Assessment of sources of PCE contamination in groundwater using a Monte Carlo method in the Functional Urban Area of Milano

Abstract

Water Framework Directive and Groundwater Directive (2006/118/CE) represent the main European regulations on water resources quality, requiring the identification and management of contamination sources threatening the achievement of acceptable groundwater quality status. The sources of groundwater contamination can be classified into two different categories: 1) point sources (PS), which are identified areas releasing plumes of high/very high concentrations (i.e. hot-spots) and 2) multiple-point sources (MPS) constituted by a series of unidentifiable small sources clustered in a large area, generating a diffuse contamination. The latter category predominates in European Functional Urban Areas and cannot be managed with the usual remediation techniques such as those conceived for large/medium contaminated sites, mainly because of the difficulty to identify the many different source areas releasing small contaminant mass. Consequently, the usual remediation procedures are not economically sustainable and often fail to provide results in an acceptable time frame. A powerful tool in this context is transport modeling, as it can be used in combination with statistical methods to assess MPS groundwater contamination even in a highly undetermined setting. Due to the uncertainty related to the exact position and intensity of MPS, a numerical model (MODFLOW/MT3DMS) was implemented in a pilot area in the North-Eastern sector of the Milano FUA. Using the inverse calibration code PEST, a model was calibrated representing the diffuse PCE source field (Italian Law limit is within 1.1 mg/l). This was then used as input to a statistical process based on the Null-Space Monte Carlo (NSMC) method, which allows to generate unlimited sets of sources, all respecting the measured concentrations. 99 different realisations were thus obtained, each attributing a contaminant inflow (with varying concentrations) to every cell of the model top layer. The model minimizing the objective function (composed by the diffuse concentration targets in monitoring wells), gives useful information to assess the overall contaminant sources distribution and to identify the probability of each domain sector to contribute to the contaminant mass inflow

Null-space Monte Carlo particle tracking to assess groundwater PCE (Tetrachloroethene) diffuse pollution in north-eastern Milan functional urban area

“…The simplified approach aims at identifying the most probable paths followed by groundwater contamination, which coincide with the groundwater flow directions. An alternative method that applies NSMC considering all the transport processes and the release of contaminant mass can be found in (Alberti et al, 2017), but its application is more complicated as it requires a transport model, which also takes a far higher simulation time compared to particle tracking.…”

Model calibration using the automatic parameter estimation procedure (PEST) of the North-eastern zone of the Milan Functional Urban Area (Italy)

“…3b). Il modello matematico è stato implementato soltanto per i primi due acquiferi che sono quelli interessati dalla contaminazione, oggetto dello studio successivamente sviluppato attraverso il modello di trasporto (Colombo 2017) e la procedura di modellazione stocastica inversa (Alberti et al 2018(Alberti et al , 2017 Conducibilità idraulica e porosità La risoluzione dell'equazione discretizzata alle differenze finite presuppone la definizione, per ogni singola cella del modello, dei parametri strutturali ed idrogeologici. Le celle del modello sono caratterizzate da condizioni e proprietà idrauliche (carichi, conducibilità idraulica e porosità efficace) di seguito descritte.…”

Statistical methods and transport modeling to assess PCE hotspots and diffuse pollution in groundwater (Milan FUA)

“…the cells that most probably can host a MPS; 2. clustered MPS: through 100 MC realizations considering variability of contaminant mass released into the shallow aquifer (computed as a product of concentration and Darcy's flow is within a range of 10 -2 -1 mg/s). Using MT3DMS, for each domain sector, it was possible to assess in a probabilistic way, the MPS contaminant distribution and the mass releases frequency (Alberti et al 2017). From many different realizations all calibrated within NSMC (Null Space Monte Carlo suite in PEST), it was therefore possible to provide results in terms of frequency of occurrence that identified areas of high probability to find MPS.…”