Sustainability of Water Quality and Ecology 6 (2015) 107–118

Contents lists available at ScienceDirect

Sustainability of Water Quality and Ecology

journal homepage: www.elsevier.com/locate/swaqe

Water quality and public health risks associated with roof

rainwater harvesting systems for potable supply: Review and
perspectives
Willis Gwenzi a,⇑, Nothando Dunjana b, Charity Pisa b, Tonny Tauro b, George Nyamadzawo a,c
a
Department of Soil Science and Agricultural Engineering, University of Zimbabwe, P.O. Box MP167, Mt. Pleasant, Harare, Zimbabwe
b
Marondera College of Agricultural Science and Technology, University of Zimbabwe, P. Bag 35, Marondera, Zimbabwe
c
Department of Environmental Science, Bindura University of Science Education, Box 1020, Bindura, Zimbabwe

a r t i c l e i n f o a b s t r a c t

Article history: Knowledge of rainwater quality is critical for safeguarding public health. The review inves-
Received 5 December 2014 tigated rainwater quality, and public health risks associated with its consumption. Land
Received in revised form 8 January 2015 use practices, roof material, weather patterns and their interactions influence rainwater
Accepted 23 January 2015
quality. Contrary to the notion that roof water is safe, data point to physico-chemical
Available online 4 February 2015
and microbial contamination of rainwater via atmospheric deposition, leaching and weath-
ering of roof materials, storage/conveyance utilities and faecal contamination. However,
Keywords:
epidemiological studies linking consumption of rainwater to public health risks are scarce
Contamination
Rainwater
especially in developing countries. This reflects the lack of epidemiological research and
Developing countries confounding factors such as high disease burden. To minimize the public health risks,
Public health risks we recommend the implementation of risk assessment framework integrating laboratory
Roof materials analytical results and sanitary inspection risk analysis. Such a framework will enable
proper prioritization and targeting of engineering/technological interventions, public
education and housekeeping programmes.
Ó 2015 Elsevier B.V. All rights reserved.

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
2. The case of roof rainwater harvesting systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
3. Nature and pathways of rainwater contamination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
4. Determinants of rainwater quality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
4.1. Catchment characteristics and land use practices. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
4.2. Roof materials and characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
4.3. Weather patterns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
4.4. Temporal controls on roof water quality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
5. Public health risk and mitigation measures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
5.1. Contaminants of public health concern . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
5.2. Strategies to safeguard roof water quality and public health. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
6. Conclusions and future research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116

⇑ Corresponding author. Tel./fax: +263 4 307304.

E-mail addresses: wgwenzi@yahoo.co.uk, wgwenzi@agric.uz.ac.zw (W. Gwenzi).

http://dx.doi.org/10.1016/j.swaqe.2015.01.006
2212-6139/Ó 2015 Elsevier B.V. All rights reserved.
108 W. Gwenzi et al. / Sustainability of Water Quality and Ecology 6 (2015) 107–118

Acknowledgement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
Appendix A. Supplementary data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116

1. Introduction

In most developing regions including sub-Saharan Africa, water availability plays a critical role in supporting livelihoods,
food security and public health (Baguma et al., 2010). The bulk of countries in sub-Saharan Africa (SSA) lie in water-scarce
river basins, while for those countries with available fresh water resources, poor distribution of portable water has resulted
in water shortages. Rapid population and industrial growth, coupled with declining and highly variable rainfall induced by
climate change exert pressure on water demand for domestic, agricultural and industrial uses especially in sub-Saharan
Africa (Barron, 2009). Current efforts to improve water supplies for domestic and industrial uses have largely focused on
exploitation of surface water and groundwater resources even in countries such as Zimbabwe where principles of integrated
water resources management are being promoted. On the other hand, water balance analysis suggest that rainwater from
impermeable roof surfaces in both urban and rural areas represents an under-utilized resource currently excluded in existing
water policies in SSA (Gwenzi and Nyamadzawo, 2014). Consequently, compared to surface water and groundwater
resources, there is relatively limited research on the quality and public health risks posed by water harvested from roofs.
Rainwater harvesting (RWH) involves the collection, storage and subsequent use of rainwater for domestic, industrial and
livelihood activities where and when it falls (Ngigi et al., 2005; Jebamalar and Ravikumar, 2011). In SSA, surface water in
lakes, reservoirs and surface runoff are often contaminated with industrial/domestic wastes, pesticides, agrochemical
released from agricultural land, roads and urban landscapes (Hranova, 2006; Gwenzi and Nyamadzawo, 2014). Attempts
to harness rainwater harvesting have been limited to agriculture (Ngigi et al., 2005, 2007; Dile et al., 2013). At present, there
are limited documented cases of RWH for domestic or industrial supply in SSA, but this trend is likely to change in the near
future due to excessive demand for limited freshwater resources in the region. Moreover, RWH is also considered a key
adaptation strategy to the impacts of climate change (Barron, 2009). Compared to centralized water supply systems,
RWH systems have the potential to provide low-cost decentralized water to urban and rural households without access
to treated water.
In principle, the collection of rainwater before it hits the ground mostly from roofs implies that is safer than surface water
in lakes and rivers, and groundwater from shallow wells. However, several recent studies suggest that roof rainwater can be
contaminated, thereby posing public health risk if consumed without treatment (Ahmed et al., 2010a,b, 2011a,b, 2012a,b,
2014; Lim and Jiang, 2013; Alves et al., 2014; Dobrowsky et al., 2014a,b; Jesmi et al., 2014; Lye, 2014). For example,
consumption of untreated rainwater has been linked to bacterial diarrheas associated with Salmonella and Campylobacter,
bacterial pneumonia due to Legionella, botulism due to Clostridium, tissue helminths, and protozoal diarrheas from Giardia
and Cryptosporidium (Lye, 2002, 2014). A study in Australia also showed that untreated roof-harvested rainwater samples
tested positive for Salmonella, Giardia lamblia, Legionella pneumophila, and Campylobacter jejuni, thereby posing public health
risks to consumers (Ahmed et al., 2010b). These several studies challenge the anecdotal view that rainwater is safe for
human consumption without treatment. Therefore, before the benefits of roof rainwater harvesting can be attained there
is need to understand the public health risks associated with consumption of such water. This is particularly important in
SSA and other developing countries, where water treatment facilities are unavailable for some poor urban and rural house-
holds. Moreover, even those who are connected often experience erratic water supplies due to demand exceeding supplies.
Understanding the determinants of roof water quality and public health risks associated with consumption of roof water is
critical for the design, operation and evaluation of roof water harvesting systems for potable water supply in developing
regions such as sub-Saharan Africa, where such data are not available. For instance, such information is critical for devising
public health measures to reduce the risks associated with consumption of roof rainwater.
Therefore, the current study reviews and synthesizes global literature drawn from journal articles, conference
proceedings and research reports covering the period 1990s to present (2014). The specific objectives of the review were:
(1) to identify possible contaminant pathways of rainwater from source to point of consumption, (2) to evaluate roof water
quality and its determinants, (3) to determine potential public health risks associated with consumption of roof rainwater,
and (4) to identify key research gaps on roof rainwater quality and associated public health risks.

2. The case of roof rainwater harvesting systems

Rainwater harvesting for domestic and agricultural uses is a very old practice dating back to 4500 BC in the Middle East
and India (Sivanappan, 1997). The practice originated in arid and semi-arid areas, but increasing water demand for industrial
and domestic uses is forcing most developing countries (Jebamalar and Ravikumar, 2011) including those in sub-Saharan
Africa to consider RWH. Rainwater harvesting involves collection of rainwater from a catchment, storage and subsequent
 W. Gwenzi et al. / Sustainability of Water Quality and Ecology 6 (2015) 107–118 109

Key:
Rainfall
1 Source of contamination

Dissolved chemical contaminants

1. Air-borne Suspended particulate matter/contaminants
1
Microbiological contaminants

Contaminant flow paths from source to point of consumption

2
3. Storage and conveyance 4. Utensils and
system
humans
2. Roof material

Human
consumption
3
4

Fig. 1. Conceptual depiction of a rainwater harvesting systems indicating the key components (catchment, conveyance and storage and subsequent use),
the nature and flow pathways of contaminants from source to point of consumption.

use for domestic and/or livelihoods uses (Fig. 1). A basic roof rainwater harvesting system typically consists of a roof catch-
ment, storage facilities, conveyance system and a delivery system and subsequent use for multiple household purposes. Roof
rainwater use for domestic consumption is not new even in developing regions like sub-Saharan Africa and parts of Asia, but
until recently it was mostly limited to communal areas and is often regarded as a safe, despite the lack of laboratory analy-
tical data. However, increasing water shortages in cities and towns have seen the practice of roof water harvesting spread to
towns. Compared to conventional piped water supplies from centralized modern water treatment systems, RWH has the
opportunity to provide an alternative source of low-cost decentralized water for poor and vulnerable communities in urban
and peri-urban areas (Gwenzi and Nyamadzawo, 2014). In large parts of Asia and sub-Saharan Africa including Zimbabwe,
Malawi, Tanzania and South Africa (Tobaiwa et al., 1991; Mamuse et al., 2003), where groundwater with high fluoride con-
centration is prevalent the use of rainwater for domestic supplies may reduce the risk of dental and skeletal fluorosis.

