Natural or artificial ground catchment and Lined sub-surface tanks

Stone masonry round berked under construction showing wires for roofing materials. Somaliland.
Photo: Eric Fewster, BushProof / Caritas.

These are natural, artificial or modified catchments that have low to relatively high runoff coefficients. Water from these catchments is captured and stored in lined sub-surface reservoirs excavated below ground level. The reservoirs are known by different names (berkeds in Somaliland, taankas in India, hemispherical sub-surface tanks in Kenya – also included in this category are excavated water cellars such as the shuijiao in China) and have been lined with many different materials. These tanks normally have a larger depth to surface ratio compared to open ponds and their scale means a roof of some description is a possibility. When the lining is constructed well, there will be no leakage, and water will either evaporate or be abstracted. These tanks are often privately-owned by one or more families, but can be communal.

Suitable conditions

• Site in a place where run-off is seen to flow after rains.
• Artificial catchments are created where infiltration of runoff zone is high.
• Care should be taken when siting in clay areas, but the type of clay is more important.
• Do not site tanks near big trees whose roots might crack the walls.
• Do not site tanks where heavy vehicles will pass close to tank wall.
• Do not site sub-surface tanks in areas of high water tables to reduce risk of flotation.

Advantages

Disadvantages

- Less evaporation than natural ponds due to less surface area to depth - Good for areas where ground would otherwise be permeable - They work well when privately owned and maintained

- Sub-surface tanks often cannot hold enough water for the whole dry season; making bigger berkeds is possible but is more difficult and expensive; if not affordable it is not replicable. - Costs currently limit the replicability of the technology for poorer families and the potential to scale things up - Considerable amounts of silt accumulate in tanks, exactly how much will depend on the area - Flotation of the tank may occur in areas with a high groundwater table - Heavy vehicles driving near to tank can cause damage - Leaks in sub-surface tanks are hard to detect - Artificial catchments take up potentially valuable land surface and are difficult to keep clean; concrete catchments tend to crack - When built in a remote area, construction is difficult due to lack of water and large distances to transport materials - Microbiological and chemical water quality is likely to not be acceptable for direct consumption

Resilience to changes in the environment

Drought

Effects of drought: Water storage used up.
Underlying causes of effects: Lack of rainfall; High evaporation rates; Leaking linings due to bad construction; Storage not sufficient for demand – tanks are too expensive for volumes of water to outlast extended dry periods.
To increase resiliency of WASH system: Build smaller tank structures but more of them over longer time, which means less reinforcement per tank, more manageable to construct and cover, and more affordable; Reduce evaporation & seepage due to poor construction & siting; Follow proper concreting guidelines; Make tanks from cheaper materials and repair more often; Improve access to micro-finance; Support the capacity of the government or private sector to be able to provide (for payment) a tankering scheme.

Drought effects on cement

Effects of drought: Badly made concrete and cracked linings (e.g. in tanks, dams, waterways, wells, and other structures).
Underlying causes of effects: Less water used for curing; Impure water used for mixing.
To increase resiliency of WASH system: Ensure adequate mixing, ratios, purity of ingredients; Minimize water content in mixture; Ensure adequate curing.

More information on managing drought: Resilient WASH systems in drought-prone areas.

Construction, operations and maintenance

General advice on cement: A common cause of cracks in structures and linings (e.g. in tanks, dams, waterways, wells) is errors in mixing and applying the cement. First of all, it is important that only pure ingredients are used: clean water, clean sand, clean rocks. The materials have to be mixed very thoroughly. Secondly, the amount of water during mixing needs to minimal: the concrete or cement needs to be just workable, on the dry side even, and not fluid. Thirdly, it is essential that during curing the cement or concrete is kept moist at all times, for at least a week. Structures should be covered with plastic, large leaves or other materials during the curing period, and kept wet regularly.

Specific advice:

The reason for constructing a sub-surface tank is to store the water. Therefore one of the most important aspects is that seepage and cracks must be avoided. Therefore good quality construction work with adequate supervision is vital to create a sound structure – this is especially important in areas with swelling soils that can affect the integrity of the lining. While ownership and management of tanks is important, such privately-owned tanks have often failed due just to the technical construction component. Construction materials vary and include the natural soil formation itself, clay, stone masonry, bricks/cement, ferrocement, anthill/lime/cement and plastic/rubber lining. Material may affect cost but choice may also depend on what is available and the type of surrounding soil. To generally prevent cracking/seepage:

