View source for OR/15/009 Review of existing studies relevant to Sierra Leone

__notoc__
{{OR/15/009}}
This section summaries the current evidence base regarding: the existing hydrogeological understanding in humid tropics, particularly regarding rapid vertical and lateral pathways; water quality issues from different groundwater sources, issues of seasonality and impacts from sanitary sources; pathogen survival in humid tropical regions.

==Hydrogeology==
===Overview of the hydrogeology of Sierra Leone===
The main hydrogeological environments of Sierra Leone are summarised in Table 3. There will be a wide variation in the properties of the aquifers within each of the major hydrogeological units, however, the main distinction is between the relatively low permeabilities of the old, hard rocks of the Precambrian Basement Complex, Saionya Scarp/Rokel River Group and Ultrabasic intrusives on the one hand, and the higher permeability and storage of the Bullom Group sands. Within the Precambrian Basement, flow is through fractures giving rapid connections over 10s of metres.

The weathered basement rocks form the most widespread and important aquifer across Sierra Leone, The weathered zone is derived from the underlying parent rock formations, under intense rainfall and large seasonal groundwater table variations. The resulting thick tropical soils form an important part of both the unsaturated zone and shallow aquifers (Akiwumi, 1987<ref name="AKIWUMI">AKIWUMI, F A. 1987. Groundwater exploration and development in the crystalline Basement rocks of Sierra Leone. Source unknown. http://www.bgs.ac.uk/africagroundwateratlas/fulldetails.cfm?id=SL4002</ref>; UN, 1988<ref name="UN">UN. 1988. Groundwater in North and West Africa. Natural Resources / Water Series No. 18. Department of Technical Co-operation for Development and Economic Commission for Africa. ST/TCD/5. </ref>). Investigations of weathered basement aquifers elsewhere have highlighted the importance of flow paths at the base of the weathered zone where groundwater flows primarily through fractures associated with the stone line at the base of the collapse zone or the basal breccia (Wright and Burgess, 1992<ref name="Wright 1992">WRIGHT, E P and BURGESS, W G (eds). 1992. Hydrogeology of crystalline Basement aquifers in Africa. Geological Society Special Publication No. 66. </ref>; Foster and Chilton 1993<ref name="Foster">FOSTER S S D. 1993. Groundwater conditions and problems characteristic of the humid tropics. Hydrology of Warm Humid Regions (Proceedings of the Yokohama Symposium, July 1993). lAHS Publication No. 216, 1993. https://www.researchgate.net/publication/255579127_Groundwater_conditions_and_problems_c haracteristic_of_the_humid_tropics</ref>). At depth, below the weathered zone, open fractures can be found associated with fault zones or the regional stress field.

Above the fracture zone at the base of the weathered zone, the weathered rock can often be clay rich with a high percentage of kaolinite clay and in certain circumstances other clay minerals such as smectite (Fookes 1997). The upper section of the weathered zone, above the clay rich kaolinite often comprises red layers of material from which the clays have been leached, leaving oxides of iron, aluminium and manganese. These can often be on the form of indurated layers or gravelly layers, and both can rapidly transmit water horizontally. For example Bonsor et al. (2014)<ref name="Bonsor">BONSOR, H C, MACDONALD, A M and DAVIES, J. 2014. Evidence for extreme variations in the permeability of laterite from a detailed analysis of well behaviour in Nigeria. Hydrological Processes 28, 3563–3573. </ref> found permeability values of >300 m/d in these shallow layers (2–4 m depth), several orders of magnitude greater than the permeability of the deeper layers.

In weathered crystalline basement, most sustainable groundwater sources tend to exploit groundwater in fractures at the base of the weathered zone. This can be in fractures 15 – 40 m depth, depending on the thickness of the weathered zone. The mean residence of time of groundwater within this zone has been measured as 30 – 60 years by Lapworth et al (2013)<ref name="Lapworth 2013">LAPWORTH, D J, MACDONALD, A M, TIJANI, M N, DARLING, W G, GOODDY, D C, BONSOR, H C, ARAGUÁS-ARAGUÁS, L J. 2013. Residence times of shallow groundwater in West Africa: implications for hydrogeology and resilience to future changes in climate. Hydrogeology Journal, 21, 3, 673-686. </ref> in similar hydrogeological and climatic environments in southern Nigeria. Shallower sources which only exploit groundwater within the upper weathered material are generally much less reliable, and tend to dry up rapidly when the rains stop and this permeable soil layer drains (Boiurgois et al., 2013<ref name"Boiurgois">BOIURGOIS, F, DE CAO, P, KORITEH ,Y, TRUAN, B, and REDON, P. 2013. Existing water access points in the districts of Bo Koinadugo and Tonkolili in Sierra Leone. ''Pro Victims'', Geneve. </ref>).

{| class="wikitable"
|+ Table 3 Hydrogeology of Sierra Leone’s main geological formations. Adapted from, Ministry of Water Resources (2015)
| '''Hydrogeological Unit (Approx % of Land Area)'''
| '''Sub-Units'''
| '''Aquifer description'''
| '''Well or Borehole depths (m)'''
| '''Well yields (L/s)'''
|-
| rowspan="3" | '''Precambrian Basement Complex (78%)'''
| Overlying valley fill deposits
| Sands, gravels and clays overlying the basement rocks, can be high permeability, flow is intergranular
| up to 15m
| Nd likely to  be 0.3 – 5 litres
|-
| Weathered zone
| Highly weathered rock often transformed to a thick tropical soil. Can have very high lateral permeability in shallow clay depleted layers
| up to 20m (max 37m)
| 0.3 – 1.5
|-
| Fractured crystalline bedrock
| At the base of the weathered zone where the bedrock is extensively fractured but not clay rich, and also in deeper fracture zones associated with faults
| 35m (average) 60m (max)
|0.3 – 1.5
|-
| rowspan="2" | '''Saionya Scarp/Rokel River Group (9%)'''
| Weathered layer
| Often clay
| nd
| nd
|-
| Fractured sediments
| Old sediments with little porosity, groundwater flows within fractures along old bedding plains
| nd
| nd
|-
| rowspan="2" | '''Bullom Group (12%)'''
| Unconsolidated sands and clays (inland alluvial & coastal)
| Groundwater flow is intergranular and storage can be high. Fracture flow is less common
| 10 to 20m
| up to 3
|-
| Interbedded sands and clays
|
| 30 – 80m
| up to 6
|-
| rowspan="2" | '''Ultrabasic Igneous Intrusives (1%)'''
| Fractured gabbros
| Groundwater flow within fractures, often does not have a thick weathered zone developed.
| nd
| nd
|-
| Weathered and fractured dolerites
|
| nd
| nd
|}


Sub-vertical features allowing rapid transit of water from the ground surface to groundwater in the shallow permeable layers in the tropical soil, and sometimes deeper to the permeable fracture zones towards the base of the weathered zone. These sub-vertical features provide vertical pathways and ''can include ''geological features, such as quartz veins or faults, or biological features such as tree roots and old termite tunnels, and anthropogenic features such as abandoned wells, unlined pit latrines or poorly constructed wells (Wright and Burgess 1992<ref name="Wright 1992">WRIGHT, E P and BURGESS, W G (eds). 1992. Hydrogeology of crystalline Basement aquifers in Africa. Geological Society Special Publication No. 66. </ref>, Hendricx and Flury 2001<ref name="Hendricx">HENDRICX, J M H and FLURY, M. 2001. Uniform and preferential flow mechanisms in the vadose zone. Chapter 5 in Conceptual Models of Flow and Transport in the Fractured Vadose Zone. Panel on Conceptual Models of Flow and Transport in the Fractured Vadose Zone, U.S. National Committee for Rock Mechanics, Board on Earth Sciences and Resources, National Research Council. National Academy Press, Washington DC, USA. http://www.nap.edu/catalog/10102.html</ref>). The unsaturated zone in Sierra Leone, which is often a useful buffer for reducing contaminant loads from ground surface to aquifer can therefore be easily bypassed.

Early experiments using lithium bromide tracers in weathered basement geology in Botswana to model the movement of contaminants from the shallow subsurface to a borehole showed rapid transport times and good recovery, implying fracture flow with little diffusion (Lewis et al., 1980<ref name="Lewis 1980">LEWIS, W J, FARR, J, and FOSTER, S S. 1980. The pollution hazard to village water supplies in eastern Botswana. ICE Proceedings, Ice Virtual Library, Vol. 69, 281-293. </ref>). Taylor et al. (2009)<ref name="Taylor">TAYLOR, R, TINDEMUGAYA, C, BARKER, J, MACDONALD, D M J, and KULABAKO, N R. 2009. Convergent radial tracing of viral and solute transport in Gneiss Saparolite. Groundwater, Vol. 48, 284-294. </ref> carried out a study in the weathered basement in Uganda using ''E.coli ''bacteriophage and forced gradient solute tracer experiments. The tracer was largely unrecovered, and rapid phage detection at the pumping well show that groundwater flow velocities exceed that of inert solutes and are consistent with statistically extreme flow pathways. This is consistent with size exclusion effects in colloid studies that show earlier arrival peaks for larger colloidal material compared to bulk solutes due to reduced matrix diffusion the and the development of preferential pathways (e.g. Sirivithayapakorn and Keller 2003<ref name="Sirivithayapakorn">SIRIVITHAYAPAKORN, S., AND A. KELLER. 2003. Transport of colloids in unsaturated porous media: A pore-scale observation of processes during the dissolution of air-water interface, Water Resour. Res., 39, 1346, doi:10.1029/2003WR002487, 12. </ref>; Lapworth et al., 2005<ref="Lapworth 2005"></ref>).

Figure 12 shows some groundwater level data collected during the Sierra Leone water security project, which has included high-resolution monitoring of water levels in hand-dug wells and boreholes in the middle Rokel river basin. Lateral movement of groundwater is determined by the combination of hydraulic gradient and permeability. Hydraulic gradients become significant in the rainy season as water tables respond to rainfall-recharge, and gradients towards zones of low relief become established.

[[Image:OR15009fig12.jpg|thumb|center|500px| '''Figure 12''' Groundwater levels for selected wells in Sierra Leone. Source, reproduced with permission from the Ministry of Water Resources (2015).]]

The graphs illustrates a number of key points:
* Water tables respond rapidly to the first rains in May;
* Water tables rise to a peak around mid- to end August, coinciding with the peak of the rains;
* Water tables recede rapidly after the peak rainfall month, despite the subsequent months having significant rainfall;
* Water tables continue to recede through the dry season, reaching their lowest levels in April.
* Average annual variation in 2013 recorded from 9 hand dug wells in the weathered basement was between 2.7-8 m, averaging 5.4 m

These observed data are consistent with the conceptual model of the hydrogeology of the weathered basement aquifer described above. Groundwater recharge is rapid during the rainy season often responding to individual rainfall events, which suggest the widespread existence of sub-vertical preferential flow pathways in the unsaturated zone. The high rate of discharge from the aquifer indicated by seasonal baseflow to the rivers, the drying up of many shallow wells and the relatively rapid decline of groundwater levels after rain can be explained by the existence of sub-horizontal preferential flow paths and zones of seasonally high permeability.

{|class="wikitable" style="width:700px" style="background:#f0f8ff; font-size:100%"
| '''Hydrogeology summary:'''<br>
Infiltration  from  the  ground  surface  to  the  sub-surface  is  likely  to  be  relatively  rapid, occurring within hours in some circumstances.<br><br>Sub-horizontal flow is likely to be rapid, as a consequence of the existence of preferential flow paths and low-permeability layers.<br><br>Groundwater in fractures below the weathered zone (typically 15 – 40 m thick) is likely to be less vulnerable to contamination due to the presence of the clay-rich layer. However, some fractures can rapidly transmit water rapidly through the hard rock.<br><br>The groundwater system can be conceptualised as two zones;
* A shallow (regolith) groundwater zone, accessed by dug wells and which is vulnerable to both diffuse surface and subsurface sources of contamination due to limited attenuation potential in the subsurface as well as rapid horizontal and vertical pathways which are likely to be regularly activated given the climate and geology of Sierra Leone.
* A deeper (hard rock) groundwater system with longer flowpaths and greater potential for natural attenuation of hazards which is accessed by boreholes.
|}

===Protection zones for groundwater points close to health centres ===
The evidence summarised above give rise to a number of questions which explore the feasibility of protecting groundwater consumers in Sierra Leone from faecal and other contamination.
# What role is there for water source protection zones?
# Is there a preference between different groundwater source types?
# Are there any specific design and construction considerations which could better protect water consumers?
# How relevant are these general considerations in the context of Ebola?

The first of these questions is the subject of this section while the second and third are addressed in the following section.  The specific case of Ebola is addressed in the report conclusions.

The use of source (water point) protection zones (SPZs) has become well-established in public water supply practice in the UK and other countries. Resources for the present study do not allow for a full review of the subject, but a limited summary of UK practice is presented here, mostly based on Environment Agency (2009)<ref>ENVIRONMENT AGENCY. 2009. Groundwater Source Protection Zones – Review of Methods. Integrated catchment science programme. Science report: SC070004/SR1. https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/290724/scho0309 bpsf-e-e.pdf</ref>.  In current practice, three zones are defined:
* '''SPZ1 – Inner Protection Zone '''is defined as the 50 day travel time from any point below the water table to the water point. This zone has a minimum radius of 50 metres.
* '''SPZ2 – Outer Protection Zone '''is defined by a 400 day travel time from a point below the water table. This zone has a minimum radius of 250 or 500 metres around the water point, depending on the size of the abstraction, except if the actual contaminant source has been verified.
* '''SPZ3 – Catchment Protection Zone '''is defined as the area around a source within which all groundwater recharge is presumed to be discharged at the source. In confined aquifers, the source catchment may be displaced some distance from the source. For heavily exploited aquifers, the final Source Catchment Protection Zone can be defined as the whole aquifer recharge area where the ratio of groundwater abstraction to aquifer recharge (average recharge multiplied by outcrop area) is >0.75. There is still the need to define individual source protection areas to assist operators in catchment management.

The determination of source protection zones is carried out through four steps:
# Data collation and conceptualisation
# Calculations, modelling and hydraulic capture zone production
# Technical review of hydraulic capture zones with modification, where appropriate, of the zone boundaries to produce the final SPZs
# Documentation and publication of final SPZs

The entire procedure is based on a number of key assumptions, which, while are realistic for high-income countries such as the UK, do not apply in countries such as Sierra Leone currently. These include:
* A mandate within a designated public sector institution to pursue this approach to water resource protection – in Sierra Leone the legal framework is yet to be approved, and the water resources department of the Ministry is in its infancy;
* The availability of trained and experienced scientists with the physical and financial resources to undertake the necessary data collection, conceptualisation, calculation, modelling and decision-making – these are not yet available in Sierra Leone;
* The existence of centralised data management systems – in Sierra Leone even basic document filing is not adequate yet to contemplate this approach.

Nearly fifteen years ago DFID commissioned BGS to undertake a piece of work with parallels the present assignment. The ARGOSS (Assessing the Risks to Groundwater from On-site Sanitation) project resulted in a manual (BGS, 2001<ref>BGS. 2001. Guidelines for assessing the risks to groundwater from on-site sanitation. BGS Commissioned Report CR/01/142. http://www.bgs.ac.uk/downloads/search.cfm?SECTION_ID=0&MIME_TYPE=0&SEARCH_T XT=argoss&dlBtn=go</ref>) containing guidance on determination of safe distances between pit latrines and groundwater wells or boreholes.

The ARGOSS guidance assumes that contamination at a groundwater source may come from either (a) seepage from pit latrines to the aquifer, then through the aquifer to the well, or (b) pollution because of poor design or construction of the well or borehole.

By way of numerous assumptions, approximations and modelling simplifications, the ARGOSS guidance provides a methodology for:
* Assessing the possible attenuation of contaminants in the unsaturated zone;
* Assessing attenuation with depth below the water table; and
* Assessing attenuation due to lateral movement of water in the aquifer.

The main drawbacks of the approach in relation to the present work are threefold:
* It did not consider multi-point source of pollution from widespread surface contamination (such as open defecation), which may well be the single most important pollutant source and pathway in Sierra Leone;
* It did not take sufficient account of preferential flowpaths and high permeability zones in the sub-surface. The model of permeability used in the approach is that there is a single ratio of horizontal to vertical permeability (ranging from 1 to 10);
* There is no allowance for seasonal fluctuation of water table in the approach – the rapidity with which pollutants can reach the water table is clearly much higher when that water table is very shallow; also the rate of movement in the aquifer will be determined largely by (relatively steep) wet season hydraulic gradients.

