CORDIS - Risultati della ricerca dell’UE

Security and decontamination of drinking water distribution systems following a deliberate contamination

Final Report Summary - SECUREAU (Security and decontamination of drinking water distribution systems following a deliberate contamination)

Executive summary:

The contamination of drinking water infrastructures (catchment areas, raw water transfer systems, treatment facilities, treated water reservoirs and distribution networks) with CBRN (chemical, biological, radionuclides) as a result of malevolent acts of sabotage represents one of the major threats that security has to face with. Rapidly restoring the functionality of such infrastructures, and the access to safe drinking water represents another major concern for regulatory agencies and water utilities. Indeed, the damage resulting from impairment of drinking water services would seriously impact the quality of life of many people not only by directly harming them but also making water systems unusable for a long period of time with a risk of societal disorder (similar situation as with any accidental contamination events or natural disasters).

The main results of SECUREAU are organised under ten items:
1. Sensors and sentinel coupons (non-specific sensors; specific sensor; biofilm sensor; sentinel coupons)
2. Optimal location of water quality sensors and data treatment
3. Optimal distribution of sentinel coupons
4. Identification of the sources of contamination and the contaminated areas
5. CBRN analytical methods
6. Modeling sorption and desorption
7. Pipe wall cleaning and decontamination
8. Handling of decontamination sludge and water
9. Validation of decontamination
10. Decision tool

The main socio economic impacts of the SECUREAU project are described in four main items:
(i) the security and safety synergy reinforcement for the benefit of EU citizens;
(ii) the development of EU industry and leadership on environmental sensors networks;
(iii) the development of modern analytical tools for contamination surface analysis and
(iv) the decontamination strategies combining complementary technics.

Project Context and Objectives:

The importance of water and water infrastructures to human and ecosystem health and to the functioning of our economy makes water systems targets for terrorism. Their contamination with CBRN agents as a result of malevolent acts of sabotage represents one of the major challenges that security has to face with. The damage is done not only by hurting people but also rendering water systems unusable for a long time with a risk of societal disorder.

In order to restore rapidly safe drinking water distribution, SECUREAU (an European project for restoring distribution systems after deliberate CBRN attack involving 14 partners from 6 countries) has defined three levels of research and development:
(i) developing early warning system (EWS) for drinking water distribution systems. It means an integrated system for deploying a monitoring technology (specific and non-specific sensors for detecting unexpected/abnormal signal variations), mathematical models for positioning these sensors, analysing and interpreting the huge mass of results given by the sensors to make decision while minimizing unnecessary concerns and inconvenience within a community.
(ii) allowing a rapid localization of the source(s) of intentional contamination thanks to accurate methods and softwares.
(iii) defining multistep strategies for cleaning the distribution system: pipe wall/biofilms /deposits. Curative treatments will concern pipes, and wastes (water bulk and deposits), as well as methods for controlling the efficacy of the decontamination procedure.

SECUREAU serves as a research project for designing and implementing an effective and timely response action after drinking water distribution system contamination which may be integrated in the management of a crisis phase (ISO 2011). The restoration of the network will have to provide maximum and rapid benefits to the users with limited environmental effects. The SECUREAU programme was organised in 8 workpackages:

1: Project management
2: Early warning systems for rapid detection of deliberated intrusion
3: Off-line rapid detection methods in water and biofilms/deposits
4: Modelling contaminant accumulation
5: Decontamination procedures of water distribution systems
6: Controlling the efficacy of decontamination of deposits
7: Case studies and data support analysis
8: Dissemination, exploitation and, transfer

As a preliminary objective, we aimed to get a "state of the art" related to early warning systems, sensors, etc.. Five main questions were clearly identified:
- What are the key characteristics (accuracy, robustness, price ...) of commercially available sensors for early warning systems (EWS)?
- How to organise pilot bench scale to control biofilm disinfection and sensors testing?
- What methods can be used or in development for cleaning contaminated networks?
- How to improve sensitivity of analytical methods for controlling rapidly water bulk (several known CBRN or surrogate)?
- How to design sensor location to monitor contaminant distribution and/or the origin of injection in the distribution systems?

During the year 2 of the project SECUREAU, five objectives were addressed:
- Definition of the key characteristics (accuracy, robustness, price...) of commercially available sensors for early warning systems (EWS) and improvement of such devices to prepare their implementation in networks and pilots during year 3 and year 4.
- Finalisation of pilot bench scale to control biofilm disinfection and sensors testing.
- Sensitivity improvement of off-line analytical methods used for detection of CBRN threat agents in water and following adsorption to biofilms and other deposits occurring in water distribution networks
- Continue the development of methods for cleaning contaminated networks (both chemical and physical methods)
- Design sensor location to monitor contaminant distribution and/or the origin of injection in the distribution systems.

Six main objectives were carried out in the third year of SECUREAU:
- Analytical methods were adapted from literature and validated for the detection of CBRN in the matrices of interest, having special attention to matrix effects. Three groups of chemicals or microorganisms or radionuclides were concerned:
paraquat, chlorfenvinphos, carbofuran, pentabromodyphenylether (BDE-100) and methylmercury in deposits and/or biofilms.
alpha, beta and gamma radionuclide emitters both in drinking water and in biofilms and deposits
various bacterial pathogens (Y. pestis, E. coli, Bacillus), including surrogate species (F. tularensis, Y. pseudotuberculosis, B. subtilis) in drinking water.
- Commercial on-line sensors (MSS, Optiquad and Neosens) and coupon monitoring devices were tested in three loop systems.
- Three new sensor prototypes measuring (i) pH, temperature, chlorine, conductivity, (ii) organic matter and turbidity, and (iii) radionuclides, were developed for early warning systems (EWS); Their implementation in two large drinking water networks (one rural, one urban) was planned as well than testing their behaviour in different pilots.
- Optimal placement of sensors was also a main objective of this work period. It was planned according to several optimisation criteria for two selected sites in Europe (one rural site and one urban site). Mathematical models were developed to define both the contamination source origin, and the contaminated area.
- Different decontamination procedures were planned to be tested and evaluated. As an example, the applicability of cavitation method to remove bacterial spores from contaminated water and pipes surfaces, by taking into account both engineering and economical aspects was planned. The development of chemical methods (dissolution, chlorination, dispersion, complexation and Fenton-like) was also taken in account. In particular, very promising results were reached with Fenton-like oxidation, which proved to be effective for degradation of SECUREAU chemical contaminants in water. Pipe deposits also revealed to act as catalysts in this process, which is already being tested at pilot scale.
- Implementation of methods and/or strategy for the treatment of contaminated water and deposits is also taken into account.

