Aid ProjectsCommercial Farm ProjectsCommunity ProjectsCourses/WorkshopsDemonstration SitesEducation CentresGabionsLandMaterialRoadsSoil ConservationStorm WaterWater Harvesting

Strawberry Fields Update: Flood Protection and Water Control in Ethiopia, Part II

We previously published a report on the development of our site’s flood control and defense infrastructure in October 2010. This is an update on that which goes on to describe some of our plans for developing that infrastructure more in the future.

Just to recap on the basics of our situation: in times of rain, the run-off from the western part of Karat Konso Town (South Ethiopia) runs down the side of the road which heads uphill to the south of our site. This flash flood creates a temporary stream which impacts the south eastern corner of the site. The flash floods can be pretty intense.

Western town watershed, running past SE corner of SFEL site

We had seen the scale that these flash floods could reach in May 2008. However the following seasons were drought seasons and in July 2009, we had become a bit complacent on this issue when we selected this location to place our car park, estimating that a 50cm high stone-cement wall would be enough to keep any flood out.

Constructing 50cm cement stone wall around car park – September 2009

But in November 2009 it suddenly rained very heavily, teaching us a lesson. The flood flattened the fence at the south eastern corner of the site, came over the 50cm high stone-cement wall, through the car park and into the lower half of our zone 1 vegetable garden, burying it all under 15cm of slimy red silt. The silt is of good quality and while the siltation increased the fertility of the garden most of the plants were killed, meaning that such floods would have to be contained in the future. Cars could also not get into the car park for days afterwards due to the mud, which lost us income.

Initially we responded to this problem by constructing a heavy dry-stone wall at 45% to the direction of the oncoming flood to push the water around the corner of the site so it would run along outside the fence rather than coming into the car park.

This was initially successful, though we soon noticed that subsequent floods, hitting this wall and running along it laterally, were digging a gulley at the base of the wall and threatening to undermine it. We also noticed that in heavy rains run-off from the neighbour’s compound as well as the adjacent ridge (Ridge I) was entering the car park and depositing the nasty, sticky, cracking, clay sub-soil of the site on the surface of the car park, which we wanted to avoid.

So, in planning our next move we considered:

  • The silt brought down from the town is fertile (red clay, fortified with animal dung and general domestic detritus from the townsfolk).
  • The silt from the adjacent site and Ridge 1 is poor quality and infertile.
  • The main body of the flood cannot be contained without a massive development of infrastructure.

So we wanted to catch as much of the town flood water and fertile town silt as we can while keeping the silt from the neighbour’s land and Ridge 1 out of the car park.

Accordingly we did the following:

  • Added 3 dry-stone gabions (cages made from wire and re-bar filled with rocks) were built projecting 2m out from the wall at 90deg.
  • Built a terrace to divert the run-off from Ridge 1 and the neighbouring site into the gap between the outer and inner car park walls, allowing the water to run away into the main body of the flood from the town.
  • Set a 20cm diameter concrete inlet pipe into the outer dry stone running through to the inner cement wall, to allow some water into the compound at the south end of the car park.
  • Raised the level of the car-park surface by 50cm with stone and gravel and terraced along the adjacent slope to form a stone lined channel between the slope and the car park.
  • Excavated a small pond at the NW corner of the car park adjacent to the lower Zone I area.
  • Built protective terraces all around the lower part of Zone 1 so that it is completely protected from run-off coming both from either Gulley 1 or from the flood from the town.

Subsequent to this we added a second wet terrace on the top edge of the lower Zone I area. The small pond and the wet terrace function as silt traps, catching fertile silt from the town watershed like paddies, in which we now grow sugar cane and banana. The water then flows on around the lower Zone II area and out of the site at the front where it re-joins the main flood at the storm drain.

The effects of establishing this system has been to accelerate the growth of trees and clumping plants on and around the cane-paddies themselves, including around lower Zone I but also around the crest of Ridge I where the water flow passes through and along the border of upper Zone I which lies to the north of Ridge 1. This has subsequently established an effective wind break for the upper and lower Zone 1 areas as well as allowing the establishment of many productive trees for shade. The result has been the creation of a sheltered and pleasant microclimate in the bottom of Gulley 1.

