Spatial Distribution of Risk Factors for
Cryptosporidium Spp. Transport in an
Irish Catchment
S. R. Samadder1*, P. Ziegler1, T. M. Murphy2, N. M. Holden1
ABSTRACT: Cryptosporidium spp. has become a major public health
concern in many parts of the globe, including Ireland, as a result of recent
reported waterborne outbreaks of Cryptosporidiosis. Continuous monitor-
ing of Cryptosporidium spp. in water supplies is not feasible, so a risk-
forecasting approach is required. This study reports a globally applicable
approach for evaluating the spatial variation in relative risk of
contaminating surface water by Cryptosporidium spp. based on a risk
potential index (RPI) as an indicator of the potential pollution of surface
water. The RPI is predicted by readily available data on land use, rainfall,
soil type, slope, soil moisture deficit, and distance from water course. A
small catchment in County Meath, Ireland, was chosen to illustrate the
analysis of the approach. Data for the study area were digitized and
rectified using surveyed ground control points to capture each of the RPI
factors, field boundaries, and land use. The six parameters were classified
and assigned a relative risk score out of 5. A Geographic Information
Systems overlay analysis then was used to calculate a cumulative relative
risk score for each month of the year. The analysis indicated that April and
June experienced a relatively low risk of Cryptosporidium spp. transport
compared with other months of the year. June had the least risk, because
more than 98% of the catchment was estimated to be of low or moderate
risk (RPI ranges 5 0 to 2). December had the highest risk of
Cryptosporidium spp. transport, because approximately 20% of the
catchment area had a moderately high to very high risk (RPI ranges 5 2
to 5). The study also made an attempt to reduce the risk of contaminating
surface water by alternative land-use practice and relocating the field
boundaries. The study demonstrated a semi-quantitative and readily
implemented method for using spatial risk assessment for planning land
management to reduce the risk of surface water contamination by
Cryptosporidium spp. from agricultural sources. The approach could be
used in any catchment with similar hydrological characteristics subject to a
surface water pathogen contamination risk from grazing livestock. Water
Environ. Res., 82, 750 (2010).
KEYWORDS: Cryptosporidium, geographic information system, Ire-
land, livestock, risk, risk potential index, surface water.
doi:10.2175/106143010X12609736966649
1
UCD Bioresources Research Centre/Biosystems Engineering, UCD
School of Agriculture, Food Science and Veterinary Medicine, University
College Dublin, Belfield, Dublin, Ireland.
2
Central Veterinary Research Laboratory, Department of Agriculture,
Fisheries and Food, Young’s Cross, Celbridge, Kildare, Ireland.
* UCD Bioresources Research Centre/Biosystems Engineering, UCD
School of Agriculture, Food Science and Veterinary Medicine, University
College Dublin, Belfield, Dublin 4, Ireland; e-mail: sukha.samadder@
ucd.ie, sukh_samadder@yahoo.co.in.
750
Introduction
Cryptosporidium parvum (C. parvum) is a microbial pathogen
that is known to infect a range of animal hosts, but is important to
human health because it is zoonotic and prevalent in both bovines
and humans. Cryptosporodiosis is an issue of global concern, with
the first waterborne outbreak of C. parvum reported in 1985 in
Texas. Many additional outbreaks have been identified around the
globe, and many people have been infected by this parasite (Kato
et al., 2004). Dairy farms are known to serve as a source of C.
parvum oocysts, because one infected calf alone can release
billions of oocysts into the environment (Walker and Stedinger,
1999). C. parvum transmitted to humans from infected cattle
causes self-limiting enteric disease (cryptosporidiosis) in immu-
nocompetent humans and chronic life-threatening infections in
people with impaired immune systems (Fayer et al., 1997).
Children are the most susceptible to Cryptosporidium infection,
and those less than 5 years of age have been shown to have the
highest prevalence (Griffiths, 1998). Most Cryptosporidium spp.
oocysts are removed from water by conventional filtration, but,
because they are resistant to conventional disinfectants, such as
chlorination, those that are not removed are of major concern to
regulatory agencies, the water industry, and consumers. DuPont et
al. (1995) reported that the infectious dose to 50% of exposed
individuals (ID50) for immunocompetent people is 132 oocysts,
but the infection may be generated by as few as 30 oocysts. The
ID50 for immunocompromised individuals has not yet been
established; however, it is thought that infection can be caused
by a single oocyst (Casemore et al., 1997).
Continuous monitoring for Cryptosporidium is difficult and
expensive and can be regarded as infeasible in surface water.
Furthermore, the results of surface water testing in agricultural
watersheds may not indicate the presence of waterborne
Cryptosporidium spp. oocysts, because the oocysts recovery
efficiencies generally do not reach 50% and often are much
lower, leaving high numbers of water samples falsely negative
(Graczyk et al., 2000; Hansen and Ongerth, 1991; Nieminski et
al., 1995; Watanabe, 1996). Consequently, forecasting the relative
risk of contamination of surface water by livestock has become
necessary.
There are many possible approaches to risk assessment, ranging
from rapidly implemented qualitative methods (e.g., The Crypto-
sporidium (Scottish Water) Directions of 2003 [Scottish Govern-
ment, 2003]) to quantitative risk methods (e.g. Pouillot et al.,
2004), but there is a need for a readily implemented method that
can capture the spatial and risk component of Cryptosporidium
Water Environment Research, Volume 82, Number 8

