Skip to main content

Impact of climate change on the transport, fate and risk management of viral pathogens in water

Final Report Summary - VIROCLIME (Impact of climate change on the transport, fate and risk management of viral pathogens in water)

Executive summary:

VIROCLIME is funded under the European Union (EU)'s Seventh Framework Programme (FP7) for Research. It is designed to determine the risk to water users resulting from altered viral pathogen dynamics attributable to predicted changes in climate at sentinel sampling sites in EU surface waters and in the Amazon basin of Brazil. Built on the concept that climate change will cause modifications to the aquatic environment which will affect exposure to all pathogenic micro-organisms transmitted by the water route, VIROCLIME addresses key issues at the interfaces between human and animal microbiology, climate change and water regulation. There were 8 participating institutes. The coordinating institution was Aberystwyth University, United Kingdom (UK) with the contact Professor David Kay (dvk@aber.ac.uk).

The project aim was achieved in several stages:

(a) determination of the virological quality of water at five globally important sites (in Spain, Greece, Sweden, Hungary and Brazil) in order to obtain the first available baseline data of the incidence and types of target viruses at these sites;
(b) use of source tracking to detect human and animal viruses and to discriminate between human and animal sources of viral water pollution to permit attribution of pollution balance at each site;
(c) hydrological modelling of the of the catchments embracing the case study sites to determine streamflow patterns, and using the models in predictive mode to illustrate river flow changes driven by climate change;
(d) use of the modelling outcomes to predict the changes in virus levels at the different sites;
(e) to use such predictions in a quantitative microbiological risk assessment (QMRA) to reveal any increase in risk associated with water use (abstraction or recreational) at the case study sites.

In addition, an extensive survey of water-related extreme events was conducted and a vade-mecum handbook produced for teams involved in managing such events.

In the virological surveillance, over 7 600 analyses were completed for different viruses and over 2 000 for bacterial faecal indicators, in a range of matrices. In surface water samples, adenovirus was detected more than other viruses (highest titre 3 × 106 genome copies per litre in river water samples and 5 × 104 GC / L in marine samples), followed by JC polyomavirus in the rivers (up to 106 GC / L) and bovine polyomavirus (BPyV) in marine recreational water samples (2 × 103) GC/L. All three viruses were found at every location. The pathogenic norovirus was detected at all riverine sites with frequencies ranging from 6% of samples positive for either genogroup I or genogroup II at the Umeå (UMU, Sweden) sites (all locations together) to 71 % samples positive for one or either genogroup at the Llobregat (UB, Cataluña) sites. If both riverine and marine sites are considered, Umeå is still the lowest (7 %) and Tisza (NIEH, Hungary) showed the highest percentage of NoV-positive samples. Other viruses were also detected. Individual site correlations between viruses and bacterial indicators (E. coli and enterococci) were generally low, though correlations between all parameters were higher when data from all matrix types were used in the analysis.

Microbial source tracking (MST) revealed a range of different animal viruses at the sites, and the probes devised could be used to discriminate between human and animal pollution. Of particular interest was the potential of the human MST probe to predict the proportion of human faecal indicator contribution within a mixed (human, porcine, avian sources) aqueous matrix. Standard operating protocols (SOPs) for all the technical procedures are on the VIROCLIME website http://www.viroclime.org

Catchment models were constructed to allow prediction of the virus concentrations under different climate change scenarios, and the results from this was used in the QMRA. Under current conditions, it was estimated that, over all sites during the EU summer bathing season, there would be between 0 and 152 cases of norovirus illness per 1 000 users. Viral projection data suggested that there would be no cases of norovirus illness at the Spanish or Greek sites. Under two of the scenarios it was suggested that there would be a small number of cases at the Brazilian site during 2100, and for the Hungarian site norovirus illness was predicted for both 2035 and 2100 under the majority of scenarios, although in all cases rates of illness were predicted to be lower than those estimated for the current situation. Dissemination of the work was effected by several means including a final conference, the website http://www.viroclime.org electronic newsletters, brochures and targeted emails, and peer-reviewed publications in high-level scientific journals. This process will continue after the end of the project.

Project context and objectives:

Pollution of water by sewage and by run-off from farms produces a perceived public health problem in many countries. In developing countries, the problem is seen in recorded outbreaks of enteric disease attributed to viral pathogens of the gastrointestinal tract; in developed regions the problem of pollution may be less but the perception that sewage pollution causes disease outbreaks is often confirmed by recorded disease patterns attributed to small rural water supplies and camping sites.

Many countries depend on tourism for much of their national income and a large part of the gross domestic product (GDP) of EU Mediterranean states, in particular, is derived from seasonal tourism around their coastlines. Here, resorts developed since the 1960 expansion in human mobility have largely been around attractive maritime areas, and large economic centres now exist which are highly dependent on the tourist trade which requires the perception of pristine coastal water for recreation. A parallel requirement is evident in the European shellfish harvesting industry, and both activities are regulated by the EU directive's compliance with which is achieved through mechanisms outlined in Article 11 of the EU Water Framework Directive (WFD) (2000). Thus, pollution directly affects the economies of these states.

Climate change will cause modifications to the aquatic environment which will affect exposure to all pathogenic micro-organisms transmitted by the water route. VIROCLIME addressed key issues at the interfaces between human and animal microbiology, climate change and water regulation, particularly:

- Viruses are pre-eminent role in water-related disease because of the almost ubiquitous presence of human-derived pathogenic viruses in sewage and waste water and because of their environmental robustness.
- There are currently few tools for routine quantitative virological monitoring of the environment to provide credible health risk quantification.
- Providing modellers with baseline virological data from different environments will underpin prediction of how virus levels may alter as a result of climate change and / or water management practices and their potential population impact on disease burden and potential health gain.

