8. Rainwater Harvesting

Annual rainfall in most of Africa is not in short supply (see my third post) The problem is the variability of this supply, which makes rain-fed agriculture hugely unreliable. Climate change threatens to make this variability even more extreme, both within and between years (Niang et al., 2014). Irrigation overcomes this variability and increases the reliability of the growing season (see my fourth post). Groundwater is undeniably an important source of irrigation water (see my fifth post)… but rainwater can be used for irrigation too! Rainwater can be ‘harvested’ during periods of excess and used to irrigate crops during dry periods, as demonstrated in the diagram below. Although not all rainwater can or should be harvested (some is needed to maintain healthy ecosystems), rainwater harvesting (RWH) alone might have the potential to solve most of Africa’s water shortages (Mati et al., 2006). 

Fig. 1 A diagram demonstrating the variability of rainfall (y-axis = rainfall) and how dry season deficits can be overcome by storing excess rainfall from the wet season (Source: NWP, 2007)

Various RWH technologies have been used for thousands of years and new ones are being developed all the time. I have come up with four main categories that relate to the location of rainfall collection and storage...

1. RWH before it reaches the ground surface

2. RWH at the ground surface before it infiltrates or flows away to the sea
3. RWH underground in the unsaturated 'vadose' zone.
4. RWH underground below the water table in the saturated zone.  

Fig. 2 A diagram I drew to try to demonstrate my four categories of rainwater harvesting technologies, which each relate to a different the location of rainwater collection and storage

These categories are discussed below with examples of their use in Kenya, a country that has been at the forefront of implementing RWH in Africa.

1. RWH before the ground surface

This primarily refers to rooftop collection. It requires an impervious roof made from non-toxic materials (e.g. galvanised iron, cement or tin) attached to a system of gutters, which slope towards a downpipe and into a large storage tank. To give an idea of scale, a roof catchment of 36.5m² would provide a water consumption of 20l/person/day (7.3m³ per person/year) in an area with an annual rainfall total of 200mm (the baseline level for a 'semi-arid' area) (Mati et al., 2006). The photo below shows a rainwater harvesting system set up at the Mundika School in Busia District, W Kenya. Enough water is now collected by this system to supply the school's kitchen garden and provide all 120 students with a school lunch everyday (Water Charity, n.y.). 

Fig. 3 Rooftop RWH system installed at Mundika School in Busia District, W Kenya (Source: www.watercharity.com)

2. RWH at the ground surface:

This refers to both conventional river dams and to collecting runoff from open surfaces (e.g roads, rocks, hillsides and open pasture lands). However, conventional dams require a large perennial river, which is often not available and only tend to be effective at a large scale due to evaporative losses from their storage reservoirs (Tuinhof et al., 2005). Such dams are expensive and can significantly limit downstream flows, which can have extremely negative effects (as shall discussed in my next post on floodplain agriculture). At a small-scale in rural SSA, surface runoff collection is arguably more feasible. In 2009, a rock-runoff cistern was built in the Soit Oudo region of Laikipia county, central Kenya (see image below). The cistern captures water running off a large rock during heavy rainfall events. To ensure the quality of the harvested water, animals are not allowed in the cistern’s catchment and a sediment trap precedes the inlet to the tank (Lancaster, 2016). 

Fig. 4 Rock-runoff cistern built in the Soit Oudo region of Laikipia County, central Kenya in 2009 (Source: www.harvestingrainwater.com)

3. RWH in the unsaturated zone

This refers to all practices that increase the rainwater stored in agricultural soils by increasing soil infiltration capacity and/or water retention capacity. These include digging furrows, terracing, and conservation agriculture (Hobbs, 2007). In this situation, the stored rainwater is not used for irrigation because it is already in the soil! The Machakos district in Kenya offers a good case study. The area is hilly and by the mid 1900s was deforested and overgrazed. When it rained water rapidly ran off the hillsides and away from the fields. In the 1990s, the community terraced the hillsides and planted fruit trees. This helped retain rainwater in the soil, boosting agricultural productivity and transforming the livelihoods of the local people (Tiffen et al., 1994).

Fig. 5 The rolling terraced hills of the Machakos District, Kenya (Source: Project Survival Media, photo by Joe Lukhovi, taken in March 2012)

4. RWH in the saturated zone.  
This refers to all technologies that use rainwater to increase aquifer recharge rates above natural levels. Although not limited to rainwater, these processes are often referred to as 'managed aquifer recharge' (MAR). MAR could allow us to increase 'sustainable groundwater abstraction' rates (Dillon, 2002; Tuinhof et al., 2005), see my last post. The most classic example of MAR is pumping surface water (which may otherwise evaporate or flow away) down an injection pipe into groundwater reservoirs, as shown in the diagram below. However, this requires technology that is often not affordable or accessible to rural small-holder farmers in SSA. PLUS extraction technologies are required to recover the water, which these farmers also may not have. 

Fig. 6 Diagram showing the process of pumping surface water down an injection pipe into a groundwater reservoir, a type of managed aquifer recharge (MAR) (Source: National Groundwater Association, 2013)

Sand dams are an alternative type of cheap, low tech MAR that create easily accessible 'raised aquifers' (Ryan & Elsner, 2016). The dams are reinforced concrete walls built across seasonal sandy riverbeds. During the rainy season sand is washed downstream and accumulates behind the dam. The accumulated sand stores some of the floodwater in a 'raised aquifer'. With each rainy season, the size of the sand reservoir increases. When mature, up to 40% of the dam's total volume is water (Maddrell & Neal, 2012)!!! The stored water is then abstracted using small wells,  ‘scoop holes’ or pipes leading to taps (Quilis et al., 2009). Sand dams have a number of advantages over conventional river dams (discussed above): (1) underground water storage significantly reduces evaporation losses, allowing them to be effective at a small-scale; (2) at full capacity they can only store around 1-3% of total yearly runoff ensuring that the local water balance remains largely unaffected; (3) underground storage also reduces contamination and the presence of mosquitoes; (4) sand filters the water, reducing bacterial threats (Huisman & Wood, 1974; Avis, 2014). 

Fig. 7 Diagram explaining how a sand dam works (Source: www.sswm.info)

In Kitui District, Kenya over 500 small-scale sand dams were constructed between 1998 and 2008 with the help of a local NGO (the 'Sahelian Solution Foundation'). More than 100,000 people now have better access to water (particularly in dry periods) and local farm yields have increased. In fact, farmers living near the dams now earn ~60% more than they did before (Lasage, 2008). Following this success another NGO (SOS Sahel) is trying to transfer these technologies from Kenya to rural Sudan. The programme began in 2010 and has so far built dams in South Kordofan and North Darfur.

Fig. 8 A sand storage dam in the Kiindu river near Kitui, Kenya (Source: Sam Sam Water)

This post has covered how humans can increase RWH to improve year round water supply. However, human intervention is not always necessary (or desirable). My next post investigates how seasonal floodwaters in African wetlands can support year round agriculture without human storage infrastructure.