spp. transfer from animals to source waters. Foster and McDonald
(2000) identified potential pollution hazards of Cryptosporidium
by creating a risk potential index (RPI) combining the annual
effective rainfall, land slope, soil hydrology, and proximity to
water course. They did not consider land-use pattern or the source
of Cryptosporidium. Jordan et al. (2007) proposed a slurry
acceptance map of Northern Ireland, in which hydrology of soil
types (HOST) classes were combined with slope, rockiness, flood
hazard, and soil moisture deficit (SMD) classes (land-use pattern
and distance to watercourse were not considered), but did not
interpret the result in terms of pathogen transport to surface
waters. An approach is presented that captures risk factors
associated with sources and transport vectors using readily
available data that can be implemented in a range of catchment
situations. It is illustrated using a small agricultural catchment in
Ireland.
Farming in Ireland is based mainly on the production of
livestock and livestock products from grass (Casey and Holden,
2006). Approximately 130 000 farmers are engaged in farming,
using 2.5 million ha of grass to rear and fatten the offspring of 1.2
million dairy and 1.1 million beef cows, which produce
approximately 2 million animals each year (Casey and Holden,
2006). In Ireland, approximately 1.2 million dairy cows on
approximately 27 000 farms consume grass in the field or as
conserved forage when housed, with additional feed as concen-
trates when needed. The weather permits 190 to 240 days of
grazing per year. All manure (slurry) produced is land-spread
(Casey and Holden, 2005). The cows graze grass from mid-March
to late October, and supplementation is achieved with silage and
concentrates. Wildlife also contributes to the load of Cryptospo-
ridium spp. oocysts in surface waters, but oocysts concentrations
in surface waters affected by dairy farms are believed to be an
order of magnitude higher than surface waters affected by wildlife
alone (Hansen and Ongerth, 1991).
The objective of this work was to develop a simple spatial risk
assessment tool based on the evaluation of monthly variation in
the spatial relative risk of contamination of surface water by
livestock using an RPI estimated for a small agricultural
catchment. The runoff of water over the land surface after a
rainfall event may transport pollutants quickly and efficiently into
the stream network. The example presented is not applicable to
free-draining catchments dominated by sub-surface and ground-
water processes. It is extremely important to assess the potential
for runoff occurring over a given area of catchment when
considering pollution sources in the catchment (Foster and
McDonald, 2000). Lewis and McGechan (1999) reported that, in
high rainfall climates, the drainage and general soil physical
characteristics are of particular importance to the generation of
surface runoff. For this study, RPI was calculated using ArcGIS
(ESRI, Redlands, California) and considering land-use pattern,
rainfall, soil type, terrain slope, SMD, and distance from the water
course.
A geographic information system (GIS) is a tool for the storage,
display, and analysis of spatial data (Bowman, 1998). Various
applications of GIS in environmental planning and management,
environmental risk assessment, and in other subjects have been
presented in literature. Watkins et al. (1996) used the data
management and spatial analysis capabilities of GIS in ground-
water modeling; Wong et al. (1997) developed a model to
estimate stormwater pollutant loading using GIS; Gunes and
August 2010
Samadder et al.
Kovel (2000) described the use of GIS in emergency management
efforts for disaster preparedness, mitigation, and response;
Kistemann et al. (2001) successfully used GIS as a tool for
microbial risk assessment in catchment areas of drinking water
reservoirs; Jordan et al. (2007) took advantage of GIS for slurry
acceptance mapping of Northern Ireland based on runoff risk
assessment; and Foster and McDonald (2000) assessed pollution
risks to water supply intakes using GIS. The GIS tools can help in
analyzing the adequacy of coordinate-based information used to
make particular decisions in a spatially sensitive manner (Parsons
and Frost, 2000). Risk managers demand access to decision-
support tools that allow them to manage and understand
complicated spatial situations (Zerger and Smith, 2003), and, for
water supply management, a readily implemented spatial risk
assessment tool would be of great benefit for supply and
catchment management.
Materials and Methods
Study Area. The study area, located in the northeast of the
Republic of Ireland (Figure 1) is a small catchment of 570 ha,
with approximately 90% of the area dedicated to agriculture, and
the remainder allocated to residential use. There are a total of 5
working farms located within its watershed—1 dairy herd and 4
beef operations. Initial contact with each farmer established herd
sizes and management practices, while a cross-sectional study on
2 of the farms (1 dairy, 1 beef) established the prevalence of C.
parvum for cattle in the catchment. This study found that 47% of
adults (age 12 to 24 months) and 72% of calves under 2 months
were infected with C. parvum, and 13% of the calves showed
clinical symptoms (McDonald et al., 2009).
Stocking density calculated from the land area used for
