Food Plants - PerennialGabionsLandPlant SystemsTrees

Desert Ways

Mongongo Tree

Whether it is an issue of conserving water of using suitable plant species, thriving in a desert environment is a masterful act of management. Permaculture co-founder Bill Mollison has spent time in many of the world’s arid regions and here shares his observations on surviving in some of them.

Building Abundance into Sandy Deserts

Why should we garden, when there are so many mongongo trees in the world? – !Kung tribesman

The mongongo tree (Ricinidendrin rautenii) grows in great groves on the crests of sand dunes in Africa’s Kalahari desert. It is a deciduous tree with two sexes. One in every 12 trees in a grove must be male to pollinate the females.

The tree is immune to drought as its roots tap the freshwater table of the sands. It is easily propagated from great truncheons of cuttings set in the sands 50–100cm deep (20-40 ins). These can be selected from female trees of good yield. Nuts buried 50cm (20 ins) in the sand will also germinate and can be used to select new varieties.

The wood, as light as balsa, is easy to work and widely used for small seats, bowls, drums and artwork. The tree has a long life and is hardy. It suffers from no serious pests and withstands frosts.

Thousands of kilograms of nuts can be harvested from a grove. Comparing the nuts with grains, Lee (1976) notes that they contain five times the calories and 10 times the protein equivalent of cooked rice or maize. The !Kung eat a handful (200g or 7oz) of nuts a day, or 1260 calories and 56g (2oz) of protein. This is equivalent to 113g (4oz) of cooked rice and 400g (14oz) of lean beef. The nuts provide about 50% of the food of the !Kung.

Groves of the mongongo trees are ideal for setting out in any dune deserts where severe frosts don’t occur, either for human food or for forage for game and pigs. It is only necessary to select a variety of males and females, take cuttings and grow them on. Extensive plantings from nuts would give future variation for plant selection. We should try to do this if we have sandy deserts or dune areas. Ref: Lee, R.B. and DeVore, I.1976, ‘Kalahari Hunter Gatherers; study of the !Kung San’, Harvard University Press, Cambridge, Mass.

Important Runoff Findings for Rocky Deserts

It may now be possible to reforest the deserts of Israel’s Negev. Recent findings from old lakes and buried soils reveal the Negev area (lat 38 N) was once forested with oak, myrtle and pines. Desertification came more than 2000 years ago with the introduction of herding and agriculture. Today, annual rain in the Negev is usually less than 200mm/yr (8ins) and, at places, more like half that. At Sede Boker in the Negev, pits and trenches dug at the top of the colluvium below the last rocky ledge grow a great variety of trees adapted to arid lands, without additional irrigation – if planted after the first winter rains. For 120mm of rains we get more than a metre of wetted soil, enough to grow figs, mulberry, jujube and citrus as well as pine, Callitris, tamarisk, juniper, oak and myrtle. It seems possible that reforestation could be achieved not only in the Negev, but in all stony deserts where such infiltration occurs.

It is rare to find well researched field data on run-off water, leading to sound strategies for run-off harvest and tree establishment. The record of Sede Boker (Sde Boqer), on heavily desertified hills with extensive cretaceous sediments (typical of many drylands), is well worth noting. The research has taken place for 15 years starting with careful rainfall measurement. It has evolved to concentrate on the fate of water in overland flow and on infiltration – and therefore on where to plant trees on hill slopes in order 1 of streams (headwaters just below the uplands and just below the ‘breakaways’ or where the small streams are well-defined).

Results demonstrated no particular rainfall pattern in the landscape and averaged much the same pattern across hills and valleys. It is a Mediterranean climate with winter rains but more than 80% of rains yield no runoff. Even in areas where stony surfaces predominate, an episode of rain must exceed 5mm (20pts) before runoff occurs. An increase in isopods (pill-bugs), together with more soil salts and a richness of species in the sparse vegetation, indicates that some areas of the slopes infiltrate water. (Neither worms nor termites have significance here as desertification is severe.)

Above the point of infiltration (where the more solid rocky surface of the convex slopes form a ledge or drop-off) and below this point (where colluviums if material wasted from the upper slope gathers and deepens downslope) there is a dramatic difference in both runoff and infiltration.

