Patrick Smeets, 1 Oktober 2007 Partner for progress Microbiologische risico s van drinkwater TU Delft
Drinkwater is toch veilig? Kiwa Water Research 2006 2
Microbiologische risico s in ontwikkelde landen door Ongelukken 1993 Milwaukee, USA Change in treatment plant operation resulted in Cryptosporidium in water 2000 Walkerton, Canada 2001 Belfast, Northern Ireland E.coli 0157H7 and Campylobacter in water due to operational errors 9 potential sources of contamination comprised the supply; sewage backflow was considered the main cause. 2002 Northeast Italy Broken sewer pipe in football field irrigation system allowed water from Bacchiglione River to enter city aqueduct. 2002 Ski resort, Sweden 2003 Divonne-les- Bains, Ain district, France Unchlorinated groundwater (GW) source. Crack in sewage pipe located 10m from a well supplying system. Cross-connection between STP effluent network and distribution system; chronic contamination of water source and inadequate treatment 403,000 people ill 4,400 hospitalised 100 people died 2,300 people ill 65 hospitalised 7 people died 191 cases of cryptosporidiosis, 41 hospitalisations. 670 reported gastrointestinal illness (GI). 32 cases were referred to hospital, 6/32 were admitted. 9/12 stool samples positive for NLV. Estimated 500 cases in the population. 786 suffered GI. Multiple enteropathogens identified, 2 people were hospitalised, 2007 Galway, Ireland Contamination with cryptosporidium 60 people hospitalised 2007 Spencer, USA Human error put lye in water 100 people medical attention at least 1 hospitalised Kiwa Water Research 2006 3
Microbiologische risico s in ontwikkelde landen Terrorisme (nog niet microbiologisch) year location Incident public health effect Intentional threats and contamination incidents 2002 Rome, Italy Plot to break in through water no, offenders arrested pipes into American Embassy 2003 China Poisoning of water reservoir with 42 people hospitalised pesticides 2003 Prague, Threat of poisoning with mercury no, blackmailer arrested Czech Republic and cyanide 2004 Carpentersville, Vandals spray fire extinguishers in no, filtration unit was not in USA filtration unit use 2005 Lake Constance, 10 litres of atrazin dropped near no, because of low solubility Germany water intake 2006 Blackstone, USA Vandals break into a water storage tank and low toxicity of atrazin no, drinking water ban proclaimed Kiwa Water Research 2006 4
Cholera in ontwikkelingslanden neemt weer toe! Kiwa Water Research 2006 5
Wat zijn de risico s en waar komen ze vandaan Virussen Bacteriën Protozoa Overige microbiologische risico s: prionen, wormen, toxines Oorsprong: Uit fecaliën van mens en dier (E. coli O157, norovirus, Cryptosporidium.) Groeien in het milieu of in drinkwater (legionella) Kiwa Water Research 2006 6
Verontreinigingsbronnen oppervlaktewater Wastewater Recreation Wildlife Kiwa Water Research 2006 7
Wat zijn de gevolgen Darmklachten, diaree, overgeven, misselijk Ernstig bij gevoelige groepen (onderdrukt imuumsysteem) Soms ook ernstige effecten, bijvoorbeeld Guillain-Barré syndrome (verlamming door Campylobacter) Dood Kiwa Water Research 2006 8
Hoe veilig moet drinkwater zijn? Géén indicatorbacteriën (E. coli) in 100 ml monster Kans op infectie <1/10.000 per jaar Hoe toon je aan dat het water veilig is? Analyse onmogelijk Pathogens in source Removal by treatment Pathogens in drinking water Risk of infection Kiwa Water Research 2006 9
Internationaal project over risico-analyse MicroRisk WRc-NSF Ltd Swedish Institute for Infectious Disease Control Partner for progress 2002-2006 11 Partijen 7 landen Kiwa Water Research 2006 10
QMRA for drinking water Quantitative Microbiological Risk Assessment Source Sedimentation Filtration Monitor 1 log 2 log 100 /L 10 /L 0.1 /L 0 500 0 1 Disinfection Distribution 2 log No cont. 0.001 /L 0.001 /L 0 1 0 1 Consumption 1 L 0.001 /d 0 2 4 6 8 Dose-Response 10% *0.1 0 1 1000 Risk of infection 0.0001 inf/d -4.0-3.1-2.2-1.4-0.5 Kiwa Water Research 2006 11
Meten in de bron Véél verschillende pathogene micro-organismen Analysemethoden moeilijk en duur Onzekerheden: Variatie: hoe hoog zijn piek concentraties? Recovery: welk percentage toon je aan? Type: is deze variant gevaarlijk voor de mens? Kiwa Water Research 2006 12