livestock production (grazing and silage) divided by the number
of cattle yielded a stocking rate of 0.86 cattle/ha. There was
limited fluctuation in the cattle population during the period of the
study. However, farm management practices were seasonal, with
cattle removed from pasture from mid-October until April.
Manure and slurry were collected from housed animals and
spread twice yearly.
Data Collection. Satellite image and topographical maps of
the study were obtained from the Ordinance Survey of Ireland
(Dublin, Ireland) (OSI). Field visits were used to confirm the land
use in each land parcel within the area under study. A soil map
and digital elevation model for Ireland were obtained from the
Environmental Protection Agency (Dublin, Ireland). Daily rainfall
data for the years 2006 and 2007 from two synoptic observation
stations near the study area (Clones Co. Monaghan, Ireland, and
Mullingar Co. Westmeath, Ireland) were obtained from Met
Eireann, the Irish National Meteorological Service (Dublin,
Ireland). Daily SMD values (Schulte et al., 2005) for well-
drained, moderately drained, and poorly drained soil for the years
2006 and 2007 were calculated from meteorological data.
Coordinates (latitude and longitude) of ground control points for
data georectification over the study area were collected by a
global positioning system. A satellite image of the study area was
rectified using the ground control points as a base map, with a
pixel resolution of 10 m for further GIS analysis of the catchment.
Risk Potential Index and Geographic Information Systems
Analysis. In this study, RPI was used as an indicator of the risk
of pollution of surface water. This was selected based on a source
R vector R target model (Ansell and Wharton, 1992), where (1)
751
Samadder et al.
Figure 1—Location of the study area in Ireland.
the direct source is infected livestock, moderated by manure and
slurry management; (2) the target is surface waters; and (3) other
factors indicate the risk of occurrence of a transport vector, via the
soil surface, between the source and target. The main factors that
influence RPI are land-use pattern (a source risk that must be
assessed on a site-specific basis), soil type (vector risk, based on
judgement or quantification of drainage processes), rainfall
(vector risk, based on observation), SMD (vector risk, reflecting
an interaction between soil type and rainfall; other soil water
status indicators could be used), slope (vector risk), and distance
from surface water (vector risk).
The following four categories of land use were defined for the
study area and scored (L, unitless, ranging from 1 to 5, Table 1),
with respect to relative source risk:
(1) Grazing with slurry and manure spreading;
(2) Grazing and manure spreading;
(3) Grazing; and
(4) Other, which included residential, marsh land, and tillage.
The scores reflect a judgement regarding the relative risk of a
source of C. parvum being present; thus, it is greater where
grazing, slurry spreading, and manure spreading coincide and least
where they do not occur. The increased risk of direct
contamination by young calves with clinical symptoms is not
considered, because they are never let out to graze in the field
while showing clinical symptoms that are associated with much
greater rates of oocyst shedding. However, field soil samples
tested for Cryptosporidium spp. from the different land uses do
not justify the relative score, because the data are only available in
terms of either negative or positive. Field boundaries were hand-
752
digitized as a GIS data layer and confirmed by field survey. Land
use was classified manually in an attribute table for each field
vector using field survey and farmer activity records.
Soil types in the catchment were classified into the 3 potential
categories of well-drained, moderately drained, and poorly drained
(defined by Schulte et al., 2005), and spatial distribution was
defined by the General Soil Map (Gardiner and Radford, 1980) and
field checking. By definition (Schulte et al., 2005), the water
content of a well-drained soil never exceeds field capacity, has a
high infiltration capacity, is unlikely to receive rainfall in excess of
infiltration rate, and therefore offers the least risk of a surface runoff
transport vector occurring. A moderately drained soil can have
water contents in excess of field capacity for up to 24 hours and
therefore both infiltration excess and saturation runoff are possible
for limited time periods. A poorly drained soil can hold water in
excess of field capacity for prolonged periods and therefore has a
greater risk of supporting surface runoff and oocyst transport. The
relative scores (So, unitless, ranging from 1 to 5, Table 1) were
assigned accordingly. Any appropriate soil indicator could be used
as the basis of scoring the soil attribute of the RPI, provided it
reflects the nature of the risk processes in the catchment.
The average daily rainfall for each month in 2006 and 2007
(Table 2) was similar to the national daily average rainfall,
ranging from 1.32 to 4.25 mm. Average rainfall data were divided
into the following 3 classes:
(1) High (.4 mm), reflecting a risk of localized high rainfall
leading to infiltration excess runoff and prolonged rainfall
leading to saturation runoff;
(2) Moderate (2 to 4 mm), reflecting a lesser risk of excess surface
water; and
Water Environment Research, Volume 82, Number 8
 Samadder et al.