The efforts spent during year 4 have permitted to continue and to improve and to conclude the work initiated during the previous periods, and also to communicate via per review publications; technical meetings; European conference. Seven main objectives were clearly identified for this year 4:
- Final experiments (focused on wall deposits and biofilms) were carried out to determine the kinetic parameters for adsorption and desorption of chemicals, radionuclides and microorganisms in biofilms and inorganic deposits. Values were obtained for sorption of: (a) paraquat to clay and iron oxide materials; (b) mercury to iron-based and calcium-rich deposits; (c) Po-210 and Am-241 to biofilms and to deposits scraped from cast iron pipes; (d) Francisella philomiragia, Yersinia pseudotuberculosis, Francisella tularensis and Bacillus subtilis to drinking water biofilms.
- The modelling part was divided in two parts. One was the software development to simulate contaminant transport along drinking water networks. The other was the development of software tools to identify the source of contamination (together with the optimization of sensor positioning, see WP2).
- Decontamination methods of water systems and handling of obtained sludges was one of the year-4 objectives. In the reporting period experiments were carried out both in laboratory scale and in three pilot loop systems. Results aimed to select the methods, which are closed to practical application, and to evaluate their applicability in real situation (economical, technical or safety reasons).
- Selected decontamination methods were tested with the aim to clean/decontaminate three pipe reactors, each of them being experimentally contaminated with either a very soluble organic pesticide (glyphosate), or spores of Bacillus subtilis, or not contaminated (then the target was the autochthonous biofilm).
- In addition controlling the efficacy of decontamination procedures was planned according three steps (i) to develop a concept for positioning coupons and sensors at representative and indicative locations in the distribution system, (ii) to optimize and to verify the decontamination efficacy by use of the sensors/coupons and (iii) to use the information as generated by the earlier tasks of WP6 in order to provide a decision tool.
- The installation of two networks (one rural and one urban) of 40 sensors each was planned to get field information and to generate a know-how to install an early warning system in real cases.
- In terms of communication, the main objective for the months 37 to 48 was to ensure a clear and effective dissemination, exploitation and transfer of the results to (i) partners inside the group, (ii) the international scientific community, (iii) the industry and (iv) the general public (in particular thanks to a European Conference).

All over the project, recommendations provided by external reviewers (Walter Biederbick and Clive Thompson) following two scientific reviews (Feb 2011 and Feb 2012) and have been taken into account.

Project Results:

Introduction: context and actions.

The contamination of drinking water infrastructures (catchment areas, raw water transfer systems, treatment facilities, treated water reservoirs and distribution networks) with CBRN (chemical, biological, radionuclides) as a result of malevolent acts of sabotage represents one of the major threats that security has to face with. Rapidly restoring the functionality of such infrastructures, and the access to safe drinking water represents another major concern for regulatory agencies and water utilities. Indeed, the damage resulting from impairment of drinking water services would seriously impact the quality of life of many people not only by directly harming them but also making water systems unusable for a long period of time with a risk of societal disorder (similar situation as with any accidental contamination events or natural disasters).

Such accidental or malevolent contamination events determine a crisis situation, which affects or is likely to affect a water utility or its provided services, and require more than the usual means of operation and / or organisational structures to deal with it. The ISO 11830 standard on crisis management describes the fundamentals of a crisis management system (ISO, 2011). ISO 11830 provides general guidance on how a crisis should be dealt with ("crisis phase"), on how to re-establish services (post-crisis phase), and on the best way to draw conclusions and revise procedures for future events.A large number of "preventive" actions have to be taken / implemented on a routine basis before the crisis, and recovery activities can start during the crisis phase. At the same time, specific scenarios for supplying potable waters (distribution of bottled waters, water tankers, emergency water treatment mobile systems, and so on) also have to be developed (Loo et al. 2012).

In this context, SECUREAU (a European project for restoring distribution system functionality after deliberate CBRN attacks involving 12 partners from 6 countries) has defined four research and development objectives:
a. tools for detecting water quality changes by combining generic non-specific and specific sensors for measuring unexpected / abnormal signal variations (and to be integrated in an early warning system); methods for identifying the best locations for sensors for full coverage monitoring; methods for treating the data collected;
b. methods for rapidly identifying the source(s) of intentional contamination thanks to accurate procedures and software;
c. multi-step strategies for cleaning distribution systems: pipe walls / biofilms / deposits and waste (water bulk and deposits extracted from the network).
d. analytical methods for confirming cleaning procedure efficiency.

1 Sensors and sentinel coupons

Fourteen commercial non-specific sensors have been evaluated regarding 12 technical criteria. Three of them were selected and are devoted to water analysis (free chlorine, pressure, temperature, conductivity, organic matter, turbidity and radionuclides). Five other devices are used for assessing the accumulation of deposits on the surface of the pipes.
Sensors devoted to water analysis have two purposes. The first is to rapidly detect any abnormal changes in water quality, and the second is to provide assistance to the operator to manage the drinking water network under normal operational conditions.

1.1. Non-specific sensors (Kapta 3000 AC4 and Kapta 3000 OT3)

The Kapta 3000 AC4 sensor measures free active chlorine, pressure, temperature and conductivity (Deliverable 2.1.2 version 2, June 2012). In the context of SECUREAU, two European drinking water systems were equipped with 80 sensors each which measure water quality online and send results every two hours to operational control centres thanks to a GSM communication module.

1.2 Specific sensor (Kapta 3000 RAD1)

A new type of online sensors for measurement of traces of gamma emitters in solution has been proposed in SECUREAU.

The detection limits (Bq/kg, measurement for 10 seconds) with the NaI(Tl) detector are presented. Using an integration time of 10 s, the sensor allows an alarm to be raised before the maximum permitted levels of radioactive 60Co and 137Cs are reached. On the other hand, the limit of detection is too high for 241Am. In order not to exceed the maximum value of 10 Bq L-1, the integration time should be around 90 min (241Am is essentially an alpha emitter and the system is not well adapted). This sensor is designed to be easily and quickly installed in drinking water distribution networks.