Hoeveel micro-organismen zitten in het monster? Gemiddeld 3 organismen in een monstervolume Waarneming varieert echter van 0 tot 7 Kiwa Water Research 2006 13
Drinkwaterzuivering Rapid Sand Filtration OZONE Slow Sand Filtration Meervoudige barrières tegen micro-organismen Kiwa Water Research 2006 14
Verwijdering door de zuivering Hoe kwantificeer je dat? Literatuuronderzoek Processen experimenteren Processen modelleren Metingen in de praktijk Kiwa Water Research 2006 15
Experimenteren Inactivatie door ozonisatie Log inactivation 5 4.5 4 3.5 3 2.5 2 1.5 1 0.5 0 0 0.05 0.05 0.1 0.1 0.15 0.15 Ct (mg*min/l) Fresh cultured Stress. environm. water Stress. Fresh sterile cultured water Stress. 6 months environm. old water Stress. Stress. sterile sterile water 6 m. old + env. water water 6 Environmental months old Kiwa Water Research 2006 16
Modelleren Model wel verifiëren met de werkelijkheid! E. coli at 10 C DE (Log inactivation) 10 9 8 7 6 5 4 3 2 1 0 0 0.5 1 1.5 2 Observed Ct (mg*l/min) T10 PLUGFLOW CSTR MIXED T10 CSTR lab E.coli T10 CSTR nat. E. CSTR colil CSTR lab E.coli lab E.coli Sample CSTR positive nat. E. colil Sample negative Verified at full-scale Kiwa Water Research 2006 17
Meten in de praktijk Invloed van variaties en metingen Treatment processes can retain and release organisms over time! A sample only represents a very small portion of the water during a short moment in time! sand filter Kiwa Water Research 2006 18
Denken in logeenheden 10 filters Elk 2 log verwijdering 2 log 2 log 100 1 faalt 0 log concentratie 100 CFU/L 2 log=99%verwijdering 1CFU/L 2 log 2 log 2 log 2 log 2 log 2 log 2 log 1 1 1 1 1 1 1 1 =11 CFU/L Kiwa Water Research 2006 19 109 10 in filtraat: A 100 CFU/L B 10 CFU/L C 1 CFU/L D 10-16 CFU/L
Lekkage bij distributie E. coli in 0.1% van de monsters Kiwa Water Research 2006 20
Drinkwaterconsumptie Telefonisch interview ( recall ) Dagboek Kiwa Water Research 2006 21
Kans op infectie Dosis-respons Wat is de kans dat je een infectie ontwikkelt wanneer je één of meer micro-organismen binnen krijgt? Relatie vaststellen met: 10 0 Vrijwilligers (studenten) Uitbraken Maximum=100% Yearly risk of infection 10-1 10-2 10-3 Human feeding trial Outbreak Maximum risk curve 10-4 10-4 10-2 10 0 10 2 10 4 Dose org/d Kiwa Water Research 2006 22
30% 25% 20% 15% 10% 5% 0% Monte Carlo analyse RAW 0.1 0.2 0.4 0.8 1.6 3.2 6.4 13 26 51 102 205 410 819 1638 3277 Campylobacter MPN/l Data MLE PDF 0.003 0.004 0.006 0.009 0.014 0.020 0.030 0.046 0.068 0.103 0.154 0.231 0.346 0.519 0.778 1.000 Kiwa Water Research 2006 23 0.05 30.0% 25.0% 20.0% 15.0% 10.0% 5.0% 0.0% 30.0% 25.0% 20.0% 15.0% 10.0% CONCENTRATION AFTER RSF REMOVAL RSF 0.0% 5.0% 0.002 0.004 0.009 0.021 0.049 0.11 0.26 0.60 1.4 3.2 7.3 17 39 89 205 471 1085 Campylobacter concentration after RSF Data Fraction passed RSF Data MLE PDF
Resultaat: FN-curve Ruwwater Na filtratie Na ozonisatie Kiwa Water Research 2006 24
Wat doe je met het resultaat? Zuiveringsstappen toevoegen? Bedrijfsvoering veranderen Meer ozon of chloor doseren Sturing verbeteren Filtratieprocessen optimaliseren.. Meer meten? Huidige situatie handhaven en borgen Kiwa Water Research 2006 25
Proces en/of bedrijfsvoering verbeteren Voorbeeld chloordesinfectie Total reduction Current 2.1 Double dose 2.3 No events 2.5 Hydraulics 2.6 Hydr.+events 4.0 Log inactivation 5 4.5 4 3.5 3 2.5 2 1.5 1 0.5 0 27 Apr 07 May 17 May 27 May 06 Jun 16 Jun Goal: 2.5 log to meet health target Kiwa Water Research 2006 26
Procesbewaking is essentieel! 5 6.5 Hours response time Log inactivation 4.5 4 3.5 3 2.5 2 1.5 1 0.5 Monitored Monitored Running Monitored average Running Monitored average Correction Running average level Correction level Achievable average 0 27 Apr 07 May 17 May 27 May 06 Jun 16 Jun Goal: achieve 2.5 log average inactivation over period Kiwa Water Research 2006 27
Procesbewaking: hoe vaak moet je meten? Risk from event < nominal risk (95% certainty) nominal log reduction Interval / year 1 1 week 2 1 day 3 3 hours 4 15 min 5 2 min Kiwa Water Research 2006 28
Verschillende landen, verschillende prioriteiten Westers ontwikkeld: verwaarloosbaar risico, hoe verifieer je dat? Beperkt ontwikkeld: hoe zorg je voor constante goede bedrijfsvoering? Primitief: hoe maak je mensen bewust van risico s (latrines boven visvijver) Source: F. Wieneke Kiwa Water Research 2006 29