We have now hit on a plan to develop this system further giving it the potential to fulfil our water needs for irrigation, not just in the rainy season, but also in the dry season. In December 2012 we built a large barrier dam just below the key-point in upper Gulley 1 of our site. We surveyed the upper gulley and roughly located the key point during a PDC we ran. The dam was built with local labour and the downhill face of the dam terraced in the same way that the local Konso community build water-holding dams in the surrounding gullies and ravines of the landscape.

In order to stabilise the dam further, banna grass was propagated into both the lower and upper faces. The banna grass has now become so prolific that the stone terracing on the lower face can hardly be seen.

The dam has yet to be sealed to make it capable of holding large volumes of water for long periods. The base clay does not hold water effectively, even if compacted, since when it dries it cracks and become porous again. Gypsum and sand were added to the dam-wall material during constriction to reduce the cracking in the wall itself. However it will need to be dug out further and lined on the base and sides in the future. It is also subject to winds from the SE coming over the prow of Ridge 1. We have planned windbreaks along the prow and leeward flank of the ridge, and have planted in different species for this, but they have yet to establish effectively.

Anyway, with the dam in place we are now on the way to establishing a water harvesting system that will be capable of supplying our irrigation water year round. The dam will have a capacity of around 200m3 when complete. It will not be able to catch this much water from its own primary catchment though, as the catchment area is too small. The idea is to harvest water from the town watershed. The two sugar cane paddies are now functioning effectively as silt traps. Our next objective then is to build a harvesting tank of around 10m3 which will fill up from water that has passed through the two paddies and dropped its silt load on the way. This tank should be placed at point X on the map below.

We would hope to be able to fill this tank around 10 to 15 times during a rainy season, hence harvest 100-150m3 of water. Another small pond has already been placed just higher than this in the gulley (marked POND in the map). This pond is placed to receive run-off from the roofs of all the buildings in the Zone 0 reception, office, kitchen, store complex which lies along the pathway to the north, however it needs to be lined with plastic. We are also planning to develop our kitchen grey water treatment system to drain into this pond. We plan to purchase a diesel pump to pump water from these two small collectors up to the large dam in the upper gulley and hence hopefully fill the dam during the course of the rainy season. Our water usage in the Zone 1 area is around 1 – 2m3 per day. Hence 200m3 represents a 3 month supply and should get us through most of the dry season if managed properly.

Building the harvesting tank will require some care. It may get silted up due to the heavy silt load from the run-on water, which carries a lot of fine material, if the cane paddies can’t trap all of it. Hence it needs to be robust enough that it can dug out with a shovel and so it won’t get punctured by someone sticking a spade into it. We need a tank which can be cleaned easily. Our experience with in-ground tanks to date is mixed. We built a stone-cement septic tank with a reinforced concrete floor in June 2011 which has worked nicely and does not leak to this day. Earth-bag is another option which may allow us to build the walls of the tank quicker and more cheaply. In this technique you fill up sacks full of earth, ram them it place around the edges of a pit then cover them in wire mesh and render over them with a layer of cement. Our experience with earth bag so far has not been very good. In July 2011 Steve Cran came and built an earth bag tank in a local school, as part of a project which was funded by the Ethiopia Permaculture Foundation (UK). The cement render cracked because the job wasn’t done well and the tank never held water.

Our septic tank: June 2011

Earth-bag tank at Gocha Primary, built July 2011 (didn’t hold water)

Having taken a bit of time to analyse the situation and what went wrong with Steve’s tank, we have put together a game plan for how we can do a better job. In planning this project we should note the following points:

  1. The base clay on our site is a very unstable material which swells and shrinks a lot during the wet-dry cycles, hence tends to crack and crumble when dry, resulting in movement in the upper layers. Hence massive objects set on this clay will tend to shift and if not integrated well structurally they tend to crack. Even if they are very well bound together some tipping is also possible unless they are deeply founded.
  2. We have an alluvial deposit at our disposal (see later on in the article), which we have trapped behind gabions along the road, comprising a mix of silt, fine gravel/course sand and large gravel with hard-core mixed in.