1.3 Biofilm sensors (MSS, Optiquad and Neosens FS-900 / SkidSens)

Allowing deposits to be monitored online is one of the key aspect as contaminants, depending on their nature, may adsorb onto pipe walls and deposits. Biofilm/deposit sensors appear very useful for evaluating the effectiveness of a cleaning procedure throughout the network, meaning they have to be installed during the pre-crisis period. Three commercially available sensors - MSS, FS-900 and SkidSens – and one prototype - Optiquad - were tested over a relatively short time period of one month to monitor the deposits formed on the pipe surfaces in a drinking water distribution network. However, their interest is quite limited in a drinking water distribution system for three main reasons:
- First, their high cost does not allow them to be used in large numbers in a network,
- Secondly, the low sensitivity of FS-900 and SkidSens does not allow them to be used in drinking water systems.
- Third, the OptiQuad sensor can only detect the first deposition events. The latter act as a screen and inhibit the spread of the signal, thus preventing the sensor from distinguishing between thin and thick deposits.

The four systems need a power supply (Optiquad: 230 V, FS-900 and SkidSens: 24 V or 230 V) and calibration step (to convert the signal to cell counts or deposit thickness, for example).

1.4 Sentinel coupons

Sentinel coupons should be installed during the pre-crisis phase in drinking water distribution networks so as to allow, during and after the crisis period, (i) the deposits on the coupons to be analysed, (ii) the adsorbed contaminants to be identified and (iii) the efficiency of the decontamination procedures to be assessed. The materials used for the coupons should not be corrodible as they will be in the pipes for years.

2 Optimal location of water quality sensors and data treatment

The installation of water quality sensors in a drinking water system allows an alert to be issued rapidly when abrupt changes in the quality of water are detected. It is necessary to ideally distribute these sensors taking into account a number of considerations. Indeed, water supply providers may consider expenses (limited budgets), efficiency (willingness to detect a contamination event within less than x hours), protection of specific groups of consumers (e.g. children, hospital patients or retirement home residents), etc. These considerations can of course be combined, always with the same purpose in mind: to alert as quickly as possible (early warning concept). Objectives and constraints need to be defined to ensure that sensors are distributed in the best possible way: some are sensor specific (technical considerations for installation), others are to be considered when it comes to cleaning, while others are peculiar to population vulnerability and financial costs.

3. Optimal distribution of sentinel coupons

Sentinel coupons are simple tools which can be used to control the efficiency of decontamination and to identify the composition of deposits throughout the network under normal operational conditions. The coupons must be positioned optimally throughout the network so as to be representative of water quality interaction with pipe surfaces while minimizing costs. Consequently, the mathematical models dedicated to the spatial optimisation of sentinel coupons differ from those dedicated to the spatial optimisation of generic water quality sensors, because the objectives and constraints are not the same.

4. Identification of the sources of contamination and the contaminated areas

If a contamination event (accidental or deliberate) occurs in a drinking water network, it is essential to identify the sources of contamination and to determine the area which is likely to be contaminated, in order to isolate and decontaminate the affected area only, as well as keep supplying drinking water in non-affected areas, depending of the nature of the contaminant.

4.1. Localization of the sources

Different strategies were implemented during this project to identify the location of the contamination sources. The approaches considered: a) a deterministic method based on successive positive readings of sensors; b) methods based on artificial neural networks (ANNs) for single and multiple contamination events; c) stochastic methods such as least-squares solving with Tikhonov regularisation or minimum relative entropy solution (MRE); and d) a method based on the analysis of flow data.

- The deterministic method based on successive positive readings of sensors is based only on the analysis of the residence time of water in pipes and it only requires a binary sensor status over time. The results for the localization of contamination sources are given sequentially, being updated each time a new sensor detect a change in contaminant concentration. Furthermore, in some situations, this method enables the verification of the occurrence false negatives and false positives. It was observed that the occurrence of false positives or negatives did not affect the results related with the real detections and it was possible to detect the sensor which suffered this anomaly.
- The method that used ANNs to identify the contamination sources in single contamination events is able to identify the correct contamination source and to predict the correct time of contamination associated to each possible contamination source, even in the case of large and highly complex DWDSs, following two approaches. The tests performed with contamination scenarios considering water demand uncertainty demonstrated that the method has shown good performance even in situations that are not described by the hydraulic model used in the development of the ANNs. Both approaches required a very low time of computation to obtain the results for real DWDSs, generally less than 5 seconds in a 3.10 GHz processor, which is a great improvement to the solution of this problem that demands the computation of the results as quick as possible.
- The method that extended the application of ANNs for multiple contamination scenarios achieved very satisfactory results for real DWDSs. The method was generally able to determine correctly the simulated source and to define a very restricted set of possible contamination sources, even when considering hydraulic scenarios with demand uncertainties. However, the estimations of the time of contamination for scenarios under demand uncertainties showed larger deviation in relation with the simulated contamination sources. The time of computation required was generally very low, which makes this method very suitable for application in real contamination scenarios.
- The source identification with general inverse problem methods developed by Irstea partner is a two-step enumeration/exploration method. Firstly, the in/out transport matrix is worked out with a backtracking method and then, minimum relative entropy method, without any assumption for the pdf (probability density function) distribution, or the least squares method with Tikhonov regularization are used to refine the results and give mean of being a source as a confidence interval for this mean. The backtracking algorithm yields good results in giving very quickly the full list of potential node sources of contamination at the different times, and the in/out (transport) matrix that gives the relation between the potential source and the detecting sensors. That matrix can then be used either on a minimum relative entropy method or a Tikhonov method. Both produces good results, the real contaminant source was always determined as potential source. The minimum relative entropy method looks discriminate more the potential nodes than the Tikhonov method (with selection criterion on strictly positive expected values – other are possible). For the Tikhonov the positivity constraint was relaxed. As a result, negative expected values are possible, and may be interpreted as non-important potential source contamination.