3. Nature and pathways of rainwater contamination

Roof runoff may contain a complex chemical and microbiological composition. Conventional chemical contaminants
include nutrients, heavy metals and organic compounds such as petroleum hydrocarbons (Hoffman et al., 1982; Heberer
2002; Haggard and Bartsch 2009), while microbial contaminants include pathogens and disease vectors. Besides convention-
al contaminants, emerging contaminants including as pharmaceuticals such as hormonal or endocrine-disrupting com-
pounds (Hoffman et al., 1982; Heberer, 2002) and nanoparticles (Gao, 2008) have recently drawn a lot of research
attention. The nature and concentrations of the contaminants in roof water will largely depend on the source, anthropogenic
activities and the flow pathways and interactions with the environment. Several studies have demonstrated roof runoff con-
tamination by physical, chemical and biological contaminants (Förster, 1998; Garnaud et al., 1999; Zhong et al., 2001) as well
as microbiological contaminants (Corden and Millington, 2001; Jones and Harrison, 2004; Evans et al., 2006). Contamination
mechanisms include air-borne and atmospheric deposition, direct leaching of metal constituents from metal sheets,
colonization by various plants and consequently direct weathering of roof material and the leaching of the accumulated par-
ticulate organic matter and flora on roof surfaces (Förster, 1999; Chang et al., 2004; Adeniyi and Olabanji, 2005). King and
Bedient (1982) attributed leaching of chemical compounds from roofs to the acidic nature of ambient rainfall while Chang
et al. (2004) report that the rough surfaces and cracks associated with wood shingle roofs trap water, aerosols and debris
which promote plant growth, particularly fungi resulting in increased organic matter retention on the roof. In Nigeria,
Adeniyi and Olabanji (2005) revealed that direct leaching of roof materials by chemical reactions was the main source of
contamination on Al and Fe-Zn sheets. In addition, concrete tiles and asbestos were found to be more prone to colonization
by various plants leading to softening of the roofing material. Although some studies did not establish clear correlations
110 W. Gwenzi et al. / Sustainability of Water Quality and Ecology 6 (2015) 107–118

between contaminants and roof types (e.g., Chang et al., 2004; Mendez et al., 2010; Meera and Ahammed, 2011), Adeniyi and
Olabanji (2005)) managed to separate their roof types into two clusters. One cluster consisting of asbestos and concrete tiles
significantly correlated with Ca and bicarbonate ions, while the other cluster comprising of thatch, Al and Fe–Zn sheets were
significantly correlated with K, Cl, sulphates and nitrate contaminants.
Fig. 1 is a conceptual depiction of the potential sources, nature and pathways of runoff contamination. In summary, the
risk for roof runoff contamination exists along the whole flow pathway of rainwater from source to point of consumption.
Sources of contaminants include those that are air-borne, atmospheric and animal/bird deposition on roof/catchment sur-
faces (Förster, 1996; Evans et al., 2006), storage and conveyance facilities, and humans and utensils at point of consumptions.
The existence of various sources and pathways of contaminants demonstrates the complexity of maintaining and safeguard-
ing public health when relying on roof water harvesting systems for domestic supply.
The quality of rainwater is a key determinant of its suitability for various uses. Although rainwater can be used for various
purposes including irrigation, livestock watering (Mwenge-Kahinda et al., 2007), laundry, bathing and toilet use (Lynch and
Dietsch, 2010; Ward et al., 2012), here we focus on its quality with respect to drinking purposes. To safeguard public health,
the water quality requirements for drinking purposes are more stringent than for others uses. The quality of rainwater shows
considerable spatial and temporal variability. Literature suggests that the spatial and temporal variability of roof water qual-
ity depends on roof materials, catchment characteristics, precipitation properties, local weather and chemical properties of
pollutants (Förster, 1996; Vazquez et al., 2003; Meera and Ahmmed, 2006; Abbasi and Abbasi, 2011). Several other studies
have investigated the impacts of these factors, and their interaction on roof water quality (e.g., Adeniyi and Olabanji, 2005;
Mendez et al., 2010; Meera and Ahammed, 2011). An understanding of any trends will provide insightful information that
can be extrapolated in the other regions where data are scarce.

4. Determinants of rainwater quality

4.1. Catchment characteristics and land use practices

The term catchment has different spatial interpretations when applied to rainwater harvesting systems and classical
hydrology. In rainwater harvesting systems, the term catchment often refers to the area (e.g., roof, pavements) that directly
contributes to the harvested runoff, while in hydrology it refers to the drainage area from which all runoff flows through a
common point. In this section, we make a distinction between the two and use the hydrological definition of a catchment.
Anthropogenic activities or land use practices in the catchment may act as diffuse and point sources of rainwater contamina-
tion. Studies comparing rainwater quality in industrialised, agricultural and natural catchments reveal a clear trend in water
quality across the land use gradient (e.g., Abbas et al., 1993; Vazquez et al., 2003; Radaideh et al., 2009). For example, a com-
parative study of rainwater quality in two catchments with contrasting land use practices revealed that industrialised areas
had poor rainwater quality than agricultural areas (Radaideh et al., 2009). Similar trends have been reported in other studies
(Abbas et al., 1993; Vazquez et al., 2003). The poor rainwater quality in industrialised catchments is attributed to air pollu-
tion and atmospheric deposition from traffic and industrial emissions and aerosols (Huston et al., 2012). A study in Brisbane,
Australia showed that atmospheric deposition was responsible for 21% of the incidence of high level of lead which were
above regulatory guidelines (Huston et al., 2012). It is noteworthy that, due to the diffuse nature of air pollution and wind
transport, the impacts of industrialization on water quality may extend beyond the spatial boundaries of the industrial activ-
ity (Uygur et al., 2010).
In urban areas the combustion of fossil fuels such as coal, oil and gas for household heating and cooking may also account
poor rainwater quality. For example, high levels of polycyclic aromatic hydrocarbons (PAHs) in roof runoff collected during
winter season were attributed to burning of fossil fuels such as coal, oil and gas for heating purposes (Förster, 1998). In
Greece, high concentrations of nitrate, nitrite and sulphate were attributed to fossil fuel combustion for heating, traffic
exhausts and industrial activity while high electrical conductivity, alkalinity, sulphate, sodium and potassium were attribut-
ed to urban development (Gikas and Tsihrintzis, 2012). Although the bulk of studies report poor quality rainwater attributed
to industrial and traffic emissions (Mouli et al., 2005; Evans et al., 2006; Rouvalis et al., 2009; Gikas and Tsihrintzis, 2012),
results have not been consistent across sites as evidenced by a few exceptions. For instance, Farreny et al. (2011) reported
low concentration of sulphates and nitrates in a study area in close to a motorway. Similarly, Mendez et al. (2011) reported
comparable water quality from full-scale and pilot scale roofs despite differences in proximity to emissions from chimneys
and traffic and overhanging vegetation. These inconsistent findings suggests that the impacts of industrial activities on roof
water quality is site specific and may depend on the magnitude of industrial emissions, rainfall characteristics and other con-
founding factors.
Agricultural activities are often associated with the use of pesticides, herbicides and fertilizers, and the generation of dust
during tillage operations which may also act as sources of contaminants in roof runoff water. High ammonia concentration in
a study area in close proximity to agricultural lands was reported by Förster (1998). In another study, agricultural activity
resulted in higher concentration of phosphamidon (a pesticide constituent), in rural agricultural catchment than in urban
areas resulting in increased toxicity of roof runoff water in rural areas (Rouvalis et al., 2009). In an area without traffic emis-
sions and agricultural activities, Sazakli et al. (2007) and Mendez et al. (2011) observed no PAHs, volatile organic compounds
(VOCs) and organochloride pesticides (OCPs). These observations evidently demonstrate the effects of anthropogenic
 W. Gwenzi et al. / Sustainability of Water Quality and Ecology 6 (2015) 107–118 111

