Turning a dry riverbed/wadi into a catchment

I'm working on restoring tree cover to a mountainside not far from my house in Israel. The land is limestone, and used to be forested, but was deforested. It's now covered in very low garrigue thorns. There is excellent terra rossa soil in pockets, but the bedrock won't hold water. There is a steep wadi nearby, which drains a fairly large area into the Jordan River Valley when it rains. It is fairly warm in the summer, hitting 90s, and gets cold but not freezing in the winter, with strong winds.

A friend and I have run an irrigation pipe down the mountain and put in 30 seedlings (plum, pear and fig.) We'll be expanding with pomegranate, carob, black locust, some eucalyptus, narcissus and prickly pear. The idea is to eventually have an entire forest which will self-propagate.

My question here is irrigation. I want to become independent of the central water system. I would like to build a series of ponds in the wadi using reinforced concrete or ferrocement, to keep water in over the dry months. What are the advantages/disadvantages? How do I do this the cheapest and simplest way possible? What's a good way to manage sediment? Can anyone give general answers/recommend books I should read?

My experience, is that water comes down the wadi and washes everything away with it. And anything left behind is swamped with sediment. Brad Lancaster's Water Harvesting for Drylands was very valuable to me. The most important lesson to me was that it's the small things that are most appropriate. All of my big structures washed away with the first rain. It was the small ones that survived.

What I found worked best was single layer rock checkdams, laid perpendicular to the water flow. They fill up with sediment after the first rain. Then add another layer. They fill up with sediment after the second rain. etc... Just keep adding layers over the years and capture the sediment and make a nice flat area in the ravine. The sediment gets saturated with water after every runoff event, and slowly releases it downhill. And retains water in the sediment for plants to use. Throw a pipe in before you start if you want, which may release some water for a while after runoffs. Springs are likely to show up downhill anyway. Especially if a bit of bedrock is exposed.

Best to start work at the top of the drainage, and work down-stream.  

Rock-filled wire-basket gabions may be a better material to work with than cement, because they can shift with the stream.

Here's a few photos from a wadi that I take care of...

First photo is a single-layer rock checkdam that was built on top of bedrock. The area behind it filled with sediment during the first big rain storm. I built a second single-layer rock checkdam on top of it and it filled with sediment during the second runoff. Water pools on the bedrock below for about a month after runoff events.

Second photo is a wire-fence checkdam. It captured trees and vegetation during the first runoff after it's construction, and more than 6 feet of sediment, so it is equivalent to a hugel check dam.

Third photo shows vegetation growing in the moist sediment collected by the wire-fence checkdam.

Fourth photo shows a wire-gabion checkdam. It likewise filled with sediment during the first runoff event after it's construction. It leaks water onto the bedrock below for a month or two after runoff events.

Thank you, great answer. Will research and implement.

>single layer rock checkdams

Just dry rock?

> wire-fence checkdam

What is the difference between this and gabions?

>Springs are likely to show up downhill anyway. Especially if a bit of bedrock is exposed.

What's a good way to capture this water and move it across the slope, do you think?

With gabions, they will fail eventually with the wire eroding. What happens then?

I would expect that the gabions will silt up becoming effectively solid :-) Its going to take some years :-)

You say you are in Israel, so I suspect that your climate is mild and you don't suffer frosts?

If so then you should seriously look into applications of vetiver grass. Planted on contour in dense hedges it traps sediment and significantly slows surface water flows. Natural terraces build up behind it in which other plants can be cultivated, and as the depth builds the vetiver continues to grow with it. Vetiver makes masses of biomass which can be cut with a chop-n-drop approach, protecting the soil from evaporation and building soil carbon, and the deep roots allow it to collect water from deep beneath the surface, even in very dry conditions - once it has got established.

In the situation you describe I would start high in the watershed, planting rows of vetiver on contour to form hedges. Once these are established it will go some way to moderating the flows in the creekbed, and will drastically reduce the soil loss from the land.

Around the creek itself I would establish vetiver on the bank to stabilise them. The roots are incredibly deep and strong and provide excellent defence against erosion. You might also try planting vetiver across the flow in some places - starting high up the water catchment. Once established these plantings will slow water and trap sediment in the channel, much as a rock or wire check dam would.

Vetiver planted across a drainage ditch to trap sediment and slow flow:


A video explaining the application. At about 1m 40s there is an example of a cross section, showing the trapped soil and the sediment build up.

The area I'm working is right below snow line (we get snow up top almost every winter.) Will vetiver grass do well there?

I would guess not in that case, sorry. It isn't frost tolerant.

Other grasses and other plants will do much the same as vetiver.  I am intending to use transplants established of chives for this purpose where I live; which is a lot more temperate and cool.

Joseph gave a lot of extremely good info above.  I would particularly stress that you should start at the top.  The more you catch water at the top, the more you will get some in the lower area in the future, and the less water pressure on any structures at the top, and as a result of the structures, the less pressure downstream.  Plant nurse trees of deep rooted and nitrogen fixers on the upper slopes of the catchment near small side channels, immediately adjacent to the narrow spots in the main channel where you plan to put check dams.  The higher up you are the more likely that a more substantial (solid, taller) dam will hold back the full force of the flow.  The more this land is forested, the more the water will infiltrate into your groundwater storage and not go into the check dam.  Any and all swales and other catchment that is possible in the upper slopes of the hill/mountain, should be done to get water away from flowing/eroding your wadi, and infiltrating instead.