4.2. Spread issues

Several approaches have been used to develop software tools for simulation of contamination transport. These approaches considered: a) an off-line software tool based on equations governing bacterial regrowth that is affected by sorption, desorption, chlorine and substrate concentration; b) a software tool considering sorption developed using MATLAB, VBA and EPANET with models for the evaluation of contaminant concentrations; c) an on-line software tool that uses flow direction data for tracking contamination spread.
- The off-line software tool is supported in a model developed in Epanet-MSX model. The model contains differential equations defining functions of attached bacteria, bulk bacteria, substrate and chlorine concentrations as functions of time and considered also the phenomena of pathogen adsorption/desorption. The model parameters are user-adjustable as various types of contaminants have different adsorption and desorption coefficients. Graphical user interface of Epanet software has been added to the model. The graphical user interface allows operator to modify model parameters, set initial conditions (e. g. contamination sources), view results in graphical or tabular form as well as visualize distribution of contamination over the network. The advantage of the off-line software tool is that it contains a comprehensive model that besides convection takes into account adsorption/desorption and regrowth of bacteria as well as chlorine and substrate concentration. The model can be used to run simulations and study proliferation of contamination, long-term effects including sorption, effects of chlorine disinfectant addition and substrate concentration. The best application of the off-line software tool is to simulate contamination scenarios, look for vulnerable parts of the networks that, if contaminated, allow the contamination to spread to a large part of the network, test effects of different chlorine concentration.
- The software tool developed on Matlab and VBA enables testing the effect of sorption phenomena on contamination spread in drinking water distribution systems and studying long-term behaviour of a partially adsorbed contaminant in a drinking water distribution system. It has been shown that the proposed method is suitable for the study of the effects of the sorption phenomena in the modelling of the transport of contaminants in real drinking water distribution system.
- The on-line software tool enables running simulations of contamination transport in a water distribution network based on flow direction data. The concept of the on-line software tool is based on the idea that in case of contamination accident the affected area of the network is mostly determined by flow directions rather than flow magnitudes. Flow direction data can be obtained by means of flow direction sensors or by hydraulic simulation. Combined approach (flow direction sensors installed in some pipes, simulation is used for other pipes) is also possible. The advantage of the method is that in case flow direction sensors are used, the software tool uses real-time data from the network and therefore provides more robust simulation results.

5. CBRN analysis

Analytical methods were developed or improved in the SECUREAU project for a number of typical CBRN pollutants. Some of them were developed for SECUREAU teams' own purposes while others could be used in the case of contamination. All the methods used provided accurate results when determining the nature and / or concentration of a contaminant in water. However, determining the nature and / or concentration of a pollutant adsorbed onto the walls of the network was systematically found to be a major difficulty.

5.1 Organic chemicals

Analytical methods for the SECUREAU's chemicals in waters were implemented and fully validated, either for rapid detection of a deliberate contamination (higher limits of detection) or for controlling the efficacy of the decontamination (limits of detection compatible with safety requirements). For paraquat, chlorfenvinphos and carbofuran, existing methods were adapted, but for BDE-100 an entirely new analytical method was developed and validated, which has the advantage over the existing ones of requiring fewer sample volume. BDE-100 method consists in a dispersive liquid-liquid microextraction technique prior to gas chromatography-mass spectrometry analysis. A detection limit of 0.5 ng/L for BDE-100 was achieved. Additionally, this method has the advantage of being extended to the quantification of other polybrominated diphenyl ethers (BDE-28, 47, 85, 99, 153, 154 and 183).

If lower limits of detection need to be achieved, a method for the detection and quantification of carbofuran and chlorfenvinphos in waters was also developed. This methodology consists of a dispersive liquid-liquid microextraction followed by gas chromatography-mass spectrometry analysis and allows the detection at lower levels of concentration. Under the optimized conditions, the detection limits were 0.04 µg/L for carbofuran and 0.02 µg/L for chlorfenvinphos.

Concerning chemicals quantification in inorganic deposits, different approaches were taken depending on the chemical and on the nature and composition of the deposit. Different inorganic deposits were tested: iron rich deposit (APP1), manganese deposit (P45), calcium carbonate rich one (BayCa) as well as kaolin:
- Paraquat revealed an extremely high affinity to kaolin and only drastic conditions, such as reflux with concentrated sulfuric acid (during 4 h) were able to remove paraquat from it. Since the interaction between paraquat and real inorganic deposits (APP1, P45 and BayCa) is lower, a simple extraction with saturated ammonium chloride is sufficient to carry the analyte from the solid to the liquid phase, where it is quantified.
- A methodology was successfully developed for the quantification of chlorfenvinphos in kaolin. Acetonitrile was used as a desorption solvent of the pesticide from kaolin. The same solvent showed to be ineffective to desorb chlorfenvinphos from the real deposits tested (BayCa and P45). This result sustains that a matrix-matched calibration approach should be used for each deposit, for quantification purposes. Carbofuran adsorption to inorganic deposits showed to be slower than chlorfenvinphos. As for chlorfenvinphos, acetonitrile was not able to desorb carbofuran from the contaminated deposits.

Chlorfenvinphos extraction from biofilms was also studied. Acetonitrile was the solvent used in the desorption step , and it was able to achieve high levels of desorption of chlorfenvinphos from the biofilm, providing that the determination of the level of contamination of a particular biofilm is possible.

Determination of methylmercury (MeHg) was started with the addtion of sulphuric acid and NaBr to liberate MeHg from the matrix and to form a neutral MeHgBr complex. MeHgBr complex was extracted either with pure hexane from water and biofilms or with 50% dichlromethane/hexane from deposits. After extraction a volatile phenyl derivative of MeHg was formed. MeHg quantification method was based on GC-HRMS technique using isotope dilution. For water samples MeHg method showed good sensitivity (LOD of 0.21 ng/l) and 96-111% recovery. The method for biofilm/deposit samples showed 15-66% recovery yield depending of deposit matrix.