activities on roof water quality. Animals and plants may also affect the composition and characteristics of pollutants within a
catchment. Bird and animal faecal deposits contribute to Escherichia coli contamination of roof runoff (Ahmed et al., 2012a,b).
Gikas and Tsihrintzis (2012) reported higher values for ammonium, soluble phosphate (PO34 ) and total phosphorus (TP) in
rural areas relative to suburban and urban areas in first flush water. They attributed their observations to greater activity of
birds and rodents (excretion) and moss and lichens growth in the less industrialised rural area. The concentration of dis-
solved phosphate was positively correlated with total coliforms (TCs), suggesting a high likelihood that the phosphate
originated from bird and rodent excreta.
The findings presented here are based largely on studies conducted in industrialized countries (e.g., Greece, Australia, USA
among others), but lacks data from developing regions such as Africa and parts of Asia, where the use of untreated roof runoff
is likely to be prevalent. The lack of data on roof water quality from developing countries largely reflects limited research on
the subject, partly due to the notion that roof runoff is generally of good quality and suitable for human consumption given
the low level of industrialization. However, results of Gikas and Tsihrintzis (2012) showing that the risk of contamination of
roof runoff could be high even in low industrialized rural areas due to animal and plant activity warrants future research on
the subject. A case study of rainwater quality conducted in Zimbabwe showing that the concentrations of heavy metals (Zn,
Cu, Pb and Ni) and basic cations (Mg, Ca, K and Na) were below the WHO guideline limits for drinking water represents one
such efforts in Africa (Dirwai, 2014).
In summary, although rainwater quality is likely to vary from location to location, some trends are evident in literature;
(1) in urban areas in developed countries heavy metals and other inorganic ions are likely to be prevalent in rainwater due to
fossil fuel combustion, and traffic and industrial emissions; (2) higher concentrations of pesticide residues and fertilizers are
likely to occur in agricultural catchments in developed countries than their developing country counterparts. This is largely
due to low pesticide and fertilizer rates used in developing countries; and (3) microbial and faecal contamination of rainwa-
ter is likely to be more prevalent in rural catchments in developing countries than developed countries. This trend could be
attributed to bird and animal activity as well as human contamination.

4.2. Roof materials and characteristics

Roof characteristics such as materials, design (e.g., slope, length), age and condition, and their interactions with weather
conditions may have a profound effect on roof water quality. Roof materials include clay tiles, aluminium and galvanized
metal sheets, concrete, polycarbonates plastic and flat gravel roofs (Farreny et al., 2011) and grass thatch, polythene plastics
and asbestos in developing regions such as Africa and parts of Asia. A number of studies have investigated the effects of
diverse roof materials on roof water quality. For example, galvanized iron sheets contributed higher zinc concentration in
roof water than asbestos cement (Handia et al., 2003). Significant differences in roof water quality were also observed among
clay tiles, metal sheets, polycarbonates plastic and flat gravel roof (Farreny et al., 2011). While some heavy metal pollution
has been associated with metal sheet roofs, other metals have been observed to occur in non-metal roof types than in metals.
Table SM1 in Supplementary material summarizes the effects of roof material on rainwater quality. Other metals that have
been studied but reported to be of low or less severe concentrations include cadmium, chromium, nickel and selenium
(Förster, 1999; Melidis et al., 2007; Mendez et al., 2011). Other studies observed that the concentrations of Mg, Na, K, nitrate,
nitrites, chloride, sulphate and ammonium were not significantly different among free fall rainwater, asbestos and Fe–Zn and
aluminium metal sheets (Adeniyi and Olabanji, 2005; Mendez et al., 2011). In other studies, high concentration of nitrates
and sulphates in roof runoff were largely attributed to fossil fuel combustion rather than roof materials (Mouli et al., 2005;
Evans et al., 2006; Melidis et al., 2007; Gikas and Tsihrintzis, 2012). A study in Texas, USA comparing the effects of asphalt
fibreglass, concrete tiles, galvanised metal sheets on the concentration of iron, manganese, chromium, nickel, zinc, alumini-
um, copper, silicon and sulphur revealed that metal roofs did not always leach higher concentrations of metals compared to
other roofing materials (Mendez et al., 2010). Empirical data on the effects of roof material on heavy metal concentration in
rainwater are mixed. Chang et al. (2004) reported the highest violations of established USEPA standards for Al, Mn, Cu, Pb, Zn
and pH with Zn and Cu being the most severe from four roof types namely, wood shingle, composition shingle, aluminium
and galvanised iron. Zinc concentrations were significantly different among the four types of roofs and free fall rainwater, but
Cu concentrations were not (Chang et al., 2004). The finding implies that roofing materials were the predominant source of
Zn but not Cu. These results confirm the findings of Mendez et al. (2010) that higher metal concentrations in roof runoff are
not always associated with metal roofs. On the contrary, an earlier study by Förster (1996) showed significantly higher cop-
per and zinc contamination as a result of contact between the roof runoff water and metal surfaces either as catchment areas,
roof side fittings or gutters.
The inconsistent effects of roof material on roof water quality possibly point to the existence of confounding factors. In
this regard, the age of the roof materials, frequency of maintenance and roof design characteristics such as slope and length
also influence roof water quality (Mitchell, 2003). For example, corrosion and weathering of zinc sheets have been shown to
result in high concentrations of zinc in rainwater (Förster, 1996; Simmons et al., 2001). In India, comparison of roof water
quality from concrete, tile, asbestos and galvanised iron reported high variability in water quality parameters among differ-
ent materials and attributed the differences mainly to the age of the roofing materials (Meera and Ahammed, 2011). Enrich-
ment factors, expressed as the ratio of physico-chemical parameters in roof runoff to that in free fall water were high in old
roof for most parameters than in new roofs (Meera and Ahammed, 2011). Similarly, higher contaminant concentrations in
112 W. Gwenzi et al. / Sustainability of Water Quality and Ecology 6 (2015) 107–118

older roofs were reported in other parts of the world such as Nigeria (Adeniyi and Olabanji, 2005). Furthermore, leaching
from roofing material such as concrete can increase the inorganic ions such as calcium in harvested water (Förster, 1998).
The design of the roof has also been shown to influence water quality. For instance, flat gravel roofs have been reported to
give highly contaminated roof water due to high particle deposition from air pollution arising from fireplaces, industry and
vehicles, roof weathering and plant colonization (Farreny et al., 2011). Roof slope and length has previously been shown to
affect harvested rainwater quality (Mitchell, 2003). Moreover, materials used to increase the durability of roofs may also act
as a source of contaminants for roof water. For example in Paris, France, registered products used for prolonging durability of
roofs gloss cleaning, de-mossing, waterproofing and painting such as benzalkonium, isothiazolinone and lead paints have
been shown to have a detrimental effect on roof water quality (Van de Voorde et al., 2009). For example, about 7–35% of
benzalkonium, isothiazolinone applied to concrete and clay tiles were recovered after 285 mm of rainfall (Van de Voorde
et al., 2009), while lead from acrylic-based paints have been shown to contribute to lead contamination of roof rainwater
(Abbasi and Abbasi, 2011).
The foregoing discussion revealed that although roof materials potentially influence roof water quality, their effects could
be confounded by age, design and maintenance status of the roof. Consequently roof water quality tends to vary among roof
materials, sites and studies. This constrains the generalization or extrapolation of research findings from previous studies.
Notwithstanding the inconsistency, old metal roofing sheets tend to be consistently associated with high concentrations
of heavy metals in roof runoff.