VIROCLIME was conceived as a deliberate integration of virology with climate change science and was constructed recognising that to achieve its objectives both disciplines must interact throughout the project. It was put together by experts in both fields so that the correct balance of the investigation could be maintained to ensure that the data accrued reflected the intended aims. The project focused on the interrelationships between (a) waterborne viruses of importance in human disease, particularly those which are considered to be emerging pathogens or to be of use as viral indicators of sewage pollution; and (b) the predicted effects of changes in climate at five sentinel sites in the EU and Brazil. It then takes this integrated information forward to consider the possible effects on the health of individuals in these areas who may use the water, either for recreation or as drinking waters.

This concept was encapsulated in the following principal project objectives:

(a) to report on the performance characterisation of methods developed in EU, international cooperation partner countries (ICPC) and United State (US) laboratories for the concentration and detection of waterborne human pathogenic viruses in environmental 'hot spots';
(b) to report on the use of improved virological tools for MST;
(c) to produce an operational model forced by environmental and water management changes at the target sites which may be calibrated to show changes in virus levels and to facilitate changes in water management strategies;
(d) to provide a report on 18-month surveillance case studies of emergent potentially pathogenic viruses at five environmentally sensitive sites in Spain, Hungary, Sweden, Greece and Brazil;
(e) to report on any relationships linking target virus incidence with that of the current faecal indicators Escherichia coli and intestinal enterococci and to assess the suitability of current faecal indicators in the face of changing climate scenarios.

Other objectives are shown associated with individual work package (WP) reports. The focus on viruses was initially driven by the observation that the revised Bathing Water Directive (2006) removed the previous enterovirus parameter which was noted in Article 14 which also recommended studies to suggest alternative replacement virological parameters better to index the risks of bathing-related illness which many epidemiological studies have attributed to viral pathogens. Part of the reason why viruses remain a potential problem is that they are resistant to environmental degradation and are more robust than many bacteria. This has been recognised for over 30 years. For example, adenoviruses have been shown to be highly stable in the environment, and to be highly resistant to diverse disinfection treatments including ultraviolet radiation, especially the adenoviruses 40 and 41 which are associated with paediatric gastroenteritis. There has, thus, been a case for considering viruses as an indicator of water quality in place of bacterial indicators. It is also the case that waterborne bacterial diseases have been well controlled partly due to the use of effective bacterial indicators of water quality; this has so far not been the case for waterborne viral diseases and they remain a public health problem.

In addition to adenoviruses and noroviruses, VIROCLIME recognised that there are several potentially pathogenic enteric viruses (e.g. rotavirus, sapovirus and hepatitis viruses) often detected in water and sewage, and the project included consideration of these in the analyses of samples taken during surveillance programmes.

If risk attributable to human waterborne viruses was to be determined it would also be necessary distinguish between human and animal faecal pollution of waters used for recreational and other purposes. VIROCLIME therefore included an MST component which developed new probes based on bovine, porcine and other viruses, as well as human types, which could be used to effect this discrimination.

Regarding climate change, VIROCLIME focused on producing basin-specific projections of the impacts of general circulation models (GCMs) 'probable scenarios' on the regime of specific rivers. Most hydrological models could already assimilate both climate observations and climate forecasts to derive flow regimes. They also integrate water management into the modelling, but problems arise when such models, which are tuned for present day water management conditions, are used in a climate change context. The two major limitations are (a) the coarseness of spatial resolution, which is generally orders of magnitude coarser than required for most river basin studies; and (b) the modelling of the hydrological components, precipitation, evapotranspiration, soil moisture and runoff is considerably less reliable than temperature and pressure.

It was likely that the key process driver of high flow sequences would dominate flux generation of virus through:

(a) spills from the sewerage system;
(b) reduced retention, and treatment, time in secondary (biological) sewage treatment works;
(c) increased flushing of urban (and to a lesser extent) diffuse pollution loadings including small domestic disposal systems;
(d) enhanced drainage from urban and industrial disposal sites; and finally
(e) in channel re-entrainment of deposited and mainly particulate-associated microbial loadings during high flow conditions.

Whilst such 'high-flow' events may represent the principal flux periods impacting on downstream receiving waters, used for bathing and shellfish harvesting, 'low-flow' events are also of major significance because these conditions represent periods of 'water stress' when treated effluents receive minimum dilution. As such, virus concentrations might be expected to peak at water supply abstraction point intakes. Thus, the key drivers of interest will be the changed periodicity of flow event sequences, not necessarily extreme events but rather the likely change in 'normally-experienced' periodicity in river flow events will be the core data requirement of this element of the project which will be delivered by the IC3 partner.

VIROCLIME built on the success of VIROBATHE, which developed a harmonised methodology for concentration and qualitative detection by adenoviruses and noroviruses, considered as putative viral indicators. Detection methods were advanced during VIROBATHE to permit the 'presence or absence' of adenoviruses to be considered as viral indicators in environmental receiving waters. The principal technical achievements included harmonised concentration methods for enteric viruses in fresh, transitional or marine water samples, a rapid molecular detection method for adenoviruses and noroviruses, and 'rolling out' the combined concentration / detection method to routine environmental monitoring laboratories. Health-related outcomes of VIROBATHE included a positive correlation between the concentration of adenoviruses and other viral (coliphages) and bacterial indicators, and demonstration that a substantial number of samples complying with the Bathing Water Directive by standard indicator values were positive for adenoviruses. VIROCLIME aimed to take this qualitative approach forward to a quantitative reproducible method which was allied to modelling and risk assessment components of the project.