5.2 Biological agents

Methods for the detection of microbiological contaminants in water have been successfully developed and validated. These methods have been tested for their reliability and accuracy, particularly in the detection of viable but non-culturable (VBNC) bacteria which can often be missed in routine measurements but could still pose a threat to public health.
- For Escherichia coli O157:H7, microscopy and molecular methods were found to be effective. The use of highly specific peptide nucleic acid (PNA) probes in a fluorescence in situ hybridisation (FISH) assay, combined with cell elongation (direct viable count, DVC), allowed detection and differentiation of viable populations (including VBNC). In addition, real time PCR (qPCR) combined with propidium monoazide (PMA) could be successfully applied, with a detection limit of 1 cell ml-1.
- Francisella philomiragia (as a surrogate for F. tularensis) and F. tularensis subsp. novicida could not be detected by DVC-PNA-FISH due to their small cellular size and extended growth time. However, the molecular technique of PMA-PCR could be successfully employed, with clear differentiation of live and dead populations in water.
- DVC-PNA-FISH was developed and validated for use in detecting Yersinia pseudotuberculosis (as a surrogate for Y. pestis) and PMA-PCR was validated directly on Y. pestis.
- To assess the applicability of these methods on a spore-forming bacterium, Bacillus cereus E33L was used as a surrogate for B. anthracis (as the closest molecular relative). PMA-PCR could not be used due to inability to ensure efficient PMA pre-treatment leading to unreliable results, although qPCR on its own could be used for general detection (with no viability assessment). In contrast, PNA-FISH combined with DVC was found to be highly effective at detecting both vegetative cells and spores, as it caused spores to germinate and show clear elongation.

5.3 Radionuclides

A range of rapid, off-line sensitive radiometric methods were selected and adapted to be able to confirm the intrusion or presence of radionuclides in drinking water networks (water and pipeline deposits). Each method was chosen for its potential to be analytically robust and for its capability in providing rapid screening to accurately detect and quantify alpha, beta and gamma emitting radionuclides, either singly or collectively. All methods were required to be free of any risk of generating false-negative or false-positive data for the broad range of SECUREAU list radionuclides (Cobalt-60, Strontium-90 + Yttrium-90, Iodine-131, Caesium-137, Iridium-192, Polonium-210, Radium-226, Americium-241, Californium-252).

- High-resolution gamma spectrometry (HRGS) is the first analytical choice to investigate unknown materials. As most of the relevant nuclides are ?-emitters the use of ?-spectrometry will ensure that these nuclides can be identified and quantified without any loss of time. The method is characterised by simple sample preparation and provides good identification and accurate quantification (with sufficient detection limits) with one measurement. Even though ?-spectrometry cannot detect all relevant radionuclides (only ?-emitters) it is highly recommended that ?-spectrometry is employed in the case of an assumed or real contamination incident on a drinking water distribution system in addition to LSCcounting or alpha spectrometry.

- LSC measurement or gross alpha-/beta-determination (GABD) is the most convenient off-line analytical method for the verification of a real or assumed attack on a water distribution system. Liquid scintillation counting coupled with spectral analysis (LSC-Spectra+) is preferred to GABD when analysing waters directly or allied with simple acid treatments (leading to dissolution/digestion/decolorisation) for biofilms and deposits. The technique is capable of identifying and quantifying the target radionuclides with limits of detection of 1 Bq/g or better with preparation and counting times taking less than 3 hours. The sensitivity is acceptable for the detection of relevant concentrations of nearly all of the SECUREAU target nuclides.

- A complementary analytical approach to the rapid LSC LSC-Spectral+ method described is alpha spectrometry that uses the specially developed ADAM method (Advanced Deconvolving of Alpha Multiplets). Simplified sample manipulation with sophisticated analysis methods is advantageous when obtaining the results rapidly is of importance. In a radiation emergency event the heavy elements, like uranium and thorium, are rapidly detected using ICP-MS technology. Liquid scintillation spectrometry has advantages for detection of alpha and beta emitting radionuclides simultaneously. However, identification of radionuclides in liquid scintillation spectrometry is challenging because spectrums and peaks overlaps. In spite of that, information from alpha and beta emitting radionuclides can be achieved. Where more detailed identification of alpha particle emitting radionuclides can be achieved by water evaporation, alpha spectrometric measurement and ADAM analysis.

6. Modeling sorption and desorption

In case of contamination of the water in a drinking water distribution network, the obvious response is discarding the contaminated water volume and to rinse the system. However, as the contaminants will interact with the surrounding surfaces (with biofilms) and deposits by adsorption and desorption, such phenomena should be assessed in order to predict the opportunity of cleaning methods.

7. Pipe wall cleaning and decontamination
Overall strategy proposed in SECUREAU was to carry out in situ cleaning (i.e. inside the pipe) and then flush neutralised sludge out of the system for safe disposal. There are some risks with most of these methods, namely, if cleaning is very intensive, pipes can be damaged and contamination will leak out to groundwater. It is therefore in cleaning step we avoided to apply "aggressive" methods such as plasma discharge. On other hand flushing should be more effective than traditional methods, which is mostly focusing on removing loose deposits. It is because corrosion layer inside the pipe will absorb some of contaminants and the only way to remove it is to flush it out with incrustation layer.

The limitations of currently used methods for cleaning drinking water distributions systems were taken into account and effective but simple solution "how to deal with adsorbed CNBR agents" were proposed. Depending of the methods, some of them were applied only in laboratory scale, other also in pilot scale, whereas one of them was tested also in full scale in an old, abandoned network.

7.1. Bacterial agent decontamination

Both spores, non-spore forming bacteria, and viruses were used as models for testing decontamination procedures which are summarized.

Shock-chlorination was studied by adding high concentration of chlorine and keeping to reach optimal CT (concentration multiplied by time) value. Results showed that 5 log10 inactivation of B. subtilis spores are achieved within 3 hours using 200 mg/L of free chlorine (pH adjusted to 6) in the bulk. Although reasonable efficacy was observed in water, biofilm were not effectively removed. Moreover very high CT value can decrease the integrity of the pipes. Experiments with non-spore bacteria showed that disinfection with shock chlorination providing sufficient contact time is an effective method for neutralising bacteria both in water and biofilm. To ensure that this kind of treatment killed also viable and not culturable bacteria the lack of metabolic activity was confirmed with molecular viability methods (e.g. FISH-DVC).

7.2. Decontamination from organic agents

Paraquat (a highly soluble (620 g/l) and highly toxic pesticide), Chlorphenvinphos, and BDE were used as model substances of organic agents in SECUREAU. To follow the principle of in situ cleaning we used Fenton's reaction to mineralised organic to inorganic carbon by means of peroxidation. The process involves a complex mechanism in which the parent molecules are oxidized into other organic compounds and ultimately into carbon dioxide and water. The organic matter oxidation is promoted by the hydroxyl radicals formed in the reaction between hydrogen peroxide and iron. The catalyst, Fe (II), is restored by the reaction of hydrogen peroxide with ferric iron.