The final reason why neither the use of source protection zones nor the application of the ARGOSS methodology is appropriate for use in Sierra Leone is the burden of data collection which would be needed. The water point mapping exercise recently undertaken in Sierra Leone revealed the existence of 18 401 hand-dug wells and 1,952 boreholes. Despite the low sanitation coverage in Sierra Leone, there are probably at least five times this total number (say 100 000) of latrine pits. Undertaking assessments of even a fraction of these would be onerous in the extreme.

It is not simply the number of water points and latrines which would make such exercises impracticable, but also the heterogeneity and highly site-specific nature of the ground conditions at each site. In a very real and practical sense, the detailed hydrogeology of each site is not only unknown, but unknowable.

For these reasons the Ministry of Water Resources guidance document 'Protection of water resources at and around Ebola care facilities' (Ministry of Water Resources, 2015b<ref>MINISTRY OF WATER RESOURCES. 2015b. Protection of water resources at and around Ebola care facilities. Government of Sierra Leone. www.salonewatersecurity.com</ref>) adopted the simple notion of siting care facilities no closer than 30m from a hand-pumped well or borehole, and 50m in the case of a motor-pumped well or borehole.

{|class="wikitable" style="width:700px" style="background:#f0f8ff; font-size:100%"
| '''Protection zone summary:'''<br>
Given the nature of Sierra Leone’s water supplies, the number of water sources involved, and the limited institutional capacity for undertaking site assessments, the implementation of a source protection zone strategy is unrealistic at the present time and remains a long- term option.  Furthermore the capacity to regulate and enforce such an approach does not yet exist in Sierra Leone.   Simpler approaches are needed.
|}

===Robustness of a single minimum separation between sources of pollution and groundwater abstraction points===
A number of studies carried out in Africa relevant to this topic are reviewed in the following section, most studies have focussed on minimum separations between pit latrines and wells. As part of their recent literature review, Graham and Polizzotto (2013)<ref name="Graham">GRAHAM, J, and POLIZZOTTO, M. 2013. Pit latrines and their impacts on groundwater quality: A systematic review. Envrionmental Health Perspectives, Vol. 121, 521-530. </ref> included an assessment of the minimum separation distances between pit latrines and groundwater receptors recommended by studies in a range of typical hydrogeological settings. Separation distances of between 10–50m were commonly recommended. However, there was no detailed consideration of higher-risk settings such as those posed by tropical soils, which cover a considerable part of Africa, or karstic settings which require considerably greater separation distances. Furthermore, the fact that multi-point source contamination is widespread, such as from open defecation and animals, the framework of safe separation distances from point sources such as pit latrines breaks down.

Microorganisms have been assumed to not survive for very long after excretion but recent studies with viruses suggest that water quality may be impaired for a considerable length of time. Using a mixture of routine culture methods and genetic detection methods, Charles and co-workers detected viruses over 300 days after they were introduced in simulated groundwater systems (Charles et al., 2009<ref name="Charles">CHARLES, K J, SHORE, J, SELLWOOD, J, LAVERICK, M, HART, A, and PEDLEY, S. 2009. Assessment of the stability of human viruses and coliphage in groundwater by PCR and infectivity methods. . Journal of Applied Microbiology, Vol. 106, 11. </ref>). Using similar survival time for viruses in groundwater systems in the Netherlands, Schijven et al (2006)<ref name="Schijven">SCHIJVEN, JF, MÜLSCHLEGEL, JHC, HASSANIZADEH, SM, TEUNIS, PFM, DE RODA HUSMAN, AM, 2006. Determination of protection zones for Dutch groundwater wells against virus contamination – uncertainty and sensitivity analysis. Journal of Water and Health 4 (3), 297– 312. </ref> calculated that protection zones of between one and two years travel time would be required to ensure an infection risk of less than one in ten thousand per person per year. This is considerably longer than the 60 day travel time that is widely applied in Europe, or shorter in other parts of the world. Although a number of assumptions have been applied to the quantitative microbial risk assessments that were used to derive the travel time, it highlights the potential inadequacy of the current protection zones and the points to the need for water treatment to ensure that it is safe to drink.

A number of approaches have been used to define the quantities and transport distances of latrine-derived microbial contaminants. The majority of these have been culture-based studies of faecal bacteria; there has only been one study of viruses related to pit latrines (Verheyen et al., 2009<ref name="Verheyen">VERHEYEN, J, TIMMEN-WEGO, M, LAUDIEN, R, BOUSSAAD, I, SEN, S, KOC, A, UESBECK, A, MAZOU, F, and PFISTER, H. 2009. Detection of adenoviruses and rotaviruses in drinking water sources used in rural areas of Benin, West Africa. Applied and Environmental Microbiology, Vol. 75, 2798-2801. </ref>).

Attenuation of microbes is likely to be dependent on the hydrological conditions both in terms of water levels and recharge rate and permeability of the aquifer, and is highly variable. Dzwairo et al. (2006)<ref name="Dzwairo">DZWAIRO, B, HOKO, Z, LOVE, D, and GUZHA, E. 2006. Assessment of the impacts of pit latrines on groundwater quality in rural areas: A case study from Marondera district, Zimbabwe. Physics and Chemistry of the Earth, Parts A/B/C, Vol. 31, 779-788</ref> found faecal and total coliforms greatly reduced beyond 5 m from pit latrines in Zimbabwe, whereas Still and Nash (2002)<ref name="Still">STILL, D, and NASH, S. 2002. Groundwater contamination due to pit latrines located in a sandy aquifer: a case study from Maputaland. Water Institute of Southern Africa Biennial Conference Durban, South Africa, Water Institute of Southern Africa, 1-6. </ref> found faecal coliforms to be attenuated to <10 cfu/100 mL after 1 metre in Maputaland, KwaZulu-Natal. In Abeokuta, Nigeria, Sangodoyin (1993)<ref name="Sangodoyin">SANGODOYIN, A Y. 1993. Considerations on contamination of groundwater by waste disposal systems in Nigeria. Environmental Technology, Vol. 14, 957-964. </ref> found coliform attenuation to be correlated both with distance from the source and with the depth of the groundwater well. In Epworth, Zimbabwe, groundwater contamination was higher in the dry season rather than in the wet, with coliforms detected up to 20 metres from the pit (Chidavaenzi et al., 1997<ref name="Chidavaenzi 1997">CHIDAVAENZI, M, JERE, M, and BRADLEY, M. 1997. Pit latrine effluent infiltration into groundwater. Proceedings of the 23rd WEDC Conference, Durban, Water, Engineering and Development Centre, 59-62. </ref>). In Benin, Verheyen et al. (2009)<ref name="Verheyen"></ref> found a positive association for detection of viruses in water sources with at least one latrine within a 50 m radius. They postulated that during the wet season viruses were transported in shallow groundwater whereas in the dry season contamination was likely to be from surface water.

In an informal settlement in Zimbabwe, Zingoni et al. (2005)<ref name="Zingoni">ZINGONI, E, LOVE, D, MAGADZA, C, MOYCE, W, and MUSIWA, K. 2005. Effects of a semi-formal urban settlement on groundwater quality: Epworth (Zimbabwe): Case study and groundwater quality zoning. Physics and Chemistry of the Earth, Parts A/B/C, Vol. 30, 680-688. </ref> found detectable total and faecal coliforms in over 2/3 of domestic boreholes and wells. In the area 75% of households used pit latrines and there were also informal trading areas. In Langas, Kenya, Kimani-Murage and Ngindu (2007)<ref name="Kimani">KIMANI-MURAGE, E, and NGINDU, A. 2007. Quality of water the slum dwellers use: The case of a Kenyan slum. Journal of Urban Health, Vol. 84, 829-838. </ref> found that 50% of wells were within 30 m of a pit latrine and that all shallow wells were positive for total coliforms with 70% >1100 mpn/100 mL; however, in Kisumu, Kenya, Wright and co-workers failed to find a significant correlation between the levels of thermotolerant coliforms in water sampled from shallow wells and the density of pit latrines (Wright et al., 2013<ref name="Wright 2013">WRIGHT, J A, CRONIN, A, OKOTTO-OKOTTO, J, YANG, H, PEDLEY, S, and GUNDRY, S W. 2013. A spatial analysis of pit latrine density and groundwater source contamination. Environmental Monitoring and Assessment, Vol. 185, 12. </ref>).

While clearly less important from a health perspective compared to microbiological contamination, chemical contaminants also pose a threat to water quality, and they are very useful tracers of microbiological contaminants, which are inherently more transient. Nutrients are also important in the fact that they are linked to the survival of pathogens in the environment. The chemical species of greatest concern from excreta disposed in on-site sanitation systems are regarded to be the macronutrients nitrate and phosphate. Pin-pointing specific sources is challenging as nitrate may be derived from numerous sources including plant debris, animal manure, solid waste and fertilisers. A common approach has been to compare areas that are similar but have different latrine densities. In an informal settlement in Zimbabwe, Zingoni et al. (2005)<ref name="Zingoni"></ref> demonstrated that the highest nitrate concentrations were associated with the highest population and pit latrine density. Similar patterns have been observed in Senegal and Southern Africa (Tandia et al., 1999<ref name="Tandia">TANDIA, A A, DIOP, E S, and GAYE, C B. 1999. Nitrate groundwater pollution in suburban areas: example of groundwater from Yeumbeul, Senegal. Journal of African Earth Sciences, Vol. 29, 809-822. </ref>; Vinger et al., 2012<ref name="Vinger">VINGER, B, HLOPHE, M, and SELVARATNAM, M. 2012. Relationship between nitrogenous pollution of borehole waters and distances separating them from pit latrines and fertilized fields. Life Science Journal, Vol. 9, 402-407. </ref>). Studies in the peri-urban areas of Kisumu, Kenya have shown that the density of latrines within a 100m radius of the sources was significantly correlated with nitrate concentrations (Wright et al., 2013<ref name="Wright 2013"></ref>). In contrast, Sangodoyin (1993)<ref name="Sangodoyin"></ref> found that nitrate concentrations were not related to distance from pit latrines in Abeokuta, Nigeria. In eastern Botswana the build-up of nitrogenous latrine effluent in soils and vertical leaching resulted in nitrate concentrations of above 500 mg/L (Lewis et al., 1980<ref name="Lewis 1980">LEWIS, W J, FARR, J, and FOSTER, S S. 1980. The pollution hazard to village water supplies in eastern Botswana. ICE Proceedings, Ice Virtual Library, Vol. 69, 281-293. </ref>).

Direct measurements and well-designed studies are sparse and rarely consider rapid flowpaths or multi-point sources of contaminations. Graham and Polizzotto (2013)<ref name="Graham"></ref> estimate lateral travel distances of 1-25 m for pit-latrine derived nitrate. Chloride is typically transported with minimal retention and frequently tracks nitrate (e.g. Lewis et al., 1980<ref name="Lewis 1980"></ref>) unless subsurface conditions promote denitrification. Ammonia does not tend to accumulate in groundwater near latrines but can accumulate and persist in anaerobic conditions and when the water table intersects the base of the latrine pit (Dzwairo et al., 2006<ref name="Dzwairo"></ref>; Nyenye et al. 2013<ref name="Nyenje">NYENJE, PM, FOPPEN, JW, KULABAKO, R, MUWANGA, A, AND UHLENBROOK, S. 2013. Nutrient pollution in shallow aquifers underlying pit latrines and domestic solid waste dumps in urban slums. Journal of environmental management, 122, 15-24. </ref>). Other contaminant tracers of waste water or faecal sources include potassium, sulphate and DOC and emerging organic contaminants (Sorensen et al., 2015a<ref name="Sorensen 2015a">SORENSEN, JPR, LAPWORTH, DJ, NKHUWA, D CW, STUART, ME, GOODDY, DC, BELL, RA, CHIRWA, M, KABIKA, J, LIEMISA, M, CHIBESA, M AND PEDLEY, S. 2015a. Emerging contaminants in urban groundwater sources in Africa. Water research, 72, 51-63. </ref>; 2015b<ref name="Sorensen 2015b">SORENSEN JPR, LAPWORTH, DJ, MARCHANT, BP, NKHUWA, DCW, PEDLEY, S, BELL, RA, CHIRWA, M, KABIKA, J, LIEMISA, M, CHIBESA, M. 2015b. In-situ tryptophan sensing: a rapid, predictor of faecal contamination in groundwater. Water research, 81, 38-46. </ref>).

{|class="wikitable" style="width:700px" style="background:#f0f8ff; font-size:100%"
| '''Hazard source-water point separation summary:'''<br>
A combination of the limited sanitation coverage, leading to an essentially diffuse (multi- point)  surface  hazard  loading,  vulnerable  shallow  geological  terrain  and climate  in  this region challenges the premise of safe lateral spacing between identified hazards, such as pit latrines, and drinking water supplies.<br><br>An alternative separation framework which is based on vertical separation, and minimises rapid  bypass  contamination  pathways  is  a  better  approach  in  this  setting  for  protecting groundwater supplies.
|}

==Water quality==
This section initially assesses the available groundwater quality literature from Sierra Leone and then reviews a wider range of water quality studies across sub-Saharan Africa (SSA) in relevant hydrogeological settings. Given the varied geology of Sierra Leone, this includes studies carried out in basement, sedimentary and volcanic terrains, with an emphasis on weathered basement settings which cover the majority of the country. The literature search included areas with annual rainfall over ca.1000 mm and also includes some studies from karstic terrains, analogous to highly vulnerable settings found in lateritic terrains during the rainy season when water levels are high and lateral flow can exceed 300 m/d in some instances (e.g. Bonsor et al., 2014<ref>BONSOR, H C, MACDONALD, A M and DAVIES, J. 2014. Evidence for extreme variations in the permeability of laterite from a detailed analysis of well behaviour in Nigeria. Hydrological Processes 28, 3563–3573. </ref>). This section provides a brief review of key water quality parameters, with a focus on microbiological contaminants which is the key water quality threat to water points, however it is recognised that other contaminants (such as NO3, As and F) are also important from a water quality perspective in the long term.

An assessment of water quality variations in relation to hydrogeology, seasonality source type and specific high risk factors are made in this section. Table A2, in the appendix, summaries case studies from hydrogeologically relevant settings in Africa (n=51). Case studies (n=18) focused on the impact of sanitary sources, principally pit latrines, on groundwater quality across SSA are summarised in Table 5 (section 2.2.6). Studies covering aspects of non-sanitary sources of contamination such as industry, historical mining legacy and waste sites/landfills are not fully reviewed however, this and water quality from a wider literature search across SSA can be found in Lapworth et al (2015a)<ref>LAPWORTH, D J, GOODDY, D, HARRISON, I, KIM, A, VANE, C H. 2005. Colloidal phase transport of pesticides: a review with special reference to major UK aquifers. Nottingham, UK, British Geological Survey, 22pp. (IR/05/131) (Unpublished) </ref>. A handful of case studies include both specific assessments of impacts of pit latrines as well as broader environmental hygiene considerations in spring catchments and well capture zones and are included in both Table A1 and Table 5.

A large proportion of studies are drawn from urban and peri-urban settings where there are generally greater risks for groundwater contamination, but the review also includes studies undertaken within rural settlements. Urbanisation processes are the cause of extensive but essentially diffuse pollution of groundwater by nitrogen and sulphur compounds, salinity as well as pathogenic bacteria, protozoa and viruses (Morris et al., 2003<ref>MORRIS BL, LAWRENCE ARL, CHILTON PJ, ADAMS B, CALOW RC, KLINCK BA. 2003. Groundwater and its Susceptibility to Degradation:A Global Assessment of the Problem and Options for Management.Early Warning and Assessment Report Series RS.03–3.United Nations Environment Programme, Nairobi, Kenya </ref>). Household attitudes to hazards posed by drinking water can enhance quality problems with poor water treatment, types of drinking water vessels/storage, hand washing practices, perceptions of safe water quality using only visual parameters (normally clarity of the water), and knowledge on waste disposal practices (Kioko and Obiri 2012<ref>KIOKO KJ AND OBIRI JK. 2012. Household attitudes and knowledge on drinking water enhance water hazards in peri-urban communities in western Kenya. J. Disaster Risck Studies, 4(1) doi: 10.4102/jamba.v4i1.49 </ref>).