Project results:

Summary of main scientific and technological results

Virology and MST

General comments

A new method (direct flocculation) for concentration of waterborne viruses was trialled (WP2) among all laboratories and found to be satisfactory for determining virus load in different environmental matrices. Although issues remain regarding its deployment as a quantitative technique for regulatory purposes, the commonly accepted repeatability criterion of within 1.0 log10 was achieved. Recovery of viruses from laboratory-seeded replicate samples of river water, sea water or mineral water resulted in recoveries within 0.9log10 genome copies (GC, adenoviruses), and within 1.3log10 GC (noroviruses). Further work, at a fundamental level concerning the relationships between virus particles and concentration matrices is required before this can be taken into the regulatory sphere. The method was used in WP3, source tracking and in the surveillance phase, WP4. In the latter, each laboratory undertook an extensive programme of water quality surveillance at their chosen Case Study site. All had more than one sampling point at each site, and sewage and sewage effluent were also sampled, though the frequency of this varied between laboratories. The four European laboratories took samples every two weeks. The Brazilian partner did fewer sampling runs but did more intensive sampling on each run. This was because of logistical and travel requirements as the laboratories in FIOCRUZ was a four-hour flight from the case study site at Manaus, Amazonia. In all, 1 472 samples were analysed and these generated over 7600 page data points. The full data are provided as D4.1 and D4.3. All laboratories analysed their samples for adenoviruses and noroviruses, and each laboratory also analysed samples of local interest. For example, UMU analysed samples for sapoviruses. The data were sent to the WP leader laboratory (NIEH) where they were collated and continuously updated spreadsheets were prepared. Throughout the programme support was provided by UMU in the provision of positive control preparations of AdV35 and norovirus. The results are summarised in the following sections and full details may be found in the VIROCLIME periodic reports and deliverables.

Overall, in surface water samples adenovirus was detected more than other viruses (highest titre 3x106 GC / L in river water samples and 5x104 GC / L in marine samples), followed by JC polyomavirus in the rivers (up to 106 GC / L) and BPyV in marine samples (2x103) GC / L. All these viruses were targeted at every location. In raw sewage the frequency of positive samples was over 70 % for all viruses except Hepatitis E. Adenovirus, Norovirus GII, and JC polyomavirus were the highest both in prevalence and concentration (up to 108 GC / L).

Case study sites

Cataluña, Spain (Partner UB)

In Cataluña, the Llobregat river rises in the Pyrenees and flows through agricultural regions and industrial, urban areas over a distance of 170 km. The river basin contains more than a half of the Catalan population of 5-million inhabitants and treated urban sewage can represent an important percentage of the river flow. It discharges in the Mediterranean Sea, the beaches of which are intensely used for recreational purposes. Three sampling sites were used for the surveillance study and a further four were used in MST. The main source of water pollution in the Llobregat basin is from human origin. Virus frequencies varied between seasons. In winter and autumn around 70 % samples were positive for HAdV, in summer this rose to 100 %. This suggests that, when river water levels are low, sewage treatment plant effluents are the biggest contribution to river water volume. The sea water sampling site also showed 100 % samples positive for human adenovirus (HAdV) during summer (when there is the main risk of exposure). The data showed low animal viral prevalence during most of the year. Porcine adenovirus (PAdV) and BPyV were detected once in winter after an extreme rain event and in all the sampling points during summer. HAdV concentrations were around 103 GC/L in all the river water points. JC polyomavirus (JCPyV), a human virus shed by most individuals, was prevalent throughout the year, but in some sampling sites during winter concentrations were lower, giving no detection values in two of the sampling sites where there was an important dilution factor. Porcine and bovine viral presence, in the samples tested over the winter reflected catchment land use above at the sampling sites where livestock and agricultural activities may have had the biggest influence giving rise to diffuse pollution events.

Patras, Greece (UPA)

The flow regime of the Glafkos River is very variable and characterised by torrential rains and floods in winter and drought during the summer. It flows into the Gulf of Patras (Ionian Sea) at Patras and, though it does not dry completely during summer, stream flow decreases dramatically.

Four sampling points were used for the surveillance programme (WP4) and two extra sampling sites for MST. Four of the sites were riverine and two were seawater sites. During the surveillance programme, HAdV and NoVs were targeted and, in the MST work, sites were monitored for the four viral markers HAdV, JCPyV, PAdV and BPyV. Temporary traveller settlements on the beach generated some faecal pollution during the summer. Seawater sampling sites showed mostly human contamination with HAdV at levels 1-1.5log10 higher than river water. The main pollution source would have been the urban area of Patras.

Data from the two main riverine sites and from both marine sites were analysed by season, namely: winter, spring, summer and autumn. Of the four UPA target viruses, HAdV was present at varying levels at all four main sites during winter and spring months at one riverine site and one marine site during summer and, curiously, at the other riverine and the other marine site during the autumn. The highest levels of HAdV were found at one of the riverine sites in summer.

For the MST studies, data obtained from the Glafkos River basin showed high variability. River sampling points showed mainly animal pollution, while analysis of samples from the marine sampling points suggested that pollution was of mainly human origin. The River usually provided little dilution of pollution inputs, so any point discharge or any rain event would quickly affect the water quality. For example, during the spring and the summer season domestic animals drink from the river, so during that period the animal markers would be expected to have had higher prevalence as well as higher concentrations. During summer, the stream flow decreases and most of the riverbed receives water from the local waste water treatment plant (WWTP). Analysis of samples taken during the winter could be interpreted differently. A small tributary of the Glafkos R, with a MST sampling point 100m upstream from the main riverine sampling point, presented evidence of animal pollution when samples were analysed using the BPyV (bovine) probe. Animal viruses were also detected in the other three Glafkos river water points, while the human markers were detected at high frequencies at the two principal ones. No HAdV were found in the extra MST sampling sites.

Umeälven, Sweden (UMU)

Umeälven River in northern Sweden is 460-km long and flows in a south-eastern direction from its origin, Lake Överuman. It is used extensively for hydroelectric power generation. Umeå was the coldest location with winter temperatures declining to -20 degrees of Celsius. Although it is covered by ice from January to the middle of April, it is prone to spring and summer floods and receives run-off water from the thawing snow. There is a WWTP in Umeå treating sewage from 100 000 inhabitants, and the area below the city is a bird reserve. There are bathing sites on the river and on the coast near its delta. Two of the sampling points were on the river, upstream and a few kilometres downstream of the WWTP, and one was sited on the sea coast at Ljumviken, a recognised bathing area. UMU analysed samples for enterovirus, sapovirus in addition to HAdV, JCPyV, PAdV and BPyV.

Human viral loads, in particular HAdV, were less prevalent all over the year and in concentrations 1.0 log10 lower than BPyV. Sapovirus and enterovirus were not found in any of the surface water samples. Animal pollution was more consistent during the year than pollution of human origin. Data showed a high presence of animal-related contamination, especially from cattle. Samples from two sampling points showed higher frequencies of animal-related viruses throughout the year than at the other sampling points, probably due to their proximity to several dairy and pig farms. During the summer-time all the cows remain outdoors which suggests that run-off may result in diffuse pollution into the riverbed or from animal waste spreading to land as a fertiliser. During springtime JCPyV (human) and PAdV (porcine) were not detected, probably due to the melting snow diluting viruses into the river water and apparently depressing viral concentrations.

Rio Negro, Manaus, Brazil (FIOCRUZ)

The city of Manaus has 1.8 million inhabitants within an area of 11.4 km2 and it is located 1 450 km inland from the Atlantic coast, in the heart of the Amazon rain forest. The river system around Manaus is affected by heavy rainfalls. The Rio Negro flows into the Rio Solimões to form the Amazon River South of Manaus. Garbage and sewage are discharged to these waters causing environmental impacts which have been verified through physical and chemical observation for many years. For both the surveillance programme and the MST work samples were collected from five points. The first point was upstream of the city and is within a recognised bathing area; two points further downstream and closer to the city were on small rivers which receive untreated sewage and then discharge into the main river, and two points were downstream of the city. For the surveillance exercise 56 samples were taken at each of the first four points, and 48 samples were taken at the fifth point. Additional samples were taken for the MST analyses.

Data were clustered in two seasons, a rain period (December - May) and a dry period (June-November). In both seasons, all sites were positive for the human viruses. For HAdV, the minimum frequency of positive samples was 68 % during the rainy season and 60% during the dry season. JCPyV was also found at high frequencies, though more variably than HAdV. Data showed a high presence of HAdV, especially downstream of the bathing area in the region where tributaries containing sewage emptied into the river with concentrations around 1.0 and 1.5 log10 higher than upstream river water samples, suggesting that the main source of contamination to be human sewage. The first sampling point, located upstream Rio Negro and upstream of the city centre presented lower viral loads.

Tisza River, Solznok, Hungary (NIEH)

The Tisza River, one of the main rivers of Central Europe, rises in the Ukraine and flows roughly along the Romanian border with Hungary until it reaches the Danube in northern Serbia. The Tisza drains an area of about 156 087 km2 and has a length of 965 km. Floods in early spring and early summer are very common. Six regular WP4 sampling sites (four river water points and two sewage treatment works) were studied for four viral markers. The level of faecal contamination was very high with high human input throughout the year. Animal contamination identified as of porcine and bovine origin was also frequently detected with higher prevalence during summer and spring.

Further details, including levels of all viruses found at all sampling points, may be found in the VIROCLIME Period 2 Report and the relevant deliverables.

Relationships to regulatory parameters

Correlation analyses did not suggest the potential to predict the concentrations of key viral pathogens in environmental matrices from either: (i) faecal indicator bacteria (the current regulatory parameters); and / or (ii) other pathogenic viruses. For this to be the case, high R2 (or explained variance) levels: i.e. which are acceptable to the regulator (here we have assumed at least exceeding 50 %); would be required. The correlation coefficients between each microbial parameter rarely achieved these levels within the same environmental matrix or indeed sampling site (i.e. river water, sea water, raw sewage and secondary treated effluents).

There are, however, some statistically significant correlations for example between the FIOs and some virological parameters: e.g. adenovirus, norovirus, JC-polyomavirus, PAdV and BPyV all correlated with E. coli and enterococci. However the highest correlation coefficient R is 0.551 translating to an explained variance (adjusted R2) of 0.228 indicating that knowledge of log10 E. coli explains 22.8 % of the variance in adenovirus in the waters tested. All other correlations and associated R2 values in river, sea water and raw sewage were lower. However, in the case of enterovirus in secondary treated effluents the correlation coefficients were higher (i.e. E. coli and enterovirus 0.576 and enterococci and enterovirus 0.602) but, again adjusted R2 values would not exceed the 50 % threshold considered necessary for serious regulatory consideration.

Correlation with environmental parameters

In addition to the search for correlation between microbial water quality parameters, the surveillance data were screened for potential associations with antecedent river flow, rainfall and temperature patterns. The initial screening for relationships in this area simply split the microbial data set into two based on whether the antecedent hydro-meteorological data for the site were above or below the mean rainfall and or river flow for the defined antecedent period. The mean log10 microbial data characteristic of each flow characteristic could then be plotted in the search for generic patterns.

The consensus view of literature sources is that FIO concentrations generally increase during higher flows. By inference, elevations in faecally-derived virus concentrations might also be expected during rainfall induced elevations inflow. However, these observations are biased to smaller catchment streams rather than large scale continental scale rivers which have been monitored in this investigation. Thus the observations on stream viral dynamics reported here are novel and of significance to the regulators and modelling communities. The effect of antecedent discharge 24, 48 and 72 hours prior to the sample being taken was also examined. It was striking that the larger river sites exhibited no pattern in the discharge split data with two Manaus (FIOCRUZ) sites showing E. coli elevation at higher flow in the 24-hour antecedent period and two showing decreases, this follows through to enterococci with similar divergent patterns for the viral parameters. The same is true of the 48-hour and the 72-hour plots.

The Glafkos river sites both showed an apparent decrease in FIO concentrations following rainfall, a surprising pattern which was also reflected in the Glafkos sea water samples. There is no clear antecedent discharge pattern driving a consistent elevation and / or decrease in any viral concentration.

The sites on the Hungarian Tisza also produced counter-intuitive data more indicative of a pollution dilution process at high flow with lower microbial concentrations generally evident for the FIOs and virological parameters.

The Swedish data from the Tvaran and Umeå rivers were no less surprising and enigmatic. Here, the FIO plots do suggest concentrations increase in response to river discharge. However, the generally low viral concentrations observed at this site makes further interpretation difficult at this time.

These observations do not suggest a generic relationship between FIO and virus concentrations and river flows, the pattern observed is heterogeneous and site specific with both increase and decrease in microbial concentrations occurring with elevated river flow. These diverse patterns persist when antecedent precipitation, instantaneous flow and temperature are compared with the microbial concentrations.

Summary of surveillance and MST programmes

Counter-intuitive data such as obtained in some of the investigations above are both interesting and valuable. They imply the existence, and simultaneous operation, of flow-driven 'concentration' and 'dilution' processes. The former 'concentration' effect is characteristic of 'transport-limited' catchment processes where a pollutant builds up awaiting an episodic energy input to transport the pollutant to a river channel. This energy input is generally provided by flowing water. The latter 'dilution' effect is characteristic of catchment-scale, 'supply-limited' process where there is a finite supply of the pollutant which is quickly exhausted as the transport process (e.g. rainfall and overland flow) becomes operative.

To the best of our knowledge, the VIROCLIME data is the most extensive surveillance data describing a range of viral parameters in a wide range of surface water matrices. The patterns obtained are therefore worthy of further study and will be published in the peer-reviewed literature. The analysis of the VIROCLIME data represents a surprising and interesting outcome for the VIROCLIME project.

Modelling

The assessment of climate change impacts on waterborne virus levels required the development of models capable of describing hydrological processes consistent with meteorological and climatic signals. This was entirely new work, demanding innovative approaches to gathering data from diverse sources in the participating countries. The need for virologists and climate change scientists to work closely together to extract from different national archives data which are relevant and of sufficient quality for high resolution modelling was a challenging task for all concerned.

The development of streamflow models for the case study catchments was designed to underpin the virus flux assessment and modelling.

Hydrological modelling

Streamflow modelling was completed for all case study catchments. The ORCHIDEE land surface model was applied to the Rio Negro, Tisza and Umeälven catchments at the standard 50-km horizontal resolution and at 20-km resolution for the Llobregat catchment. For the Glafkos catchment, because of its small area (approximately 100 km²), a dedicated lumped rainfall-streamflow model accounting for evapo-transpiration was developed. To check the ability of the hydrologic models to reproduce observed streamflow, simulations were run for the 20th century using atmospheric forcing fields derived from the ERA-40 reanalysis and station data. The inter-annual streamflow variability was predicted satisfactorily for all catchments, despite some biases. However, seasonal cycles were reproduced less accurately.

Climate change streamflow projections were produced by forcing the hydrologic models with the outputs of three GCMs run under the A2 emissions scenario of the IPCC. Streamflow inter-annual trends for the 21st century from the climate change simulations were computed. Nearly all predicted changes with respect to the 20th century simulations were affected by significant uncertainties. Of particular relevance were the predicted changes in extreme streamflow values, which were likely to have a significant impact on water-borne viral concentration predictions. Frequencies of extreme flows are expected to increase for all catchments. However, significant uncertainties are associated with these predictions.

Future water management scenarios for the 21st century were simulated by defining a water reservoir regulating streamflow for each catchment. These scenarios were conceived to simulate possible policies that could alter the high and low flow regimes, without affecting the overall water balance.

Virus concentration modelling

Statistical analysis and modelling was used to deliver this element. The models generated could be used as ‘site-specific operational models’ of viral concentration, with all the caveats that fitting of models of this nature imply. The data essentially fell in the category of 'non-linear ecological short time series' which are very noisy (or chaotic) and sometimes not regularly sampled over time, which makes them particularly challenging to analyse and model.

The exploratory data analysis used the selection process of both response viral data and environmental explanatory data for modelling and the model selection methodology using the Akaike Information Criterion (AIC) was brought together as a set of site-specific models for viral/FIOs concentration in a collection of tables where the models coefficient could be expressed as well as their associated p values. Multiple regressions to fit the chosen explanatory variables to the response variables, where some of them are log-transformed were performed.

It was then possible to use the best fitted models to predict viral concentrations under different water management and climate change scenarios. The modelling work was limited to virus variables available at each location site and only Enterococcus was modelled as the faecal indicator organism relevant for water quality evaluation (i.e. given its importance as a recreational water health risk indicator as outlined in the World Health Organisation (WHO) Guidelines published in 2003). For viral load projections, locations were selected that were either sea or river. Raw sewage, sewage effluents, and treatment plant locations were not taken into account. To produce site-specific projections of viral load for the locations of interest under both climate change scenario and water management scenarios the fitted models were fed with environmental explanatory variables from simulations of three different GCMs set to simulate conditions of the IPCC climatic scenario A2 and under four different site-specific water management scenarios. These projections were delivered as text files for their use in the risk assessment work.

Extreme events and risk assessment

Public Health Wales completed a systematic review of the literature on waterborne disease outbreaks following extreme water-related climatic events in period 1. Heavy rainfall and flooding were found to be by far the most frequently implicated events in waterborne disease outbreaks. However, outbreaks following tidal surges and seawater inundation affected the greatest number of people and caused the most deaths. Many of the outbreaks involved more than one waterborne pathogen, and most were thought to be due to a lack of clean drinking water, usually as a result of sewerage systems becoming inundated or unable to handle the increased input.

A survey of European public health agencies was undertaken in period 1 to gather any information on the health effects of extreme climatic events seen in Europe. Public Health Wales held a workshop to discuss what should be included in a protocol to offer guidance to public health teams on the investigation of extreme climatic water events in real time, to help measure the health effects of these events. It was recognised that the value of the workshop and its associated protocol would be further enhanced by also providing the protocol in a handbook form rather than only as a Workshop report. Comment was sought from VIROCLIME partners and participants at a protocol development workshop held in Cardiff UK on May 2010. The production of the full operating handbook was done on December 2012 (month 24) creating an additional deliverable (D7.6). Ratification by STMB was obtained by email.

A report describing the process of QMRA was completed by Aberystwyth University. This summarised information about the recreational case study sites studied in WP4 and outlined the literature-derived data that would be used for the overall assessment. QMRA consists of four steps namely, hazard assessment, exposure assessment, dose-response assessment and risk characterisation. The principal hazard to be assessed was that due to noroviruses as they are the main cause of acute non-bacterial gastroenteritis, they are globally important, they have been recognised as causing a number of recreational waterborne outbreaks of illness and they were identified as the most common cause of waterborne outbreaks related to extreme water events (D7.1). For these reasons, they were monitored at all the case study sites. In order to standardise across the case study sites a single hypothetical population of bathers was assumed to be exposed over the bathing season, with virus pollution based on monitoring results and ingestion of water by bathers based on data from the literature.

For the QMRA itself, the current situation was estimated using a standard hypothetical population for each of the case study areas and the monitoring data acquired during the project (WP 4). The likely change in risk as a result of a number of climate change and water management scenarios was modelled for 2035 and 2100 based on data from WP5. Norovirus and adenovirus were isolated from each of the case study recreational water sites, although the frequency with which they were isolated during the bathing season (when people might be exposed to the contamination) varied between the sites from 0 to 38 % for norovirus and from 10 to 100 % for adenovirus. From the frequency of contamination and the geometric mean concentration of virus, it was estimated that during the bathing season there would be between 0 and 152 cases of norovirus illness per 1 000 users. Viral projection data, based on four different river flow / water management scenarios (a severe increase in stream flow, a moderate increase in stream flow, a moderate decrease in stream flow and a severe decrease in stream flow, relative to the current extreme high flow) and three different climate change models, were used to estimate levels of noroviral infection at four of the recreational sites in the future. Predictions suggested that there would be no cases of norovirus illness at the Spanish or Greek case study sites. Under two of the scenarios it was suggested that there would be a small number of cases at the Brazilian site during 2100 (less than 3 cases / 1 000 users). For the Hungarian case study site, norovirus illness was predicted for both 2035 and 2100 under the majority of scenarios, although in all cases rates of illness were predicted to be lower than those estimated for the current situation. It is speculated that the decrease in illness rates seen in 2035 and 2100 may be due to a reduction in the frequency of summer rainstorms.

Potential impact:

Dissemination activities

Potential impact of VIROCLIME

VIROCLIME should have a major impact as we believe it to be unique amongst projects funded by the European Commission (EC) conducting microbiological surveillance for18 months in four European countries (North: Sweden, West: Spain, Central: Hungary, East: Greece) and from one of the most globally important aquatic systems, the Amazon. It is the first study of such geographical diversity and integration to target several different waterborne viruses and comparing levels with those of faecal indicator bacteria. Several different viruses were studied. The project collected data from all parts of the sewage entrainment process. It is the first time that such a complete dataset has been generated and it is one of the first projects that an interdisciplinary integrated approach was been used where virologists, epidemiologists, public health specialists, statisticians, and climatic modellers have participated.

There are few reports in the literature on the potential impacts of climate change on human health, which add to the fact that there is every reason to believe that if climate change were to occur, the consequences could be significant for many countries, especially for developing countries. Some potential consequences could include the following (references to these points are provided in the period 2 and final reports):

- vulnerability of human settlement to rapid climate change,
- vulnerability of human settlement to sea-level rise,
- vulnerability of human settlement to tropical cyclones,
- vulnerability of human settlement to flood,
- vulnerability of human settlement to drought or water shortages,
- vulnerability of human settlement in some countries to loss of biomass,
- vulnerability of human settlement to rapid thawing of the permafrost,
- vulnerability of human settlement to health problems associated with climate change.

Owing to the complexity of developed countries and the fact that many factors largely independent of climate change employment changes, technological innovation, changes in terms of trade and currency values, and land use policies - will affect human habitat, it appears quite difficult to isolate changes which are due to global warming from other changes that might occur.

VIROCLIME can generate important information on the impact of climatic effects on water and waterborne viruses. The most difficult task is to correlate analogous studies (assessment of effects from historical and geographical analogies) with the future climate change projections. The project has performed these studies and the outcomes can be used to inform policy debates in each country. This is one of the most significant Project outcomes. Through this report and subsequent publications in the scientific literature, policy makers of each country will be informed for the fate of waterborne viruses, their relationship with bacteria and the possible consequence of climate change on their presence in the water following possible future scenarios.

For example, heavy rain may decrease water quality. Sometimes, frequent flooding has threatened the health of people in developing countries, directly or indirectly. Permafrost degradation may cause leaching from disposed wastes, resulting in contamination of the groundwater. If global warming worsens the water quality or increases inundation, diarrhoea, cholera and dysentery epidemics could spread in developing countries and in the subarctic area. The results of such changes have been imprinted by the public health specialists of the Project and a protocol to deal with such situations has been produced. This protocol will be available to incident control teams and others directly concerned with management of water-related extreme events, as well as policy makers in ministries of environment.

Concerning the epidemiology of viral diseases, VIROCLIME provides important information and new knowledge to the policy makers as well as to the research community from different European countries on the presence of different target viruses and their possible effect in the future due to climate change. Different scenarios have been developed and these scenarios will increase the existing scare knowledge on this matter for policy makers. Also the first publication of the extreme events and waterborne viruses is the first one published so far on this topic.

VIROCLIME is the only a project aiming at the determination of the climate change on waterborne viruses. The possible recipients of the project outcomes could be scientists working in the specific area, researchers in environmental non-governmental organisations (NGOs), scientists working in policy-maker organisations such as ministries of environment in each country, municipalities, prefectures, bodies such as European Centre for Disease Prevention and Control (ECDC) or European Environment Agency (EEA), Environment Protection Agency (EPA) and of course EC. This Project reflects the strong economic and scientific base that can provide Europe with a range of opportunities to address the potential health impacts of climate change. Equally it sounds a warning that institutional, political and financial constraints may hinder such opportunities. It is clear that we need to overcome these constraints and avoid a future of lost opportunities which may have significant consequences.

The project logo, website will remain active for at least two years for the policy makers to be informed. Several photographs have been taken during the sampling period in all different countries and all of them are available from the website as well as the Flick website. These photographs will give the scientific community a more vivid way of seeing the things during the sampling period.

Plan for exploitation of foreground

Dissemination measures (Sect. A)

The dissemination measures that will be followed are divided into the following subsections.

Publications

10 to 12 peer-reviewed articles will be published. The published articles will be produced until October 2013. So far, one publication has been performed. Some potential journals would be: Applied and Environmental Microbiology, Journal of Applied Microbiology, Water Research, Journal of Public Health, Epidemiology and Infection, Journal of Virology, Food and Environmental Virology.

The publication strategy/plan is following the corresponding scientific parts of the project. So there will be publications in each of the following areas. Some details in each proposed publication are given below:

Methods (including MST markers)
This will concern with the Standardisation of methods for analysing viruses and MST for multi-laboratory studies. This will include experiments with the swimming pool, preparation of SOPs, intra-laboratory analysis, inter-laboratory analysis with mineral water and sewage, and with bovine and porcine markers, and repeatability.

Surveillance
This publication will include the summary of surveillance data for all viruses the viral indicator / pathogens HAdV, and also the pathogens analysed. All viruses will be included in publication except MST viruses that are in a separate paper. Also FIOs and seasonability will be included. The data will presented in the context of possible future changes in climate and policy changes.

Surveillance statistics
This publication will include extended statistical analysis considering bacterial indicators and new viral indicators, environmental parameters etc.

MST
This publication will include the identification of the origin of faecal contamination by using human and animal markers in diverse geographical areas. The summary of the MST data collected in all labs and correlation of viral MST markers and bacterial standards or climatologic conditions will be included. Also the analysis of the proportionality of sources of faecal contamination using human and animal viruses as MST tools will be considered in the article.

Case-study- site specific
A publication from each site by site description of the study will be performed. This means at least five publications. This publication will include any specific pathogens analysed in each location and if possible hydrological models. The cooperation between laboratories for comparative discussion will be performed.

QMRA
A publication on quantitative microbial risk assessment data proposed by the project will be produced.

Modelling
The final publications will include the predictions/projections of virus dissemination in water taking into account the output from the climate change models.

Report for policy makers
Reports specific for policy makers will be produced in several languages (English, Spanish, Swedish, Hungarian, Greek, and Portuguese). The report (D10.4) has been prepared based on the Periodic Reports of the Project and will be disseminated by the Pertner responsible for the Dissemination WP (UPA - see website for contact details) by the end of 2013 to the corresponding bodies such as local water companies, ministries of environment etc. The report will be open for the public as well so it will be uploaded in the website. The report will be prepared in paper copy as well as in digital copy (DVD) including all the photos presented in the final conference relating with sampling in each country. Also the report will be uploaded in the website for other policy makers that they have not participated in the conference. The report will include the total project results including statistical analysis and tables for each participating country as well as recommendations and conclusions. This report will be a summary of an extended report with all the data that will be produced and will be sent to the EC and corresponding bodies such as ECDC and United States Environmental Protection Agency (USEPA).

Meetings with policy makers in each country
Project dissemination will continue through the different meetings with policy makers that will be organised in each country. The scientific partners of each country will contact the policy makers of their region and they will present their country results and present them the short report of the project. These meetings will be organised separately or during workshops. Policy makers will include representatives from ministries of environment, local government, municipalities, ministries of health, prefectures, etc. A list of the contacted policy makers will be included by each partner and will be uploaded in the website. The policy makers will also receive the protocol for the investigation and management of extreme water events that has been developed and their technical opinion / evaluation will be asked in order to finalise and uploaded in the website by the end of 2013. The reports from each contact person will be uploaded in the restricted website of the Project and will be available to the EC.

