Strawberry Fields Update – Flood Protection and Water Control in Ethiopia
One of the biggest challenges of doing Permaculture in a semi-arid place like Konso is the drought-flood hydrology besets in degraded dry-lands. The whole of south Ethiopia has now been so deforested, added to the fact that the global climate is getting completely messed up, that rainfall is now completely unpredictable. The old folks are always talking about it here – “you can’t tell when it will rain any-more, it’s not like the old days….” That makes planning plantings much harder for one thing. The other thing is that when it does rain, it pours.
Our site at Strawberry Fields is placed (purposefully) at the bottom of a watershed and at the junction of this watershed and a larger watershed which carries run-off down the main road from the town.
Rough Topographic sketch of the site at SFEL. Shows approximate
positions of the 3 ridges (R1,R2, R3 and 3 primary gulleys G1, G2 and G3
as well as the Main Gulley on as well as the 2 main flows of run-off
effecting the site.
This means that it gets somewhat inundated at times of heavy rain, which is not necessarily a bad thing, provided we have the infrastructure in place to both protect what is vulnerable from the flooding and catch and store as much of the water resource as possible for dryer times ahead. We have been investing considerable time, thought, sweat, and cash in developing that infrastructure, and we are now eager to see what nature throws at us during the impending short rains, which are now due.
To give you a rundown of the flood defense / water harvesting infrastructure I will first orient you with regards to the site. The site covers 2.2 hectares, located on the edge of the main road from Karat Konso to Arbaminch, at a bend in the road to which the road runs straight down-hill for about 400m from the Town.
In times of rain, the run-off from the western part of the town runs down the ditch on the side of the road and hits the south-western corner of the site. We had not experienced this phenomenon for the first two years of the project’s lifetime due to a lack of rain. But in November 2009 it suddenly rained heavily. The flood flattened the fence at the south western corner of the site (which is where our car-park is located), came over a 50cm high stone-cement wall which was expected to divert the flood, came through the car park and into the lower half of our zone 1 vegetable garden – burying it and the car-park 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 (including our only strawberries!), 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. The flood also removed our wooden foot-bridge at the front entrance and flattened the fence adjacent to it at the point where it joins with the other main water-course affecting our site (more on this one later).
Schematic diagram of flood in November 2009
Our car-park after the first flood in November 2009
The lower half of our zone 1 garden after the first flood in November 2009
Our immediate response to this was to repair our fence at the two points where it had been flattened. The fence is a living fence comprising dry-land species which can be cut and re-planted and will take root simply through being in contact with the soil. This is a normal way of fencing in Konso. Secondly we constructed a dry-stone wall at 45 degrees 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 although when another big flood came some of the water did enter the site lower down, though there was no damage except for mud being deposited on the paths. It was also noticed that water was entering the car-park from a second source, run-off from the site adjacent to ours, as well as from the western slope of one of our own ridges (imaginatively called Ridge 1). The soil on our site and on the adjacent site is a poor quality cracking clay. It contains almost no nutrition, gets very sticky when wet so smothers most things to death quite effectively – bye-bye papayas – but is also very quick to let moisture escape from itself once the rain has passed. Basically it sucks, so being smothered with deposit from this source is worse than the silt from the town.
So, in planning our next move we had to consider the following issues:
- 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.
- The main body of the flood cannot be contained without a massive development of infrastructure.
- We want to catch as much of both the town flood and the silt as we can while not catching the silt from next door. It was also observed that while the outer dry-stone wall was effectively repelling the flood, the water flow running along the face of it was starting to dig and undermine it.
Accordingly we did the following:
- 3 dry-stone gabions (cages made from wire and re-bar filled with rocks) were built projecting 2m out from the wall at 90 degrees.
- A 20cm diameter concrete inlet pipe was set into the wall to allow some water into the compound at the south-west corner.
- A channel was dug along the side of the car park to carry water from the inlet to a small pond which was dug at the far end of the car-park adjacent to the lower part of zone 1.
During the following rains the largest flood smashed the outer dry-stone wall, again came over the inner stone cement wall and buried the car park, entirely filled the small pond up with silt and damaged the fences on the way out of the site near the front entrance.
Subsequently we have responded by:
- Raising the inner stone cement wall by another 50cm to 1m height – which will keep the flood out of the car park if the outer wall is breached.
- Rebuilding and re-enforcing the outer dry-stone wall with re-bar at the critical point of impact of water flow.
- Joining the inlet pipe from the outer dry-stone wall to an inlet set into the inner wall, so that the latter excludes water from the adjacent site, while allowing water from the road into the site.
- Raising the level of the car-park surface by 50cm with stone and gravel and terracing along the adjacent slope to form a stone lined channel between the slope and the car park.
- Re-excavating the pond at the corner of the car park.
- Building protective terraces all around the lower part of zone 1 so that it is completely protected from run-off coming both from gulley 1 or from the road flood.