Overall, compared to other regions globally there have been relatively few studies carried out in Africa. The review draws mostly on research articles but also includes some reports and book chapters. It is recognised that these have been published for a range of purposes, with this in mind, the studies can be categorised into three broad groups and are identified by the notation (1, 2  or 3) in Tables A1 and Table 5:
# Case-studies presenting data from a limited number of sites (n<20), limited temporal resolution as a single survey or use only basic chemical indicators and limited analysis of the results.
# Case studies which either draw from larger data sets or include both chemical and microbiological indicators but have limited data analysis regarding sanitary risk factors.
# Case studies with greater temporal resolution or are accompanied by a more thorough analysis of the data, for example using statistical techniques to understand the significance different risk factors on water quality observations.

Studies from group 3 provide the greatest insights regarding pollution sources, pathways and risk factors and can be considered as benchmark studies. It is clear from looking across the published literature that there has been a large number of groundwater quality related studies in southern Nigeria which account for some 30% of the published studies overall and most fall into either category 1 or 2 studies. Most of these studies are located near Lagos, Abeokuta and Ibadan in the south west, the Delta area in the south and Calabar. Other notable examples of locations that have a larger number of case studies include Lusaka in Zambia, Kampala in Uganda, Dakar in Senegal, and Addis Ababa in Ethiopia. These all have relevance to Sierra Leone given the varied hydrogeology, climate and socio-economic conditions found across Sierra Leone. Kampala (basement setting), Addis Ababa and Lusaka (basement and karstic settings) all have vulnerable hydrogeological settings analogous to those found in many parts of Sierra Leone. Dakar has comparable shallow coastal sedimentary aquifer systems. Studies in Southern Nigeria have both comparable hydrogeology and climate.

===Baseline hydrochemistry and non-sanitary sources of contamination in Sierra Leone ===
While perhaps less important compared to microbiological contamination from a health perspective in the short term, in the long term a range of water quality issues need to be considered for Sierra Leone, these are briefly reviewed in the following section.

Wright (1982a)<ref>WRIGHT, R C. 1982a. Seasonal variation in water quality of a West African river (R. Jong in Sierra Leone). Rev. Hydrobiol. Trop. 15,3, 193-199. </ref> presents chemistry results from a seasonal study of the River Jong (also referred to as River Njala) in Sierra Leone. The waters are characterised as having very low SEC (range 13-30 mS/cm) but showed pronounced seasonal trends associated with changes in baseflow and throughflow contributions. Baseflow chemistry is characterised by higher pH (6- 6.5) HCO3, Si, Na, Ca, K and Mg, and lower Fe and turbidity.

A draft report by Mott MacDonald Int. (1991)<ref>MOTT MACDONALD INTERNATIONAL. 1991. Draft Country report for Sierra Leone, Sub-Saharan Africa hydrological assessment, west African countries, World Bank-UNDP-ADB, pp177. </ref> states that fluoride concentrations were found to be >5 mg/L in around 20% of wells in the Bombali and Kambia area. Risks related to potentially elevated trace element concentrations in groundwater sources due to geogenic sources in mineralised zones (e.g. arsenopyrite mineralisation in the shallow weathered schist terrain for example) as well as heavy metal contamination (e.g. mercury) associated with mining waste processing activities are highlighted by Akiwumi (2008)<ref name="Akiwumi 2008">AKIWUMI, F A. 2008. An assessment of hazards from gold mining in Sierra Leone. Paper of the Applied Geography Conference 2008. p 10-18. </ref>.

While there are very few published results for trace element analysis from groundwaters in these settings in Sierra Leone compared to other West African contries (e.g. Babut et al., 2003<ref>BABUT, M, SEKYI, R, RAMBAUD, A, POTIN-GAUTIER, M, TELLIER, S, BANNERMAN, W and Beinhoff, C. 2003. Improving the environmental management of small-scale gold mining in Ghana: A case study of Dumasi. Journal of Cleaner Production 11, 2, 215-221. </ref>) data from soil, stream sediment and whole rock analysis suggest that groundwaters in the main gold bearing terrains (schist and granites of the Ankobra, Pra and Tano River basins) could have naturally elevated arsenic concentrations (Akiwumi 2008<ref name="Akiwumi 2008"></ref>).

There is no As data available for groundwater in Sierra Leone, but there could be As related water quality concerns considering concentrations found in analogous settings in Ghana and Burkina Faso, where elevated As concentrations have been reported (up to 1640 mg/L) associated with geogenic sources and mining activities (Smedley 1996<ref>SMEDLEY, PL.1996. Arsenic in rural groundwater in Ghana: part special issue: hydrogeochemical studies in sub-saharan Africa. Journal of African Earth Sciences, 22(4), 459-470. </ref>; Smedley et al., 2007<ref>SMEDLEY, PL, KNUDSEN, J, and MAIGA, D. 2007. Arsenic in groundwater from mineralised Proterozoic basement rocks of Burkina Faso. Applied Geochemistry, 22(5), 1074-1092. </ref>). While certainly not considered as a major issue compared to faecal/sanitary sources of contamination in shallow wells in Africa, locally contamination from mine waste could lead to elevated trace element pollution in groundwater sources, as well as fish, particularly given the relatively low buffering capacity and low pH values found in these granitic terrains (Akiwumi 1987<ref name="Akiwumi 1987">AKIWUMI, F A. 1987. Groundwater exploration and development in the crystalline Basement rocks of Sierra Leone. Source unknown. http://www.bgs.ac.uk/africagroundwateratlas/fulldetails.cfm?id=SL4002</ref>; Ouedraogo and Amyot 2013<ref>OUEDRAOGO, O., and AMYOT, M. 2013. Mercury, arsenic and selenium concentrations in water and fish from sub-Saharan semi-arid freshwater reservoirs (Burkina Faso). Science of the total environment, 444, 243-254. </ref>).

{| class="wikitable"
|+ Table 4 Studies investigating groundwater contamination from pit latrines in analogous regions in SSA (n=19)
| '''Region/Country (rural/urban)'''
| '''Subsurface conditions'''
| '''Sample sites (n)'''
| '''Water quality parameters'''
| '''Sampling time frame'''
| '''Conclusion'''
| '''Reference'''
|-
| <sup>3</sup>Kulanda town in Bo, Sierra Leone
| Weathered Granitic Basement
| Wells (33), lined and unlined
| FC, SEC, NO3, Turb, inorganic majors, pH
| Wet season
| No statistical significance found for pit latrine distance, lowest p value (0.06) for distance from field. Low pH concern for corrosion.
| Jimmy et al. (2013)
|-
| <sup>3</sup>Kamangira, Zimbabwe (rural)
| Sandy soils over fractured basement
| Installed test wells (17)
| NH4, NO3, turb, pH, Conductivity, TC, FC
| Feb-May 2005
| Low FC >5m from PL, N conc. usually below WHO standards
| Dzwairo et al. (2006)<ref name="Dzwairo"></ref>
|-
| <sup>3</sup>Epworth, Zimbabwe (urban)
| Fine sandy soils over fractured basement
| New and existing wells (18) and boreholes (10)
| Na, Zn, Cu, Fe, PO4, NO2, TC, FC
| N/A
| Elevated N and Coliforms in most of study area
| Zingoni et al. (2005)<ref name="Zingoni"></ref>
|-
| <sup>3</sup>Epworth, Zimbabwe (urban)
| Fine sandy soils over fractured basement
| Installed wells
| N, SO4, FC
| 2-8 week intervals 1998-1999
| Rapid reduction in Coliforms, S and N 5-20 m from PL
| Chidavaenzi et al. (2000)<ref name="Chidavaenzi">CHIDAVAENZI, M, JERE, M, and BRADLEY, M. 1997. Pit latrine effluent infiltration into groundwater. Proceedings of the 23rd WEDC Conference, Durban, Water, Engineering and Development Centre, 59-62. </ref>
|-
| <sup>2</sup>Lusaka, Zambia (urban)
| Thin soils and karstic Dolomite
| Existing wells (NA)
| NO3, Cl, FC
| November 2003, March 2004, October 2004
| Greatest FC loading from PL and other waste sources in wet season and dilution of N pollution
| Nkhuwa (2006)<ref name="Nkhuwa 2006">NKHUWA, D C W. 2006. Groundwater quality assessment in the John Laing and Misisi areas of Lusaka. 239-252 in Groundwater pollution in Africa. XU, Y, and USHER, B (editors). (Leiden: Taylor & Francis/Balkema.) </ref>
|-
| <sup>3</sup>Dakar, Senegal (urban)
| Fine-course sands over sediments
| Existing wells (47)
| Broad hydrochemistry, FC
| July and November 1989
| Nitrate strongly linked to PL proximity
| Tandia et al. (1999)<ref name="Tandia"></ref>
|-
| <sup>2</sup>NW Province, South Africa (rural)
| N/A
| Existing wells (9)
| NH4, NO3, NO2
| June-July
| High contamination <11 m from PL
| Vinger et al. (2012)<ref name="Vinger"></ref>
|-
| <sup>3</sup>Mbazwana, South Africa (urban)
| Sands
| Installed test wells (5)
| FC and NO3
| Bimonthly 2000-2002
| Low nitrate (<10 mg/L) and FC (<10/100mL) >1m from PL
| Still and Nash (2002)<ref name="Still"></ref>
|-
| <sup>2</sup>Bostwana, Mochudi/Ramotswa (rural)
| Well-poorly drained soils
| Existing wells (>60)
| P, N, stable isotopes and Cl
| N/A
| Variable N leaching from PL
| Lagerstedt et al. (1994)<ref name="Lagerstedt">LAGERSTEDT, E, JACKS, G, and SEFE, F. 1994. Nitrate in groundwater and N circulation in eastern Botswana. Environmental Geology, Vol. 23, 60-64. </ref>
|-
| <sup>2</sup>Botswana (rural)
| fractured basement
| Existing well and observation well (2)
| Broad Hydrochemistry, ''E. coli''
| October-February 1977
| Contamination of wells near latrine with ''E. coli ''and nitrate
| Lewis et al. (1980)<ref name="Lewis 1980"></ref>
|-
| <sup>3</sup>Various, Benin (rural)
| N/A
| Existing wells (225)
| Andenovirus, rotavirus
| Wet/dry season 2003-2007
| Viral contamination is linked to PL proximity
| Verheyen et al. (2009)<ref name="Verheyen"></ref>
|-
| <sup>3</sup>Langas, Kenya (urban)
| N/A
| Existing wells (35)
| TC,FC
| January-June 1999
| 97% wells positive for FC, 40% of wells >15m from PL
| Kimani-Murage and Ngindu (2007)<ref name="Kimani"></ref>
|-
| <sup>3</sup>Kisumu, Kenya (urban)
| Sedimentary
| Existing wells (191)
| FC, NO3, Cl
| 1998 to 2004
| Density of PL within a 100 m radius was significantly correlated with nitrate and Cl but not FC (''PC'')
| Wright et al. (2013)<ref name="Wright 2013"></ref>
|-
| <sup>3</sup>South Lunzu, Blantyre, Malawi (urban)
| Weathered basement
| Borehole, springs and dug well (4)
| SEC, Cl, Fe, FC,FS
| Wet and dry season on two occasions
| Groundwaters highly contaminated due to poor sanitation and domestic waste disposal. 58% of residence use traditional PL
| Palamuleni (2002)<ref name="Palamuleni">PALAMULENI, L G. 2002. Effect of sanitation facilities, domestic solid waste disposal and hygiene practices on water quality in Malawi’s urban poor areas: a case study of South Lunzu Township in the city of Blantyre. Physics and Chemistry of the Earth, Parts A/B/C, Vol. 27, 845- 850. </ref>
|-
| <sup>3</sup>Uganda, Kampala (urban)
| Weathered basement
| Piezometers (10)
| NO3, Cl, PO4
| March-August 2010 biweekly sampling
| PL found to be a significant source of nutrients (N) compared to waste dump
| Nyenje et al. (2013)<ref name="Nyenje"></ref>
|-
| <sup>3</sup>Uganda, Kampala (urban)
| Weathered basement
| Installed wells and spring (17)
| SEC, pH, P, NO3, Cl, FC and FS
| March-August 2003, weekly and monthly
| Widespread well contamination linked to PL and other waste sources
| Kulabako et al. (2007)<ref name="Kulabako">KULABAKO, N R, NALUBEGA, M, and THUNVIK, R. 2007. Study of the impact of land use and hydrogeological settings on the shallow groundwater quality in a peri-urban area of Kampala, Uganda. Science of The Total Environment, Vol. 381, 180-199. </ref>
|-
| <sup>3</sup>Uganda, Kampala (urban)
| Weathered basement
| Springs (4)
| FC, FS, NO3, NH4
| Wet and dry season for 5 consecutive weeks
| Widespread contamination from PL and poor animal husbandry, both protected and unprotected sources unfit for drinking
| Nsubuga et al. (2004)<ref name="Nsubuga">NSUBUGA, F B, KANSIIME, F, and OKOT-OKUMU, J. 2004. Pollution of protected springs in relation to high and low density settlements in Kampala—Uganda. Physics and Chemistry of the Earth, Parts A/B/C, Vol. 29, 1153-1159. </ref>
|-
| <sup>3</sup>Uganda, Kampala (urban)
| Weathered basement
| Springs (25)
| FC, FS
| Monthly September 1998- March 1999
| Spring contamination linked to local environmental hygiene and completion rather than on-site sanitation (''LR'')
| Howard et al. (2003)<ref name="Howard">HOWARD, G, PEDLEY, S, BARRETT, M, NALUBEGA, M, and JOHAL, K. 2003. Risk factors contributing to microbiological contamination of shallow groundwater in Kampala, Uganda. Water Research, Vol. 37, 3421-3429. </ref>
|-
| <sup>3</sup>Lichinga, Mozambique
| Mudstone
| Lichinga (25)
| TTC, EF (Enterococi)
| Monthly for 1 year
| Higher risk at onset of the wet season and end of the dry season. Predominant source was from animal faeces rather than PL or septic tanks (''LR'')
| Godfrey et al. (2006)<ref name="Godfrey">GODFREY, S, TIMO, F, and SMITH, M. 2006. Microbiological risk assessment and management of shallow groundwater sources in Lichinga, Mozambique. Water and Environment Journal, Vol. 20, 194-202. </ref>
|}
PL = Pit latrine, FC = Faecal coliform (values given as 0 are below detection limit of method), SEC= Specific electrical conductivity, TTC= Thermotolerant coliforms, TC = Total coliform, FS = Faecal strep, Turb=turbidity, ''LR''=logistic regression, ''PC''= Pearson’s correlation. Concentrations in mg/L unless otherwise stated.


Jimmy et al (2013)<ref name="Jimmy">JIMMY, D H, SUNDUFU, A J, MALANOSKI, A P, JACOBSEN, K H, ANSUMANA, R, LESKI, T A, BANGURA U, BOCKARIE A S, TEJAN E, LIN, B, and STENGER, D A. 2013. Water quality associated public health risk in Bo, Sierra Leone. Environmental monitoring and assessment, 185, 1, 241- 251. </ref> and Ibemenuga and Avoaja (2014)<ref name="Ibemenuga">IBEMENUGA, K N, and AVOAJA, D A. 2014. Assessment of groundwater quality in wells within the Bombali district, Sierra Leone. Animal Research International, 11, 1, 1905-1916.</ref> present some hydrochemical data for urban and rural wells in Sierra Leone, but limited analysis and interpretation of results. Overall, both studies conclude that microbiological contamination is more of a health risk to users compared to inorganic contaminants. Both studies do show a significant proportion of sites with low (<6.5) pH values in groundwater sources. While this is perhaps not critical from a drinking water quality perspective it does have implications for microbiological survival and corrosion potential for infrastructure such as piped water sources and borehole casing. There is anecdotal evidence from the early 1990s in Sierra Leone that suggests this may have happened in boreholes drilled by JICA which failed within a few years of installation. High iron is also a widespread problem in wells and boreholes in Sierra Leone<ref>Pers. coms., Peter Dumble, PDHydrogeology, February 2015</ref>

===Sanitary sources of contamination in Sierra Leone ===
There are four studies available that contain microbiological and chemical water quality data (nitrate) from wells and springs on the basement terrain of Sierra Leone related to sanitary sources of contamination, these are summarised in Table 5. Two early papers by Wright (1986 and 1982b) were seasonal studies carried out in rural settlements in South-Eastern Provinces in Sierra Leone. Both of these papers investigated a range of drinking water sources (wells, springs, streams and swamps), and the temporal changes in FC (faecal coliforms), FS (faecal streptococci) as well as ''E. coli'', ''S. faecalis, C. perfringens ''and ''Salmonella ''spp. Both of these studies showed gross levels of microbiological contamination in unprotected springs and wells throughout the year (e.g. FC >30k cfu, mean 3k in Wright (1982b)<ref name="Wright 1982b">WRIGHT, R C.1982b. A comparison of the levels of faecal indicator bacteria in water and human faeces in a rural area of a tropical developing country (Sierra Leone). Journal of Hygiene, 89, 01, 69-78. </ref>), with higher levels of contamination for FC in groundwater towards the start of the wet season compared to larger rivers and comparable contamination to smaller surface water sources.