Table 1—Parameters, classes, and their relative score. The integration of soil type and rainfall through the calculation
of average daily SMD for each month for well-drained,
Relative score moderately drained, and poorly drained soil was derived for
Parameter Classes of risk (out of 5) 2006 and 2007 using synoptic data (Table 3). Monthly average
Land-use pattern Grazing + slurry + manure 5 daily SMD for the catchment varied from 28.55 mm (near
Grazing + manure 4 saturation, gravity moveable water available to cause runoff) to
Grazing 3 30.06 mm (quite dry soil, no gravity moveable water available),
Other 1 and SMD was categorized as high (.100 mm), medium (15 mm
Rainfall High (.4mm) 5 to 100 mm), and low (,15 mm), as was done by Lea (1979). A
Moderate (2mm to 4mm) 3 high relative score was assigned to low SMD, and a low relative
Low (,2mm) 1 score was assigned to high SMD (Smd, unitless, ranging from 1 to
Soil type Well-drained 1
5, Table 1), to reflect the probability of gravity moveable water,
Moderately drained 3
Poorly drained 5 and hence a transport vector, being present. While the relative risk
Terrain slope Steep (.7%) 5 of a low SMD for a well-drained and poorly drained soil is not
Moderate (3% to 7%) 3 likely to be the same, the interaction is captured by combining
Mild (,3%) 1 with a ranking for soil type and rainfall amount. Daily average
Soil moisture deficit High (.100mm) 1 SMD data for each month were entered into GIS attribute tables to
Medium (15mm to 100mm) 2 generate monthly SMD layers as a time series. Any appropriate
Low (,15mm) 5 soil water status indicator could be used as the basis of scoring the
Distance from the High (. 250m) 1
soil–water interaction attribute of the RPI.
water course Medium (50m to 250m) 2
Low (,50m) 5
The slopes in the study area were divided into steep (.7%),
moderate (3 to 7%) and shallow (,3%), as was done by Jordan et
al. (2007). Steep slope was assigned the highest score, because, if
runoff is generated, it is more likely to be maintained as a surface
(3) Low (,2 mm), reflecting conditions when rainfall will more pathway down-slope by faster-moving water (resisting infiltra-
readily infiltrate into the soil and any runoff will be of small tion). On a shallow slope, surface runoff is more likely to dry up
volume and less likely to flow continuously from source to as water infiltrates into the soil (Sl, unitless, ranging from 1 to 5,
target. Table 1). The average slope in each field was calculated using the
Spatial Analyst tool in ArcGIS (ESRI) from the Digital Terrain
As most rainfall events are associated with frontal systems and Model of Ireland (McGinnity et al., 2002) and saved as the slope
last between 4 and 7 hours in Ireland, daily rainfall provides a layer for further analysis. This ranking is consistent with that used
broad outline of the risk that may arise. A relative risk score was by Jordan et al. (2007), based on the finding of Farquharson et al.
assigned (R, unitless, ranging from 1 to 5, Table 1). No field data (1978) and Heathwaite et al. (2003), and applies in situations
were available to calibrate this relationship empirically, so where soil properties accentuate runoff (impermeable layers and
sensitivity to boundary value was evaluated (61 mm rainfall for slow permeability) and the water table is shallow. In some
each boundary). Daily average rainfall data for each month were situations, a different ranking may be required, if shallow slopes
entered into GIS attribute tables to generate monthly rainfall are at a high risk of runoff, as a result of landscape position, return
layers as a time series. flow, or surface ponding during rainfall.
The distance from each cell of each land parcel (typically a
field) from the nearest surface water drainage line was classified
Table 2—Average daily rainfall of the year 2006 and as high (.250 m), medium (50 to 250 m), and low (,50 m), and a
2007 (mm). relative score was assigned (D, unitless, ranging from 1 to 5,
Table 1). No field data were available to calibrate this relationship
Average of 2 years of daily rainfall empirically, so sensitivity to boundary value was evaluated
Stations (640 m for each boundary). The UK Codes of Good Agricultural
Average Practice (Department for Environment Food and Rural Affairs,
Month Clones Mullingar grand total 2003) for the prevention of pollution of water recommends that
slurry is never spread within 10 m of a waterway or within 50 m of
January 2.45 2.52 2.48 a spring, well, or borehole (Jordan et al., 2007). Irish legislation
February 1.82 2.06 1.94 (SI 101 of 2009) requires buffers of between 3 m (next to drains)
March 3.32 2.98 3.15
and 200 m (around water supplies of .100 m3/d). Drainage lines
April 1.43 1.22 1.32
May 3.04 3.19 3.11
were extracted from the Digital Terrain Model for Ireland using
June 1.70 2.09 1.90 the Spatial Analyst tool in ArcGIS and were confirmed by field
July 3.21 3.25 3.23 survey. The distance of each cell of parcel from the nearest down-
August 3.09 2.71 2.90 slope drainage line was calculated and used as the distance layer
September 2.71 2.98 2.84 for analysis.
October 2.82 2.47 2.65 All the data layers were converted to a raster layer data rectified
November 2.60 2.45 2.52 to the satellite image base map. Each pixel on each layer then was
December 3.97 4.54 4.25 assigned its relative risk score (derived from Table 1) to create
Average grand total 2.69 2.71 2.70
monthly RPI maps of the catchment. Each factor in the RPI