• Round tanks are inherently stronger than rectangular ones. Hemispherical and cylindrical designs are commonly used.
• Type of tank will vary depending on the swelling ability of the surrounding soil – a problem in clay areas, but type of clay is more important – montmorilonite, calcium-containing clays (in marls/gypsum sediments) and black cotton soils are all prone to swelling and can crack sub-surface tank walls that are not built robustly enough. Therefore it is important to construct the right type of tank for the area. When in doubt, avoid making sub-surface ferrocement or anthill/lime/cement tanks in unstable soil.
• Admixtures can be added to the concrete mix in order to reduce the amount of water needed. Research has shown that superplasticizers work best by reducing the amount of water that needs to be added when mixing concrete, which results in 35% less shrinkage. The resulting end material is stronger and can reduce the amount of micro cracks in mortar by half compared to normal mortar while resulting in 76% fewer leaks. In general, the amount of plasticizer to be added should not be greater than 2% of the dry material weight. A plasticizer that can be used that is possibly available is household washing up liquid. In hot climates though, more research is needed in the field application of plasticizers, since the reduction of water used (and increased strength of product) may not be that great due to more water needed to prevent drying out between mixing and application.

Choose a tank to build

Key construction issues for good workmanship (which also relate to preventing cracking/seepage) and costs for specific lining types are detailed below:
Stone Masonry:

• In Somaliland it cost between US$39 - $43 per m3 of storage for a new berked and US$8 per per m3 for a rehabilitated berked, excluding about 30-45% of local contribution (e.g. 493m3 new berked = $19,550; rehabilitated existing berked = 4,000 USD). In India, stone masonry sub-surface tanks cost US$28 per m3 of storage (35m3 tank cost US$990).
  
• Floor to be made from concrete which needs to be laid with vibration in order to be sure they are leak-proof.
• In clay areas, be sure to build the tank robustly enough to resist cracking. Sample dimensions & mixtures for walls and floor for stone masonry tanks in an area of swelling clay are:

1. Walls: 0.4m wide, 2 blocks thick.
2. Floor leveling mixture: 0.05m thick, ratio 1:4:6 (not used in rehabilitated berkeds as level floor foundation already exists).
3. Floor: 0.16m thick, unreinforced concrete, ratio 1:21⁄2:4.
   

Bricks/cement:

• In Kenya, brick/cement tanks cost US$37 per m of storage (21m3 tank cost US$780).
• In Sri Lanka, brick tanks cost 28 per m3 of storage (5m3 tank cost US$140).
• In clay areas, be sure to build the tank robustly enough to resist cracking.

Ferrocement:

• Ferrocement tanks in other areas seem to cost in the range of US$30.5 (60m3 tank cost US$1,830) to US$32 per m3 storage (60m3 tank cost US$1,900) including all costs.

Anthill/lime/cement – in Kenya, anthill material and lime has been added to reduce cost of lining. The lower capital cost though does mean more maintenance work. However there does not seem to be too much field data on how these function in the longer term.

• The following plaster mixture proved to work better: 4 parts anthill soil, 1 part cement, 2 parts lime, 6 parts river sand.

Plastic/rubber: Choosing a plastic/rubber lining is not an option in most circumstances due to combinations of fragility, expense and feasibility for welding together sheets for larger ponds. Below the pros and cons of various liners are discussed:

• There are five main types of liner constructions: Polyethylene, Polythylene, PVC liners, EPDM/rubber, and polypropylene. As an indication of costs, an EPDM/rubber sourced in the UK is around $6.5 per square metre not including shipping costs.
• Choice of lining needs further consideration:

1. Needs to be food grade since the water it stores is for drinking.
   
2. Cost: Polyethylene and Polythylene liners typically cost half that of Polypropylene and EPDM. PVC and PVC-E liners are the next step up from Polyethylene and polyethylene in terms of cost. Compared to other liners, PVC is somewhat more affordable, while being somewhat puncture resistant at the same time but in terms of durability, the typical 20-mil (i.e. 0.020") thick PVC is somewhat mediocre in terms of durability. EPDM is more expensive than Polyethylene, polyethylene, and PVC liners, but can last for up to 20 years. Polypropylene is an expensive material but can last for up to 40 years.
   
3. Durability:
   1. Polyethylene and Polythylene liners typically only last one season.
   2. Polyethylene will readily conform to any shape, however, it does not have the sturdiness that is required for a permanent pond liner.
   3. Polythylene, on the other hand, is extremely rigid and can be stiff to work with. Polythylene can be damaged easily by rocks, and has to be handled with care. If polyethylene is damaged, it cannot be seamed together without expensive welding equipment.
   4. PVC and PVC-E liners are the next step up from polyethylene and polyethylene, and they can last for up to 10 years.
   5. Polypropylene is an expensive material but it is the most durable pond liner material in existence because it can last for up to 40 years. Polypropylene, however, is not as flexible as EPDM liners.
   6. EPDM (Ethylene Propylene Diene Monomer) rubber liners are recommended for most pond installations because of their delicate balance between longevity, flexibility, affordability, and their lack of toxic plasticizers. Because EPDM liners are rubber-based, they are extremely flexible (much more so than PVC liners) - the extra flexibility of EPDM comes in handy when working with irregular folds and shelves that are commonly found in a pond. They also do not contain any plasticizers that can make the liner brittle and crack with age. A 45-mil EPDM liner can last for up to 20 years because of its natural resistance to UV, and its puncture resistance.