4.3. Weather patterns

Antecedent and prevailing weather conditions such wind and rainfall characteristics influence physico-chemical and
microbiological quality of roof runoff (Sazakli et al., 2007). The effects of weather conditions on roof water quality are related
to the temporal variability of roof water quality and first flush phenomenon (Section 4.4). Rainfall amounts and timing are
critical factors influencing roof water contamination. Studies have shown that the highest concentrations of constituents are
often detected during low rainfall events preceded by long dry spells of atmospheric deposition (Yaziz et al., 1989). For
example, high rainfall events cause contaminant dilution especially when occurring later in the season (Jiries et al., 2002;
Teemusk and Mander, 2007). On the other hand heavy storms falling early in the season after atmospheric deposition of con-
taminants facilitate detachment and transport of contaminants into the roof water. Bangira et al. (2007) noted that very high
lead (Pb) concentration when the rainfall amount was below 20 mm per day and a similar observation was reported in a
separate study conducted in Jordan (Radaideh et al., 2009). Further, the rate of pollutant build-up on roof surfaces is relative-
ly high soon after a storm event and reduces as the antecedent dry days increases (Egodawatta et al., 2009). Note that
although the factors influencing the spatial and temporal variability of rainwater were considered separately, in reality these
multiple factors interactively control water quality.
Wind characteristics particularly direction and speed are important meteorological factors influencing roof runoff water
quality through detachment, transport and deposition of contaminants of anthropogenic and natural origins. For example,
based on the ratio of Na+: Cl , Evans et al. (2006) predicted that Cl concentrations would be highest for events dominated
by prevailing winds originating from south east both antecedent to and during storm events; heterotrophic plate counts
(HPCs) were demonstrated to be largely determined by wind speed irrespective of direction during storms, while on the
other hand, Pseudomonas counts appeared to be under much stronger influence of wind direction (Evans et al., 2006). Jato
et al. (2000) showed some positive correlation of pollen from Pinus spp. with temperature, hours of sunshine and winds from
forested areas. Evans et al. (2006) attributed the significantly high Cl concentrations in rainwater to aerosols of sea origins
and transportation by the south easterly winds both before and during storm events. On the other hand, total bacterial load
in roof runoff was attributed to greater uplift of organisms from sources due to high wind velocities (Evans et al., 2006). In
addition, the effects of temperature and water availability on the growth of the source material can affect the concentration
of biological aerosols, and subsequently their leaching and weathering from the roof material, and consequently their con-
centration in roof runoff (Jones and Harrison, 2004).
A review on the mechanisms of meteorological contamination reveals that different contaminants undergo different pro-
cesses during deposition, removal and movement of bioaerosol contaminants (Jones and Harrison, 2004). Aerodynamic drag,
replacement by similarly charged particles and the impact of other particles knocking the particle might remove some par-
ticles from the surfaces where they were deposited. Maximum wind speeds have to exceed a threshold. Jones and Harrison
(2004) concluded that temperature and water availability can affect biological aerosol concentrations through their effect on
the growth of the source material. Moreover, temperature and moisture are likely to affect concentration of chemical
contaminants by influencing leaching and weathering of the roof material. Other studies concluded that wind speed is more
important for uplift of contaminants from sources and their deposition at the catchment surface, while wind direction most-
ly influences the composition of the contaminant load because it determines the source (Evans et al., 2006).

4.4. Temporal controls on roof water quality

The physico-chemical and microbiological quality of roof runoff may exhibit temporal variability at different scales
(Simmons et al., 2001; Sazakli et al., 2007). Seasonal variations in physico-chemical and microbiological quality of roof
water have been reported in a few studies. For example, higher electrical conductivity and chlorides in winter than in
 W. Gwenzi et al. / Sustainability of Water Quality and Ecology 6 (2015) 107–118 113

summer (Sazakli et al., 2007). The seasonality in roof runoff quality has been attributed to severe weather conditions such
as strong winds which deposited sea material on the catchment surfaces in winter, while lower values of microbial
indicators in winter were attributed to growth inhibition due to unfavourable conditions. Similar observations which attest
to seasonal variability of roof runoff pollution have been reported elsewhere (Simmons et al., 2001; Evans et al., 2006).
Although information of roof water quality at large temporal scales such as years, decades and millennia are scarce, such
variability may reflect changes in weather patterns, climate and land use practices (e.g., industrialization) within the
catchment.
The most documented case of temporal variability of roof water quality is at event scale. Several studies have shown that,
in general, the concentrations of contaminants in roof runoff are highest at start of the rainfall event compared to subsequent
events of similar magnitudes (Förster, 1996; Mendez et al., 2011). The presence of high concentrations of contaminants in
the first spill relative to subsequent spills is referred to as the first flush phenomenon. The concentration of contaminants
in roof water also depends on the timing and amount of rainfall. The first flush phenomenon has been extensively reported
in literature (Förster, 1996; Mendez et al., 2011; Gikas and Tsihrintzis, 2012). For example, significantly higher concentra-
tions of metals (Al, As, Cu, Fe, Pb and Zn) (Mendez et al. (2011), electrical conductivity (EC), alkalinity, Mg, Na, K, Ca and sul-
phates (Gikas and Tsihrintzis, 2012) have been reported in the first flush than subsequent samples. However, a few
exceptions have been reported in literature: for example Gikas and Tsihrintzis (2012) observed similar concentrations of
total coliforms and other microbial contaminants in both first flush and subsequent samples collected thereafter. Therefore,
Gikas and Tsihrintzis (2012) concluded that in some cases the first flush system could safeguard rainwater from microbial
contamination. Two possible reasons could account for this anomaly: (1) the magnitude of the rainfall event associated with
the first flush was inadequate to clean the roof/catchment, and (2) the rainwater could have contained microbiological con-
taminants from air-borne sources. Indeed, studies have shown that the quantities of the first flush quantities could be as
lows as between 0.11 and 2.6 mm (Van Metre and Mahler, 2003; Mendez et al., 2011; Gikas and Tsihrintzis, 2012). In regions
such as water-limited environment with distinct wet and dry season and highly seasonal rainfall, the first flush phenomenon
often termed seasonal flush refers to differences in quality of roof runoff between the first and subsequent events during the
season (Stenstrom and Kayhanian, 2005). In this case, the long dry periods contribute to contaminant build-up thus creating
large discharge of contaminants with the onset of the wet season (Stenstrom and Kayhanian, 2005). Using data collected over
extended periods of time in Los Angeles, Stenstrom and Kayhanian (2005) reported the occurrence of seasonal first flush
events and observed that the seasonal first flush had higher contaminant concentration than subsequent rainfall events.
However, seasonal first flush data from roofs remain limited, with most studies concentrating on the conventional first flush
water quality relative to later stages with a single rainfall event.
On the other hand, in regions where rainfall is highly episodic and sparsely distributed in time, the first flush may refer to
the difference in quality between the first and the next subsequent event. In other words, the interim period between the
first and subsequent event may lead to build-up of contaminants on roofs/catchment surfaces. Therefore, depending on
the magnitude of the first flush and ambient air quality, rainwater quality parameters within WHO guidelines for drinking
water have been reported, while in other studies roof runoff failed to meet drinking water guidelines. These inconsistent
findings stress the need to determine the minimum acceptable threshold for first flush and the need for roof runoff quality
testing to safeguard public health.
The first flush occurrence is influenced by various factors among them roof type and its imperviousness (Stenstrom and
Kayhanian, 2005). Compared to pervious surfaces, highly impervious surfaces create velocities that easily scour adsorbed
contaminants from surfaces and easily transport them almost immediately (Stenstrom and Kayhanian, 2005). These first
flush values have given rise to runoff of varying quality; in some cases the quality parameters were within regulatory guide-
lines for potable water, while in other cases the quality was above the regulatory guidelines. There are no widely accepted
principles for the determination of the first flush. Key determining factor on the amount of first flush commonly cited in lit-
erature include; (1) the intended purpose or use of the water collected, and (2) the treatment processes that will be carried
out on the harvested water to minimize public and health risks of using such water. For example, Gikas and Tsihrintzis
(2012) showed that first flush of 0.11–0.13 mm was not adequate to significantly prevent microbial contamination thus
the harvested water violated portable water quality standards, but such water could be useful for other non-potable uses.
The amount of rainfall anticipated for the season also influences the amount of water that can be diverted as first flush.
For example, if a dry season is anticipated then there is increased pressure to harvest as much water from the little rainfall
which could mean reducing the amount of water diverted as first flush.

5. Public health risk and mitigation measures

5.1. Contaminants of public health concern

Rainfall water provides the cleanest naturally occurring water supplement for households and institutions especially in
developing countries. It is generally believed to be chemically and biologically low-risk water (Hollander et al., 1996;
Thomas and Martinson, 2007). Rainfall water harvesting is practised in areas where surface water and groundwater are
not readily available or are of poor water quality. Roof rainwater has received a significant attention as a potential alternative
source of potable and non-potable water where water is scarce (Dillaha and Zolan, 1985; Hollander et al., 1996; Meera and
114 W. Gwenzi et al. / Sustainability of Water Quality and Ecology 6 (2015) 107–118