In groundwater sources, detailed seasonal studies point to increased risk of enhanced microbiological contamination from faecal coliforms and associated pathogens (e.g. ''Salmonella ''spp) during the onset of the dry season and the start of the wet season, and then a reduction in faecal coliform counts as the wet season progresses. The author suggest that this may be a dilution control, and the fact that shallow groundwater sources are the only reliable sources of drinking water in the dry season means there is a higher risk for users during this period. None of the settlements had sanitation facilities, and open defecation was cited as normal practice and none of the wells had any protection (unlined/covered), so surface sources and contamination from runoff would have likely been significant.

Two recent studies have carried out single campaign water quality surveys at the start of the wet season in wells in two different regions of Sierra Leone (Jimmy et al. (2013)<ref name="Jimmy"></ref>; Ibemenuga and Avoaja (2014)<ref name="Ibemenuga"></ref>). Both papers conclude that microbiological contamination was the greatest health risk associated with drinking water (compared to major ion chemistry); however, neither study carried out trace element analysis for arsenic or heavy metals. Jimmy et al (2013)<ref name="Jimmy"></ref> carried out a survey of lined and unlined wells (n=33) in the Kulanda township of Bo, and investigated the importance of different risk factors, including well depth, proximity to pit latrines, proximity to fields and well completion. Unlined wells were found to have poorer water quality compared to lined wells (Figure 13) and the shallow sources were more contaminated compared to deeper sources with regards to FC, nitrate and SEC (Figure 14). The relationship between NO3 and FC and distance from nearest toilet shows generally lower concentrations where distances are >40m, but a high degree of variability for sites with toilets/pit latrines <40m, although this relationship is not statistically significant.

[[Image:OR15009fig13.jpg|thumb|center|300px|'''Figure 13''' Box-plots of faecal coliform (FC), nitrate and SEC distributions in lined and unlined wells in Bo, Sierra Leone, data from Jimmy et al. (2013)<ref name="Jimmy"></ref>. Tukey box-plot used.]]

[[Image:OR15009fig14.jpg|thumb|center|400px|'''Figure 14''' Relationship between water quality parameters and well depth, distance from nearest toilet and distance from field, data from Jimmy et al. (2013)<ref name="Jimmy"></ref>.]]

Using logistic regression, the distance from field boundaries were found to have the lowest p value (0.06) for predicting the presence of FC, with much larger p values for other predictors, such as distance from nearest toilet. The significance of this result should be treated with caution as the data set is small and the distance from the nearest toilet (e.g. rather than more conventional measures such as density within a particular search radius) may not be the best criteria to use for this type of analysis.

Ibemenuga and Avoaja (2014)<ref name="Ibemenuga"></ref> present water quality results (FC, SEC and major ions, F, and some trace elements e.g. Fe and Cu) from a larger sample size (n=60) from rural settlements in the Bombali region of Sierra Leone. Overall, mean levels of FC were comparable (mean 16.6 cfu/100mL, range BDL-80 cfu) with those from Bo (mean 19.5 cfu/100mL, range BDL-75 cfu) (Jimmy et al (2013)<ref name="Jimmy"></ref>). Median and mean nitrate and SEC values were comparable (Table 5, Figure 15) but Figure 15 does show that this study had a consistently smaller inter-quartile- range compared to the study carried out by Jimmy et al (2013)<ref name="Jimmy"></ref>. There is very little detail given in the paper regarding well construction and other risk factors in this study so it is difficult to draw any firm conclusions. The generally lower level of contamination could be due to the fact that these are from smaller settlements with lower levels of diffuse contamination in the subsurface. Both studies had ca. 60% of groundwater sources that were contaminated with detectable FCs and a similar overall FC distribution. Compared to the earlier studies by Wright (1986<ref name="Wright 1986">WRIGHT, R C. 1986. The seasonality of bacterial quality of water in a tropical developing country (Sierra Leone). Journal of Hygiene, 96, 01, 75-82. </ref>; 1982b<ref name="Wright 1982b">WRIGHT, R C.1982b. A comparison of the levels of faecal indicator bacteria in water and human faeces in a rural area of a tropical developing country (Sierra Leone). Journal of Hygiene, 89, 01, 69-78. </ref>), where sources had no protection, the level of faecal contamination found in the shallow wells in these two studies were two orders of magnitude lower on average, suggesting that well construction and protection is a highly significant factor controlling pollution pathways.

[[Image:15009_fig15.jpg|thumb|center|500px|'''Figure 15''' Comparison of water quality data from shallow wells from Bo and Bombali district, Sierra Leone. Data from Jimmy et al. (2013)<ref name="Jimmy"></ref> and Ibemenuga and Avoaja (2014)<ref name="Ibemenuga"></ref>.]]

===Chemical and physical indicators of groundwater quality degradation from comparable hydrogeological settings===
Figure 16 shows the distribution of case studies used in this water quality review. Most are from West Africa, although there are also a number from East Africa including case studies in Uganda, Malawi and Kenya and Zimbabwe.

[[Image:15009_fig16.jpg|thumb|center|500px|'''Figure 16''' Location of case studies used in this review. Background map showing regional scale aquifer productivity from MacDonald et al. (2012)<ref name="MacDonald">MACDONALD, AM, BONSOR, HC, Ó DOCHARTAIGH, BÉ AND TAYLOR, R. G. 2012. Quantitative maps of groundwater resources in Africa. Environmental Research Letters, 7(2), 024009. </ref>.]]

====PHYSICAL INDICATORS====
Total dissolved solids (TDS) and specific electrical conductivity (SEC) are the most commonly applied physical water quality indicators in groundwater studies and are often used in combination with more specific indicators such as dissolved chemistry or microbiology (see Table A1). They have a major advantage of being field methods, which are relatively easy to deploy and versatile, enabling the user to carry out an initial assessment of water quality rapidly, and with minimal cost. The baseline quality of groundwater, with relatively low total dissolved solids (TDS) in most basement and alluvial settings, makes TDS a good indicator of contaminant loading (Figure 17).

[[Image:15009_fig17.jpg|thumb|center|500px|'''Figure 17''' Summary water quality results for SEC from shallow groundwater studies carried out across hydrogeologically relevant terrains in SSA based on climate and geology. Data extracted from tables and figures in peer reviewed literature, some summary statistics (mean) are not available from the literature. W=wells, B=boreholes, *Data from Sierra Leone, note log scale on x-axis.]]

===Nitrate and chloride===
Nitrate and chloride are the most widely used chemical indicators of anthropogenic pollution. Nitrate data has been reported in over 80% of the groundwater studies summarised in Table A1. Summary statistics for a number of studies in basement and sedimentary settings in Africa is presented in Figure 18. The relatively simple sample preservation and analysis required makes these parameters attractive for initial water quality screening. Overall, nitrate concentrations ranged from Below Detection Level (BDL) to >500 mg/L (as NO3), although typical maximum concentrations were generally below 150 mg/L (Figure 18). The WHO guideline value for nitrate is 50 mg/L as NO3. The WHO has not published a health-based guideline for chloride, but suggests that concentrations over 250 mg/l can give rise to a detectable taste.

Both tracers have been used in a broad range of geologic and climate zones to investigate pollution from on-site sanitation, waste dumps, as well as urban agriculture (Table A1). Nitrate concentrations show a high degree of variability both within studies and between studies that have been reviewed. Two principle factors that affect nitrate occurrence are firstly the prevailing redox conditions in groundwater, and secondly the residence time and vulnerability of the groundwater body. There are several examples of low nitrate groundwater in Table 1 which show evidence of faecal contamination (Gelinas et al., 1996<ref name="Gelinas">GELINAS, Y, RANDALL, H, ROBIDOUX, L, and SCHMIDT, J-P. 1996. WELL WATER SURVEY IN TWO DISTRICTS OF CONAKRY (REPUBLIC OF GUINEA), AND COMPARISON WITH THE PIPED CITY WATER. Water Reserch, Vol. 39, 2017-2026. </ref>; Mwendera et al., 2003<ref name="Mwendera">MWENDERA, E J, HAZELTON, D, NKHUWA, D, ROBINSON, P, TJIJENDA, K, and CHAVULA, G. 2003. Overcoming constraints to the implementation of water demand management in southern Africa. Physics and Chemistry of the Earth, Vol. 28, 761-778. </ref>; Nkhuwa, 2003<ref name="Nkhuwa 2003">NKHUWA, D C W. 2003. Human activities and threats of chronic epidemics in a fragile geologic environment. Physics and Chemistry of the Earth, Vol. 28, 1139-1145. </ref>) which has implications for the potential for denitrification in shallow groundwaters. Nitrate has been used successfully to characterise urban loading to groundwater from a range of sources including pit latrines (Cissé Faye et al., 2004<ref name="Cisse">CISSÉ FAYE, S, FAYE, S, WOHNLICH, S, and GAYE, C. 2004. An assessment of the risk associated with urban development in the Thiaroye area (Senegal). Environmental Geology, Vol. 45, 312- 322. </ref>), landfills (Ugbaja and Edet, 2004<ref name="Ugbaja">UGBAJA, A, and EDET, A. 2004. Groundwater pollution near shallow waste dumps in Southern Calabar, South-Eastern Nigeria. Global Journal of Geological Sciences, Vol. 2, 199-206. </ref>; Vala et al., 2011<ref name="Vala">VALA, R M K, TICHAGWA, L, MUSIBONO, D E, and LUKANDA, V M. 2011. Environmental and health concerns regarding the quality of water in a poor suburb of Kinshasa in the Democratic Republic of Congo. Water Science & Technology: Water Supply, Vol. 11, 266-273. </ref>) and applied to look at impacts on groundwater quality across different population densities (Goshu and Akoma, 2011<ref name="Goshu 2011">GOSHU, G, and AKOMA, O C. 2011. Water quality assessment of underground and surface water resources of Bahir Dar and Periurban areas, north-west Ethiopia. Global Journal of Environmental Sciences, Vol. 10, 11-21. </ref>; Goshu et al., 2010<ref name="Goshu 2010">GOSHU, G, FARNLEITNER, A, MANAFI, M, and BYAMUKAMA, D. 2010. The bacteriological quality of traditional hand dug wells and protected hand pumps in Bahirdar Town and peri-urban areas, Northern Ethiopia. Proceedings of the First National Research Symposium on: Sustainable Development: A great concern in Africa, Debre Markos, Ethiopia, 247-259. </ref>; Orebiyi et al., 2010<ref name="Orebiyi>OREBIYI, E O, AWOMESO, J A, IDOWU, O A, MARTINS, O, OGUNTOKE, O, and TAIWO, A M. 2010. Assessment of pollution hazards of shallow well water in Abeokuta and environs, Southwest, Nigeria. American Journal of Environmental Sciences, Vol. 6, 50-56. </ref>). There are other sources of N loading to groundwater in growing urban areas including the impact of deforestations, and these can be delineated using N:Cl ratios and in one example by using d15N analysis (Faillat, 1990<ref name="Faillat">FAILLAT, J P. 1990. Sources of Nitrates in Fissure Groundwater in the Humid Tropical Zone - the Example of Ivory-Coast. Journal of Hydrology, Vol. 113, 231-264. </ref>).

[[Image:15009_fig18.jpg|thumb|center|500px|'''Figure 18''' Summary water quality results for NO3 from shallow groundwater studies carried out across hydrogeologically relevant terrains in SSA based on climate and geology. Data extracted from tables and figures in peer reviewed literature, some summary statistics (mean) are not available from the literature. W=wells, B=boreholes, *Data from Sierra Leone, note log scale on x-axis.]]

A series of geochemical transformations can occur in water with a high carbon concentrations and a progressive decline in redox potential, leading to the removal of nitrate by denitrification, the mobilisation of manganese and iron and the reduction of sulphate. Borehole mixing processes can cause dilution and overall lower nitrate concentrations while still having significant microbiological contamination. Lagerstedt et al. (1994)<ref name="Lagerstedt"></ref> and Cronin et al. (2007)<ref name="Cronin">CRONIN, A A, PEDLAY, S, HOADLY, A W, KOUONTO KOMOU, F, HALDIN, L, GIBSON, J, and BRESLIN, N. 2007. Urbanisation effects on groundwater chemical quality: findings focusing on the nitrate problem from 2 African cities reliant on on-site sanitation. Journal of Water and Health, Vol. 5, 441-454. </ref> successfully used NO3:Cl to fingerprint different sources of urban and peri-urban pollution in groundwaters in SSA. This has a certain appeal due to its simplicity; however, prevailing redox conditions and mixing processes need to be considered when using this approach. Many studies have effectively used nitrate in combination with other basic physical indicators such as SEC or TDS and turbidity to assess contamination and map areas of relatively high and low pollution.

===Ammonium and phosphate ===
It is evident from the literature that only a minority of case studies (ca. 20%) contain data for NH4 and close to 30% contain data for PO4 (see Table A1). In part this is due to the more involved analytical procedures for NH4, the high detection limits for PO4 by ion chromatography and the fact that these parameters need to be analysed rapidly after sampling to ensure valid results. The WHO have not published health-based guidelines for ammonia and phosphate, but P is often the limiting nutrient in the aquatic environmental and therefore concentrations >20 mg/L are considered high in surface water bodies.

Both species are closely associated with contamination from pit latrines and leaking sewer systems. Examples of ammonia and phosphate contamination from the cities of Lusaka, Abeokuta and Calabar are shown in Table 1 (Berhane and Walraevens, 2013<ref name="Berhane">BERHANE, G, and WALRAEVENS, K. 2013. Geological and geotechnical constraints for urban planning and natural environment protection: a case study from Mekelle City, Northern Ethiopia. Environmental Earth Sciences, Vol. 69, 783-798. </ref>; Cidu et al., 2003<ref name"Cidu">CIDU, R, DE WAELE, J, DI GREGARIO, F, and FOLLESA, R. 2003. Geochemistry of groundwater in an intensely urbanised karst area (Lusaka, Zambia) GeoActa, Vol. 2, 35-42. </ref>; Taiwo et al., 2011<ref name="Taiwo">TAIWO, A M, ADEOGUN, A O, OLATUNDE, K A, and ADEGBITE, K I. 2011. Analysis of groundwater quality of hand-dug wells in peri-urban area of Obantoko, Abeokuta, Nigeria for selected physico-chemical parameters. The Pacific Journal of Science and Technology, Vol. 12, 527-534. </ref>. Ammonium concentrations in groundwater range from BDL-60 mg/L, although most case studies had maximum concentrations below 10 mg/L. The highest concentrations were reported in Lusaka, Zambia where karstic limestone aquifer which underlies much of the city and very rapid transport times in the groundwater are implicated. Both indicators do not behave conservatively in soils and groundwater, for example NH4 is positively charged and therefore has a strong affinity for negatively charged surfaces such as clays, for this reason, as well as microbiological processing, attenuation is particularly high in the soil zone.

Phosphate concentrations range from BLD-86 mg/L, although very few studies report  values >20 mg/L. Phosphate has very limited mobility in the subsurface and has a strong affinity to iron oxy-hydroxides as well as carbonates, background concentrations are usually low, e.g. <0.2 mg/L, concentrations in urban groundwater are also usually low unless there is either a very high loading or very rapid groundwater flow for example in fractured basement or karstic limestone (Cidu et al., 2003<ref name="Cidu">CIDU, R, DE WAELE, J, DI GREGARIO, F, and FOLLESA, R. 2003. Geochemistry of groundwater in an intensely urbanised karst area (Lusaka, Zambia) GeoActa, Vol. 2, 35-42. </ref>; Nkansah et al., 2010<ref name="Nkansah">NKANSAH, M A, BOADI, N O, and BADU, M. 2010. Assessment of the quality of water from hand-dug wells in Ghana. Environmental Health Insights., Vol. 4, 7-12. </ref>; Zingoni et al., 2005<ref name="Zingoni"></ref>).