August 2010 753

Samadder et al.

Table 3—Average daily SMD for the years 2006 and 2007 (mm).

Average of two years daily soil moisture deficit

Well-drained soil Moderately drained soil Poorly drained soil

Stations Stations Stations
Average Average Average
Month Clones Mullingar grand total Clones Mullingar grand total Clones Mullingar grand total

January 0.59 0.57 0.58 21.53 21.35 21.44 28.37 28.49 28.43
February 2.19 1.31 1.75 0.99 20.16 0.41 25.72 26.69 26.20
March 2.33 1.54 1.94 0.13 20.42 20.15 26.47 27.24 26.86
April 15.54 17.59 16.56 15.19 17.45 16.32 9.44 12.01 10.72
May 17.54 14.20 15.87 16.53 13.26 14.89 13.91 10.63 12.27
June 30.66 28.94 29.80 30.66 28.69 29.67 31.91 28.20 30.06
July 28.25 31.67 29.96 27.34 30.35 28.84 26.68 28.23 27.45
August 12.58 26.67 19.62 11.69 26.05 18.87 9.95 24.29 17.12
September 7.56 11.45 9.50 6.47 10.73 8.60 0.35 9.10 4.73
October 3.13 2.17 2.65 1.725 0.96 1.34 25.38 23.70 24.54
November 0.82 0.67 0.75 21.02 21.05 21.04 27.94 27.83 27.89
December 0.55 0.49 0.52 22.57 22.55 22.56 28.53 28.57 28.55
Average grand total 10.17 11.49 10.83 8.82 10.21 9.51 4.19 5.88 5.03

calculation was given equal weighting, with the vector factors
being moderated by the source factor; thus,
 
SozRzSmdzSlzD : L
RPI~ ð1Þ
5 5
The final RPI value ranged from 0–1 to 4–5, where 4–5 denotes a
zone where the source and transport factors are most likely to
coincide, thus indicating a high risk of surface water contamina-
tion by Cryptosporidium from agricultural sources; 0–1 denotes a
zone with least risk; and values in between reflect a progression
on increased risk, as a result of the coincidence of risk factors.
Where the source risk is deemed small, the vector risk is reduced
accordingly.
A low RPI was predicted for large areas of the catchment (51 to
75%) from April to August (Figure 3), because the transport
vector was unlikely to be present, even if there was a source in the
field. As the RPI was calculated from six layers of L, So, R, Smd,
Sl, and D, and only Smd and R vary for each month, the change in
spatial distribution of RPI is a function of rainfall distribution
throughout the year, as moderated by the weather variables
influencing SMD (wind, temperature, relative humidity, and
incident radiation). There are areas (approximately 75%) in the
western half of the catchment that are less prone to risk (Figure 3),
as there is little chance of a source being present (mostly tillage

Results and Discussion

An area of ‘‘very high’’ risk (indicated by an RPI range of 4–5)
can occur only when there is likely to be a transport vector (high
rain, poorly drained soil, low SMD) near a surface water course
(,50 m), on a steep slope that can be subject to grazing, slurry
spreading, and manure spreading. This was predicted only for one
very small area of the catchment (0.3%) in December, and, while
the land could be used for spreading, according to legislation (SI
101 of 2009), it is prohibited in winter months, so the real
likelihood of a source of Cryptosporidium being present would be
very low. An area of ‘‘low’’ risk (indicated by an RPI range of 0–
1) can occur only when there is little chance of a source being
present, there is little chance of a transport vector (high SMD, low
rainfall), the soil can readily accept any rainfall that might occur,
and the site has little slope and is distant from any water course.
The intermediate RPI ranges 1–2, 2–3, and 3–4 were classified as
moderate, moderately high, and high risk, respectively. Figure 2
was prepared to establish the fixed risk distribution, as a result of
the interaction of slope, soil type, and distance to the water course
(fixed transport vector). The area was estimated as follows:
approximately 28% as fixed low risk; approximately 56% as fixed
moderate risk, 14% as fixed moderately high risk, approximately
2% area as fixed high risk, and a very small area (0.2%) as fixed Figure 2—Fixed risk distribution as a result of slope, soil
very high risk. type, and distance to water course.