Ahammed, 2006; Ahmed et al., 2010b). The general notion is that roof harvested rainwater is safe to drink and this is
supported by limited epidemiological evidence (Gould and McPherson, 1987; Heyworth et al., 2006).
During water harvesting, precipitation landing on roof is conveyed through gutters or pipes to a storage tank. This water
is mainly used for domestic purposes and watering animals at household level. A lot of attention is being focused on rainfall
water harvesting systems and this has resulted in a decrease in the confidence in quality of harvested rainwater. Good qual-
ity water is free from disease causing organisms, harmful chemical substances and radioactive material. It should also be
aesthetically appealing and free from objectionable colour or odour (Issaka, 2011).
Rainwater tanks can provide an excellent habitat for mosquito breeding. Some mosquito types can be vectors of
arboviruses for example the mosquito species for dengue virus found in the tropics and subtropics (Commonwealth
Australia, 2010). Rainwater tanks are potential breeding site for the dengue virus vectors and therefore it is recommended
they be screened (WHO, 1997). Hanna et al. (1998) reported an outbreak of Dengue in the Torres Straits Islands in 1996–
1997 while a survey by Ritchie et al. (2002) conducted in the same islands, and revealed that the adult mosquitos were
present in rainwater tanks that had no screens or had damaged screens.
Dust on roof tops coming from industrial areas may contain heavy metals which find their way into storage tanks during
rainwater collection. These have negative effects on human and animal health. Heavy metals combine with body bio-mole-
cules like proteins and enzymes mutilating their structure and hindering their biological functions (Duruibe et al., 2007).
Ingestion of cadmium, lead, zinc, copper and aluminium can lead to gastrointestinal disorders, diarrhoea, stomatitis, tremor
and haemoglobinuria (Duruibe et al., 2007). Salem et al. (2000) reported a strong relation between heavy metal contaminat-
ed water and chronic diseases such as renal failure, liver cirrhosis, hair loss and chronic anaemia.
Generally industrial and traffic emissions are unlikely to cause significant impacts on rainwater collected in tanks as com-
pared to surface water (Commonwealth Australia, 2010). However, in other studies urban emissions had no significant effect
on lead concentrations in rainwater (Fuller et al., 1981; Coombes, 2002). In Adelaide, the testing of rainwater from household
tanks near industrial precincts was undertaken as part of two investigations into the impacts of contaminants associated
with local emissions. The concentrations of Pb, Mn, Ni, Zn and hydrocarbon in rainwater samples were consistently below
the guideline values for drinking water (Commonwealth Australia, 2010). Investigating the concentration of Pb in rainwater
from 13 sites, Huston et al. (2012) concluded that high concentrations of Pb in rainwater originated predominantly from roof
materials, and to a lesser extent, atmospheric deposition.
High nitrate (NO3 ) levels exceeding 50 mg L 1 NO3 or 10 mg L 1 as NO3 -N in drinking water causes methaemoglobi-
naemia or blue-baby syndrome in newly born babies (WHO, 2011). Nitrates occur in roof water storage facilities via dust
and atmospheric deposition on roof surfaces. Once in the human body, nitrates are reduced to nitrite, which is toxic. Nitrites
react with haemoglobin to form methaemoglobin, resulting in asphyxia (Knepp and Arkin, 1973). Although cases of
methaemoglobinaemia have been reported due to consumption of groundwater with high nitrate concentrations, no such
cases have been attributed to consumption of contaminated roof runoff. This possibly suggests that although nitrate con-
tamination of runoff may occur, the concentrations are below the guideline limits causing methaemoglobinaemia.
Table SM2 in Supplementary material presents an overview of the potential risk of roof runoff contamination by various
sources.
Anecdotal evidence in sub-Saharan Africa and a few studies have concluded that roof rainwater is chemically and micro-
biologically safe for human consumption (e.g., Dillaha and Zolan, 1985; Gould and McPherson, 1987; Rodrigo et al., 2007).
For example, results of bacteriological analyses conducted on water samples from 13 roof tanks in Botswana showed that
rainwater collected from corrugated iron roofs and stored in covered tanks was of high quality and safe for drinking com-
pared with traditional water sources (Gould and McPherson, 1987). By contrast, runoff from ground catchment systems
showed microbial contamination and posed a serious health hazard. However, several recent studies drawn from various
countries including Australia (Ahmed et al., 2010a,b, 2011a,b, 2012a,b, 2014; Lye, 2014), South Africa (Dobrowsky et al.,
2014a,b), India (Jesmi et al., 2014) and Brazil (Alves et al., 2014) have reported the microbiological contamination of roof
rainwater exceeding WHO guidelines for drinking water. For example, polymerase chain reaction (PCR) assays of samples
from 29 rainwater tanks in South Africa detected both indigenous and pathogenic organisms from eight genera (Klebsiela,
Enterobacter, Legionella, Salmonella, Shigella, Yersinia and Giardia) (Dobrowsky et al., 2014a). Microbiological contamination
was attributed to animal droppings, gardens and dust and lack of first flush diversion devices (Dobrowsky et al., 2014a,b).
Several studies conducted on microbiological quality of rainwater in Australia showed the prevalence of pathogenic
organisms (Simmons et al., 2001; Lye, 2002; Ahmed et al., 2010a,b, 2011a,b, 2012a,b, 2014). The pathogens including
Aeromonas spp., Salmonella spp, Giardia spp, Campylobacter spp, Cryptosporidium spp and virulence genes in E. coli and Ente-
rococcus spp. Cryptosporidium spp may cause gastrointestinal illness in humans with nausea, vomiting and or diarrhoea
occurring after ingestion of contaminated water, while Legionella pneumophila may cause respiratory infection, pneumonia
and fatality rate can be 50% in immune-compromised patients (Yu, 2000). These pathogens generally originate from human
and animal wastes that contaminate rain water during catchment and conveyance. Ahmed et al. (2010a) reported the pres-
ence of pathogens in approximately 5% of the samples, in a study in South East Queensland where 119 water tanks were
sampled. In the same study, Ahmed et al. (2010a) reported that the risk of infection from ingestion of Salmonella spp. and
G. Lamblia via drinking undisinfected rainwater is high. Consequently, gastrointestinal diseases such as Salmonellosis and
Giardiasis are expected to be high, with infection incidences ranging from 5.0  100 to 2.8  101 and 1.0  101 to
6.4  101 cases per 10,000 urban persons per year for Salmonellosis and Giardiasis, respectively (Ahmed et al., 2010a).
Besides consumption of microbially contaminated rainwater, studies conducted in Netherlands also reported public health
 W. Gwenzi et al. / Sustainability of Water Quality and Ecology 6 (2015) 107–118 115

risks associated with inhalation and ingestion of pathogenic organisms (Legionella pneumophila and Campylobacter jejuni) in
rainwater used in splash parks (De Man et al., 2014a,b).
Overall, except for a few earlier studies (e.g., Dillaha and Zolan, 1985; Rodrigo et al., 2007), the bulk of recent studies
showed a high risk of microbial contamination of rainwater (e.g., Ahmed et al., 2011a,b, 2012a,b, 2014; Lim and Jiang,
2013; Dobrowsky et al., 2014a,b). The conflicting conclusions between earlier and recent studies suggest that rainwater
quality varies among sites and could also partly reflect differences in parameters analyzed as indicators of microbiological
contamination and assaying techniques. For example, earlier studies (e.g., Dillaha and Zolan, 1985) based on traditional cul-
turing techniques largely relied on faecal bacteria such as E. coli and enterococci as proxy indicators of faecal pollution or
presence of pathogens. Although high levels of faecal coliforms (>1000 E. coli per 100 ml) is linked to incidences of diarrhoea
in children (Issaka, 2011), other studies have shown that faecal bacteria has poor correlation with the presence of pathogenic
microorganisms especially viruses and protozoa (Ahmed et al., 2009, 2010b). Traditional cultural techniques are generally
laborious and costly have been criticised for their limitations including underestimation of the bacterial concentration
due to the presence of injured or stressed cells, while certain pathogens can be viable but not cultivable (Ahmed et al.,
2009, 2010a). To overcome this drawback, recent advances in molecular techniques such as polymerase chain reaction
(PCR) amplification of 16s RNA of isolates (e.g., Dobrowsky et al., 2014a,b) enables direct, rapid, specific and sensitive
detection of many pathogens (Ahmed et al., 2009; Dobrowsky et al., 2014a,b). Moreover, PCR enables the quantification
of pathogens that are otherwise difficult and/or laborious to culture using traditional culturing methods.
The foregoing discussion evidently suggests that the potential risk of human poisoning and fatalities due to consumption
of contaminated rainwater exists. The risk is particularly high in developing countries where rainwater is often consumed
with minimum or no treatment. In general, there is a paucity epidemiological data directly linking disease outbreaks and
fatalities to consumption of rainwater except for case studies from Australia (Lye, 2002; Ahmed et al., 2010a; Lye, 2014).
It is also noteworthy that the bulk of the studies documenting microbial contamination are drawn from developing coun-
tries, while cases studies from water-limited sub-Saharan Africa are still limited except for South Africa. This is largely
because research on rainwater quality, epidemiology and their relationship is generally lacking in most developing countries.
Moreover, due to high disease burden in most developing countries, even when data are available, the relationship between
rainwater quality and epidemiological data is confounded by several factors. As Heyworth et al. (2006) pointed out, there is
need for studies to assess the risk factors for illness which are directly related to consumption of rainwater. Currently, the
biggest challenge in utilizing rainwater is the lack of local guidelines for specifying rain water quality. In most countries,
general lack of experience and technical knowledge have often led to reluctance by local agencies to approve plans such
as rainwater collection systems for buildings for community with access to centralized water treatment systems (Lye, 2014).