Sheet size: ideally we should not have to weld sheets together in situ.

• The main advantage of polypropylene is that it comes in large sheets larger than 50’ x 100’. If you are building an extremely large pond, polypropylene may be a viable option.
• A limitation of EPDM is its size they typically arrive in sheets ranging from 5’ x 10’ to a 50’ x 100’ roll.

Limiting evaporation

Sand-filled sub-surface tank, Botswana Image courtesy of WEDC. © Ken Chatterton. In: Hussey, S.W. (2007) Water from sand rivers: guidelines for abstraction. WEDC, Loughborough University, UK.

Sub-surface tanks are usually small enough that it is viable to have a roof to limit evaporation (and improve water quality if possible, which means less algae build-up). Shading can reduce evaporation by around 30%. Placing local bush or grass materials on a frame of wires doesn’t seem to work well because they get blown off, and also still let light in, which creates algae growth. Corrugated iron roof on wooden frame works well but is expensive (about $20 per m2 in Somaliland). In addition, if the tank is not fenced, animals walking on the roof can damage it. The challenge is to make a roof that is cost-effective for small-scale farmers – one idea is to investigate income-generating roofs since that can help pay for the structure (e.g. passion fruit). Excavated water cellars by their nature have small area roofs.

Deeply dug tanks mean less evaporation, but will provide more water quantity to last longer into the dry season. Perhaps a rule of thumb should be that depth should be greater than the maximum PET rate for the area in question. For example, the average length and width of berkeds from projects in Somaliland were 11.4 and 6.3 metres, while average water-holding depth was 2.9 metres, whereas PET rates ranged from 1.75 – 2.25 metres per year. Problem: deeper tanks could mean more investment.

Another method to reduce evaporation and at the same time improve water quality is to use a lined, sand-filled tank. It appears that plastic pond liners are in general more tolerant to earth tremors than solid lining like concrete – in some situations when the rains might cause swelling of the surrounding ground which might move the existing wall in a similar way, the plastic lining might in fact still be functional. In such a case it might be good to try out the following method which has been tried in Botswana in a lined rectangular tank:

• Use a plastic lining to create an impermeable layer on top of the existing lining. Protect the lining with a sand layer both on the floor of the berked before the lining is laid (evens out floor, protects against sharp objects, dried clay fragments etc), and also on top of the plastic after it is laid (to protect from flotsam and when people walk on it).
• Create an abstraction point (using sand to filter the water).
• Fill the remaining volume with sand.
• Add a coarse mesh after the silt trap before inlet to prevent large debris from entering the tank.

Water extraction

For open water catchments, like from rock surfaces or stored behind earth dams, direct abstraction (pump or pipe taking water off) works well. Abstraction method should minimize disturbance of the settled water, thus reducing treatment requirements later.

• Direct abstraction is one option, via a bank-mounted pump (Small and efficient motor pumps or Handpumps) which uses a floating intake to reduce sediment intake. An outlet pipe and strainer through the dam wall to the downstream side is another option, but these have potential problems of weakening the dam wall. In addition, piping will have to be secured externally when traversing rocks, so care has to be taken to secure pipes with anchor posts.
• With preventive methods to reduce turbidity (silt trap, extraction method) the water is still turbid & contaminated and will require treatment.
• For direct abstraction, promotion of household water treatment is advocated. Choice of household water treatment technology should be based on efficiency of removing contaminants present in the water. For open water that may be prone to cyanobacterial blooms, a Concrete Biosand Filter is a good choice due to its ability to remove cyanobacterial toxins. Other technologies however may be more suitable for mobile communities (e.g. Sodis or a Ceramic pot filter, depending on turbidity levels). For reservoirs near urban environments or where the runoff area has intensive agriculture practised in its vicinity, diversification of water resources is a good idea to provide alternatives for direct drinking purposes. Strengthening controls and restrictions on use of illegal substances will also help.