754 Water Environment Research, Volume 82, Number 8


Figure 3—Monthly RPI.
and grazing); there is little chance of a transport vector (high
SMD, low rainfall); the soil can readily accept any rainfall that
might occur; and the site has mild slope and is distant from any
water course. The areas having the least transport vector risk, if
used for high source risk management (slurry and manure
spreading), would lead to the least chance of contamination of
surface waters. The areas with the highest risk generally are steep
slopes that are near surface water drainage lines. Even though
these should not be used directly for spreading (based on good
agricultural practice), as a result of the risk of a continuous
transport vector over these sites, it would be appropriate not to
spread slurry or manure in adjacent areas, to minimize surface
water contamination.
In January, approximately 35% of the catchment was estimated
to be low risk, approximately 55% as moderate risk, approxi-
mately 9% as moderately high risk, and approximately 1% as high
risk (Figure 3). In February, the moderate risk area increased to
approximately 61%, the moderately high risk reduced to 4%, and
the high risk reduced to 0.4%. The reduction in risk in February
reflected a decrease in rainfall. In Ireland, the biggest difference
between mean monthly rainfall values occurs as a decrease from
January to February; thus, the transport risk was reduced. With an
increase in rainfall in March, the percentage area in each risk
category returned to similar values, as estimated for January with
the same spatial distribution. By April, the RPI risk was reduced
to low (69%), moderate (30%), and moderately high (1%), as a
result of increasing SMD and less rainfall. In May, a small
(approximately 1%) area of high risk emerged, with approxi-
mately 4% of the catchment being deemed moderately high risk,
44% as moderate risk, and approximately 51% as low risk, as a
result of a substantial increase in rainfall and reduced SMD. With
August 2010
Samadder et al.
the onset of full summer, approximately 75% of the catchment
was estimated to be low risk, 24% as moderate risk, and 1% as
moderately high risk. The dominance of low to moderate risk was
maintained through July and August. With the onset of autumn in
September, October, and November, the risk was forecast to
increase (35% low, 55% moderate, 9% moderately high, and 1%
high risk) to a maximum in December (35% low, 45% moderate,
18% moderately high risk, approximately 2% high risk, and 0.3%
very high risk). Monthly water sample testing for Cryptosporid-
ium for the study area confirmed that December had the highest
(80 oocysts/50 L of water sample) concentration of oocysts
compared with other months of the year.
Sensitivity to boundary values was evaluated for the months
December to March for distance to water course (640 m for each
boundary) and rainfall (61 mm rainfall for each boundary), as no
field data were available to calibrate the relationship empirically
(Figure 4). The analysis revealed that there was no shift in the RPI
range of 0–1 (35% in all case) for all 4 months analyzed. There
was a small amount of shift (1 to 7%) in moderate and moderately
high risk zones (RPI 5 1–2 and 2–3) towards positive and
negative sides, as a result of the effect of the change in boundary
values. Negligible change (0 to 0.4%) in shift was observed in
high and very high risk areas.
To evaluate whether existing land-use practice could be
changed to reduce risk, a virtual change in land management
was imposed, assuming the total land area per farm remained the
same; this restricted farming activities within 50 m of water
courses (tillage or silage production was allowed) and moved
slurry and manure spreading activities into fields with low fixed
risk (moderate slope, greater than 50 m from a water course, and
soil either well-drained or moderately drained). The recalculated
755
Samadder et al.
Figure 4—RPI variation resulting from risk boundaries.
RPI with this alternate land-use options for the months January,
February, May, July, and December (Figure 5) shows that
improvements can be made. These 5 months represented the
risk-reduction options for the whole year, except April and June,
as these 2 months were at relatively low risk and were excluded
from the risk-reduction analysis. January represented the risk
distribution of March, September, October, and November, as risk
distribution was same as that of January. Risk distribution in
Figure 5—Reduced RPI distribution.
756
January was evaluated to be 46% (increased by 11%, as a result of
the change in land use) as low risk, 50% (5% reduction) as
moderate risk, and 4% (5% reduction) as moderately high risk. In
February, low risk was increased by 11%, moderate risk was
reduced by 7%, and moderately high risk was reduced by 4%.
Moderate risk in May was reduced by 9% and moderately high
risk was increased by 1%, but high risk was reduced completely
(1%). In July, high risk (RPI 5 3–4) was reduced completely, and
Water Environment Research, Volume 82, Number 8