5.2. Strategies to safeguard roof water quality and public health

Understanding the factors controlling rainwater quality and contamination pathways is critical for minimization of rain-
water contamination and safeguarding public health. Measures to minimize rainwater contamination and safeguard public
health can be categorized into three broad categories; engineering/technological interventions, public education and house-
keeping. Consequently, devices that divert the first few mm of rainfall from entering the storage tanks have been designed to
reduce roof runoff water contamination. Despite significant reduction of certain contaminants through use of first flush
devices, certain contaminants may still occur in significantly high quantities even after diversion of first flush water. There

Sanitary Inspection Risk Score

(Susceptibility of rainwater to
contamination by human and animal
faeces)

0-2 3-5 6-8 9-10

(counts/100 ml)

<1
Classification
E. Coli

1-10

11-100

>100

Key:
Low Risk: Intermediate risk: High Risk: Very High Risk:
No action Low action High action Urgent action
required priority priority required

Fig. 2. A sanitary inspection risk assessment framework for prioritizing control/remedial actions of household roof water harvesting systems based on a
grading system of laboratory analytical results (E. coli counts) and sanitary inspection risk score (Gwenzi et al., 2013 adapted from WHO, 2011).
116 W. Gwenzi et al. / Sustainability of Water Quality and Ecology 6 (2015) 107–118

are strategies required to further safeguard public health risks associated with consumption of contaminated roof runoff.
Engineering and technological measures include the selection of appropriate of roof materials, roof designs and first flush
devices, and use of technological interventions including use of water treatments systems (e.g., chlorination, filtration) to
remove specific contaminants. Public education entails raising awareness of the potential public health risks associated with
roof runoff and possible mitigation measures among communities relying on roof water harvesting for potable uses.
Housekeeping issues include regular maintenance and cleaning of various components of roof water harvesting systems
include roofs, conveyance and storage facilities and prevention of contamination at the point of use. Housekeeping may
entail a detailed sanitary inspection of household roof water harvesting systems to ascertain possible sources and mechan-
isms of contamination to derive a sanitary inspection risk score (SIRS) (WHO, 2011). Combined with laboratory analytical
results, SIRS enables prioritization/targeting of mitigation/control measures. Fig. 2 presents a risks assessment framework
on how the laboratory analytical results (e.g., E. coli counts) and the SIRS for a roof water harvesting system can be used
to plan and prioritize a control program when water fails to meet WHO drinking water guidelines (WHO, 2011). Effective
mitigation of roof water contamination and protection of public health will entail using a combination of measures available
to users. Table SM3 in Supplementary material summarises some of the measures for minimizing contamination of roof
water to safeguard public health.

6. Conclusions and future research

The current review highlighted the contamination pathways and determinants of rainwater quality and the associated
public health risks and control measures. The review pointed out the risk of rainwater contamination by physico-chemical
and microbial contaminants from source to point of use. Contamination mechanisms include air-borne/atmospheric depo-
sition, leaching and weathering of roof materials, faecal contamination by animals and humans. Weather conditions, land
use practices, roof materials, temporal patterns of hydrological factors, and their interactions are key determinants of rain-
water quality. Although several studies have documented rainwater contamination, detailed epidemiological studies linking
disease outbreaks and fatalities to consumption of rainwater are still limited. Exceptions are two studies conducted in Aus-
tralia that directly linked outbreak of diarrheal diseases to consumption of rainwater from residential roof water harvesting
systems. The bulk of the literature used in the current study was drawn from industrialized countries except a few from
Nigeria and South Africa. By comparison there is limited data on roof water quality from developing regions such as sub-
Saharan Africa. Limited capacity of city authorities to provide reliable water implies that households in most developing
countries will continue to rely on roof water for domestic supplies. Moreover in such regions, communities are most likely
to use untreated roof rainwater for drinking purposes due to highly centralized water treatment systems. Moreover, while
literature covers a diverse range of root materials commonly used in developed countries, no roof water quality data were
available on other traditional roofing materials such as asbestos and grass thatch, plastics and asbestos sheets which are still
common among poor households in both urban and rural communities in developing countries. Therefore future research in
developing countries should focus on identifying nature and sources of contaminants, evaluating effect of different roofing
materials and land use practices on rainwater quality. The link between roof water quality from various roof materials and
disease outbreaks also warrants investigation. In developed countries attention should focus on emerging contaminants such
as PAHs, air-borne carcinogens and emerging contaminants such as endocrine-disrupting compounds and nanoparticles
associated with recent advances in pharmaceuticals and nanotechnology.

Acknowledgement

We thank our respective employers for providing in-kind support during the study. WG originated the idea, designed and
implemented the research, and wrote the manuscript. All other authors contributed equally in design and implementation of
research, and compilation of manuscript. All authors have approved the manuscript for submission. We are grateful to the
Editor-in-Chief and the two anonymous reviewers whose insightful comments improved the overall quality of the
manuscript.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.swaqe.
2015.01.006.

References

Abbas, M.S., Bruns, R.E., Scarminio, I.S., Ferreira, J., 1993. A multivariate statistical analysis of the composition of rainwater near Cubatao, SP, Brazil. Environ.
Pollut. 79 (3), 225–233.
Abbasi, T., Abbasi, S.A., 2011. Sources of pollution in rooftop rainwater harvesting systems and their control. Crit. Rev. Environ. Sci. Technol. 41, 2097–2167.
Adeniyi, I.F., Olabanji, I.O., 2005. The physico-chemical and bacteriological quality of rainwater collected over different roofing materials in Ile-Ife,
southwestern Nigeria. Chem. Ecol. 21, 149–166.
Ahmed, W., Brandes, H., Gyawali, P., Sidhu, J.P., Toze, S., 2014. Opportunistic pathogens in roof-captured rainwater samples, determined using quantitative
PCR. Water Res. 15 (53), 361–369. http://dx.doi.org/10.1016/j.watres.2013.12.021.
 W. Gwenzi et al. / Sustainability of Water Quality and Ecology 6 (2015) 107–118 117