We will do the following to make sure the tank is structurally integral, stable and hence does not crack and remains sealed:

  1. Dig the bottom of the tank deep (2m) to be sure the base is stable and founded on well compacted soil.
  2. Put down a solid concrete foundation, comprising rubble-hardcore (40cm) and a concrete slab floor (10cm), as we have used on our shower houses. This deep foundation will not shift or tip and the strength of the reinforced slab means that even if there is any pressure differential the slab will not crack.
  3. Use stable soil in the earth bags. We will get this by digging out the alluvial deposit behind our gabion along the road-front off our site. We will wash the silt fraction out of this and then sift it to separate particles <10mm from >10mm. The <10mm fraction comprising sand and gravel (course but not sharp, since it is alluvial) will be used to fill the earth bags. This material is stable and will not swell or shrink with any change in moisture level.
  4. We will use rows of barbed wire set in between the layers of rammed bags to prevent any lateral slippage of the bags over each other. We will also set strands of wire projecting out laterally from between the rows of bags which will firmly grip the cement render layer that we apply over the bags. We will lay down 10mm wire mesh over the surface of the bags, once in place, which will be pinned in place by these projecting barbed wire strands. The mesh will also run across the floor of the tank and be rendered over on top of the concrete foundation.
  5. The pit we dig out will be of dimensions 2.5m deep by 3.8m by 3.8m. However when we set the bags in place we will leave a gap between the wall of the pit and the outer edge of the bags. This will be back-filled with the larger grade >10mm material sieved from the alluvial deposit. Since this material does not swell or shrink with moisture level changes it will buffer the walls of the tank from pressure changes in the surrounding clay which may result from swelling and shrinking.
  6. We will render over the surface of the bags with a 3:1 sand-cement mix. We will make sure the cement cures properly by keeping it damp for at least 2 weeks after it sets.

Taking these precautions in mind we are confident of getting a good tank that holds water at a lower cost than with stone-cement walls, using less cement and less labour.

With the water harvesting tank in place the next stage in the development of the system will be to dig out and line the high dam. This will require a large amount of labour so will be best done by giving a contract to a team of locals from one of the nearby villages. This will need its own budget and we probably won’t be able to get it done this year. However with the harvesting tank in place and shown to be working we will know that it’s worth proceeding with the big dam project and can budget accordingly.

Meanwhile, Further Down the Watershed….

The system described so far has been effective in protecting the car-park and lower Zone 1 area from flood destruction while allowing us to catch some water and bring it to where it is needed in useful quantities. However this system alone has been insufficient to control all the destructive effects of the flash floods coming down from the town watershed. The storm drains installed by the road construction company on the road front shortly before the project began in 2007 were set with too low a spillway, so that a gulley soon started to develop along the edge of the road between the road and the site boundary. In the photos below the storm drain can be seen before project commencement on the left and in November 2009, when one large flood had removed a large amount of material around the drain itself. This cavity subsequently began to be eaten back under the foot bridge along the road front to the south.

We were constantly being advised by locals to build a bridge (concrete slab culvert) into the car park, and drivers were complaining that the entrance was becoming problematic for their vehicles. But instead of splashing out huge amounts for cement and re-enforcement, we decided to install a series of gabions along the road-front. Initially we put a single 50cm gabion under the foot-bridge, however this turned out to be insufficient so we later added a 1m gabion above this, and a second 1m gabion just down-stream of the vehicle entrance.

The result is that we have raised the stable ground level by 2.5m above that of the storm-drain spillway. This allows cars to drive straight into the car park off the road without any problem. We had to heap a berm of material across the vehicle entrance so that flood water does not enter the car park, which the cars drive over. This has proven effective so far in keeping water out.