===Microbiological contaminants ===
Studies have shown that greater than 90% of thermotolerant coliforms (TTCs) are ''E. coli ''(Dufour, 1997 cited in Leclerc et al. (2001)<ref name="Leclerc">LECLERC, H., MOSSEL, D., EDBERG, S AND STRUIJK, C. 2001. Advances in the bacteriology of the coliform group: their suitability as markers of microbial water safety. Annual Reviews in Microbiology 55(1), 201-234. </ref>) and as high as 99% in groundwater impacted by poor environmental sanitation in Africa (Howard et al. 2003<ref name="Howard"></ref>). Despite this there have been some doubts about the reliability of TTCs to indicate faecal contamination in water. Although the TTC group includes the species ''E.coli'', which is generally considered to be specific for faecal contamination, it also includes other genera such as ''Klebsiella ''and ''Citrobacter ''which are not necessarily of faecal origin and can emanate from alternative organic sources such as decaying plant materials and soils (WHO 2011<ref name="Who">WHO. 2011. Guidelines for drinking-water quality (4th edition), WHO Press, Geneva, Switzerland. </ref>).

Human faeces harbour a large number of microbes, including bacteria, archaea, microbial eukarya, viruses, protozoa, and helminths (Graham and Polizzotto, 2013<ref name="Graham"></ref>). In the context of this review there have been no studies that have assessed protozoa or helminths, which exhibit little movement in groundwater due to their size (Lewis et al., 1982<ref name="Lewis 1982">LEWIS, W J, FOSTER, S S, and DRASAR, B S. 1982. The risks of groundwater pollution by on-site sanitation in developing countries: a literature review. (Duebendorf, Switzerland: International Reference Centre for Waste Disposal.) </ref>). The characteristics of microorganisms and the aquifer and soil environment that affect microbial transport and attenuation in groundwater are shown in Table 4.
<center>
{| class="wikitable"
|+ Table 5 Factors affecting transport and attenuation of microorganisms in groundwater (from Pedley et al. (2006)<ref name="Pedley">PEDLEY, S, YATES, M, SCHIJVEN, J F, WEST, J, HOWARD, G, BARRETT, M, SCHMOLL, O, CHILTON, J, and CHORUS, I. 2006. Pathogens: health relevance, transport and attenuation. Protecting groundwater for health: managing the quality of drinking-water sources. (Geneva: WHO.) ISBN 92-4-154668-9</ref>)
| '''Characteristics of the microorganism'''
| '''Aquifer/soil (environment) properties'''
|- valign="top"
| Size<br>shape<br>Density<br>Inactivation rate (die-off)<br>reversible adsorption<br>Physical filtration
| Groundwater flow velocity<br>Groundwater flow velocity<br>Dispersion<br>Pore/aperture size (intergranular or fracture)<br>Kinematic/effective porosity<br>Organic carbon content<br>Temperature<br>Chemical properties of groundwater (pH etc.)<br>Minerals composition of aquifer/soil material<br>Predatory microflore<br>Moisture content<br>Pressure
|}
</center>
In-situ sanitation, largely in the form of pit latrines, is often considered the dominant cause of microbiological contamination and a major cause of nutrient loading to water resources in SSA. This is a very well-studied area and a worldwide review has been published recently by Graham and Polizzotto (2013)<ref name="Graham"></ref>. The main findings from relevant studies carried out in SSA have been collated in Table 5 and are summarised below along with other studies specifically targeting contamination from sanitary sources. Given the low sanitation coverage, and the wet climate of Sierra Leone, surface sources such as open defecation are also significant, but it is noteworthy that there have been very few studies that have considered this a major source of contamination in this region.

[[Image:15009_fig19.jpg|thumb|center|500px|'''Figure 19''' Summary water quality results for faecal coliforms from shallow groundwater studies carried out across hydrogeologically relevant terrains in SSA based on climate and geology. Data extracted from tables and figures in peer reviewed literature, some summary statistics (mean) are not available from the literature. W=wells, B=boreholes, S=springs, *Data from Sierra Leone.]]

Figure 19 shows summary statistics for FC contamination (cfu /100 mL) found in shallow groundwater sources from representative case studies across SSA, from both sedimentary and basement settings. There is evidence of widespread contamination in shallow groundwater sources with mean FC ranging from >10-10 000 (cfu/ 100mL). There is no significant difference between the level of contamination found in sedimentary and basement terrains from shallow wells (see Figure 19).

===Seasonal trends in groundwater quality ===
A recent review by Kostyla et al (2015) found significant seasonal trends of greater faecal contamination in developing countries during the wet season irrespective of source type, climate and population. However, there are relatively few studies that have undertaken regular water quality monitoring over extended periods or have carried out detailed seasonal comparisons in Africa. An early study by Wright (1986)<ref name="Wright 1986"></ref> in Sierra Leone showed that wells and springs had a pronounced seasonality, with higher counts for FC and FS progressively during the dry season and reduced counts at the start of the wet season, dilution was implied as a controlling factor, which is likely given the strong seasonality in rainfall and the fact that sanitation was non-existent and open defecation was practised. A study by Howard et al. (2003)<ref name="Howard"></ref> is one notable example where detailed seasonal monitoring of microbiological indicators was carried out over a twelve month period to characterise the risks factors for spring contamination in Kampala, Uganda. Significantly higher contamination was observed after rainfall events and there was strong evidence that rapid recharge of the shallow groundwater causes a rapid response in spring quality (Barrett et al., 2000<ref name="Barratt">BARRETT, M H, JOHAL, K, HOWARD, G, PEDLEY, S, and NALUBEGA, M. 2000. Sources of faecal contamination of shallow groundwater in Kampala. 691-696 in Groundwater: past achievements and future challenges. SILILO, O T N, APPLEYARD, S, and BARRETT, M (editors). (Rotterdam: Balkema.) </ref>). Godfrey et al (2006)<ref name="Godfrey"></ref> collected data on TTC and enterococci monthly for a year, these results showed that microbiological contamination was enhanced in the rainy season and in the lead up to the rains, which could also be liked to well use and demand during this period as was suggested in the study by Wright (1986)<ref name="Wright 1986"></ref>. A recent study by Nyenje et al (2013)<ref name="Nyenje"></ref> showed that nitrate concentrations up-gradient and down- gradient of pit latrines over a four month period showed large seasonal changes, the data suggest that dilution from intense rainfall and recharge may be an important control.

Higher maximum FS (faecal streptococci) counts were found in the wet season compared to the dry season for studies in Uganda (Kulabako et al., 2007<ref name="Kulabako"></ref>) and Malawi (Palamuleni, 2002<ref name="Palamuleni">PALAMULENI, L G. 2002. Effect of sanitation facilities, domestic solid waste disposal and hygiene practices on water quality in Malawi’s urban poor areas: a case study of South Lunzu Township in the city of Blantyre. Physics and Chemistry of the Earth, Parts A/B/C, Vol. 27, 845- 850. </ref>). Higher maximum SEC were observed in all three case studies in the wet season, however median values are comparable. Changes in nitrate show a mixed picture with higher maximum concentrations in two studies from Uganda and DRC (Kulabako et al., 2007<ref name="Kulabako"></ref>; Vala et al., 2011<ref name="Vala"></ref>) during the wet season, while in the case study from Zimbabwe (Mangore and Taigbenu, 2004<ref name="Mangore">MANGORE, E, and TAIGBENU, A. 2004. Land-use impacts on the quality of groundwater in Bulawayo. Water SA, Vol. 30, 453-464. </ref>) lower maximum values were found (Table 1A). Median values for nitrate were lower in the wet season for both the Uganda and Zimbabwe case studies, which may indicate a dilution effect, while the higher maximum concentrations may be explained as a result of a pulse of contaminants at the start of the rainy season, evaporative effects concentrating N during the dry season or the rise in groundwater table picking up a plume of high N water in the unsaturated zone. Understanding seasonal trends in nitrate are complicated by the changes in redox conditions, particularly in low lying areas which are prone to flooding in the wet season which are not uncommon in SSA, e.g. Lusaka, Zambia. These may shift from an oxidising regime in during low water table conditions which retains NO3 to a reducing regime where denitrification can take place during inundation (Sanchez-Perez and Tremolieres, 2003<ref name="Sanchez">SANCHEZ-PEREZ, J M, and TREMOLIERES, M. 2003. Change in groundwater chemistry as a consequence of suppression of floods: the case of the Rhine floodplain. Journal of Hydrology, Vol. 270, 89-104. </ref>; Spalding and Exner, 1993<ref name="Spalding">SPALDING, R F, and EXNER, M E. 1993. Occurrence of Nitrate in Groundwater - a Review. Journal of Environmental Quality, Vol. 22, 392-402. </ref>).

{| class="wikitable"
|+ Table 6 Comparison of microbiological water quality from multiple groundwater sources including boreholes, wells and springs.

| '''Town/city/area'''
| '''Country'''
| '''Geology/sites'''
| colspan="3" | '''Water Quality (cfu/100 mL)'''
| '''Contamination'''
| '''Reference'''
|-
| Oju area
| Nigeria
| Sedimentary n=30
| Borehole: FC BDL-500 typically <200
| Improved well: FC 50-500 typically >200
| Trad. Well: FC>500
| Borehole<<improved well<<traditional well
| Bonsor et al. (2014)<ref name="Bonsor"></ref>
|-
| Yaounde
| Cameroon
| Basement n=40
| Spring: FC 2-72 FS 0
| colspan="2" | Well: FC 7-100 FS 0-100
| Spring<Well
| Ewodo et al. (2009)<ref name="Ewodo">EWODO, M. G, EKWELGEN, C, NTEP, F, AND EKODECK, G.E. 2009. Impact of urbanisation on the Mingosso watershed in the Yaounde periurban zone. African Journal of Environmental Science and Technology, 3(10). </ref>
|-
| Kumasi
| Ghana
| Basement n=9
| Well: FC  mean >30k, EC=0-1152
| colspan="2" | Borehole FC mean>20k EC 0-36
| Borehole<Well
| Obiri-Danso et al. (2009)<ref name="Obiri">OBIRI-DANSO, K. S, ADJEI, B, STANLEY, K, AND JONES, K. 2009. Microbiological quality and metal levels in wells and boreholes water in some peri-urban communities in Kumasi, Ghana. African Journal of Environmental Science & Technology, 3(3), 59-66. </ref>
|-
| Blantyre
| Malawi
| Basement n=9
| Borehole: FC 0-30 FS 0
| Spring: FC 530-9500 FS 0-7000
| Wells: FC 3500-11k FS 250-2650
| Borehole<<Spring<Well
| Palamuleni (2002)<ref name="Palamuleni"></ref>
|-
| Njala
| Sierra Leone
| Basement n=8
| Spring: FC 50-30k FS 8-2500
| colspan="2" | Wells: FC 125-63k FS 5-2500
| Spring<Well
| Wright (1986)<ref name="Wright 1986"></ref>
|-
| Kampala
| Uganda
| Basement n=16
| Spring: FC 29-10k FS 6-8.3k
| colspan="2" | Wells: FC 0-26^6 FS 0-26^8
| Spring<<Wells
| Kulabako et al. (2007)<ref name="Kulabako"></ref>
|-
| Harare
| Zimbabwe
| Basement n=29
| Borehole: FC 0-30k
| colspan="2" | Well: FC 0-30k
| Borehole<Well for FC
| Zingoni et al. (2005)<ref name="Zingoni"></ref>
|-
| Douala
| Cameroon
| Sedimentary n=4
| Spring: FC 1-950 FS 0-420
| colspan="2" | Borehole: FC 1-2.3k FS 0-1.4k
| Spring<Borehole
| Takem et al. (2010)<ref name="Takem">TAKEM, GE., CHANDRASEKHARAM, D, AYONGHE, SN, AND THAMBIDURAI, P. 2010. Pollution characteristics of alluvial groundwater from springs and bore wells in semi-urban informal settlements of Douala, Cameroon, Western Africa. Environmental Earth Sciences, 61(2), 287- 298. </ref>
|-
| Kabwe
| Zambia
| Karstic n= 75
| Borehole FC<2-630
| colspan="2" | Well: FC <2-28k
| Borehole<Well
| Lapworth et al (2015b)<ref name="Lapworth 2015b"></ref>
|}
FC= Feacal coliforms, FS=Feacal strep., EC=Entrococci, TC=Total coliforms

===A comparison of results from wells, springs and boreholes===
The vast majority of the studies that are included in this review contain data from shallow hand dug wells (ca. 60%), this is true of most published water quality studies in Africa, a further 22% include data from boreholes and 18% include results from springs. A small number of studies have compared a range of different groundwater sources, usually two different sources; boreholes vs wells (n=8) and wells vs springs (n=5) and boreholes vs springs (n=4). Table 6 summarises the results from comparative studies with two or more groundwater source types.

As you might expect, overall wells are generally the most contaminated groundwater source type compared to springs and boreholes. Open and unlined wells are consistently of poorer quality compared to lined or ‘improved’ wells (e.g. Godfrey et al 2006<ref name="Godfrey"></ref>; Jimmy et al., 2013<ref name="Jimmy"></ref>; Lapworth et al., 2015b<ref name="Lapworth 2015b">LAPWORTH, D J, NKHUWA DCW, SORENSEN J, BELL R, PEDLEY, S, READ D. 2015b. Summary data report for Kabwe groundwater quality assessment 2013-14. British Geological Survey Draft Open report OR/15/012 [Draft]. 50pp. (unpublished) </ref>). In some studies springs have been found to be better quality compared to boreholes (Takem et al. 2010<ref name="Takem"></ref>) and others cases the trend is reversed (Palamuleni 2002<ref name="Palamuleni"></ref>) or both sources were found to have comparable levels of contamination by FC (e.g. Abiye 2008). It is important to note that many of these studies contained very few observations for each source type and generalisations should be treated with caution however, together they form a more compelling body of evidence. Overall there is no clear patterns that emerge regarding water quality in different hydrogeological settings, i.e. basement or sedimentary, comparable mean and maximum levels of contamination are found for FCs and nitrate. With perhaps the exception of highly karstic settings for microbiological and nitrate the following order of water source quality (best to worst) is found as follows: boreholes >> improved wells = springs > traditional wells. Improved wells do not generally exhibit the same level of gross contamination observed in traditional wells and springs. However, in the majority of studies, wells (both improved and unimproved), are found to have water with unacceptable levels of contamination with faecal coliforms by WHO standards (and typically > 100 cfu/ 100 mL) in at least some part of the year and often throughout the year.

There is some evidence that the water quality of wells may be affected by usage rates, i.e. with fewer groundwater sources being relied on towards the end of the dry season there is greater risk of contamination, e.g. from materials used for drawing water, especially for unimproved sources (Godfrey et al. 2006<ref name="Godfrey"></ref>; Wright et al., 1986<ref name="Wright 1986"></ref>). For boreholes this contamination pathway is generally not a major risk factor and this supports the generally better quality found in these types of sources. The lower storage volume of shallow boreholes compared to wells may also be an important factor.

{|class="wikitable" style="width:700px" style="background:#f0f8ff; font-size:100%"
| '''Water quality summary:'''<br>
There is little convincing evidence that water quality is consistently better at distances >30 m from individual pit latrines – what evidence exists suggests a link with density of contaminant sources.<br><br>The water quality of  shallow  groundwater  accessed  by shallow  wells  is  often  of  very poor quality, based on faecal coliform and nitrate data, for at least some of the year in most settings, and all year in many cases.<br><br>Water  quality  from  boreholes  is  generally  of  better  quality  compared  to  wells  and  springs probably because  it  accesses  deeper  groundwater  and  has  better  protection  around  the  well head.<br><br>Wells are highly vulnerable to microbiological hazards, particularly surface material introduced by rope and buckets. There are significant seasonal changes in  water quality in wells, with generally poorer water quality observed at the end of the dry season and during the onset of the wet season.<br><br>Seasonal pressures on particular water sources may increase the likelihood of water quality deterioration in wells (and spring collectors).
|}

==Pathogen survival==
===Introduction===
Pathogens contaminate the subsurface from many different sources: leaking sewers; septic tanks; surface application of faecal sludge in agriculture; surface waste; and pit latrines to name but a few. Once in the soil layer, or having reached the groundwater, pathogens are subject to a broad range of environmental factors that dictate the survival time of the pathogen and the distance that it can migrate from its source. In addition, the nature of the pathogen itself determines its interaction with the environment, and thus its survival and mobility in the subsurface. In the broadest sense, there are three main groups of pathogens that that are of concern in groundwater: viruses; bacteria and protozoa (Table 7). The characteristics of each group of pathogens are quite distinct, which contributes to their different behaviours in the environment.