In the highest reaches of your catchment and side channels, a single rock, placed in the channel will hold back sediment.  Place rocks in 'V' shaped lines that point upstream with small gaps between the stones.  The largest rock in the center/top.  Then do another one close downstream.''

 

At my place, I don't try to capture the torrential flash floods that rip through the wadi. They are too extreme. They carry too much sediment. What I can capture, is a heck of a lot of water filled sediment. And I can aspire to fill the preexisting sediments with water. Then the water in the soil can be used by plants, that grow roots that can help stabilize the soil, and they grow stems that can slow the flow of the next flash flood.

The wire fence is a singe layer of wire. A gabion is a wire basket filled with rocks. Gabions are much more stable.

At my place in the desert I have two materials that are great for making small check dams. Rocks, and tree branches obtained while doing fire-suppression pruning. Both go into the wadi to help slow the flash floods.

I also make swales at my place. The swales are not put in the wadi. They are put in areas where there is only gentle sheet flow of water. As with the checkdams, the purpose of the swales is not to capture water for irrigation, it is to put water into the soil. Plants can use the water from the soil, or the water can seep into the ground and perhaps it will come out down-slope as a spring or seep.

Hopefully, by the time a gabion fails, there is enough vegetation growing in the area to retain the soil.

In my ideal world, I would like the flash flood water in the wadi to be an inch deep and 60 feet wide, rather than a foot deep and 5 feet wide. Much of my work in the wadi, is designed to spread a shallower flow across a wider surface area.

The book "Growing Food in a Hotter, Drier Land" by Gary Paul Nabhan may be of interest to you. He's studied desert farming in the American Southwest and has a lot of info about capturing runoff in dry climates.

The book "Growing Food in a Hotter, Drier Land" by Gary Paul Nabhan  

 I don't know this book, but I have met Nabhan, read one of his other books (The Desert Smells Like Rain), and visited the site of his project in Patagonia Az back in 2004.  If this book is anything like his previous work and projects then it is very worthwhile.  

Just going over some course notes from my Geoff Lawton PDC.

He mentions building check dams, and slightly upstream having the water diverted laterally via drains; these take the excess water that is held in the check dam and spread across the slope via swales, spreading it gently down slope via spreader banks (long low sills), or dumping it into ponds off to the side.   He suggests willows or casuarinas to be planted initially, to create stable root systems to support the soil/land, and to create quick shade which reduces evaporation and creates a favorable microclimate for secondary species.

A perforated drainage pipe can also be placed on the lower inside of a check dam and protected by large rocks (preferably round rocks that will allow water to move around them.  This pipe will take water into it, through the large round rocks, and divert it laterally to fill a long swale or pond, or, if it is clean, go into pipes directly.  Generally, the water is put laterally into a swale or diversion drain that will dump into a pond; the pond filters and settles the silts and organics, and a pipe then drains the pond to your water system/irrigation lines.

I would like the flash flood water in the wadi to be an inch deep and 60 feet wide, rather than a foot deep and 5 feet wide. Much of my work in the wadi, is designed to spread a shallower flow across a wider surface area.

 In permaculture terms from the PDC this would be one of the basic principles of water usage:  Slow it down, spread it out, and soak it in.  Generally the longer you can have the water stay on your property and soak deep into your soils, and be utilized by plants, the better.

My situation is that I want to do most of my growing on the slopes above the wadi, not in the wadi itself. The area gets decent rainfall half the year, and those slopes are overgrown with thorns and briars. Because of the steepness of the wadi, significant portions of those slopes sit below its upper/mid portions. So while I will probably use gabion check dams to spread out the flow and catch silt (and maybe grow vegetables on the silt in the rainless half of the year,) I need to think about how to get that water that runs off collected and piped across the slope (building hundreds of meters of terraces/swales is not currently an option because of manpower and budget restrictions.) I can see lots of ruined Second Temple period terraces, and some of them do seem like they once served this function, but rebuilding them is outside my grasp right now...

I can understand the entire terrace rebuilding project to be beyond your immediate reach, but I think something like that is going to be the only way to harvest and retain the water. Starting with one check dam where it can feed the highest terrace with modest repair work would get you started; is there any other way to start retaining water?

To go from the highest terrace to my project would mean building about 400 meters worth of terrace.

Baruch Kogan wrote:To go from the highest terrace to my project would mean building about 400 meters worth of terrace.

Rebuilding 5 meters of terrace per day gets the project done in less than 2 months... Building 5 meters per week gets it done in 1.5 years.

I don't see how a forest can be self-perpetuating unless it is growing where water flows naturally. You might consider relocating your project to where it can take advantage of the wadi and terrace structure.

If moving the wadi water to the project is beyond your means, then is it within your means to use the water where it currently exists: In the wadi? The nearer to the wadi your project is, the easier it seems to use the water from it.

I built 120 meters of swale by hand in a couple days...

I found a video of a wire gabion that I built in a ravine in the desert. It filled with sediment during the first run-off after construction.

This is fabulous information