moderately high risk (RPI 5 2–3) was reduced by 0.4%, but
moderate risk (RPI 5 1–2) was increased by 6%. In December,
very high risk (RPI 5 4–5) and high risk (RPI 5 3–4) were
reduced completely, and moderately high risk (RPI 5 2–3) and
moderate risk (RPI 5 1–2) were reduced by approximately 8%
and 2%, respectively.
Conclusions
This study presents an approach to identify when a water course
is at risk of being contaminated by Cryptosporidium spp. The
approach considers RPI to cause Cryptosporidium spp. transport
to the water course. The RPI is influenced by land-use pattern,
rainfall, soil type, terrain slope, SMD, and distance of the pasture
from the water course. The study reports the distribution of RPI of
an Irish catchment for the months January to December, which
varies each month because of varying rainfall amounts and SMD.
Analysis revealed that the months April and June are of relatively
low risk of Cryptosporidium transport through runoff compared
with other months of the year. June has the least risk of runoff, as
the percentage of low-risk-prone area (RPI 5 0–1) is highest, and
the maximum RPI range is 2–3 for that month. December has the
highest risk of Cryptosporidium transport to the water course, as
this month has the highest RPI distribution. An attempt was made
to reduce the risk of contaminating the water course by restricting
farming activities within 50 m of the water course (tillage or silage
production was allowed) and shifting slurry and manure spreading
activities into the field within farm boundaries with low fixed risk,
where slope is mild, distance is higher than 50 m, and soil is well-
drained or moderately drained. The study suggests that slurry and
manure spreading should be restricted in April and June, as the
risk of runoff is least in these 2 months. Slurry spreading should
be done on the areas having RPI ranges of 0–1 (least risk). The
areas with the highest RPI scores (3–4 and 4–5) should be
restricted by land boundaries to protect runoff towards the water
course to reduce pollution risk.
The present study provides an example of the implementation of the
RPI concept to a small catchment in Ireland. The information will help
planning and implementing of alternative land-use practice and
relocating field boundary options to reduce the risk at streams
considering the risk scores in each month. The approach is globally
applicable for small catchments dominated by surface water processes.
Credits
The authors thank the European Commission Framework
Programme 6 Marie Curie Transfer of Knowledge project
‘‘Cryptonet.ie’’ (MTKD-CT-2005-029454) for funding this work.
Submitted for publication August 7, 2009; revised manuscript
submitted December 11, 2009; accepted for publication January
25, 2010.
References
Ansell, J.; Wharton, F. (1992) Risk: Analysis, Assessment and Manage-
ment; John Wiley & Sons: Chichester, United Kingdom.
Bowman, D. (1998) Civil Engineering Data Meets GIS. ASCE J.
Computing Civil Eng., 12 (1), 5–7.
Casemore, D. P.; Wright, S. E.; Coop, R. L. (1997) Cryptosporidiosis—
Human and Animal Epidemiology. In Cryptosporidium and Crypto-
sporidiosis, Fayer, R. (Ed.); CRC Press: Boca Raton, Florida, 65–92.
Casey, J. W.; Holden, N. M. (2005) Analysis of Greenhouse Gas
Emissions from the Average Irish Milk Production System. Agric.
Syst., 86, 97–114.
August 2010
Samadder et al.
Casey, J. W.; Holden, N. M. (2006) Quantification of GHG Emissions
from Sucker-Beef Production in Ireland. Agric. Syst., 90, 79–98.
Department for Environment Food and Rural Affairs (2003) Manure
Management Plan: A Step by Step Guide for Farmers. Department for
Environment Food and Rural Affairs: London, United Kingdom,
http://www.defra.gov.uk/corporate/docs/forms/agri_env/nvz/manureplan.
pdf (accessed July 20, 2009).
DuPont, H. L.; Chappel, C. L.; Sterling, C. R.; Okhuysen, P. C.; Rose, J.
B.; Jakubowski, F. (1995) The Infectivity of Cryptosporidium Parvum
in Healthy Volunteers. N. Engl. J. Med., 332, 855–859.
Farquharson, F. A. K.; Mackney, D.; Newson, M. D.; Thomasson, A. J.
(1978) Estimation of Runoff Potential of River Catchments from Soil
Surveys, Special Survey No. 11. Soil Survey of England & Wales:
Harpenden, United Kingdom.
Fayer, R.; Speer, C. A.; Dubey, J. P. (1997) The General Biology of
Cryptosporidium. In Cryptosporidium and Cryptosporidiosis, Fayer,
R. (Ed.); CRC Press: Boca Raton, Florida, 1–42.
Foster, J. A.; McDonald, A. T. (2000) Assessing Pollution Risks to Water
Supply Intakes Using Geographical Information Systems (GIS). J.
Environ. Modell. Softw., 15, 225–234.