7.3. Decontamination from inorganic agents

Mercury was used as model substance of inorganic agent in SECUREAU and several methods were tested (water flushing with chlorinated or non chlorinated water; ice pigging). The results showed that in smooth pipeline materials (such as PEX) large part of Hg2+ was sorbed to the biofilm surface. Then an effective removal of biofilm also removed large major portion of Hg2+. A 5 min high water flow flushing reduced the concentration of Hg2+in the biofilm only by 58 %. Use of low flow chlorinated water (10 mg/l Cl2) flushing for 1 hour reduced the Hg2+concentration in the removed biofilm more 92 %. Total removal rate taking into consideration the fraction remaining in the pipe surface after biofilm removal was 82% with chlorine flush. Flushing time with chlorine should have been longer as Hg2+concentration in both water and biofilm were still clearly decreasing after 1 hour. Higher chlorine concentrations, higher flushing time and possibly also higher flow rate might have helped to reduce the Hg2+ concentration in biofilms and water to an acceptable level. Also, more effective biofilm and corrosion scale removal methods such as ice pigging + Comprex would most likely be very useful. Decontamination of corroded cast iron tube by mixing with ice cubes was found ineffective as 40 % of Hg2+ was bound in the deeper layers of the corroded surface already after 2 hours of incubation that were not removed by shear forces of modest ice flow. Highly effective means that remove the corroded surface of pipelines, such as Comprex with small stones added, should be used for decontamination.

7.4. Decontamination from radioactive agents

Radionuclides can be introduced as in water supply by dissolving in water and depending on their type and species they can strongly absorbed adsorb on the surface of pipes (e.g. cast iron). Thus, their should be desorbed and later concentrated for the safe disposal.

In SECUREAU, chemicals (calcium acetate, sodium bicarbonate, sodium citrate, tri-n-octylamine, ethylenediaminetetraacetic acid and sodium hypochlorite as an traditional disinfectant) as release agents of radiological agents (Americium, Polonium, Radium and radiocolloids) from pipe material (stainless steel, polyvinyl chloride, high density polyethylene, polypropylene) and real pipe deposits was tested.

8. Handling of decontamination sludge and water

In SECUREAU, the applicability of the following sludge handling technologies were investigated: i) separation of solids from water with flocculation and sedimentation, ii) removal of radioactive substances with sorbents; and iii) removal of organic substances with advanced oxidation.

9. Verification of decontamination
In SECUREAU, several solutions and strategy for effective in situ cleaning based on using simple reagents were selected:

- Chemicals such as pesticides as well as pathogens and autochthonous bacteria could be removed using H2O2 for generating free radicals (Fenton's reaction). Metal pipe surface (iron and copper) can be used as catalysts of the reaction. In some cases, the treatment can be carried at pH near the neutrality;
- Removal of resistant microorganisms such as Bacillus spores from pipe surfaces could be achieved by alternating treatment with sodium hydroxide and chlorination;
- For removal on radionuclides, desorption by sodium bicarbonate solution and flushing of the system for safe storage or utilisation offer a remediate option.
- Ice slugs and gravel in combination with water flushing are effective methods for removing both loose deposits and corrosion layer of pipes. Such a treatment is needed as soon as chemicals are suspected to diffuse and sorb easily in the deposits.
- For separation of sludge from water chemical coagulation with addition of flocculants was shown to be effective for handling of decontamination waste.
Whatever the solutions, verification of the decontamination level of the pipe wall is requested either by deposits, biofilms or scale analysis (from in situ samples or sentinel coupons analysis); or by using sensor signals.

9.1. Off-line detection of contaminants in deposits
9.1.1. Chemicals
Acquisition of sufficient amounts of biofilms and the interference of abiotic deposit components with their heterogeneity is a considerable challenge but could be achieved. Methods were developed for paraquat, chlorfenvinphos and methyl mercury in deposit and for methyl mercury in biofilms. Validation included the linearity of the response, the limits of detection and quantification, the precision as expressed by the variation coefficient, the accuracy evaluated by the percentage of recovery from artificially contaminated samples and by the matrix-matched calibration. Paraquat as a representative for herbicides with highly polar properties is strongly adsorbed to the deposits. Therefore, drastic methods were applied for extraction, indicating a low recontamination potential. Chlorfeniphos was more easily extracted and, thus, is more likely to be leached out of a deposit after adsorption. Methyl mercury was determined in both biofilms and deposits and could be reliably detected.

9.1.2. Pathogens
Two general detection strategies have been pursued and compared with more conventional culture methods. One was cultivation based. This method of culture detection of microorganisms is the so-called "gold-standard". However, it is well known for a number of limitations, the most important one being that lack of culturability does not mean lack of presence and possible recovery of microorganisms. Such transient loss of culturability has been termed a "viable but non-culturable" (VBNC)-state and is understood as a stress response. Two main alternative culture-independent and molecular-biological approaches were assessed for application on water samples; the microscopy based combined techniques of cell elongation (DVC) and peptide nucleic acid/fluorescence in situ hybridization (PNA-FISH), and the combination of propidium monoazide with real-time PCR (PMA-PCR). In SECUREAU, these approaches were applied both on drinking water biofilms and on loose deposit samples. Tested pathogens and surrogate bacteria could be detected, and viable populations quantified, using the DVC-PNA-FISH method. Tested pathogens and surrogate bacteria could be detected, and viable populations quantified, using the DVC-PNA-FISH method. These methods are not trivial and require validation. Quantitative polymerase-chain reaction (qPCR) was suited to detect low levels of target bacteria when DNA extraction had been successful, although it cannot distinguish between viable and dead cells. Both of these methods (DVC-PNA-FISH and qPCR) allowed to detect lower levels of target spike bacteria than culturation methods alone, indicating the presence of VBNC cells. In addition, a new protocol for the permeabilization of Bacillus spores was developed and tested. The application of this method will allow rapid and accurate staining of spores and permit easier detection in these complex sample types. The advantages and disadvantages of these methods on a range of samples is now understood and will enable these techniques to be for pathogen detection and quantification in water, biofilm and deposit samples.