The gabions have had several positive spin-offs resulting from the huge amount of alluvial deposit and associated moisture which has been trapped along the road-front, rather than being eroded away. This spoil is a sandy-gravel mix with some silt in it. The material is ideal for back-filling terraces and laying down for pathways. Since it is a stable material it will also be good for filling earthbags! We can dig as much of it out as we want and the next flood that comes along will deliver us a new load of bulk material – for free!

The back-fill from the gabions has built right back up the watershed, so that the outer terraced wall we installed for flood defense has now been partially buried. Water is therefore now able to enter the space between the two flood defense walls. However, with the inner cement wall raised to 1m level, no flood has managed to breach the inner wall to date. The site fence, which is a living fence made up of local drought tolerant species, has become a silt trap, leading to a fertile silt deposit in the site corner just outside the car park. Bananas, cassava and sugarcane have become well established and we have even managed to establish young avocados inside this microcosm, which usually struggle in this climate.

The lowest gabion now sits in front of two and a half meters of backfill material, so following rains water continues to seep out from this spot for weeks on end. The result is a damp, cool and shaded microcosm. The gabion’s northerly aspect shades the space in front of them during the winter months and we were delighted to notice that a guava tree had self established at the corner of the lowest gabion, but we’ll now see if it can survive the dry season without being browsed off by herds of goats.

At the foot of the lower gabions we have also planted “shumboko” (Spanish cane) (a deep rooted mat-forming reed species) as well as Agave, along the eastern site border adjacent to the storm drain. These stabilise the soil and prevent erosion. Despite this there is some loss of material to the left hand side of the lowest gabion, which is where the flood from the main gulley watershed leaves the site and joins with the flood from the town water-shed. The lowest gabion was also getting slightly undermined by the water splashing down from it and digging a splash pool at its base. This has necessitated us to add another smaller step 3m in front of this lowest gabion. Which we actually did yesterday:

The Next Stages in flood control and water harvesting system Development

In the short term therefore we have several tasks in the ongoing development of this system which we would like to implement in order to make effective use of the rains this coming rain season.

  1. Establish the harvesting tank mentioned earlier (point X on Gulley 1 map)
  2. The berm at the car park entrance, mentioned above, was placed to stop flood water from entering the car park. This berm has been getting eroded however by the water running along the outer car park wall. Hence we would like to add another gabion cage at the upstream edge of the car park entrance. This will not cut right across the stream flow, but be a partial gabion to push the flow away from the car-park entrance. A second gabion cage will be added at the other side of the stream flow to stop the water from eating out a corresponding chunk from the road edge, which lies opposite.
  3. Before installing these gabion cages however, we would like to lay down an additional in-flow pipe to the sugar cane paddy. This one will enter not from above the car park, but from the side, running from just above the car park entrance, under the car park surface to enter the sugar cane paddy from the side. Having a second inflow pipe will increase the frequency of filling the harvest tank and we should then be able to harvest 10m3 of water even from a relatively small rain event.
  4. We also need to fix up the guttering on our Zone 0 complex which includes our restaurant, offices, reception, store and kitchen. With better guttering installed we also need to put in drain pipes which will lead the roof run-off water directly down to the small harvesting pond. That pond also needs to be plastic lined.

So we are hoping to get stuck into all these jobs and finish some or all of them during our upcoming 4-week internship program in April, 2013. If you are interested to learn about systems like this, both in theory and practice, why not join up and help us implement the design. Actually there’s more to this design story, but I’ll tell you about that later. We are doing a PDC from April 1 – 13th and an internship to follow April 15th – May 11th. The price for the full program combo is only $1,600 and for the internship alone $1000. Sign up and pay up before Feb 15th and get the 10% early bird discount.

Hope to see you here!


  1. Great project and hard work, this site will be education for me and give me some ideas, since we are planning permaculture project in Somalia.

Leave a Reply

Your email address will not be published. Required fields are marked *

Related Articles

Back to top button