGoogle Scholar
  13. Eidsvik J, Murerji T, Bhattacharrjya D (2015) Value of information in the earth sciences. Cambridge University Press, Cambridge, UKGoogle Scholar
  14. Friedel S (2003) Resolution, stability and efficiency of resistivity tomography estimated from a generalized inverse approach. Geophys J Int 153:305–316.  https://doi.org/10.1046/j.1365-246X.2003.01890.x CrossRefGoogle Scholar
  15. Haimes YY (2016) Risk modeling, assessment and management. Wiley, Chichester, UKGoogle Scholar
  16. Hassaf H, Saadeh M (2009) Geostatistical assessment of groundwater nitrate contamination with reflection on DRASTIC vulnerability assessment: the case of the upper Litani Basin, Lebanon. Water Resour Manag 23(4):775–796.  https://doi.org/10.1007/s11269-008-9299-8 CrossRefGoogle Scholar
  17. Heckerman D, Breese JS (1996) Causal independence for probability assessment and inference using Bayesian networks. IEEE Trans Syst Man Cybernet 26(6):826–831.  https://doi.org/10.1109/3468.541341 CrossRefGoogle Scholar
  18. Henriksen HJ, Rasmussen P, Brandt G, Von Bülow D, Jensen FV (2007) Public participation modeling using Bayesian networks in management of groundwater contamination. Environ Modell Soft 22(8):1101–1113.  https://doi.org/10.1016/j.envsoft.2006.01.008 CrossRefGoogle Scholar
  19. Hermans H, Irving J (2017) Facies discrimination with electrical resistivity tomography using a probabilistic methodology: effect of sensitivity and regularisation. Near Surf Geophys 15(13):25.  https://doi.org/10.3997/1873-0604.2016047 Google Scholar
  20. Jensen FV, Nielsen TD (2007) Bayesian networks and decision graphs, 2nd edn. Springer, Heidelberg, GermanyGoogle Scholar
  21. Kingston GB, Maier HR, Dandy GC (2008) AI techniques for hydrological modeling and management I: simulation. In: Robinson NL (ed) Water resources research progress. Nova, New York, pp 15–65Google Scholar
  22. Kjærulff UB, Madsen AL (2013) Bayesian networks and influence diagrams: a guide to construction and analysis, 2nd edn. Springer, Heidelberg, GermanyGoogle Scholar
  23. MacLean L, Ziemba WT (eds) (2013) Handbook of the fundamentals of financial decision making. World, SingaporeGoogle Scholar
  24. Mohajerani H, Kholghi M, Mosaedi A, Farmani R, Sadoddin A, Casper M (2017) Application of Bayesian decision networks for groundwater resources management under the conditions of high uncertainty and data scarcity. Water Resour Manag 31(6):1859–1879.  https://doi.org/10.1007/s11269-017-1616-7 CrossRefGoogle Scholar
  25. Pang-Ning T, Steinbach M, Kumar V (2006) Introduction to data mining. Pearson, LondonGoogle Scholar
  26. Pellegrini L, Vercesi PL (1995) Considerazioni morfotettoniche sulla zona a Sud del Po tra Voghera (PV) e Sarmato (PC) [Considerations about the morphotectonic setting of the area located south of the River Po between Voghera (PV) and Sarmato (PC)]. Atti Ticinensi Sci Terra 38:95–118Google Scholar
  27. Pilla G, Sacchi E, Ciancetti G (2007) Studio idrogeologico, idrochimico ed isotopico delle acque sotterranee del settore di pianura dell’Oltrepò Pavese (Pianura lombarda meridionale) [Hydrogeological, hydrochemical and isotopic study of the underground water in the Oltrepò Pavese plain (southern Lombardy plain)]. Giornale di Geologia Applicata 5:59–74.  https://doi.org/10.1474/GGA.2007-05.0-05-0167 Google Scholar
  28. Pilla G, Torrese P, Bersan M (2010) Application of hydrochemical and preliminary geophysical surveys within the study of the saltwater uprising occurring in the Oltrepò Pavese plain aquifer. Boll Geofis Teor Appl 51(4):301–323Google Scholar
  29. Pilla G, Torrese P, Bersan M (2015) The uprising of deep saline paleo-waters into the Oltrepò Pavese aquifer (northern Italy): application of hydro-chemical and shallow geophysical surveys. In: Lollino G, Arattano M, Rinaldi M, Giustolisi O, Marechal JC, Grant GE (eds) Engineering geology for society and territory, vol 3. Springer, Cham, pp 393–397.  https://doi.org/10.1007/978-3-319-09054-2_82 Google Scholar
  30. Raiffa H (1968) Decision analysis: introductory lectures on choices under uncertainty. Addison-Wesley, BostonGoogle Scholar