Ahmed, W., Gardner, T., Toze, S., 2011a. Microbiological quality of roof-harvested rainwater and health risks: a review. J. Environ. Qual. 40 (1), 13–21.
Ahmed, W., Goonetilleke, A., Gardner, T., 2010a. Implications of faecal indicator bacteria for the microbiological assessment of roof-harvested rainwater
quality in Southeast Queensland, Australia. Can. J. Microbiol. 56, 471–479.
Ahmed, W., Hodgers, L., Masters, N., Sidhu, J.P., Katouli, M., Toze, S., 2011b. Occurrence of intestinal and extraintestinal virulence genes in Escherichia coli
isolates from rainwater tanks in Southeast Queensland, Australia. Appl. Environ. Microbiol. 77 (20), 7394–7400. http://dx.doi.org/10.1128/AEM.06047-11.
Ahmed, W., Sidhu, J.P.S., Toze, S., 2012a. An attempt to identify the likely sources of Escherichia coli harbouring toxin genes in rainwater tanks. Environ. Sci.
Technol. 46, 5193–5197.
Ahmed, W., Sidhu, J.P., Toze, S., 2012b. Speciation and frequency of virulence genes of Enterococcus spp. isolated from rainwater tank samples in Southeast
Queensland, Australia. Environ. Sci. Technol. 46 (12), 6843–6850. http://dx.doi.org/10.1021/es300595g.
Ahmed, W., Vieritz, A., Gardner, T., Goonetilleke, A., 2009. Microbial risks from rainwater tanks in Southeast Queensland. Water 36 (8), 80–85.
Ahmed, W., Vieritz, A., Goonetilleke, A., Gardner, T., 2010b. Health risk from the use of roof-harvested rainwater as potable or nonpotable water, determined
using quantitative microbial risk assessment in Southeast Queensland, Australia. Appl. Environ. Microbiol. 76 (22), 7382–7391.
Alves, F., Köchling, T., Luz, J., Santos, S.M., Gavazza, S., 2014. Water quality and microbial diversity in cisterns from semiarid areas in Brazil. J. Water Health
12 (3), 513–525. http://dx.doi.org/10.2166/wh.2014.139.
Baguma, D., Loisk, I.W., Jung, H., 2010. Water management, rainwater harvesting and predictive variables in rural households. Water Resour. Manag. 24 (13),
3333–3348.
Bangira, C., Manyevere, A., Mubvumba, P., 2007. The spatial and temporal variation of pH and lead in rain water in Harare city, Zimbabwe. J. Sustainable Dev.
Africa 9 (1), 1–19.
Barron, J. (Ed.), 2009. Rainwater harvesting: a lifeline for human well-being. A Report Prepared for UNEP by Stockholm Environment Institute. UNEP/SEI,
Nairobi, p. 80.
Chang, M., McBroom, M.W., Beasley, R.S., 2004. Roofing as a source of nonpoint water pollution. J. Environ. Manage. 73, 307–315.
Coombes, P., 2002. Rainwater tanks revisited: new opportunities for urban water cycle management. Dissertation, University of Newcastle.
Commonwealth of Australia, 2010. Guidance on use of rainwater tanks. Commonwealth of Australia, Canbera.
Corden, J.M., Millington, W.M., 2001. The long term trends and seasonal variation of the aeroallergen Alternaria in Derby, UK. Aerobiologia 17, 127–136.
De Man, H., Bouwknegt, M., van Heijnsbergen, E., Leenen, E.J., van Knapen, F., de Roda Husman, A.M., 2014a. Health risk assessment for splash parks that use
rainwater as source water. Water Res. 1 (54), 254–261. http://dx.doi.org/10.1016/j.watres.2014.02.010.
De Man, H., Leenen, E.J., van Knapen, F., de Roda Husman, A.M., 2014b. Risk factors and monitoring for water quality to determine best management
practices for splash parks. J. Water Health 12 (3), 399–403. http://dx.doi.org/10.2166/wh.2014.127.
Dile, Y.T., Karlberg, L., Temesgen, M., Rockström, J., 2013. The role of water harvesting to achieve sustainable agricultural intensification and resilience
against water related shocks in sub-Saharan Africa. Agric. Ecosyst. Environ. 181, 69–79.
Dillaha, T.A., Zolan, W.J., 1985. Rainwater catchment water quality in Micronesia. Water Res. 19, 741–746.
Dirwai, T., 2014. Quantifying roof runoff and its chemical quality in a suburban area in Harare, Zimbabwe. Dissertation, University of Zimbabwe.
Dobrowsky, P.H., De Kwaadsteniet, M., Cloete, T.E., Khan, W., 2014a. Distribution of indigenous bacterial pathogens and potential pathogens associated with
roof-harvested rainwater. Appl. Environ. Microbiol. 80 (7), 2307–2316. http://dx.doi.org/10.1128/AEM.04130-13.
Dobrowsky, P.H., van Deventer, A., De Kwaadsteniet, M., Ndlovu, T., Khan, S., Cloete, T.E., Khan, W., 2014b. Prevalence of virulence genes associated with
pathogenic Escherichia coli strains isolated from domestically harvested rainwater during low- and high-rainfall periods. Appl. Environ. Microbiol. 80
(5), 1633–1638. http://dx.doi.org/10.1128/AEM.03061-13.
Duruibe, J.O., Ogwuegbu, M.O.C., Egwurugwu, J.N., 2007. Heavy metal pollution and human biotoxic effects. Int. J. Phys. Sci. 2 (5), 112–118.
Egodawatta, P., Evan, T., Ashantha, G., 2009. Understanding the physical processes of pollutant build-up and wash-off on roof surfaces. Sci. Total Environ.
407 (6), 1834–1841.
Evans, C.A., Coombes, P.J., Dunstan, R.H., 2006. Wind, rain and bacteria: the effect of weather on the microbial composition of roof-harvested rainwater.
Water Res. 40, 37–44.
Farreny, R., Morales-Pinzon, T., Guisasola, A., Taya, C., Rieradevall, J., Gabarrell, X., 2011. Roof selection for rainwater harvesting: quantity and quality
assessments in Spain. Water Res. 45, 3245–3254.
Förster, J., 1996. Patterns of roof runoff contamination and their potential implications on practice and regulation of treatment and local infiltration. Water
Sci. Technol. 33, 39–48.
Förster, J., 1998. The influence of location and season on the concentrations of macroions and organic trace pollutants in roof runoff. Water Sci. Technol. 38,
83–90.
Förster, J., 1999. Variability of roof runoff quality. Water Sci. Technol. 39, 137–144.
Fuller, C.O., Martin, T.J., Walters, R.P., 1981. Quality aspects of water stored in domestic rainwater tanks (a preliminary study). Engineering and Water
Supply Department, South Australia.
Gao, L., 2008. Potential environmental implications of manufactured nanomaterials: toxicity, mobility and nanowastes in aquatic and soil systems. PhD
Dissertation, University of Florida, Florida, USA.
Garnaud, S., Mouchel, J.-M., Chebbo, G., Thevenot, D.R., 1999. Heavy metal concentrations in dry and wet atmospheric deposits in Paris district: comparison
with urban run-off. Sci Total Environ. 235, 235–245.
Gikas, G.D., Tsihrintzis, V.A., 2012. Assessment of water quality of first-flush roof runoff and harvested rainwater. J. Hydrol. 466, 115–126.
Gould, J., McPherson, H., 1987. Bacteriological quality of rainwater in roof and ground catchment systems in Botswana. Water Int. 12 (3), 135–138.
Gwenzi, W., Nyamadzawo, G., 2014. Hydrological impacts of urbanization and urban roof water harvesting in water-limited catchments: a review. Environ.
Process. http://dx.doi.org/10.1007/s40710-014-0037-3.
Gwenzi, W., Manjengwa, F., Takawira, A., 2013. Chivi water, hygiene and sanitation borehole drinking water quality. A Technical report submitted to
Zimbabwe Red Cross Society, Harare, Zimbabwe. pp. 26.
Haggard, B.E., Bartsch, L.D., 2009. Net changes in antibiotic concentrations downstream from an effluent discharge. J. Environ. Qual. 38, 343–352.
Handia, L., Tembo, J.M., Mwiindwa, C., 2003. Potential of rainwater harvesting in urban Zambia. Phys. Chem. Earth 28 (20), 893–896.
Hanna, J.N., Ritchie, S.A., Merrit, A.D., van den Hurk, A.F., Phillips, D.A., Serafin, I.L., Norton, R.E., McBride, W.J.H., Gleeson, F.V., Poidinger, M., 1998. Two
contiguous outbreaks of dengue type 2 in north Queensland. Med. J. Aust. 168, 221–225.
Heberer, T., 2002. Occurrence, fate, and removal of pharmaceutical residues in the aquatic environment: a review of recent research data. Toxicol. Lett. 131,
5–17.
Heyworth, J.S., Glonek, G., Maynard, E.J., Baghurt, P.A., Finlay-Jones, J., 2006. Consumption of untreated tank rainwater and gastroenteritis among young
children in South Australia. Int. J. Epidemiol. 35 (4), 1051–1058.
Hoffman, E.J., Latimer, J., Mills, G.L., Quinn, J.G., 1982. Petroleum hydrocarbons in urban runoff from a commercial land use area. J. Water Pollut. Cont. Fed. 5
(11), 1517–1525.
Hollander, R., Bullermann, M., Grob, C., Hartung, H., König, K., Lücke, F.-K., Nolde, E., 1996. Microbiological and hygienic aspects of the use of rainwater as
process water for toilet flushing, garden irrigation and laundering. Gesundheitswesen 58, 288–293 (in German).
Hranova, R. (Ed.), 2006. Diffuse pollution of water resources: principles and case studies in the southern African region. Taylor and Francis/Balkema, London.
Huston, R., Chan, Y.C., Chapman, H., Gardner, T., Shaw, G., 2012. Source apportionment of heavy metals and ionic contaminants in rainwater tanks in a
subtropical urban area in Australia. Water Res. 46 (4), 1121–1132.
Issaka, Z., 2011 Appropriate rainwater harvesting and domestic water quality a case study of central Gonja district. MSc thesis. Faculty of Mechanical and
Agricultural Engineering, Kwame Nkrumah University of Science and Technology, Ghana.
Jato, M.V., Rodriguez, F.J., Seijo, M.C., 2000. Pinus pollen in the atmosphere of Vigo and its relationship to meteorological factors. Int. J. Biometeorol. 43, 147–153.
118 W. Gwenzi et al. / Sustainability of Water Quality and Ecology 6 (2015) 107–118