The pond at the corner of the car-park sits next to the lower part of the zone 1 garden. This is now under drip irrigation, so getting buried again would be a disaster. However it is now effectively guarded on all sides by protective terraces. The out-flow from the pond runs around the lower zone 1 garden, joining with run-off coming down gulley 1 and exiting the site next to the front entrance.
There is the possibility to dig a second pond in the bottom of gulley 1 to receive over-flow from the car-park pond.
The second main source of run-off affecting the SFEL site flows down the “Main Gulley”. The SFEL site actually contains 3 primary rain catchments which site between ridges 1, 2 and 3. These feed into the main gulley, which crosses the site from west to east. The main gulley has an up-stream catchment area of about 25Ha. Since the 25Ha is not tarmacked, the amount of run-off coming from the catchment is less than that coming down from the town. However since the 25Ha is all farmland which is usually dug up, weeded and planted with sorghum at the start of each rainy season, the amount of silt that is carried in the run-off is greater. During the last rains the pond of around 120m3 was entirely filled up with silt. This silt is a resource in itself, being extremely fertile with better water-retaining capacity than the cracking clay soil of the site itself, but having it fill up our ponds is not the best way to use it, as it then has to be dug out and carted off to where we want to use it which is labour intensive. Far more efficient is to get the water to dump the silt where it will be of most use and leave our pond empty, for catching water.
To do this we have built a series of wood and brush check dams. The trick with these is to make them very strong. This is done by digging big posts as deep as possible into the ground and weaving branches between them to form barriers running across the gulley bottom. Where possible established living trees are incorporated into the structure as these are dug in far deeper and more firmly than any dead post you can put in yourself. Brush wood and general organic trash – dry grass, acacia branches etc. – are then thrown in behind the barrier to resist the water flow and catch the silt. The strength of the check dam must be huge as in a gulley of 1m depth and 2m width with a back-fill of 10m the dam will have to hold back 20m3 of wet mud, which weighs 20,000kg. Not to mention the water flowing over all that mud. The dam is not permanent either. Termites, fungi etc start to attack it immediately. Once it fills up and the mud dries out the dam will be at least half under ground. The trick is to put that 20m3 of juicy fresh wet silt to use immediately, get plant roots binding it together and holding it in place. Get branches and stems growing out of it to take over the role of the wooden dam as it rots away. Elephant grass, sugar cane, cassava, bananas, mulberry stakes and sweet potato all go into the silt in rows, running across the gulley, which is now becoming flatter and wider. The water flow is spread across the gulley, split up into different streams which are now meandering into and out of each other. Rather than rushing to carry off all our topsoil, the water is now depositing all the topsoil the farmers up-stream have kindly agreed to donate to us.
Our initial check dams were made using earth. Pits were dug in the bottom of the gulley floor to 1.5m deep, 2m long and 2m wide. The soil was heaped to form a dam on the down-stream side of the pit. Spillways were dug to the side of the dam and lined with stone. These initial dams were quite successful. Though some of them were not big enough and were wiped away by the floods, others stood up and forced the water out of the gulley. The backed up water deposited its silt behind the dams to level off the gulley in some parts, catching large volumes of silt. However the water, once backed up began to flow around one side of the dam quickly digging a new gulley. Thus it became clear that shifting earth about, while a good initial technique, was not a long term solution. It is also very labour intensive in a setting where we use almost no machinery.
Anther technique which we have also used which seems more effective is to construct a wooden check dam. First stakes are driven deep into the ground at the low point in the gulley. It is best to do this at a point where there are some trees growing. If the stakes go in behind tree roots it is better as these will secure the stakes permanently at the bottom. Tree trunks can also be incorporated into the structure, which of course makes it far stronger and more durable. Thinner branches are cut from trees and woven between the stakes into a lattice. It can also be bound with wire to make sure it is completely secure. Then thin sticks can again be woven between these branches in a vertical direction. This structure is strong enough to stand up to the full force of the flood. Now brush, grass and general organic detritus can be piled behind the dam to catch the silt while the water flows thorough. This dam will again back up many cubic meters of silt behind it. Once the bottom of the gulley has filled in the water will start to flow around it, so it has to be extended so that it covers right across the bowl of the catchment floor. What we have achieved by that stage is to spread and slow the water flow so that we are getting continuous deposition of silt rather than erosion.
The next stage is to plant into the accumulated silt behind the dams, get it full of roots which will hold onto it during subsequent run-off, as well as producing food. The wooden structures will quickly rot and be eaten by termites so the aim is to replace them with living materials that will continue to grow indefinitely. Elephant grass is also good to get a fast living dam on top of the old dead structure, but it does not produce food. We put it in initially but then also added sugar cane. Behind them we silt loving species like banana, cassava, sweet potato and also some maize to see how it does. Following the last rains we have now added 5 new check dams to those that were already constructed further up-stream, as well as extending and re-enforcing the older ones.
We are now waiting for the rains to begin in earnest so all this new infrastructure can be put to the test!