On the upper rock slopes, an average of 24 episodes of runoff annually provides 3.5 litres of water per sq m per year. On the colluvium, only 2-3 episodes of runoff occur and this yields only 0.5 litres per sq m per yr.
Hence, water yielded by the upper slopes is 12 times more frequent, and seven times heavier (a total of 84 times more plentiful) than runoff on the colluvial boundary. Also because it does not seem to enhance the stream flow (which occurs 2-3 times a year), it is clear this water infiltrates at the base of the rocky slopes and is absorbed by the colluvial mantle. Israeli scientists have found that this runoff pattern holds true for stony deserts in South America and Africa.
Professor Moshe Sharter of the Jacob Blaustein Institute of Desert Research, Ben Gurion University at Sede Boker, explained all of this to me on the site itself.

Liman or Valley Catchment in Israel

Linman is an ancient and, more recently, widely used method of patch reforestation in Israel successful in growing a wide variety of trees. It is essentially a bund of earth about 2m (6ft) high and 4-6m (13-20ft) through at the base which is carefully sculpted across the valley flood-flow.

Rainfall of 70-100mm (3-4ins) and floods two of three times a year, yield enough water in perhaps 10% of the landscape to fill small fields – up to 4 acres (1.6ha) – with floodwater. Any surplus can be split to another liman.

At Wadi Mashash, Professor Area Rogel – who worked with Michael Evenari in the Negev – grows olives, grapes, figs, acacia, eucalyptus, leucaena, carob, oaks, stone fruits and almonds in a succession of liman. Crops or intercrops comprise sorghum, wheat, adapted maize and legumes.

While building the liman, Rogel uncovered buried stone walls, about a metre (3ft) wide and deep across the valleys. It is thought these walls were once a series of “leaky” liman, now silted over, built more than 2000 years ago by Jewish settlers.

To ensure about a metre (3ft) of water a year, liman might not use all of the flood plains. But we must allow about 20 acres (8ha) of runoff to each acre of liman. This proportion can be increased (1 acre of crop to 5 acres of runoff) as rainfall increases toward 500mm (20ins) a year.

The flood levels are clearly marked on the inner walls of the liman by a line of weeds and straw trash carried by floods. Great care must be taken to spill into grassed or stony spreader drains as any concentration of floodwaters causes rapid gullying in valley sediments.

In the last 40 years, hundreds of liman have been have been made and planted to trees, mainly as shelter or forage for herds. Wadi Mashash, however, uses very limited grazing. Even so, many trees are bark damaged by goats. One senses that any reforestation must be accompanied by protection from the numerous herds of goats that traverse the Negev.

Small valleys may be fully bunded across, but liman in the sides of large valleys may have a low diversion bank directing water to the upstream entry of the liman. Spillways are usually close to the entries, often in the opposite side of the liman.



  1. great to see that the mongongo nut can be propagated.Would you please write about taro/rice padie method of consruction and use

  2. Fascinating!! Perrenial tree crops providing staple carbs, proteins and essential nutrients seems a vital subject/objective for any self-suffient food system… Over use of grains appear to be an inefficient use of land and ecological overstretch etc… Would love to hear Bill and others clarify these and similar issues if possible (I must research existing publications). Seems to be issues vital to permacultural praxis…

  3. Another good perennial tree for carbs is chestnut! Thanks for the article bill!

    I read about the negev desert in brad lancaster’s book “rainwater harvesting for drylands and beyond” – very inspiring since I live in a mediterranian climate on the central coast of california.

    I’m inspired to build swales and be a steward to the watershed, but am wondering in areas where there is little to no runoff, they may not be worth the energy.. still something I am looking into. Example, from what I’m gathering, sandy soils may not be worth the effort, but clay soils would be.

    As the opportunities that are available to me are in suburbia or small (1acre-ish) lots, there is not much room for earthworks, and so I find myself choosing between using irrigation to increase yields in the small space and simplifying and planting climate adapted or drought tolerant perennials.

    does anyone here (incluidng bill) have an opinion on drip irrigation? it is a good way to conserve moisture, but is it sustainable? I am reading how they do recycle drip tubing now, but can that plastic be recycled back into more drip or is it of lesser quality? I’ve also heard about bio plastics that can be made to be robust, but not sure of the energy required to process it.

    So my feeling is that vegetables and most water intensive plants be done in the green belt outside of a town, where there is land to build dams/swales or there are rivers to utilize. while perennials can be planted all over the city (and outside as well), utilizing roof gutters to flow into deeply mulched basins.

    Ahh sustainability questions… what a journey it will be.


Leave a Reply

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

Related Articles

Check Also
Back to top button