Jesmi, Y., Rahiman, K.M., Hatha, A.A., Deepu, L., Jyothi, S., 2014. Risk assessment of rooftop-collected rainwater for individual household and community use
in central Kerala, India. J. Environ. Health 76 (6), 114–121.
Jebamalar, A., Ravikumar, G., 2011. A comparative analysis of hydrologic responses to rainwater harvesting - a case study. Indian J. Sci. Technol. 4 (1), 34–39.
Jiries, A., Hussein, H., Halaseh, Z., 2002. The quality of water and sediments of street runoff in Amman, Jordan. Hydrol. Process. 15 (5), 815–824.
Jones, A.M., Harrison, R.M., 2004. The effects of meteorological factors on atmospheric bioaerosol concentrations-a review. Sci. Total Environ. 326, 151–180.
King, T.L., Bedient, P.B., 1982. Effect of acid rain upon cistern water quality. In: Proceedings of an International Conference on Rainwater Cistern Systems.
University of Hawaii at Manoa, pp. 244–248.
Knepp, H., Arkin, S., 1973. Ammonia toxicity levels and nitrate tolerance of channel catfish, the progressive fish. Culturist 35 (4), 221–224.
Lim, K.Y., Jiang, S.C., 2013. Reevaluation of health risk benchmark for sustainable water practice through risk analysis of rooftop-harvested rainwater. Water
Res. 47 (20), 7273–7286.
Lye, D.J., 2002. Health risks associated with consumption of untreated water from household roof catchment systems. J. Am. Water Resour. Assoc. 38 (5),
1301–1306.
Lye, D., 2014. , Rooftop Runoff as a Source of Contamination: A Review. USEPA. 2014 Rainwater Resources. http://www.rainwaterresources.com/rooftop-
runoff-source-contamination-review. Accessed 6 October 2014.
Lynch, D.F., Dietsch, D.K., 2010. Water efficiency measures at Emory University. J. Green Building 5 (2), 41–54.
Mamuse, A., Sibanda, F., Watkins, R.T., Nyatsanza, F.F., Makwarimba, F.W., 2003. Prevalence and causes of dental fluorosis among school children in Gokwe
North District, Zimbabwe. Department of Applied Geology, Curtin University of Technology, Perth.
Meera, V., Ahammed, M.M., 2011. Factors affecting the quality of roof-harvested rainwater. In: Proceedings of the 12th International Conference on
Environmental Science and Technology. 8–10 September 2011, Rhodes, Greece.
Meera, V., Ahmeed, M., 2006. Water quality of rooftop rainwater harvesting systems: a review. J. Water Supply: Res. Technol. 55 (4), 257–268.
Melidis, P., Akratos, C.S., Tsihrintzis, V.A., Trikilidou, E., 2007. Characterization of rain and roof drainage water quality in Xanthi, Greece. Environ. Monit.
Assess. 127, 15–27.
Mendez, C.B., Afshar, B.R., Kinney, K.A., Barrett, M.E., Kirisits, M.J., 2010. Effect of roof material on water quality for rainwater harvesting systems. Texas
Water Development Board Report. Texas Water Development Board, Austin, Texas, p. 46.
Mendez, C.B., Klenzendorf, J.B., Afshar, B.R., Simmons, M.T., Barrett, M.E., Kinney, K.A., Kirisits, M.J., 2011. The effect of roofing material on the quality of
harvested rainwater. Water Res. 45, 2049–2059.
Mitchell, K., 2003. A study of roof runoff quality in Auckland New Zealand, implications for stormwater management. Prepared for Auckland Regional Council.
Mouli, P.C., Mohan, S.V., Reddy, S.J., 2005. Rainwater chemistry at a regional representative urban site: influence of terrestrial sources on ionic composition.
Atmos. Environ. 39, 999–1008.
Mwenge-Kahinda, J., Taigbenu, A.E., Boroto, R.J., 2007. Domestic rainwater harvesting to improve water supply in rural South Africa. Phys. Chem. Earth 32,
1050–1057.
Ngigi, S.N., Savenije, H.H., Gichuki, F.N., 2007. Land use changes and hydrological impacts related to up-scaling of rainwater harvesting and management in
upper Ewaso Ng’iro river basin, Kenya. Land Use Policy 24 (1), 129–140.
Ngigi, S.N., Savenije, H.H.G., Thome, J.N., Rockström, J., Penning de Vries, F.W.T., 2005. Agro-hydrological evaluation of on-farm rainwater storage systems
for supplemental irrigation in Laikipia district, Kenya. Agr. Water Manage. 73 (1), 21–41.
Radaideh, J., Al-Zboon, K., Al-Harahsheh, A., Al-Adamat, R., 2009. Quality assessment of harvested rainwater for domestic uses. Jordan J. Earth Environ. Sci. 2
(1), 26–31.
Ritchie, S., Montgomery, B., Walsh, I., 2002. Production of mosquitoes in rainwater tanks and wells on Yorke Island, Torres Strait: preliminary study. Environ.
Health 2 (3), 13.
Rodrigo, S; Sinclair, M; Cunliffe, D and Leder, K. A critical assessment of epidemiological studies for the investigation of the health risk of drinking untreated
rainwater [online]. In: Rainwater and Urban Design 2007. [Barton, A.C.T.]: Engineers Australia, 2007: [936]-[943]. Availability: http://search.
informit.com.au/documentSummary;dn=889900316059723;res=IELENG.
Rouvalis, A., Karadima, C., Zioris, I.V., Sakkas, V.A., Albanis, T., Iliopoulou-Georgudaki, J., 2009. Determination of pesticides and toxic potency of rainwater
samples in western Greece. Ecotoxicol. Environ. Saf. 72, 828–833.
Salem, H.M., Eweida, E.A., Farag, A., 2000. Heavy metal in drinking water and their environmental impact on human health. In: Proceedings of the
ICEHM2000 Conference, Cairo University, Egypt, September, 2000, pp. 542–556.
Sazakli, E., Alexopoulos, A., Leotsinidis, M., 2007. Rainwater harvesting, quality assessment and utilization in Kefalonia Island, Greece. Water Res. 41, 2039–
2047.
Simmons, G., Hope, V., Lewis, G., Whitmore, J., Gao, W.Z., 2001. Contamination of potable roof-collected rainwater in Auckland, New Zealand. Water Res. 36,
1518–1524.
Sivanappan, R.K., 1997. State of the art in the area of water harvesting in semi arid parts of the world. Presented at the Workshop on Water Harvesting for
Supplementary Irrigation of Staple Food Crops in Rainfed Agriculture, Stockholm University, Stockholm, Sweden.
Stenstrom, M.K., Kayhanian, M., 2005. First flush phenomenon characterization. California department of transportation. Division of environmental analysis,
Sacramento, CA, p. 69.
Teemusk, A., Mander, U., 2007. Rainwater runoff quantity and quality performance from greenroof: The effects of short-term events. Ecol. Eng. 30 (3–2),
271–277.
Thomas, T.H., Martinson, D.B., 2007. Roofwater harvesting: a handbook for practitioners. In IRC Technical paper series (No. 49). IRC.
Tobaiwa, C., Musiyambiri, M., Chironga, L., Mazorodze, O., Sapahla, S., 1991. Fluoride levels and dental fluorosis in two districts in Zimbabwe. Cent. Afr. J.
Med. 7 (11), 353–361.
Uygur, N., Karaca, F., Alagha, O., 2010. Prediction of sources of metal pollution in rainwater in Istanbul, Turkey using factor analysis and long-range transport
models. Atmos. Res. 95 (1), 55–64.
van de Voorde, A., Ghassan, C., Marie-Christine, G., 2009. Rainwater harvesting: what are the potential effects of roof maintenance on runoff quality? France
as an example. Daniel Thevenot. In: 9th World Wide Workshop for Young Environmental Scientists WWW-YES-Brazil-2009: Urban Waters: Resource or
Risks? October 2009, Belo.
Van Metre, P.C., Mahler, B.J., 2003. The contribution of particles washed from rooftops to contaminant loading to urban streams. Chemosphere 52, 11127–
11741.
Vazquez, A., Cotstoya, M., Pena, R.M., Garca, S., Herrero, C., 2003. A rainwater quality monitoring study of the composition of rainwater in Galicia.
Chemosphere 51 (5), 375–385.
Ward, S., Memon, F.A., Butler, D., 2012. Performance of a large building rainwater harvesting system. Water Res. 46 (16), 5127–5134.
WHO, 1997. Dengue haemorrhagic fever: diagnosis, treatment, prevention and control, control, second ed. World Health Organization, Geneva.
WHO, 2011. Guidelines for drinking-water quality. Available from: http://www.who.int/entity/water_sanitation_health/publications/2011/dwq_
guidelines/en/index.html (accessed June 2014). .
Yaziz, M.I., Gunting, H., Sapari, N., Ghazali, A.W., 1989. Variations in rainwater quality from roof catchments. Water Res. 23, 761–765.
Yu, V.L., 2000. Legionella pneumophila, . fifth ed. In: Mandell, G.L., Bennett, J.E., Dolin, R. (Eds.), Principles and practice of infectious diseases fifth ed., vol. 2
Churchill Livingstone, Philadelphia, PA, pp. 2424–2435.
Zhong, Z.C., Victor, T., Balasubramanian, R., 2001. Measurement of major organic acids in rainwater in Southeast Asia during burning and non-burning
periods. Water Air Soil Pollut. 130 (1–4), 457–462.