9.1.3. Radionuclides
The methods for detection of radionuclides as adapted were acceptably sensitive, even at activities below the recommended safety guidelines.
- Liquid scintillation counting coupled with spectral analysis together with acid treatment of biofilms and deposits are shown to be effective. The technique offers a rapid approach for the identification and quantification of radionuclides in pipeline deposits in an emergency situation.
- The analytical methods are suitable for rapid off-line determination of radionuclides both in abiotic deposits and biofilms.
- Simplified sample manipulation with sophisticated alpha spectrometric analysis (ADAM method) is advantageous if it is necessary to obtain rapid results.

In general, the suitability of the methods for detection of radionuclides down to levels permitted by the safety guidelines was successful.

9.2. Sensors
To obtain information about surface deposits in drinking water pipes, the most direct way is to sample pieces of the pipe and analyze the deposit on their inner surface. However, this requires a construction pit and the replacement of a piece of the pipe. Furthermore, for assessment of cleaning success, another piece of pipe has to be extracted and analyzed.

For deposit analysis, basically three options are possible as already described for fouling monitoring in water systems:
i) Sampling of surfaces by gaining physical access and removing the deposit by scratching or wiping from a defined surface area with subsequent analysis in the laboratory.
ii) Exposure of test surfaces ("sentinel coupons") which are located at representative sites and removed after given periods of time or after treatments, again with subsequent analysis in the laboratory.
iii) Installation of sensors which possibly provide information on deposit formation and removal in situ, on-line, in real time and non-destructively, which was considered an optimal way of monitoring.
To overcome these problems, the use of so-called sentinel coupons is one possible option. Such a coupon monitoring device should fulfill the following requirements: small and easy to exchange surface pieces; easy to sample; possibility for multiple samplings; possibility for easy offline analyses in the laboratory; surface characteristics, geometry and hydrodynamics similar to and representative for real pipe surface.

10. Decision tool
The main goal of drinking water supply is to provide to customers an uninterrupted supply of water with a high aesthetic quality and with a minimal risk to human health. However, because of potential hazards in the raw water, treatment and distribution parts, water supply systems and consumers may get exposed to a wide variety of risks. In addition, crisis situations as e.g. a terrorist attack on the water supply system are potential risks with high impact. As a result, water companies and authorities have to continuously adapt to manage those risks and to be prepared in such crisis situations.

Potential Impact:

- 1. Security and safety synergy reinforcement for EU citizens benefit
SECUREAU project opens the way for a new drinking water network every-day management, as well as a new way for rapid detection and cleaning of contaminated networks. The alliance done between new information and communication technologies and drinking water expertise allowed to re-invent drinking water supply in several ways. SECUREAU project provides solutions designed for "security" applications (detect terrorism attack). Nevertheless, the Early Warning System (EWS) based on the multi-parameter sensor network developed has another purpose: this second use is dedicated for day to day operations. Allowing abnormal change detection of water quality, the EWS allows operators to have a better control of the water quality impact of standard waterworks: washing tanks, pipe renewal, valves closing... Moreover, this system is able to detect incidents such burst, under/over chlorination dosing, pressure control system failure... all the major events affecting the water quality and the water supply services. Finally, such tools could be used to improve asset management strategies. Usually the pipe renewal schemes are only driven by failure rate or burst model forecast; a water quality criteria could be now added to sort the list of pipes to be renewed in priority. As investments of municipalities are under constraint, optimizing the renewal effort without decreasing the safety of the water supply for the European citizen is a major output from SECUREAU.

- 2. Development of EU industry and leadership on environmental sensors networks
Another benefit of the project is the sensor approach. Sensors suitable to determine surface and deposit contamination have been evaluated. Removable surfaces ("sentinel coupons") have been exposed which had the same history as the water system they survey. These can be sampled at suitable points in time, e.g. to verify a contamination and to verify cleaning success. The technical development of such sensors is a typical strength of creative SMEs and will provide them with a selection advantage on the market. More sophisticated devices allow for on-line, real-time, in situ and remote sensing and to survey very valuable systems. Finally, mathematical modeling tools have been developed which allow for optimal positioning of sensors. Such models improve the application of the sensors dramatically as they can be positioned in the most representative locations and improve their early-warning capacity. SMEs or stakeholders which implement these models will be competitive for protection of water systems. As soon as the sensors can be produced more cheaply, they will be powerful instrument to detect biofouling in industrial systems (e.g. heat exchanger, membrane treatment) early, allowing for timely countermeasures and optimization of efficacy.
Additionally the high level of technology developed within SECUREAU project in addition of full demonstration reinforced the industrial tissue of EU.

- 3. Development of modern analytical tools for contamination surface analysis
SECUREAU had a conceptual impact on any water system decontamination approach by taking the role of surfaces and their deposits into account. They act both as sink and source of the waterborne contaminants. Due to sorption processes, contaminants can be sequestered from the water phase. Due to desorption and deposit decomposition processes, sorbed contaminants can be mobilized in the water phase. Therefore, a new paradigm has to be implemented into practice: do not neglect the surfaces.

- 4. Combined decontamination strategies
Decontamination strategies needed first to define precisely the extent of contaminated zones. Then the research studies for localization of contamination and spread issues in drinking water distribution systems brought new knowledge with significant technical, public health and economic impacts.

- 5. Dissemination activities and exploitation of results
The important SECUREAU outcomes have been disseminated through 12 main actions, reported in the Deliverable 8.3.2 (public deliverable, available on the SECUREAU web site) and summarised hereinafter:

- European and national level conferences
The partners involved in SECUREAU have demonstrated a strong participation in conferences, both nationally and internationally by more than 42 oral presentations.
A strong effort has been done by SECUREAU partners to communicate efficiently, outside their country, mainly in Europe (9 countries, 37 conferences) even if four conferences were done outside Europe, one in Russia and three in USA.

- Poster presentation
A total of 21 posters were presented, mainly during two conferences: the 4th Water Contamination Emergencies (WCEC4) held in 2010 and the 5th Water Contamination Emergencies (WCEC5) held in 2012, both in Mulheim an der Ruhr (Germany). Three posters have been recognised as the best scientific poster presentation and received the Poster Award 2010 (in the WCEC4 conference, first price for Sandra Wilks and Bill Keevil, SOTON, UK and second price for Martin Strathmann, IWW, Germany) and the Poster Award 2012 (in the WCEC5 conference, for Florence Gosselin UL, F – Luis Miguel Madeira, Uporto, P – Talis Juhna, RTU, LV).