  31. Regione Lombardia, Eni Divisione Agip (2002) Geologia degli acquiferi padani della Regione Lombardia [Geology of the Padanian plain aquifers of Lombardy]. S.EL.CA, Florence, ItalyGoogle Scholar
  32. Reynolds JM (2011) An introduction to applied and environmental geophysics, 2nd edn. Wiley, Chichester, UKGoogle Scholar
  33. Ricci PF (2006) Environmental and health risk assessment and management. Springer, Heidelberg, GermanyGoogle Scholar
  34. Schachter RD (1986) Evaluating influence diagrams. Oper Res 34(6):871–882.  https://doi.org/10.1287/opre.34.6.871 CrossRefGoogle Scholar
  35. Shenoy PP (1992) Valuation-based systems for Bayesian decision analysis. Oper Res 40(3):463–484.  https://doi.org/10.1287/opre.40.3.463 CrossRefGoogle Scholar
  36. Shenoy PP (1994) A comparison of graphical techniques for decision analysis. Eur J Oper Res 78(1):1–21.  https://doi.org/10.1016/0377-2217(94)90118-X CrossRefGoogle Scholar
  37. Strimling D (2008) Operations research and decision analysis: software tools and applications. In: Linkov I, Ferguson E, Magar VS (eds) Real-time and deliberative decision making: application to emerging stressors. Springer, Dordrecht, The Netherlands, pp 285–303Google Scholar
  38. Tartakosvky DM (2013) Assessment and management of risk in subsurface hydrology: a review and perspective. Adv Water Resour 51(2013):247–260.  https://doi.org/10.1016/j.advwatres.2012.04.007 CrossRefGoogle Scholar
  39. Terek M (2005) Decision trees and influence diagrams in decision analysis. J Appl Math Stat Inform 1:121–335. http://jamsi.fpv.ucm.sk/docs/v01n01_04_2005/v01n01_09_Terek.pdf. Accessed November 2017
  40. Torrese P, Pilla G, Bersan M, Rainone ML, Ciancetti G (2009) Mapping the uprising of highly mineralized waters occurring along a fault zone in the “Oltrepò Pavese” plain upper aquifers (northern Italy) using VLF-EM survey. In: Proc. of SAGEEP 22nd annual symposium on the Application of Geophysics to Engineering and Environmental Problems. Environment and Engineering Geophysical Society, pp 749–760.  https://doi.org/10.4133/1.3176764
  41. Torrese P, Pilla G (2015) Comparison between VLF-EM and resistivity anomalies associated with salt paleo-water contaminations in an alluvial aquifer: the san re test site (norther Italy). In: Proceeding of SAGEEP 28th annual symposium on the Application of Geophysics to Engineering and Environmental Problems. Society of Exploration Geophysicists and Environment and Engineering Geophysical Society, pp 631–637.  https://doi.org/10.4133/SAGEEP.28-090
  42. Torrese P, Pilla G (2016) Time-lapse 3D ERT imaging of a brackish paleo-water contamination at the alluvial aquifer of the San Re test site (northern Italy). In: proceeding of the symposium on the application of geophysics to engineering and environmental problems. Society of Exploration Geophysicists and Environment and Engineering Geophysical Society, pp 635–643.  https://doi.org/10.4133/SAGEEP.29-101
  43. Vick SG (2002) Degrees of belief: subjective probability and engineering judgment. ASCE, Washington, DCGoogle Scholar
  44. Woodard DB, Wolpert RL, O’Connell MA (2010) Spatial inference of nitrate concentrations in groundwater. J Agric Biol Envir St 15(2):209–227.  https://doi.org/10.1007/s13253-009-0006-x CrossRefGoogle Scholar
  45. Zhou B, Dahlin T (2003) Properties and effects of measurements errors on 2D resistivity imaging surveying. Near Surf Geophys 1(3):105–117. http://portal.research.lu.se/portal/files/2917689/4064844.pdf. Accessed 22 June 2018CrossRefGoogle Scholar

Copyright information

© Springer-Verlag GmbH Germany, part of Springer Nature 2018

Authors and Affiliations

  • Enrico Cameron
    • 1
    Email author
  • Giorgio Pilla
    • 2
  • Fabio A. Stella
    • 3
  1. 1.GeoStudio – Environmental and Geological Consulting OfficeMorbegnoItaly
  2. 2.Department of Earth and Environmental SciencesUniversity of PaviaPaviaItaly
  3. 3.Department of Informatics, Systems and CommunicationUniversity of Milano BicoccaMilanItaly

Personalised recommendations