Tank modifications and maintenance

• In case of cracked linings, the following could be tried to salvage the tanks:
  • If the crack is only at the base, covering the tank base with clay and compacting it might work. Addition of powdered anthills or lime is said to make this lining more robust. If the cracks are also found in the walls, then rehabilitation or an alternative lining might be a solution. Taking the example of berkeds in Somaliland, many remain unused due to previous poor workmanship, yet rehabilitation is expensive ($8 per m3), requires skill and is not always successful. In some cases it is also not even possible to rehabilitate – some cracked berkeds can be rehabilitated if the original walls were made solidly enough, but otherwise there are many berkeds that can never be rehabilitated. In such cases, plastic linings might be worth trying.
• Where catchments have low runoff coefficients, this can be increased by modifying the existing surface or creating an artificial surface:

1. In China, soils with reasonable infiltration capacity had a runoff coefficient of 2%, which was increased to 20% after the soil was compacted.
2. Artificial lining of catchments is a possibility. Various catchment types and their runoff coefficients are: concrete (73-76%), cement-soil mix (33-42%), buried plastic sheet (28-36%).

• Shallow drainage canals can be dug to direct the runoff into the tank.
• Silt intake into sub-surface tanks ideally should be limited – how much silt will accumulate will depend on the area. In China, 80m3 had accumulated in 4 years. Ideas to limit silt include:

1. Keeping a good cover of grasses or vegetation in the run-off area.
2. Silt trap prior to tank intake. However, experience from Somaliland shows that silt traps (small mini reservoirs prior to main tank) are not very effective. A better method might be to replicate silt traps used in Charco dams where perennial vegetation is grown between small dams in the intake channel to encourage deposition.
3. Alternatively where vegetation may not grow due to climate, stones similar to roughing filter can be used on intake to increase sedimentation before water enters the tank. A roughing filter operates through increasing the surface for sedimentation and could be designed into the berked intake where stones of 3 different sizes between 25 and 5mm are used in 3 separate sections. But if it is to function properly its area needs to be designed based on flow rates and inflow water quality. The filtration rate should be calculated by flow (m3/hr) divided by surface area (m2) and then different filtration rates are suitable for different water qualities – this information may be hard to estimate in the field though.

• Create a fence around the tank to prevent children from possibly falling in, and to prevent large vehicles from driving too close and damaging the lining.
• Support the notion of private ownership & management. A fence can be constructed to improve private ownership.
• Fish can be introduced to eat mosquito larvae, while at the same time providing a source of nutrition.
• Support the capacity of the government or private sector to be able to provide (for payment) a tankering scheme to fill tanks during the driest parts of the year.

Costs

Access to finance is a main obstacle to promotion of rainwater harvesting for households, and is important so that users can replicate the technology – there are few examples on a global level with micro-credit for rainwater harvesting. The cost of underground tanks can be high and variable in cost per m3 of storage (averages around US$30-40 per m3 of storage or more depending). Sub-surface hemispherical tanks made from stone masonry and bricks/cement in Ethiopia costs 113 - 219 Euro per m3 of storage including all costs. May need to pay for trucked water for construction and for a higher solidity of tanks (in clay areas). However there are several ways to save money.

Save on costs

High cost of tank construction will decrease water availability because smaller tanks can be made. Ways to increase storage are to build with cheaper lower quality materials, use less material for construction and reduce labour costs. In this way, sub-surface tanks can become a more realistic option. Ideas include:

• Use existing soil as a natural lining if it is relatively impermeable. In China, clay has been used to line excavated water cellars (called Shuijiao) in areas where the natural soil (loess) is already fairly impermeable. The lining process is difficult and time-consuming and has been replaced largely by ferrocement or plastic. However, it proves that in some areas it is possible to construct a low-cost tank. In Somaliland, similar water cellars were observed that were excavated in impermeable stable soil formations – runoff water entered through a small inlet channel. Other tanks are sealed with a 10cm unreinforced cement lining – it seems that 30m3 is the most economical size, as it is the most volume without needing reinforcement – such tanks cost US$189 or about US$6.3 per m3 of storage (materials only presumably).
• In Kenya, tank linings have been made with powdered anthill material and lime which substitute some of the cement and bricks, bringing cost to US$9.8 per m3 of storage.
• Reducing the size of structures. More manageable to construct in terms of cash flow, and easier to cover. This way, tanks are more affordable to families, and more tanks can be added in subsequent years, thus spreading out costs.
• However, care needs to be taken with cheap linings – in some areas with swelling clay and differential settlement, linings can easily crack, as has often been observed in some areas. The areas where cheap linings might work therefore may be site-specific, and depend on the clay content of the soil. For plastic linings, experience from India shows that these can be punctured by rodents, crabs or insects if there is no rodent/insect-proof layer before the plastic.

Manuals, videos and links

• Sand Abstraction. Sourcebook of Alternative Technologies for Freshwater Augumentation in Africa. United Nations Environment Programme.
• Evaluation of the Sand Abstraction Systems for Rural Water Supply: the case of Lupane District, Zimbabwe. Vol 3, No. 1, January 2011.

Acknowledgements

• CARE Nederland, Desk Study: Resilient WASH systems in drought-prone areas. October 2010.