- Peer review and technical papers
By the end of 2012, 20 peer review and technical papers were submitted (11) or published (9) in journals or part of books.

- Workshop, seminars, technical meetings
Six partners have participated to oral presentations during 15 technical manifestations, both linked to specific WP or more generally concerned the overall SECUREAU project. Participation to such events permits both to diffuse information to selected public, as explained below through few examples:
- Meetings occurred at Ministère de l'écologie and Ministère de la Santé et du Développement Durable in France (two meetings in 2011 and two in 2012) to transfer specific, classified results to national French security agencies. One meeting was also done with the "détachement central interministériel d'intervention technique DCI-IT" in 2012. The same strategy was applied in Finland, with a specific seminar organised by STUK (December 2012) and one organised by SECUREAU in Germany, during the WCEC5 meeting;
- Results and contacts were also taken during meetings held by ASTEE (Association Scientifique et Technique pour l'Eau et l'Environnement, France), mainly with industrial partners;
- Presentation devoted mainly to SME was done, for example under the Pôle de compétitivité Alsace/Lorraine sur la qualité et la gestion de l'eau Hydreos (conference held by October 2012 in Metz, France);
- Cluster meetings with other European projects / coordinators also occurred, for example during meeting held in Stockholm (SRC'09 conference, Sept 2009), in Ispra (ERNCIP conference, Dec 2012), by direct contacts;
- Presentation to Technical committees was done during WCEC conferences (Mulheim, Germany) and ERNCIP conference (Ispra, Italy)...

Such actions were important (i) to feed the network of contacts and competences, (ii) to diffuse selected information to specific target identified (national authorities; SME; stakeholders; (iii) and to fulfil SECUREAU obligation written in the DoW.

- Participation in forum and organisation of workshop
(a) World forum for security and fire prevention (Essen, Germany)
SECUREAU was present at the "World forum for security and fire prevention" (Essen, Germany, Sept 2012) on the European Commission stand. It was the opportunity to have contacts with the other European projects on the EC stand (8 projects), the EC colleagues and industrial visitors, regarding technical SECUREAU solutions exposed. Many tools were presented at this occasion : sensors developed by VERI (Endetec) and IWW, coupon sentinels, models developed by UPORTO (slides were presented continuously on a 42" flat screen), Propella reactors...

Around 10 to 20 contacts or visits per day came in the SECUREAU area, including visit of M. Ralf Jäger (Minister of Municipal Affairs of Nordrhein-Westfalen), Paul Weissenberg (Deputy Director-General of the Enterprise and Industry Directorate-General), Marco Malacarne (Head of Unit for "Security Research and Development" in Directorate-General for Enterprise and Industry of the European Commission)... A 15 minutes conference on SECUREAU was also presented during this forum.

(b) 5th Water Contamination Emergencies: managing the threats
SECUREAU has co-organised with the WCEC5 the "5th Water Contamination Emergencies: managing the threats". During this conference, SECUREAU group has organised a closed workshop (requiring security clearance to hear confidential information). It was a very successful meeting: 150 participants from 26 countries. Around 25 participants were affiliated to Universities, around 25 to national or international health centres and around 100 to the industry. More than 50 people attended the session 8 "Security SECUREAU session" devoted only to the presentation of classified results. See Deliverable 8.5 (on-line on the SECUREAU web site) for more information regarding this workshop.

- Masters – PhD thesis – Post doctoral
A total of 13 young researchers were supported throughout SECUREAU budget (Five master thesis, Seven PhD and one post-doctoral formation).

- Website
The website (see online) is a tool used for many purposes:
- Dissemination of non-confidential information (scientific objectives, expected impacts, structure of the consortium, public deliverables, main events with SECUREAU members participation),
- Document storage area devoted only for SECUREAU members (login + password needed) for scientific documents (e.g. deliverables not public, meeting reports, DoW, slides presented during the meetings...) as well as administrative documents (e.g. financial guide, audit and certification guidelines, project reporting guide, Grant agreement, consortium agreement...);
- Dissemination of scientific communications, conference and abstracts of publications. The scientific community is the main target of this action.
- Industrial and final products
Two main items has been developed or improved during SECUREAU: sensors and mathematical models and are reported above.

- Press releases, interviews
Few actions related to press release or interview occurred within SECUREAU project.
An article entitled " SECUREAU : faire face à une attaque terroriste sur les réseaux d'eau potable " was published (see online) on the French Lorraine web site. Two video interviews were realised, one for the European Commission (2009) followed by a DVD freely distributed (Video illustration on the Security Research Theme) and one done by the Latvian National TV Chanel, dedicated in general to RTU research and in particular to SECUREAU project.
Last, SECUREAU is also referenced (see online) on the Wicri/Lorraine web site, which refers all research activity in which one partner (industrial, research organisation, SME...) came from Lorraine (France).

- Methodological guide for end-users
The Methodological guide for end users (Deliverable 8.4 public document available on the SECUREAU web site) aims to highlight, along with Deliverable 6.8 (Decision tool Suitable for Assessment to Approval of Successful decontamination), some of the methods selected in SECUREAU for pre-crisis, crisis and post-crisis phases. Their main interests or novelty and their limits are also discussed in the document.

- Classified results
Special attention has been paid to the secured dissemination of classified results. First, preconisations given both by the European Commission and by the national security agencies were strictly respected: no classified information on the Web site, even on a restricted area; no diffusion of this information on the EC Participant Portal)... Specific meetings were organised to diffuse this classified information, already described above in the Chapter 4: Workshop, seminars, technical meetings and chapter 5: Forum participation and workshop organisation.

- Deliverables produced
The Deliverables produced are sorted by (i) 8 public deliverables (available freely on the SECUREAU web site), (ii) 35 restricted and confidential deliverables, i.e. not public documents, not available, and (iii) 4 classified deliverables. Systematically these documents were reviewed by the steering committee and by the dual use advisory board for classified deliverables. 70 % of these documents were also revised by external experts during two scientific reviews (one held for 2009-2010 period and the second one for the 2011 period).

List of Websites: