Drainage of Urban Water

Team members: Mezule, L., Hamzah, F., Vang, Ó., Ortells, H.S., Verhaelen, K., & Niño, G.V.

Problem Statement

The increase of rainfall due to climate change might collapse the sewer systems of cities that will not be able to collect and transport such amount of water over a short period of time. Cities in The Netherlands are increasingly designing sustainable urban drainage systems (SUDS) within urban areas to relieve the burden of the sewer system during such events. One of the examples is the Westergasfabriek park in Amsterdam, which is a former gasworks factory on the west edge of Amsterdam and provide space for creative and cultural businesses (Figure 1). There are trees, meadows and streams, dozens of offices, spaces for large and small events, bars, restaurants, a cinema, a theatre and much more (Westergras).

Figure1. Map of Westergasfabriek park and illustrations of the water source (A and B).

The water source intended for recreation in the park consists of rainwater collected by runoff. This place has become a popular place for children to play (Fig. 1). However, since the park is open not only for the people but also to animals, due to runoff of rainwater the pathogens from animal (dogs, birds etc.) feces might be transported into the water source.

The most common pathogen microbes that can be found in (surface) waters are cited in Table 1. Most of them are zoonotic, e.g. Campylobacter or Cryptosporidium, but also Legionella or cyanotoxines produced by Cyanobacteria are common.

Table 1. Waterborne pathogens (adapted from Coulliete et al.).

Pathogen	Illness	Source
E.coli O157:H7	Bloody diarrhea, Hemorrhagic colitis, Hemolytic Uremic Syndrome	Human and animal feces, Manure
Campylobacter	Gastrointestinal illness, Guillian-Barré syndrome	Animal feces
Sallmonella typhimurium	Septicemia	Human and animal feces
Cryptosporidium	Gastrointestinal illness	Human and animal feces
Cyclospora	Vomiting, fever, flu-like symptoms	Human feces
Adenovirus	Conjunctivitis, Respiratory infection, Gastroenteritis	Human feces
Hepatitis E virus	Liver disease	Human feces
Norovirus	Gastroenteritis	Human feces
Polyomaviruses	Respiratory, kidney and brain illness Merkel skin cancer	Human feces and urine
Legionella	Pneumonia	Water aerosols
Cytotoxins (Cyanobacteria)	Hepatotoxicity Neurotoxicity
Helicobacter	Peptic ulcers	Waterborne

In the present study, Cryptosporidium parvum has been selected as a target microorganism for several reasons:

It is a zoonotic parasite, shed to the environment in the feces of several animals, including dogs, cats, birds and rodents, which can be easily found in and around the park (WHO)
Cryptosporidium is also very prevalent in the human population: 2.6-21.3% of the population in African countries, 3.2-31.5% in Central and South America, 1.3-13.1% in Asia, 0.1-14.1% in Europe and 0.3-4.3% in North America (data based on detection of oocyst in fecal specimens) (Fayer).
It is present in the environment in the oocyst stage, which is very resistant to adverse environmental conditions, e.g. several months in cool moist conditions, 6 months between 10 and 20 °C, 3 months between 25 and 30 °C and it is very resistant to chlorine and other cleaning products (WHO)
The patent period in animals oscillates between 1 and several weeks, which means that after infecting an animal, they are shed for a relatively long period (WHO).
It is highly infective, with an ID50 of 9-1024 oocysts, and is immediately infective after it is secreted to the environment (Fayer).
It exist a well-described dose-response relationship model for human Cryptosporidium infection (WHO).

Children playing in the pond water may contract cryptosporidiosis. What is the risk of children playing in these waters?

Hazard Identification

Cryptosporidium is a protozoan parasite which can infect both animals and humans. The parasite infects the epithelial cells of the digestive and respiratory tract causing diarrhea. After ingestion of an oocyst four sporozoites are released and infect the epithelial cells. Cryptosporidium only survives outside a host in the oocyst state where it can live month in cold water (Fayer). 49 outbreaks have been reported between 1984 – 1999. The largest outbreak in Milwaukee (USA) in 1993 (Mac Kenzie) caused infections to more than 400.000 persons. C. parvum is the most frequently zoonotically acquired Cryptosporidium (Fayer). However, it has been reported that C. baileyi, C. canis, C. felis, C. hominis, C. meleagridis, C. muris can also infect humans (Fayer).

Signs and Symptoms in humans

Cryptosporidium cause cryptosporidiosis, which is a self-limiting gastrointestinal illness in immunocompetent humans. The incubation period lasts around 7 days (range 1-14 days) and the disease has a duration of 6-9 days. Longer duration is often found in AIDS patient where the effect of the infection can be lethal. All age groups can be infected but young children are most susceptible. Infection can happen through consumption of water or food contaminated with animals and humans feces (WHO; Fayer; Coulliette et al.). Diarrhea is the main symptom, and is usually accompanied by abdominal cramps, anemia, low-grade fever, nausea, vomiting and weight-loss. No curative therapy exists for individuals with cryptosporidiosis to the date (Coulliette et al.).

Prevalence

It has been reported that more than 155 different mammals can be infected with Cryptosporidium. For this case study the excretion of human pathogenic Cryptosporidium from dogs, birds, ducks and rabbits is regarded. In European countries the prevalence in humans varies between 0.3 - 4.3% from the population (Fayer).

Exposure Assessment

Figure 2. Area of interest of the Westerngasfabriek Park

The amount of Crypostporidium present in the park depends on the concentration of the pathogen in animal feces (dogs, rabbits, birds and ducks), the amount of visitors with dogs and the prevalence of the parasite in the Dutch animals.

Excretion in animal feces

The number of animals that visit the park every day, amount of feces that these animals deposit in the park and concentration of oocysts in those feces, is shown in table 2. The maximum amount of visitors Westergasfabrik park can hold during events is 18000, however, the estimated amount of people visiting the park during weekends in summer months is around 1500 (Westergras); due to lower frequency of visits during the working days, the average frequency of visitors can be assumed to be 700 people per day. Families usually consist of 3 members; therefore it can be assumed that an average of 200 children are visiting the park daily. According to the Dutch news (Dutch News) the Netherlands has a total population of 1.8 million dogs therefore an average of 10% of people have dogs.

Table 2. amount of feces per animal, number of animal visiting the park, calculated total amount of feces and concentration of oocytes in feces.

Animal	Amount of feces (gpapd)*	Number of animals (apd)*	Total amount of feces (per day)	Concentration oocytes in feces (per g)	Reference
Dog	220	70	15400	10-5000	Cox et al.
Rabbit	100	100	10000	0-77	Cox et al.
Geese	150	50	7500	370+/-197	Graczyk et al.
Duck	153.6	50	7680	50+/-270	Kuhn et al; Assumption
Birds	151.8	100	15180	210	Assumption

With this data, it is possible to calculate number of Cryptosporidium oocysts that can be found in the park per day.

Pathogen inactivation and removal

This section depicts the pathogen characterization relevant for exposure of children playing in the described WADI; including the die off rate of Cryptosporidium and the removal due to the performed reed pond filtration. Cryptosporidium parvum oocysts follow a first order die off, whereby temperature is believed to be the most lethal factor. Here, the survival of Cryptosporidum in water and fecal material is of relevance. The exponential relationship between the die off rate and temperature and parameterization in water and feces is defined by Peng et al.. Cryptosporidium is known to be very sensitive to desiccation, and experiments show that they can only survive for about 2 h in dry state (Fayer). However, since the moisture content in feces after one week of storage at ambient temperature is still about 70 % (Ferguson) this factor is considered as negligible.

Runoff

To estimate the runoff of rainwater into the WADI data on the rainfall during the summer months in the area of Amsterdam (meteorological station Schiphol) was collected using the KNMI website. Data on the daily precipitation events from 1st of June 2009 until the 31st of August was used and it was found that a Weibull distribution fits best. The duration of precipitation was fitted using a triangular distribution. The average duration of precipitation (day) is 0.045 and the maximum precipitation is 0.375 mm.

Moreover, the area needs to be defined from which water will flow into the WADI. This data can be found in the following table:

Table 3. area of the different parts of the WADI

Area	Stream 1 (m²)	Reed (m²)	Stream 2 (m²)
Concrete	0	2500	0
Grass	10,000	2,500	12,000
Water (m²)	2,500	4,000	6,500

Besides the amount of precipitation and the area, the run off coefficients, which depend on the kind of soil, need to be considered. We can find two types of surface in the considered area: Concrete and lawn. The runoff coefficient for concrete is 0.29 and for lawn (heavy soil average 2-7%)0.22.

Background concentration

The average concentration of Cryptosporidium oocysts in river water according to dispersion models developed by Medema and Schijven ranged from 4.5 – 5.4 oocysts/L, however, the waters where there was no influence of river water (as in Westergasfabrik park) contained as many as 0.3 – 0.38 oocysts/L.

Pathogen concentration in water

Considered parameters and their distributions

In the previous sections the parameters relevant for the calculation of the final concentration of oocysts in the water are described. Relevant for the concentration of oocysts on the ground are: the total amount of feces excreted from the considered animals, the concentration of oocysts per gram of feces, the prevalence of excretion of oocysts for these animals and the die-off rate of oocysts in feces. Moreover, we assumed that 90 % of the dog owners would pick up the dog excrement, meaning that only 10 % will remain on the lawn. The amount of feces and therewith oocysts transported into the water after a rainfall event is depended on the run off coefficient and the intensity of the rainfall. The data for these parameters were fitted to the following distributions (Table 4).

Table 4. Exposure parameters and distributions

Parameter	Distribution
Background concentration	Single value
Total amount of feces excreted in g	Sum of the mean value of the amount of feces excreted per animal
Oocysts per gram of feces	Poisson
Prevalence of excretion of oocysts	Poisson
Die-off rate of oocysts in feces	Single value
Daily precipitation in m³	Weibull distribution
Duration of precipitation per day	Triangular
Run off coefficient	Single value (for concrete and lawn)
Die-off rate of oocysts in water	Single value

For the data on prevalence and concentration of oocysts per gram of feces, average values of all animals were necessary to calculate the final concentration of oocysts in the water. A Poisson distribution was fitted to the data of each animal. The mean of e.g. the oocyst concentration of each animal was chosen as the describing parameter of the Poisson distribution. To result in a distribution on the average values of the concentration of oocysts in the total amount of feces, the distributions on the oocysts concentration per animal were added and divided by the total number of distributions. A Monte Carlo simulation with 10.000 repetitions was used to finally result in the distribution of the average concentration of oocysts for all animals per gram of feces. The same procedure was performed for the prevalence of oocysts excretion.

Equations used to calculate pathogen concentration in water

The pathogen concentration is different in the 3 areas surrounding the main stream of water. We name those areas Stream 1, Reed and Stream 2. Figure 1 represents the water flow along these 3 areas.

Figure 3. Flow model (for nomenclature see box 1) S1: stream 1, S2: stream 2.

The runoff flow and runoff concentration of C.parvum can be calculated using the equation 1 and 2, respectively:

Equation 1:

\begin{align*} Q_{RO} = \frac{P \times A_{land} \times RO_c}{1d} \end{align*}

Equation 2:

\begin{align*} C_{RO} = \frac{C_F \times M_F}{Q_{RO} \times t_P} k_W \end{align*}

To calculate the parasite concentration in each area, the following mass balances are applied:

Equation 3:

\begin{align*} C_{S1}=\frac{(QC_0+(Q_{RO,S1}\times C_{RO,S1}))\times k_{decay,w}}{Q} \end{align*}

Equation 4:

\begin{align*} C_R = \frac{((QC_{S1}+(Q_{RO,R}\times C_{ROR})\times k_{decay,w})\times(1-F_{removal})}{Q} \end{align*}

Equation 5:

\begin{align*} C_{S2} = \frac{(QC_{R}+(Q_{RO,S2}\times C_{RO,S2}))\times k_{decay,w}}{Q} \end{align*}

Equation Notation:

Q = Flow Rate (m3/s)

C = Concentration of C.parvum (particles/m3)

Aland = Area (m2)

M = Mass

kdecay = decay constant of cryptosporidium in water

t = time (s)

Fremovla: Fraction of removal

Sub indexes:

S1 = Stream 1

S2 = Stream 2

R = Reed

RO = Runoff

Ingestion Rates

Cryptosporidium oocysts are transmitted via the fecal oral route; the only exposure route for consumption of the organism was considered the deliberate or accidental ingestion of contaminated water (WHO). The ingestion routes considered were direct swallowing of water, swallowing of droplets during splashing and through hand – mouth contact. The suggested volume of water ingested by non-adults ranges from 30 – 45 ml per 45 minutes of swimming (here playing). For further analyses the mean of 37 ml was used (Dufour). The mean scenario: A non-adult is assumed to ingest 0.037 L (37 mL) of water swallowed by a splash in a day. The person is assumed to play for 2 hours a day on the weekends over a five month period (40 days in a year).

Dose

The dose ingested per person (children) per day is equal to the water swallowed times the concentration of Cyrptosporidium in the water. To calculate the dose, three different rainfall scenarios were considered: the falling of 15mm (average precipitation amount), the falling of 40 mm (maximum precipitation) and a Weibull distribution of the rainfall for the whole period. The risk of infection for children was calculated separately for stream 1 and 2.

Dose Response Assessment

The best fit dose-response relation for infection with Cryptosporidium is well described by Teunis et al. and it is an exponential model (see equation 6).

Equation 6:

\begin{align*} P = 1 - e^{(-rd)} \end{align*}

Where P is the probability of infection (in children), r is the probability of a microorganism to survive inside the host and cause an infection (for Cryptosporidium it is 0.004005) (Teunis et al.) and d is the dose.

Risk Characterization

The risk of infection in the different scenarios and streams is shown in figure 4 and table 5. The risk was found to be higher after the 15 mm rainfall event than after the 40mm. It could be expected a higher risk after higher precipitation, as more water is available to drag the oocysts present on the ground. However, the more amount of water produces higher dilution of the parasite in the streams and, therefore, lower concentration and lower dose. This explanation is supported by figure 5, which shows the sensibility analysis for stream 2 after 15mm rainfall event. It demonstrates that the duration of precipitation is indirectly related with the risk of infection. The model should be modified to include the effect of the amount of water on the amount of oocysts dragged.

Table 5. Risk of infection and probability of risk for the 3 scenarios and the 2 streams.

	Risk	Probability
15mm
S1	6.3E-04	0.36
40mm
S1	8.3E-04	0.35
S2	2.8E-04	0.35
Summer
S1	6.3E-04	0.35
S2	2.5E-04	0.35

The sensitivity analysis (figure 5) shows that the prevalence of infection of the different animals is the factor that contributes more to the uncertainty of the model, specially the prevalence in ducks. The other contributing factor is the duration of the precipitation which, as stated before, has an indirect effect on the risk.

Figure 4. Monte Carlo analysis results for the risk of infection for children at Westergasfabriek Park, summer 2009. A and B show the risk of infection after the 15mm rainfall: C and D after the 40 mm rainfall; E and F show the results for the whole summer (Weilbull distribution). A, C and E, represent the risk in stream 1; B, D and F represent the risk in stream 2.

Figure 5. Sensitivity analysis for the risk of infection in stream 1 after the 15mm rainfall scenario.

Although the risk could be considered low, not all the possible risk factors were considered due to data gaps. Several assumptions were done based on literature and some under common knowledge inside the group; this would give certain bias to the model design. Further definition of the model would give more specific results.

Risk Management & Communication

As recommendations we consider some measures that can be executed to reduce the quantity of feces dropped in the park and the risk of infection:

Awareness campaigns for people to pick up the feces of dogs and pets brought to the park
Signalization and campaigns to reduce the feeding of wild animals
Close or signalized areas of risk when sudden and high rainfalls are presented after dry long periods.

References

Teunis, P. F. M., Nagelkerke, N. J. D., & Haas, C. N. (1999). Dose response models for infectious gastroenteritis. Risk Analysis, 19(6).
Coulliete, A. D., Alan-Yillmaz, A. ., Dreelin, E. A., McNinch, R. M., Fong, T. T., & Rose, J. B. Drinking Water Safety in the 21st Century.
/HSE/WHS, W. . (2009). Risk Assessment of Criptosporidium in Drinking Water. Geneva, Switzerland: WHO.
Fayer, R. . (2004). Cryptosporidium: a Water-borne Zoonotic Parasite. Veterinary Parasitology, 126.
Kenzie, W. R. M., Hoxie, N. J., Proctor, M. E., Gradus, M. S., Blair, K. A., Peterson, D. E., … P, D. J. (1994). A Massive Outbreak in Milwaukee of Cryptosporidium Infection Transmitted through the Public Water Supply. The New England Journal Medicine, 126(4).
Fayer, R. ., & Xiao, L. . (2008). Cryptosporidium and Cryptosporidiosis (2ndnd ed.). CRC Press, IWA publishing.
Westergras. Retrieved from www.westergasfabriek.nl
Cox, P. ., Griffith, M. ., Deere, D. ., & Ferguson, C. . (2005). Concentrations of Pathogens and Indicators in Animal Feces in the Sydney Watershed. Applied and Environmental Microbiology, 71(10).
Graczyk, T. ., Fayer, R. ., Trout, J. M., Lewis, E. J., Farley, C. A., Sulaiman, I. ., & Lal, A. A. (1998). Giardia sp. Cysts and Infectious Cryptosporidium parvum Oocysts in the Feces of Migratory Canada Geese (Branta Canadensis). Applied and Environmental Microbiology, 64(7).
Kuhn, R. C., Rock, C. M., & Oshima, K. H. (2002). Occurrence of Cryptosporidium and Giardia in Wild Ducks along the Rio Grande River Valley in Southern New Mexico. Applied and Environmental Microbiology, 68(1).
Peng, X. ., Murphy, T. ., & Holden, N. M. (2008). Evaluation of the Effect of Temperature on the Die-Off Rate for Cryptosporidium parvum Oocysts in Water, Soils, and Feces. Applied and Environmental Microbiology, 74(23).
Ferguson, C. M., Davies, C. M., Kaucner, C. ., Krogh, M. ., Rodehutskors, J. ., Deere, D. A., & Ashbolt, N. J. (2007). Field scale quantification of microbial transport from bovine faeces under simulated rainfall events. Journal of Water and Health, 0.5(1).
Dufour, A. P., Evans, O. ., Behymer, T. D., & Cantu, R. . (2008). Water ingestion during swimming activities in a pool: A pilot study. Journal and Water and Health, 0.4(4).
Medema, G. J., & Schijven, J. F. (2001). Modelling the sewage discharge and dispersion of cryptosporidium and giardia in surface water. Water Research, 35(18).