Gardiner, M. J.; Radford, T. (1980) General Soil Map, Soil Survey Bulletin
No. 36. An Forás Talúntais: Dublin, Ireland.
Graczyk, T. K.; Evans, B. M.; Shiff, C. J.; Karreman, H. J.; Patz, J. A.
(2000) Environmental and Geographical Factors Contributing to
Watershed Contamination with Cryptosporidium Parvum Oocysts.
Environ. Res. Sect. A, 82, 263–271.
Griffiths, J. K. (1998) Human Cryptosporidiosis: Epidemiology, Trans-
mission, Clinical Disease, Treatment, and Diagnosis. Adv. Parasitol.,
40, 37–85.
Gunes, A. E.; Kovel, J. P. (2000) Using GIS in Emergency Management
Operations. ASCE J. Urban Plann. Dev., 126 (3), 136–149.
Hansen, J.; Ongerth, J. E. (1991) Effects of Time and Watershed
Characteristics on the Concentration of Cryptosporidium Oocysts in
River Water. Appl. Environ. Microbiol., 57, 2790–2795.
Heathwaite, A. L.; Fraser, A. I.; Johnes, P. J.; Hutchins, M.; Lord, E.;
Butterfield, D. (2003) The Phosphorus Indicators Tool: A Simple
Model of Diffuse P Loss from Agricultural Land to Water. Soil Use
Manage., 19, 1–11.
Jordan, C.; Higins, A.; Wright, P. (2007) Slurry Acceptance Mapping of
Northern Ireland for Runoff Risk Assessment. Soil Use Manage., 23,
245–253.
Kato, S.; Jenkins, M.; Fogarty, E.; Bowman, D. (2004) Cryptosporidium
Parvum Oocyst Inactivation in Field Soil and Its Relation to Soil
Characteristics: Analyses Using the Geographic Information Systems.
Sci. Total Environ., 321, 47–58.
Kistemann, T.; Dangendorf, F.; Exner, M. (2001) A Geographical
Information System (GIS) as a Tool for Microbial Risk Assessment
in Catchment Areas of Drinking Water Reservoirs. Int. J. Hyg.
Environ. Health, 203, 225–233.
Lea, J. W. (1979) Slurry Acceptance. In Soil Survey Applications, Technical
Monograph No. 13, Jarvis, M. G., Mackney, D. (Eds.); Soil Survey of
England & Wales: Harpenden, United Kingdom, 83–99.
Lewis, D. R.; McGechan, M. B. (1999) Watercourse Pollution Due to
Surface Runoff Following Slurry Spreading. Part I: Calibration of the
Soil Water Simulation Model SOIL for Fields Prone to Surface
Runoff. J. Agric. Eng. Res., 72, 275–290.
McDonald, S.; Berzano, M.; Ziegler, P.; Peng, X.; Samadder, S. R.;
Murphy, T.; Holden, N. M. (2009) How Sensitive is Qualitative Risk
Assessment at the Small Agricultural Catchment Scale? Joint Meeting
of the Irish Society of Parasitology and British Association of
Veterinary Parasitologists, Nov. 27, Central Veterinary Research
Laboratory, Young’s Cross: Celbridge, Co., Kildare, Ireland.
McGinnity, P.; Mills, P.; Roche, W.; Muller(2002) WATER FRAME-
WORK DIRECTIVE—A Desk Study to Determine a Methodology for
the Monitoring of the ‘Morphological Condition’ of Irish Rivers. Irish
Environmental Protection Agency: Johnstown Castle, Ireland.
757
Samadder et al.
Nieminski, E. C.; Schaer, F. C.; Onghert, J. E. (1995) Comparison of Two
Methods for Detection of Giardia Cysts and Cryptosporidium Oocysts
in Water. Appl. Environ. Microbiol., 53, 627–676.
Parsons, R. L.; Frost, J. D. (2000) Interactive Analysis of Spatial
Subsurface Data Using GIS-Based Tool. ASCE J. Computing Civil
Eng., 14, 215–222.
Pouillot, R.; Beaudeau, P.; Denis, J-B.; Derouin, F. (2004) A Quantitative
Risk Assessment of Waterborne Cryptosporidiosis in France Using
Second-Order Monte Carlo Simulation. Risk Analysis, 24, 1–17.
Schulte, R. P. O.; Diamond, J.; Finkele, K.; Holden, N. M.; Brereton, A. J.
(2005) Predicting the Soil Moisture Conditions of Irish Grasslands.
Irish J. Agric. Food Res., 44, 95–110.
Scottish Government (2003) The Cryptosporidium (Scottish Water)
Directions. http://www.scotland.gov.uk/Publications/2004/01/18727/
31490 (accessed July 10, 2009).
758
SI 101 of 2009 (2009) European Communities (Good Agricultural Practice
for Protection of Waters) Regulations 2009. Statutory Instruments,
Government of Ireland: Dublin, Ireland.
Walker, F. R.; Stedinger, J. R. (1999) Fate and Transport Model of
Cryptosporidium. Environ. Eng., 4, 325–333.
Watanabe, M. E. (1996) New Cryptosporidium Testing Methods. Environ.
Sci. Technol., 30, 532–535.
Watkins, D. W.; McKinney, D. C.; Maidment, D. R.; Lin, M. D. (1996)
Use of Geographic Information Systems in Groundwater Flow
Modeling. ASCE J. Water Resour. Plann. Manage., 122 (2), 88–96.
Wong, K. M.; Strecker, E. W.; Stenstrom, M. K. (1997) GIS to Estimate
Storm-Water Pollutant Mass Loadings. ASCE J. Environ. Eng., 123
(8), 737–745.
Zerger, A.; Smith, D. I. (2003) Impediments to Using GIS for Real-Time
Disaster Decision Support. J. Comput. Environ. Urb. Syst., 27, 123–141.
Water Environment Research, Volume 82, Number 8