Groundwater has been identified as the vehicle of pathogen transmission in numerous outbreaks of waterborne disease. In the USA and elsewhere, summaries of the sources of waterborne disease highlight the importance of groundwater. Statistics collected by the US water-borne Disease Outbreak Surveillance System between 1971 and 2008 show that 30% of the 818 outbreaks of disease were a result of supplying untreated drinking water from groundwater sources (Wallender et al. 2013<ref name="Wallender">WALLENDER, E K, ET AL., 2013. Contributing Factors to Disease Outbreaks Associated with Untreated Groundwater. Groundwater, 52, 6, pp.886–897. </ref>; Craun et al. 2010<ref name="Craun">CRAUN, G F, ET AL. 2010. Causes of outbreaks associated with drinking water in the United States from 1971 to 2006. Clinical Microbiology Reviews, 23, 3, pp.507–528. </ref>). Over a similar time period in Norway 44% of water-borne disease outbreaks could be linked to groundwater sources (Kvitsand & Fiksdal 2010). There are fewer examples of outbreaks attributable to groundwater being reported in Sub-Saharan Africa (SSA), possibly due to the numerous confounding factors present in SSA that complicate the exposure- risk relationships (Payment & Hunter 2001<ref name="Payment">PAYMENT, P and HUNTER, P R. 2001. Endemic and epidemic infectious intestinal disease and its relationship to drinking-water. In L. Fewtrell and J. Bartram, eds. Water Quality: Guidelines, Standards and Health. London: IWA Publishing, pp. 61–88. </ref>), but the widespread and high levels of contamination in water from hand-dug wells and shallow boreholes in urban and rural areas means that inevitably they will be source of disease transmission (Kimani-Murage & Ngindu 2007)<ref name="Kimani"></ref>. Given the  extent to which these sources are used in SSA, the burden of disease attributable to the consumption of contaminated groundwater will be high.

The duration and extent of the recent outbreak of Ebola in Western Africa has raised questions about the possible importance of environmental sources of the virus and whether it might persist in body fluids long enough to present a risk of transmission by indirect routes, including the contamination of water. Furthermore, the sudden and necessary diversion of medical attention towards the control of Ebola in the affected countries has caused concern amongst some that the classical water-related diseases are being ignored, and that there might be a silent increase in their prevalence.

This section of the report summarises the current knowledge about the survival of pathogens in the sub-surface, and the factors that contribute to their dispersal through groundwater. Our review will draw upon two relatively recent published reviews of the fate and transport of pathogens in groundwater (Pedley et al. 2006<ref name="Pedley"></ref>; Tufenkji & Emelko 2011<ref name="Tufenkji">TUFENKJI, N and EMELKO, M B. 2011. Fate and Transport of Microbial Contaminants in Groundwater. Encyclopedia of Environmental Health, pp.715–726. </ref>) to create the foundation of this report, and expand upon the reviews with more recent significant findings. Both of these reviews provide the reader with a link to the early, but still relevant literature.

For the purpose of this report we will concentrate our discussion upon the viral and bacterial pathogens – in particular those pathogens that may inform conclusions about the potential for Ebola virus and ''Vibrio cholerae ''to survive and migrate through groundwater – in the context of the environmental conditions that exist in Sierra Leone. This section will firstly cover a short summary of the physical and chemical characteristics of the cholera vibrio and the Ebola virus so that parallels can be drawn with alternative microorganisms that may be used as their surrogates in a risk assessment. This will be followed by a review of the factors that influence the length of time that bacteria and viruses survive in the subsurface and the characteristics of the pathogens and the environment that control the movement of the pathogens through groundwater. Taken together, the survival and transport data can be used to estimate the extent to which the pathogens may disperse in groundwater from a particular source.

{| class="wikitable"
|+ Table 7 Characteristics of the major pathogen groups

| '''Pathogen group'''
| '''Characteristics'''
|-
| Bacteria
| Bacteria are prokaryotic microorganisms, which means that they do not have a defined nuclear membrane (lack an identifiable nucleus) or other organised intracellular structures. Although their size varies considerably between species, individual cells range in width between 0.5µm and 5.0µm. Bacteria are ubiquitous, and can colonise the most extreme environments. The vast majority are harmless saprophytes. They can have a variety of shapes, and some species are motile. Apart from a few exceptions, the bacterial cell contains all the cellular components necessary for to metabolise nutrients to generate energy, and for it to replicate. This characteristic means that some pathogens may be able to maintain themselves in the environment when the conditions are favourable to their replication. Some species of bacteria produce spores that are highly resistant to environmental stress and may survive for years, even decades. Other bacteria may enter a dormant state when the environmental conditions are unfavourable. The significance to human health of this dormant state is being investigated.
|-
| Viruses
| Viruses are obligate intracellular parasites, which means that they have an absolute requirement to infect a host cell in order to replicate. Outside the cell they are dormant. Hence, once a virus has been expelled from the host into the environment it cannot replicate itself. Viruses are orders of magnitude smaller that bacteria – between 20nm and 300nm – and have a much simpler structure. For some virus species, for example the enteroviruses, their simplicity makes them particularly resistant to environmental stress, and they can survive considerable amounts of time when the environmental conditions are favourable. Some pathogenic viruses capture a portion of the host cell membrane when they replicate (e.g. measles virus, mumps virus, influenza virus), but others remain uncoated (e.g adenovirus, norovirus, poliovirus). The surface of the latter group carries a charge derived from the relative levels of ionisation of the amino and carboxyl groups in the proteins that encase the nucleic acid. The nett charge on the surface is a function of the composition of these proteins and the pH and the ionic strength of the surrounding medium.
|-
| Protozoa
| Protozoa are single-celled, eukaryotic microorganisms. Unlike bacteria, the cell has a defined nucleus surrounded by a nuclear membrane, and the identifiable intracellular organelles. There are a number of pathogenic species, although Giardia, Cryptosporidium and Entamoeba are the most frequently referenced in relation to water- borne disease. However, recently there has been a growing interest in the waterborne transmission of ''Toxoplasma gondii''. Protozoa are ubiquitous in water and soils. Most of the enteric protozoa produce cycts, or oocysts, as part of their life cycle. Cysts are a dormant form of the organism that play an important role in the transmission of the pathogen. Cysts (and oocysts) are highly resistant to environmental stress, remaining viable for several months at low temperatures, and can often survive the normal doses of chlorine used to disinfect drinking water. Cysts vary in size depending on the species of protozoa, but for Giardia, Cryptosporidium and Entamoeba they are in the range of 4µm to 20µm in diameter.
|}

===Characteristics of Ebola virus and ''Vibrio cholera''.===
====EBOLA VIRUS====
Ebola virus belongs to the family ''Filoviridae''. The genome of the virus is a single strand of negative- sense RNA.  The virus has a pleomorphic structure which folds to form the characteristic 'U'  or '6' shapes that are seen in electron micrograph images. The virus capsid is surrounded by a lipid membrane that it derives from the infected host cell. The virion is approximately 80nm in diameter, but can vary considerably in length, occasionally reaching 14µm (Anon 2014<ref name="Anon">ANON, 2014. Ebola (Ebola Virus Disease). Available at:http://www.cdc.gov/vhf/ebola/about.html [Accessed January 28, 2015]. </ref>).

Three characteristics are particularly important when looking for possible surrogates to help gauge the survival and transport potential of the virus: the presence of a modified lipid envelope; the capsid structure; and the RNA genome. The virus envelope plays a crucial role in the process of infection by attaching to the surface of the target cell and then fusing with the cell membrane to release the virus capsid and nucleic acid into the cell cytoplasm, where new copies of the virus will be generated. But membranes can be fragile, and the envelope surrounding the virus may be particularly vulnerable to environmental conditions.

Very few survival studies have been attempted with Ebolavirus, and the ambient conditions used for the experiments were not representative of the environmental conditions in SSA. Consequently, only limited conclusions can be drawn from these studies.  In the dark and at 20°C–25°C, Ebolavirus infectivity was reduced by 1 log10 in 35 hours and by 4 log10 in 6 days when incubated on a number of surfaces (Sagripanti et al. 2010<ref name="Sagripanti">SAGRIPANTI, J L, ROM, A M, and HOLLAND, L E. 2010. Persistence in darkness of virulent alphaviruses, Ebola virus, and Lassa virus deposited on solid surfaces. Archives of Virology, 155, pp.2035–2039. </ref>). This inactivation rate is similar, although slightly longer, than the inactivation rate published by Smither et al (2011)<ref name="Smither">SMITHER, S J, ET AL. 2011. An alternative method of measuring aerosol survival using spiders’ webs and its use for the filoviruses. Journal of Virological Methods, 177(1), pp.123–127. Available aTemplate:Doi</ref> who was studying survival in aerosols. Inactivation rates in the environment would be expected to be greater due to the sensitivity of the virus to UV irradiation (Sagripanti and Lytle, 2011<ref name="Sagripanti 2011">SAGRIPANTI, J L, and LYTLE, C D. 2011. Sensitivity to ultraviolet radiation of Lassa, vaccinia, and Ebola viruses dried on surfaces. Archives of Virology, 156, pp.489–494. </ref>). Given the difficulty of working with Ebolavirus, estimates of environmental survival times might be derived from studies of other virus groups with similar characteristics: single-stranded, negative sense RNA; pleomorphic capsid; and lipid envelope. Viruses of the family paramyxoviridae (for example, measles virus, mumps virus, Hendravirus and Nipahvirus) share these characteristics. Ecologically, Nipahvirus and Hendravirus have even greater similarities to Ebolavirus, being a recently emerged zoonotic disease with a natural reservoir in bats (Fogarty et al. 2008<ref name="Fogarty">FOGARTY, R, ET AL. 2008. Henipavirus susceptibility to environmental variables. Virus Research, 132, pp.140–144. </ref>; Scanlan et al. 2014<ref name="Scanlan"></ref>). Laboratory studies have shown that the survival of Hendravirus is inversely related to temperature. At 4°C, 22°C and 56°C, the half-life of the virus was 308, 50.2 and 1.85 hours respectively (Scanlan et al. 2014<ref name="Scanlan">SCANLAN, J C, ET AL. 2014. Survival of Hendra Virus in the Environment: Modelling the Effect of Temperature. EcoHealth. </ref>). Hendravirus is also highly sensitive to desiccation, surviving for less than two hours under these conditions (Fogarty et al. 2008<ref name="Fogarty"></ref>). Environments with a low relative humidity (between 20% and 30%) generally favour the survival of enveloped viruses (Tang 2009<ref name="Tang">TANG, J W. 2009. The effect of environmental parameters on the survival of airborne infectious agents. Journal of the Royal Society, Interface / the Royal Society, 6 Suppl 6(September), pp.S737–S746. </ref>), suggesting that the Ebolavirus might be less stable in regions with high relative humidity.

No reports of the water-borne transmission of paramyxoviruses could be found; however, several reports have been made of the potential for avian influenza virus – another RNA enveloped virus to be transmitted through water (Hinshaw et al. 1979<ref name="Hinshaw">HINSHAW, V S, WEBSTER, R G and TURNER, B. 1979. Water-Borne Transmission of Influenza A Viruses? Intervirology, 11(1), pp.66–68. Available at:http://www.karger.com/DOI/10.1159/000149014.</ref>; Achenbach & Bowen 2011<ref name="Achenback">ACHENBACH, J E and BOWEN, R A. 2011. Transmission of avian influenza a viruses among species in an artificial barnyard. PLoS ONE, 6, 3. </ref>; Brown et al. 2007<ref name="Brown">BROWN, J D, ET AL. 2007. Persistence of H5 and H7 avian influenza viruses in water. Avian Diseases, 34, pp.406–411. </ref>). Avian strains of influenza maybe an anomaly among enveloped viruses because the host lives on or near water and the virus may have evolved to use water as a transmission route.

====V.CHOLERAE====
Cholera is a disease of antiquity that was first described over 2000 years ago, although the causative agent was not identified until the mid/late 19th Century. ''V.cholerae ''is a Gram-negative bacillus (Gram-negative is one of two outcomes of a diagnostic staining technique widely used in microbiology. Under the microscope Gram-negative cells are red whereas Gram-positive cells are dark blue/purple. Bacillus simply means rod-shaped) that has a very characteristic “comma” shape when viewed under the microscope. The bacterium is a facultative anaerobe, which means that it can grow in environments with and without oxygen (Valdespino & Garcia-Garcia 2011<ref name="Valdespino">VALDESPINO, J L, and GARCIA-GARCIA, L. 2011. Cholera: Environmental Risk Factors. Encyclopedia of Environmental Health, pp.641–649. </ref>). ''V.cholerae ''has a single flagellum (a hair-like structure) at one end of the cell that is used to propel the cell through water. Consequently, ''V.cholerae ''is motile and can move itself within its immediate environment (Janda 1998<ref name="Janda">JANDA, J M. 1998. Vibrio, Aeromonas and Plesiomonas. In A. Balows and B. I. Duerden, eds. Topley and Wilson’s Microbiology and Microbial Infections. Volume 2: Systematic Bacteriology. London: Arnold, pp. 1065–1089. </ref>). The relevance of this facility to the potential dispersal of the organism in groundwater is unknown, but it is very unlikely to make a significant contribution particularly in flowing water systems.

Members of the genus ''Vibrio ''are normal inhabitants of marine environments, and water bodies that are immediately in contact with marine environments such as estuaries. Unlike some other vibrio species ''V.cholerae ''does not have an absolute requirement for a saline environment, so this organism can also be isolated from freshwater where the saline conditions are replaced by warmth and organic nutrients (Janda 1998<ref name="Janda"></ref>; Jutla et al. 2013<ref name="Julta">JUTLA, A, ET AL. 2013. A water marker monitored by satellites to predict seasonal endemic cholera. Remote sensing letters (Print), 4(February 2015), pp.822–831. Available at:http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=3714106andtool=pmcentrezandrende rtype=abstract. </ref>; Rebaudet et al. 2013a<ref name="Rebaudet 2013a">REBAUDET, S, ET AL. 2013a. Cholera in Coastal Africa: A systematic review of its heterogeneous environmental determinants. Journal of Infectious Diseases, 208(Suppl 1), pp.S98–S106. </ref>). Several publications since the mid-1990s have developed the idea of cholera epidemiology being linked to coastal aquatic environments and the abundance indigenous phytoplankton (cited in Jutla et al. 2013<ref name="Julta"></ref>). The apparent correlation with phytoplankton abundance has spurred a number of studies into the use of satellite imagery to monitor the phytoplankton and provide an early warning of cholera outbreaks (Jutla et al. 2013<ref name="Julta"></ref>). Two recent reviews of the distribution of cholera outbreaks in Africa suggests that the coastal link may not be as strong as it is in other parts of the world (Rebaudet et al. 2013a<ref name="Rebaudet 2013a"></ref>; Rebaudet et al. 2013b<ref name="Rebaudet 2013b">REBAUDET, S, ET AL. 2013b. Environmental determinants of cholera outbreaks in inland africa: A systematic review of main transmission foci and propagation routes. Journal of Infectious Diseases, 208(Suppl 1). </ref>): only a minority of total recorded cases could be attributed to coastal areas (Rebaudet et al. 2013b<ref name="Rebaudet 2013b"></ref>). Although the Great Lakes Region and the Lake Chad basin appear to have a particularly high number of cases, and that outbreaks seem to occur with the rainy season, ''V.cholerae ''has rarely been isolated from water samples (Rebaudet et al. 2013b<ref name="Rebaudet 2013b"></ref>). These authors report only one incidence of the vibrio being isolated from a well.

Certain species of bacteria can survive prolonged exposure to adverse environmental conditions by producing spores. These spores can survive for years, sometimes decades, and germinate under the right conditions. Gram-negative bacteria do not adopt this strategy, but appear to undergo a cellular transformation that puts them into a dormant state where they are metabolically viable but cannot be grown by standard laboratory culture methods. This state is known as Viable but Non- Culturable (VNC). ''V.cholerae ''has been show to enter a VNC state in water where it may play an important role in the initiation of epidemics (Alam et al. 2007<ref name="Alam">ALAM, M, ET AL. 2007. Viable but nonculturable Vibrio cholerae O1 in biofilms in the aquatic environment and their role in cholera transmission. Proceedings of the National Academy of Sciences of the United States of America, 104, 45, pp.17801–17806. </ref>). During outbreaks of cholera in crowded urban slums it is inevitable that the local environment will become contaminated with the pathogen, and it is highly likely that it will contaminate vulnerable shallow wells (Momba et al, 2006<ref name="Momba">MOMBA, M N B, MALAKATE, V K, and THERON, J. 2006. Abundance of pathogenic Escherichia coli, Salmonella typhimurium and vibrio cholerae in Nkonkobe drinking water sources. Journal of Water and Health, 4, pp.289–296. </ref>; Rebaudet et al. 2013b<ref name="Rebaudet 2013b"></ref>). ''V.cholerae ''can survive in freshwater, particularly when the water is contaminated with organic nutrients, as might be expected of many of the wells in urban slums. Its potential for long-term survival in these conditions is increased if the bacterium enters a VNC state, and the VNC state may present a risk to human health. The presence of biofilms on the walls of the well may also create an environment that allows the cholera vibrio to extend its survival time (Alam et al. 2007<ref name="Alam"></ref>). However, this is speculative and there appears to be very little firm evidence to suggest that groundwater is an important vehicle for the transmission of cholera, and no evidence was found to indicate that ''V.chloerae ''is transported through groundwater.

===Factors affecting the survival and mobility of bacteria and viruses in the subsurface ===
Studies of virus survival and mobility in the subsurface have been carried out using non-enveloped viruses. Ebola is an enveloped virus so it is difficult to say how relevant the findings from current literature will be. Groundwater is widely used as a source of water for drinking, agriculture and industry. Globally, it is estimated that two billion people rely on groundwater (Tufenkji & Emelko 2011<ref name="Tufenkji"></ref>). Groundwater has always been considered to be of a better quality than surface water due to the protection given by the soil layers that restrict the ingress of microbial pollutants. From this perspective, groundwater is often consumed untreated or is given a minimum amount of treatment, such as chlorination. But frequently this is insufficient to prevent outbreaks of disease. Techniques that help to reduce the likelihood of contamination at the point of abstraction, such as groundwater protection zones that are applied in many developed countries, depend on an understanding of the survival and mobility of pathogens in the groundwater systems that are being used. To this end there has been a substantial amount of work carried out in the laboratory and at field sites to build an understanding of the most important factors that contribute to pathogen survival and transport (Table 8). The data from these studies inform models that can then be used to predict the risks of contamination at different points away from the source.

{| class="wikitable"
|+ Table 8      Factors that influence the survival and mobility of bacteria and viruses in the subsurface (adapted from Pedley et al. (2006)<ref name="Pedley"></ref>.
| rowspan="2" | Factor
| colspan="2" | '''Viruses'''
| colspan="2" | '''Bacteria'''
|-
| '''Influence on survival'''
| '''Influence on migration'''
| '''Influence on survival'''
| '''Influence on migration'''
|-
| Temperature
| Longer survival at low temperatures
| Unknown
| Longer survival at lower temperatures
| Unknown
|-
| Microbial activity and diversity
| Varies: Some viruses are inactivated more readily in the presence of certain microorganisms; some are protected; and for some there is no effect.
| Retard migration via attachment to biofilms
| The presence of indigenous microorganisms appears to increase the inactivation rate of enteric bacteria. Community diversity rather than microbial density is important and certain species may have a greater inhibitory effect. Several mechanisms will be involved. Biofilms may harbour pathogens and either extend or limit their survival.
| Biofilms
|-
| Moisture content
| Most viruses survive longer in moist soils and even longer under saturated conditions; unsaturated soil may inactivate viruses at the soil water interface.
| Virus migration usually increases under saturated flow conditions.
| Most bacteria survive longer in moist soils than in dry soils.
| Bacterial migration usually increases under saturated flow conditions.
|-
| pH
| Most enteric viruses are stable over pH range of 3 to 9; however, survival may be prolonged by near neutral pH values.
| Low pH typically increases virus sorption to soils; high pH tends to cause desorption and facilitates greater migration.
| Most enteric bacteria will survive longer at near neutral pH.
| Low pH encourages adsorption to the soils and the aquifer matrix; the tendency of bacteria to bind to surfaces and form biofilms may reduce detachment at high pH.
|-
| Dissolved Oxygen
| Possible decrease inactivation in anaerobic water
| Unknown
| Faster death rates at low DO levels.
| Varies. Some bacteria are retained more strongly under low DO conditions, whereas others migrate farther.
|-
| Salt species and concentration
| Certain cations may prolong survival depending upon the type of virus.
| Increasing ionic strength of the surrounding medium generally increases sorption.
| Generally unknown. Cholera vibrios have a preference for saline conditions, but are able to survive In freshwater. Saline groundwater may extend the survival of ''V.cholerae.''
| Increasing ionic strength of the surrounding medium generally increases sorption.
|-
| Association with soil/aquifer matrix
| Association with soil generally increases survival, although attachment to certain mineral surfaces may cause inactivation.
| Viruses interacting with soil particles are retained at the point of attachment.
| Adsorption onto soil surfaces reduce inactivation rates. The number of bacteria on surfaces may be several orders of magnitude higher than the concentration in the aqueous phase.
| Interaction with the soil inhibits migration.
|-
| Soil properties
| Probably related to the degree of virus sorption.
| Preferential flow pathways through soils (Artz et al. 2005<ref name="Artz">ARTZ, R R E, ET Al. 2005. Soil macropores and compaction control the leaching potential of Escherichia coli O157:H7. Environmental Microbiology, 7, pp.241–248. </ref>). Small differences in the internal structure of soil cores can have a big effect on migration. Soils with charged surfaces, such as clays, adsorb viruses.
| Probably related to the degree of bacterial adsorption.
| Preferential flow pathways through soils (Artz et al. 2005<ref name="Artz"></ref>).  Small differences in the internal structure of soil cores can have a big effect on migration. Soils with charged surfaces, such as clays, adsorb bacteria.
|-
| Bacteria/virus type
| Varies between different virus types.  Possible that the process of inactivation is gradual and may be reversible under certain conditions (Alvarez et al. 2000<ref name="Alvarez">ALVAREZ, M E, ET AL. 2000. Inactivation of MS-2 phage and poliovirus in groundwater. Canadian Journal of Microbiology, 46, pp.159–165. Available at:http://www.nrcresearchpress.com/doi/abs/10.1139/w99-128#.VMjrRtLkd4c.</ref>).
| Sorption to soils is related to physico-chemical differences in the secondary and tertiary capsid structure, the presence or absence or absence of a membrane envelope, and amino acid sequence.
| Varies between different species. Some species are able to enter a dormant state (Viable Non- Culturable) that may extend their survival in the sub surface. Some indications that VNC cells may be of health significance.
| Some species of bacteria are more capable of binding to surfaces; variation may also occur between strains of the same bacterial species. Some bacterial species are motile and may respond to physical or chemical stimulants. Motility unlikely to be significant in the dispersal of bacteria at scale.
|-
| Organic matter
| Organic matter may prolong survival by competitively binding at air-water interfaces where inactivation can occur.
| Soluble organic matter competes with the viruses for adsorption on to sol particles which may result in increased virus migration.
| The presence of organic matter may act as a nutrient source for bacteria, promoting growth and extending survival.
| Organic matter may condition solid surfaces and promote bacterial adsorption.
|-
| Hydraulic conditions
| Unknown
| Virus migration generally increased at higher hydraulic loads and flow rates.
| Unknown
| Bacterial migration generally increased at higher hydraulic loads and flow rates.
|-
| Clay minerals and colloids
| In combination with other factors, virus survival is affected by the type of clay mineral.
| Clay minerals strongly adsorb viruses and will restrict the mobility of viruses. Attachment to colloids may further restrict mobility; however, there is evidence that colloids may increase the mobility of attached pathogens in groundwater
| Unknown
| Bacteria can adsorb to clay minerals, which may restrict their mobility on the subsurface. Attachment to colloids may further restrict mobility
|}

====TEMPERATURE====
Temperature is probably the most important factor influencing the inactivation of bacteria and viruses in the environment (Pedley et al 2006<ref name="Pedley"></ref>). Inactivation rates at particular temperatures are different for bacteria and viruses and will vary considerably between different species of bacteria and viruses; however, the general trend is for a direct correlation between temperature and inactivation rate. With a few exceptions (for example, Vinten et al. 2002<ref name="Vinten">VINTEN, A J A, ET AL. 2002. Fate of Escherichia coli and Escherichia coli O157 in soils and drainage water following cattle slurry application at 3 sites in southern Scotland. Soil Use and Management, 18, pp.223–231. Available at: http://doi.wiley.com/10.1079/SUM2002114\nISI:000178379200009. </ref>) bacteria and viruses tend to survive longer at lower temperatures. This trend is more apparent for some organisms than for others. Figure 20 and Figure 21 below show the link between water temperature and the inactivation rate coefficient for bacteriophage MS2 and Poliovirus 1. The data for these figures was taken from a table of compiled inactivation rate coefficients published in Pedley et al (2006)<ref name="Pedley"></ref>. Poliovirus 1 and MS2 were the only two microorganisms, including bacteria, for which a reasonable number of studies had published the inactivation rates at different temperatures. The trend of higher inactivation rates at higher temperatures is clear for MS2 but much less so in the case of Poliovirus 1. The reason for this difference is not clear, but it may result from the biological differences between the two viruses (suggesting that Poliovirus survival in water is less dependent on temperature), or from differences between experiment designs when using the two viruses.

[[Image:15009_fig20.jpg|thumb|center|500px|'''Figure 20''' Effect of temperature on the inactivation rate of bacteriophage MS2 in water (reproduced from data in Pedley et al 2006)<ref name="Pedley"></ref>.]]

The effect of temperature on the mobility of bacteria and viruses in the subsurface is not known, although there is a suggestion from the literature that the retention of bacteria may be greater at higher temperatures (Tufenkji & Emelko, 2011<ref name="Tufenkji"></ref>).

[[Image:15009_fig21.jpg|thumb|center|500px|'''Figure 21''' Effect of temperature on the inactivation rate of Poliovirus 1 in water (reproduced from data in Pedley et al 2006)<ref name="Pedley"></ref>.]]

====MICROBIAL ACTIVITY AND DIVERSITY====
Soils are populated with a very complex and diverse range of microorganisms, with some estimates being as high as one million distinct genomes in pristine soils (Torsvik et al. 1990<ref name="Torsvik">TORSVIK, V, ET AL. 1990. High diversity in DNA of soil bacteria . High Diversity in DNA of Soil Bacteria. Applied and environmental microbiology, 56, 3, pp.782–787. </ref>; Bunge et al. 2005<ref name="Bunge">BUNGE, J, EPSTEIN, S S and PETERSON, D G. 2005. Computational improvements reveal great bacterial diversity and high metal toxicity in soil. Science, 309(August), pp.1387 – 1390.</ref> Such a diverse indigenous micro-flora represents a significant barrier for any introduced and non-native species to become established. van Elsas et al., (2012)<ref name="van Elsas">VAN ELSAS, J D, ET AL. 2012. Microbial diversity determines the invasion of soil by a bacterial pathogen. Proceedings of the National Academy of Sciences of the United States of America, 109(4), pp.1159–64. Available at:http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=3268289andtool=pmcentrezandrende rtype=abstract [Accessed January 29, 2013]. </ref> studied the survival of ''E.coli ''O157 in soils that had been enriched with increasingly complex mixes of indigenous soil microorganisms and found an inverse relationship between the soil species diversity, although not density, and the survival time of the introduced species.  But the relationship may not entirely be a result of community complexity, as there are indications from comparative studies of livestock bedding that certain microbial species may have a greater influence over the survival of ''E.coli ''O157 than others (Westphal et al. 2011<ref name="Westphal">WESTPHAL, A, ET AL. 2011. General suppression of Escherichia coli O157:H7 in sand-based dairy livestock bedding. Applied and Environmental Microbiology, 77(6), pp.2113–2121. </ref>).

Stated simply, indigenous microorganisms out-compete the pathogens (Toze 2003<ref name="Toze">TOZE, S, 2003. Pathogen survival in groundwater during artificial recharge. IAHS-AISH Publication, pp.70–84. </ref>), but this disguises a multitude of different process that might occur in the soils. Competition for nutrients is very likely to be a factor mediated by the diversity of indigenous species being able to exploit all nutrient sources. However, the importance of particular species suggests other mechanisms of suppression, such as the presence of antibiotics (Ramette et al. 2003<ref name="Ramette">RAMETTE, A, MOËNNE-LOCCOZ, Y, and DÉFAGO, G. 2003. Prevalence of fluorescent pseudomonads producing antifungal phloroglucinols and/or hydrogen cyanide in soils naturally suppressive or conducive to tobacco black root rot. FEMS Microbiology Ecology, 44, pp.35–43. </ref>), and predation by protozoa might influence the survival time of pathogens.

The relationship between microbial activity and virus survival is not straight forward. Predation by prokaryotic and eukaryotic cells in soils and aquifers, and the harsh environments created by indigenous microbial species will reduce the number of viruses, but the magnitude of the effect may be dependent on the particular virus type (Hurst et al. 1980<ref name="Hurst">HURST, C J, GERBA, C P and CECH, I. 1980. Effects of environmental variables and soil characteristics on virus survival in soil. Applied and Environmental Microbiology, 40, pp.1067– 1079. </ref>; Matthess et al. 1988<ref name=">MATTHESS, G, PEKDEGER, A, and SCHROETER, J. 1988. Persistence and transport of bacteria and viruses in groundwater: a conceptual evaluation. Journal of Contaminant Hydrology, 2, pp.171– 188. </ref>), with some viruses being more susceptible than others.

Biofilms will develop naturally on any surface that is moist and is exposed to microorganisms. Biofilms can vary in size and complexity from a single layer of cells over the surface to a thick glutinous film that is easily visible to the naked eye. Thicker biofilms generate different environments as distance from the surface increases: the surface areas may be aerobic and relatively nutrient rich whereas the deeper biofilm will be anaerobic and nutrient poor. Biofilm growth on soil particles and the aquifer matrix may incorporate pathogens and potentially extend their survival time in the aquifer. Alam et al. (2007)<ref name="Alam"></ref> were able to maintain VNC forms of ''V.cholerae ''in biofilms for 495 days and still recover viable cells following passage through animals. In contrast Banning et al. (2003)<ref name="Banning">BANNING, N, TOZE, S and MEE, B J. 2003. Persistence of biofilm-associated Escherichia coli and Pseudomonas aeruginosa in groundwater and treated effluent in a laboratory model system. Microbiology, 149, pp.47–55. </ref> suggest that biofilms may limit the survival of pathogens in groundwater by effectively competing for nutrients.

====MOISTURE CONTENT====
In many settings, an increase in the soil moisture correlates with longer survival times of bacteria and viruses, but there are exceptions. In soils with a low moisture content, the inactivation rate of poliovirus decreased as the moisture content increased to 15%. As the soil moisture content was increased above 15% the inactivation rate of the virus started to increase (Hurst et al, 1980<ref name="Hurst"></ref>) Furthermore, greater migration has been observed under saturated conditions.

====PH====
Most bacteria and viruses tend to survive longer within the pH range 6 to 8. However, enteric microorganisms must be able to survive exposure to stomach acids before being carried into the intestine.  Most enteric viruses are stable over a pH range of 3 to 9.

pH has a strong influence over the adsorption of bacteria and viruses to surfaces. In general, adsorption of bacteria and viruses to the aquifer matrix and soils increases as the pH decreases. Higher pH values can result in desorption and remobilisation of some viruses.

====DISSOLVED OXYGEN====
The role of dissolved oxygen in the survival of mobility of pathogens in the sub-surface has not been well characterised. There is some evidence to suggest that dissolved oxygen levels may be linked to the retention of bacteria on surfaces and their survival time; however, the evidence is contradictory and depends on the bacterial species (Tufenkji & Emelko 2011<ref name="Tufenkji"></ref>).

====IONIC STRENGTH====
The ionic strength of water has a significant impact on the survival and transport of bacteria and viruses in the sub-surface, but the magnitude and direction of the effect is influenced by the species of bacteria or virus, the nature of the aquifer matrix and the type of ions in the water. Certain cations have been shown to prolong the survival of some virus types. In contrast, some enteric bacterial species survive longer in freshwater than seawater, which shows that a high salt concentration can have disinfecting properties. Increasing ionic strength generally increases the adsorption of viruses and bacteria to the soil/aquifer matrix (Bellou et al, 2015<ref name="Bellow">BELLOU, MI, SYNGOUNA VI, TSELEPI, MA, KOKKINOS PA, PAPARRODOPOULOS SC, VANTARAKIS A,. CHRYSIKOPOULOS CV. 2015. Interaction of human adenoviruses and coliphages with kaolinite and bentonite. Science of The Total Environment, 517, pp.86–95. </ref>; Knappett, et al, 2008<ref name="Knappett">KNAPPETT, P.S.K., EMELKO, M.B., ZHUANG, J., MCKAY, L.D. 2008. Transport and retention of bacteriophage and microspheres in saturated, angular porous media: effects of ionic strength and grain size. Water Research 42, 4368–4378. </ref>; Walshe et al, 2010<ref name="Walshe">WALSHE, GE., PANG, L, FLURY, M, CLOSE, ME., FLINTOFT, M, 2010. Effects of pH, ionic strength, dissolved organic matter, and flow rate on the co-transport of MS2 bacteriophages with kaolinite in gravel aquifer media. Water Res. 44, 1255–1269. </ref>), although the opposite has been reported with the bacteriophage strains MS-2 and φX174 when passed through columns of Al-oxide coated sand (Zhaung and Jin, 2003<ref name="Zhaung">ZHUANG, J. AND JIN, Y. 2003. Virus retention and transport through Al-oxide coated sand columns: Effects of ionic strength and composition. Journal of Contaminant Hydrology, 60, pp.193–209. </ref>). Studies have shown that viruses can desorb from surfaces as a result of sudden changes in the ionic strength of the suspending medium, for example following rainfall events (Hurst and Gerba 1980<ref name="Hurst"></ref>; Bales et al, 1993<ref name="Bales">BALES, RC, LI, S, MAGUIRE, K.M., YAHYA, MT, GERBA, CP, 1993. MS-2 and poliovirus transport in porous media: hydrophobiceffects and chemical perturbations. Water Res. Res. 29,957–963. </ref>; Busalmen and Sanchez 2001<ref name="Busalmen">BUSALMEN, JP, AND DE SÁNCHEZ, SR. 2001. Influence of pH and ionic strength on adhesion of a wild strain of Pseudomonas sp. to titanium. Journal of Industrial Microbiology and Biotechnology, 26(5), 303-308. </ref>; Krauss and Griebler, 2011<ref name="Krauss">KRAUSS, S. & GRIEBLER, C. 2011. Pathogenic Microorganisms and Viruses in Groundwater, Georessource Wasser-Herausforderung Globaler Wandel. 69pp, ISBN 9783942044226</ref>). This may be significant in areas with high annual rainfall or where long dry spells that are interrupted by sudden and heavy downpours.

====SOIL PROPERTIES====
Most of the information regarding the effect of soil properties on the survival and mobility of pathogens has come from studies of ''E.coli ''O157 and various ''Salmonella ''species. The focus on these pathogens reflects the concerns about the contamination of groundwater from the surface spreading of manure on agricultural land. Vinten et al. (2002)<ref name="Vinten"></ref> found some variation in the survival times of ''E.coli ''and ''E.coli ''O157 in soils from different locations and between soils in laboratory and field conditions. Survival times were short (half-life between 1.8 and 2.9 days), but a small proportion of the population had a half-life in the soil of between 15 and 18 days (Vinten et al. 2002<ref name="Vinten"></ref>).

The migration of microorganisms through the soil is influenced by the soil structure. Using soil columns and ''E.coli ''O157 as the representative pathogen, Artz et al. (2005)<ref name="Artz"></ref> showed that migration rates were substantially reduced in compacted soils, but were significantly increased in the presence of earthworm burrows. Similarly, root systems can provide pathways for the rapid migration of microorganisms through soils (Kemp et al. 1992<ref name="Kemp">KEMP, J, ET AL. 1992. Leaching of genetically modified Pseudomonas fluorescens through organic soils: Influence of temperature, soil pH, and roots. Biology and Fertility of Soils, 13, 4, pp.218–224. Available at [doi|10.1007/BF00340579] </ref>), although conversely, these authors quote others who suggest that the production of polysaccharides by plant roots may retard the transport of microorganisms through the soil.

Clay minerals in soils are known to adsorb viruses, and can either increase or decrease their survival in the subsurface. The interaction between clay minerals and viruses is described in more detail in a later section. In a recent study soils collected in the South East of the UK, several factors in addition to clay mineral content were found to strongly influence the ability of soils to attach two strains of bacteriophage (Chi-Hiong 2013<ref name="Chi-Hiong">CHI-HIONG, F L. 2013. Effects of Soil Property Interactions on the Removal of Bacteriophages. University of Surrey. </ref>). Although the two virus types of bacteriophage showed some differences in their preference for soil properties, aluminium and soil pH were particularly important for the attachment of both virus types.

====BACTERIA/VIRUS TYPE====
The survival times of bacteria and viruses can vary considerably between the different groups of microorganisms, and within the groups between different species. In a summary of inactivation studies compiled by Pedley et al (2006)<ref name="Pedley"></ref> the inactivation rate coefficients in groundwater ranged between 0.0058 day-1 for the bacteriophage MS2 at 7°C to 5.3 day-1 for ''V.cholerae ''at 9-13°C. Although these are extreme values from single publications, they highlight the general observation that viruses have a slower inactivation rate than bacteria. Despite the wide variation in survival times, the sort of inactivation rates to expect are 0.03 log10 per day for enteric viruses and 0.09 log10  per day for enteric bacteria (Tufenkji & Emelko 2011<ref name="Tufenkji"></ref>).

Microorganisms vary considerably in size (Table 9) and are known to overlap with the pore sizes of rocks and some soil types (Figure 22).
<center>
{| class="wikitable"
|+ Table 9      Approximate sizes of selected bacteria and viruses (adapted from Pedley et al. (2006)<ref name="Pedley"></ref>)
| '''Class'''
| '''Microorganism'''
| '''Size'''
|-
| rowspan="5" | Virus
| Bacteriophage (common surrogates of enteric viruses)
| 20-200nm diameter
|-
| Poliovirus
| 30nm diameter
|-
| Adenovirus
| 80nm diameter
|-
| Hepatitis A virus
| 27-32nm diameter
|-
| Ebola virus
| 80nm diameter, but can be up to 14µm in length. Pleomorphic and enveloped.
|-
| rowspan="5" | Bacteria
| Bacterial spores (''Bacillus ''spp; ''Clostridium ''spp)
| 1µm
|-
| ''E.coli''
| 0.5µm x 1µm – 2µm
|-
| ''Salmonella typhi''
| 0.6µm x 0.7µm – 2.5µm
|-
| ''Shigella ''spp.
| 0.4µm x 0.6µm – 2.5µm
|-
| ''Vibrio cholerae''
| 0.5µm x 1.4µm – 2.6µm
|}
</center>
[[Image:15009_fig22.jpg|thumb|center|500px| '''Figure 22''' Examples of pathogen diameters compared to aquifer matrix apertures, colloids and suspended particles (adapted from Pedley et al. 2006<ref name="Pedley"></ref> and Lapworth et al. 2005<ref name="Lapworth 2005">LAPWORTH, D J, GOODDY, D, HARRISON, I, KIM, A, VANE, C H. 2005. Colloidal phase transport of pesticides: a review with special reference to major UK aquifers. Nottingham, UK, British Geological Survey, 22pp. (IR/05/131) (Unpublished) </ref>).]]

Where the size of the microorganism is larger than the pore spaces in the soils and aquifer matrix the mobility of the organism will be restricted by filtration or straining. This mechanism is particularly important for limiting the mobility of larger pathogens, such as the protozoa, but if the soil or aquifer system has particularly small pore sizes, bacteria and viruses may also be retained by filtration. However, where fissures of sufficient size exist enteric viruses and other and other faecal-derived microorganisms can penetrate aquifers to quite significant depths (Powell et al. 2003<ref name"Powell">POWELL, K L, ET AL. 2003. Microbial contamination of two urban sandstone aquifers in the UK. Water Research, 37, 2, pp.339–352. </ref>).

The method of replication is a key difference between bacteria and viruses that has implications for their survival in the subsurface. Viruses are obligate intracellular parasites, which means they must infect a host cell in order to be able to reproduce themselves: there is no exception to this rule. Outside the host, virus particles are inert and do not carry any of the cellular material that is necessary for the production of energy or the synthesis of the biomolecules that produce the daughter viruses: these services are provided by the host cell after infection. Human pathogenic viruses, therefore, cannot replicate in the environment. Once released from the host their numbers will decline at a rate determined by the particular virus type and the nature of the environment. In contrast bacteria are reproductively self-sufficient. Each cell has a complete set of metabolic systems that allow it to reproduce itself in a favourable environment. Hence, there is the potential for bacterial pathogens to increase in numbers – if only temporarily – when they are released into the environment, and their die-off rate will be determined by the balance between their rate of inactivation and replication. For most bacterial pathogens in the environment, the former greatly exceeds the latter.

Some species of bacteria, for example ''V.cholerae ''(Janda 1998), are motile. These cells can propel themselves through the suspending medium using hair-like surface structures called flagella: some species have a single flagellum at one end of the cell, whereas others have multiple flagella in different arrangements on the cell surface. Cells tend to swim in the direction of their long axis at about 35 diameters per second (Tufenkji & Emelko 2011<ref name="Tufenkji"></ref>). Under the microscope, the direction of movement of a single cell appears random, but many motile species of bacteria can orientate their movement in response to a particular stimulant, such as a chemical, light, or magnetic field. Viruses are not motile.

====ORGANIC MATTER====
Dissolved organic matter adsorbs to surfaces of grains and inhibits microbial attachment, thus lowering retention. Organic layer may prime surfaces for the development of biofilm (Wimpenny 1996<ref name="Wimpenny">WIMPENNY, J W T. 1996. Ecological determinants of biofilm formation. Biofouling, 10(1-3), pp.43–64. </ref>) that may eventually restrict pore space and increase straining, also biofilm may increase the potential for pathogens to be eliminated by grazing by indigenous biofilm organisms (Banning et al. 2003<ref name="Banning>BANNING, N, TOZE, S and MEE, B J. 2003. Persistence of biofilm-associated Escherichia coli and Pseudomonas aeruginosa in groundwater and treated effluent in a laboratory model system. Microbiology, 149, pp.47–55. </ref>; Tufenkji & Emelko 2011<ref name="Tufenkji"></ref>). Organic layers may provide hydrophobic binding sites for the adsorption of viruses with hydrophobic groups on their surface.

====HYDRAULIC CONDITIONS====
There is no evidence to suggest that the hydraulic conditions have a noticeable effect on the survival of bacteria or viruses, but at higher hydraulic loads and faster flow rates there is less retention of bacteria and viruses on solid surfaces and an increase in the dispersal of the pathogens (Pedley et al. 2006)<ref name="Pedley"></ref>.

====CLAY MINERALS AND COLLOIDS====
This section will overlap to an extent with the soil section above, but the purpose here is to concentrate on particular clay minerals that are of relevance to Sierra Leone, and the colloids produced from these minerals. Studies have concentrated on a limited number of clay minerals, particularly kaolinite and different forms of smectite, especially montmorillonite (Chi-Hiong 2013<ref name="Chi-Hiong"></ref>). Virus attachment to clay minerals is complex, and there is evidence that different viruses may interact with the minerals in different ways (Chrysikopoulos & Syngouna 2012<ref name="Chrysikopoulos">CHRYSIKOPOULOS, C V and SYNGOUNA, V I. 2012. Attachment of bacteriophages MS2 and QX174 onto kaolinite and montmorillonite: Extended-DLVO interactions. Colloids and Surfaces B: Biointerfaces, 92, pp.74–83. Available a [doi|10.1016/j.colsurfb.2011.11.028] </ref>; Lipson & Stotzky 1985a<ref name="Lipson 1985a">LIPSON, S M and STOTZKY, G. 1985a. Infectivity of reovirus adsorbed to homoionic and mixed- cation clays. Water Research, 19, 2, pp.227–234. Available at:http://www.sciencedirect.com/science/article/pii/0043135485902040 [Accessed February 3, 2015].</ref>; Lipson & Stotzky 1985b. Lipson & Stotzky (1985) found differences in the relative levels of attachment of Reovirus and coliphage T1 to kaolinite and montmorillonite, but did not observe competition for binding sites when the two viruses were added together, suggesting that variations in surface properties of the clay minerals is important for the specificity of virus attachment. Attachment to the clays was pH dependent, with a higher level of attachment of Reovirus at lower pH values (Lipson & Stotzky 1985a<ref name="Lipson 1985a"></ref>), but adsorption was not blocked when the positively charged sites on the minerals were chemically blocked.

Bacteria and viruses can attach to clay colloids in groundwater, through hydrophobic interactions (Chrysikopoulos & Syngouna 2012<ref name="Chrysikopoulos"></ref>). In studies using glass beads to simulate the aquifer matrix, the flow of bacteria and viruses through the column was shown to be retarded when bound to clay colloids (Vasiliadou & Chrysikopoulos 2011<ref name="Vasiliadou">VASILIADOU, I A, and Chrysikopoulos, C V. 2011. Cotransport of Pseudomonas putida and kaolinite particles through water-saturated columns packed with glass beads. Water Resources Research, 47(May 2010), pp.1–14. </ref>; Syngouna & Chrysikopoulos 2013<ref name="Syngouna">SYNGOUNA, V I, and CHRYSIKOPOULOS, C V. 2013. Cotransport of clay colloids and viruses in water saturated porous media. Colloids and Surfaces A: Physicochemical and Engineering Aspects, 416, pp.56–65. Available aTemplate:Doi</ref>). The mechanism proposed by these authors to explain this observation is that the bacteria and viruses attach to the colloids, which then attach strongly to the glass beads. If these laboratory findings do mimic the interactions taking place in soil and aquifer systems, the presence of clay colloids derived from kaolinite and montmorillonite may limit the dispersal of pathogens.

===Implications for Sierra Leone ===
Temperature has a strong influence on the survival times of bacteria and viruses in water. At the average temperature of the groundwater in Sierra Leone (ca. 26°C), the inactivation rates of pathogens are likely to be quite high, so the survival times will be short relative to colder environments. Nevertheless, the survival times will vary considerably between different pathogens and it is likely that virus pathogens will survive longer than bacterial pathogens and faecal indicator organisms.

The chemical and physical characteristics of soils and the aquifer matrix have an important role in the adsorption of pathogens onto these surfaces. Bacteria and viruses are adsorbed by the types of clay minerals found in Sierra Leone which will restrict the migration of the pathogens from the source. However, it is likely that the groundwater will contain colloidal material from the clay minerals, which may enhance the migration of pathogens through the aquifer.

Sierra Leone has witnessed a number of cholera outbreaks, and it seems feasible that ''V.cholerae ''will contaminate groundwater at these times. The literature is limited, but ''V.cholerae ''has been isolated from wells in SSA. ''V.cholerae ''is known to transform into a dormant state (VNC) in adverse environmental conditions, and there is some evidence to suggest that it remains a risk to human health when in this state. The significance of the VNC state for the survival and dispersal of ''V.cholerae ''in groundwater is unclear, but it may allow the bacterium to travel further than would be anticipated from the viable cell, and it will be very difficult to detect in groundwater samples using standard microbiological methods.

Potentially high levels of organic contamination in groundwater from latrines and surface wastes may counteract the capacity of the soils and the aquifer matrix to adsorb pathogens and allow them to migrate further than they would in a clean environment. However, it may also create conditions that help to reduce the survival times of pathogens.

Pathogen survival and transport in the subsurface has been studied and reported on for several decades. Most of the work has been done in developed countries with mainly temperate climates. The data shows that the fate of pathogens in the subsurface is determined by a complicated and poorly understood set of interactions between several known factors and possibly as many unknown ones. Consequently, it is extremely difficult to predict the behavior of a pathogen when it is released into the environment, even in regions where these studies have been done. There is significantly less information about the environmental survival of pathogens in SSA, which is a knowledge gap that does need to be filled before a reasonable attempt can be made to assess the risks from pathogens moving through groundwater.

{|class="wikitable" style="width:700px" style="background:#f0f8ff; font-size:100%"
| '''Pathogen survival summary:'''<br>
Based  on  average  groundwater  temperatures  in  Sierra  Leone  (26°C),  inactivation  rates  for pathogens are likely to be high. Nevertheless, survival times will vary considerably between different pathogens and overall viruses are likely to survive longer than bacteria.<br><br>Physical  and  chemical  processes  within  the  soil  attenuate  pathogens  and  restrict  migration. However,  colloidal  attachment  may in  some  cases  enhance  migration  due  to  size  exclusion effects (i.e. reduced diffusion) along preferential flowpaths.<br><br>The dormant state that ''V.cholerae ''can exist in (VNC) suggests that its survival and dispersal in the subsurface could be greater than would be expected for viable cells, and necessitates the use of sequencing techniques for detection.<br><br>Fate and transport of pathogens are determined by interactions between multiple factors, e.g. initial  pathogen  levels,  nutrient  levels,  temperature,  completion  for  resource  with  other groundwater micro-macro fauna to name a few. There are very few studies that have considered pathogen survival in conditions relevant to Sierra Leone.
|}

==References and footnotes==

[[category: OR/15/009 Threats to groundwater supplies from contamination in Sierra Leone, with special reference to Ebola care facilities | 03]]