Questionnaire
The evaluation of the results of the project by policy-makers will be performed. The questionnaire will be filled by the policy makers in each country. The processed outcomes of the evaluation of the questionnaire will be uploaded in the website. The evaluation will be performed by the end of 2013. These evaluation reports will be uploaded in the restricted website.

Final conference
The final conference was held in Cardiff on March 2013. The conference addressed the research community as well as the policy makers. The conference programme is included in the WP9 deliverable report.

Final report
A final newsletter will be sent to recipient list. This newsletter will present the final outcomes of the project to the recipients of the previous newsletters (mainly NGOs).

Website activity
The website will be open to the public for at least two years after the end of the project (end of March 2015). Every publication or other report will be uploaded in the open access or restricted website and remains there until the end of March 2015.

Exploitable foreground (Sect. B)

Amongst other benefits, the information obtained from the Ppoject comprises probably the most complete series of data for waterborne viruses that have ever been collected from various European waters. These data, along with their statistical analyses and the climatic data and public health evaluation constitute a complete set of scientific information on waterborne viruses. As the data were collected recently, significant exploitation has to await publication in 2013. The acquired data will be available to many different stakeholders such as environmental scientists, climatologists, virologists, public health specialists and various bodies such as ECDC, WHO, EPA, etc.

The total purpose of the exploitation foreground strategy is the same that initially proposed, which is to disseminate and use it as widely as possible. The purpose of this strategy will be to create a complete plan for the exploitation of the project results and data. A few of these steps have been already performed and others will follow after the conference.

As all the project results will be published, there are no main significant non-publishable (confidential) parts of the project except for the guidelines and report for policy makers who are responsible for adapting the project results to their existing strategies for public health issues. In more detail, there are several parts of this strategy that can be emphasised. The principal exploitable areas are outlined below.

Knowledge on the presence of waterborne viruses (general advancement of knowledge)

The purpose of this part is the increase of knowledge in the scientific discipline of water virology. The foreground acquired in this project will be exploited by several publications in scientific research peer reviewed articles. The outcomes published will be used by the researchers in the areas of chemical engineering, microbiology, water virology, epidemiology and public health. One paper has already been published and the rest are prepared according the publication plan (section A, above).

The data will be included in the report submitted to the policy makers in the European area and to global regulatory and policy agencies in order to inform debate and discussion regarding framework planning relating to (for example) climate change impacted events, water quality indicator selection policy, and strategies to respond to agriculture-related water pollution events. The data will be made available, as summary reports, to policy makers during 2013 and will be available on the VIROCLIME website until 2015. In parallel, the data will be available (after publication) to the stakeholders of corresponding are as in each country from where VIROCLIME participants came, and a report /evaluation will be requested from them. Finally a complete report of the European data (excluding the data from Amazon River) will be sent to ECDC and EPA for exploitation to their projects before September 2013. This data can be used then from these bodies in order to evaluate that and use it for the launching of new proposals.

Models for prediction of climate change (exploitation of results through EU policies)

During the project new models for the prediction of the consequences of climate change on waterborne virus levels were produced. These models could be used by EU to establish preventive policies. These are the first models produced based on virus data. The models for the possible scenarios could be used by the EU to inform strategy for the member countries. These models will also be disseminated to policy makers (ministries of environments) to each country as well as to central European environmental bodies such as EPA, ECDC. These models can be used in a restricted format. These models have not yet been sent to any scientific body as the final forms of them will be presented in the final conference. These models will be sent to EU, WHO, ECDC, EPA and USEPA. As ECDC has previously prepared corresponding models, based mainly on bacteriological data, this should be very significant and valuable for them to compare these and extract new policies (if needed).

Guidelines on the presence of waterborne viruses (exploitation of results through (social) innovation)

The purpose of this part is the production of a report which will include appropriate guidelines for public health officers and local policy makers. These guidelines could help the policy makers to encounter with the extreme climatic events relating to waterborne viruses and take the appropriate preventive measures for the public health protection. This report will be translated in several languages and can be used in different European countries. Also as in the project, three different sizes of water bodies in different countries have been included (large, medium, small), different guidelines based on the impact of climate change on waterborne viruses will be produced. Also specific corresponding guidelines for the presence of waterborne viruses in the Amazon region will be generated. These guidelines will be also uploaded in the website by the end of October 2013. The aim of these guidelines is that the public will be informed for their area. These guidelines will be based on the predictive models that climatists of the Project have produced and they will be finalised after the final conference.

Schedule for the exploitation of the foreground

1. Actions that have been made
- Three newsletters have been sent to more than 250 recipients NGOs. The aim was to notify to them, the website as the major point for their project information.
- A letter to a few targeted research bodies was been sent to inform them about the end of the Project and the forthcoming conference and the website.
- Five articles have been published.
- Two social media (Facebook and Flick) have been used to upload information for the Project. The Facebook group is called 'Viroclimers'.
March 2013: Final VIROCLIME conference

2. Scheduled actions for the exploitation of the foreground
All actions are scheduled to be completed by the end of 2013. The website, including the principal contact details, will remain active until 2015.
- The results of the conference will help to the evaluation of the project. A questionnaire will be provided to the participants.
- March - October 2013: Publications according the plan.
- End of 2013: Report to policy makers for each country.
- End of 2013: Report to stakeholder-research scientific bodies.
- End of 2013: Guidelines for the policy makers.
- End of 2013: Evaluation from policy makers through a questionnaire.
- End of 2013: A letter to NGOs will be sent asking them to tell their opinion for the project. Their opinions will be collected in a report that it will be uploaded in the website.

Contact details: Professor Apostolos Vantarakis, University of Patras, Greece
avantar@med.upatras.gr

List of websites: http://www.viroclime.org

Related documents