02 April 2020

Our challenges with Cape Flats sand in our gardens.

We have a real challenge on the Cape Flats. The soil is hydrophobic! Here is an image showing nice wet surface, but I walked on it and the top few millimetres simply came off on the soles of my shoes leaving dry, dry sand below it. The water simply does not penetrate the sand.

The water simply sits on the surface with its convex meniscus as though it is on a greased surface. Its water holding properties are next to nothing! It is low in nutrients - a feature of Fynbos soils. The book describing The Vegetation of South Africa, Lesotho and Swaziland, and the Fynbos Biome in particular, show for:
  • In the Fynbos Biome description, for the Pediment and Valley Floor Soil of the Fynbos: A great range of sandy soils, usually acid and highly leached;
  • FFs 5 Cape Flats Sand Fynbos: Geology & Soils Acid, tertiary, deep, grey regic sands, usually white, often Lamotte form. Land types mainly Ga, Hb and Db;
  • FS 6 Cape Flats Dune Strandveld: Built mainly of Tertiary to Recent calcareous sand of marine origin and overlying metasediments of the Tygerberg Formation (Malmesbury Group, Namibian Erathem). Outcrops of Sandveld Group limestone (hardpan) are found on the False Bay coast (Cape Peninsula and especially the Wolvengat area). Dominant land type Ha (about 50%), with Hb and Ga playing subordinate roles.

The vegetation and soil maps and descriptions can be viewed using the Biodiversity GIS map viewer.

I was wondering about the acidity or alkalinity of the 'soil' in our garden. It is complimentary to refer to Cape Flats sand as soil. It does have some organic matter - I add compost to it, and there is decaying litter in it - but this organic matter seems to add to the hydrophobic properties of the sand.

I decided to look at the soil texture using the sediment method https://www.thoughtco.com/quick-sediment-testing-particle-size-1441198. See also the mason jar soil test https://preparednessmama.com/jar-soil-test/.

I tried the vinegar - there is distinct bubbling or fizzing, suggesting that this is not acidic soil.

I tried the bicarbonate of soda - the powder was very slow wetting up, but there may have been some degree of fizzing, but it is really not distinct. I was not sure if one should try to first dissolve the bicarb in water, hot or cold, or simply pour the half cup of powder onto the mud! I kind of tried both powder and solution.

In conclusion, I reckon my sand is alkaline (basic) rather than acidic.

I bought some dolomitic agricultural lime because I seem to recall hearing many years ago that this should help for water to penetrate our sand. Well, I tried sprinkling some on the soil and then pouring on water, but that did not really seem to make a difference when compared to pouring water on bare sand. I then prepared a flat area of sand in a basin, and made ten depressions and poured on 20 ml of water onto each depression. I returned after nearly 20 minutes, expecting the water to still be sitting on the surface, but the water of all but two depressions had penetrated.

Water penetration in depressions

This was evidently not a true reflection because it had rained in the past 36 hours and there is a bit of moisture in the sand, but it does show how the sand is hydrophobic (repels the water that simply sits on the surface as though on an oily surface).

20 ml of water with or without lime poured into each depression.

30 seconds later, from a slight angle, showing the water sitting on the sand, 
and the whitish lime more evident in 5 depressions.

24.5 minutes later, all water penetrated except for one depression without lime. 
No apparent difference with using lime. The white lime is visible in 5 depressions.

I later repeated this with new depressions so as to get more frequent photos, shown later.

I then poured some water on the soil surface.

 a small depression for pouring water -  and water being poured

After 6 seconds, 13 seconds

After 45 seconds and 80 seconds

All had penetrated the soil after 2 minutes and 20 seconds. 
The soil has some moisture from the recent rain.

I made more depressions and poured in water

This allowed me to take more photos to record the penetration of the water.
Water in new depressions, soon after pouring in, and after about 7 seconds

96 seconds                            5 min 38 sec

7 min 1 sec                                            9 min 16 sec

10 min 5 sec, then the last water sank into the soil after 11 min 33 sec

I added a second lot of water in each depression

Soon after adding the water to wet depressions and after 6 sec.

After 3 min 7 sec and then 3 min 21 sec.
This is far more rapid than the penetration in the first instances, taking more than 24 minutes.
The red rings indicate the lime treatment for first watering, 
and the green the lime treatment for second watering.

I then inundated the sand that had been used for the wet depressions

In the process of inundation - notice the convex meniscus. 
To the rights is 1 min 49 sec after inundation. 
Note that sand at top and right is dry after the water receded after it had been covered.

2 min 33 sec and 4 min 31 sec

All water has entered the sand after 6 min 14 sec

Although the water has all penetrated the sand, it does not move easily within the sand. 
This is 7 min 22 sec after inundation.

This is 9 min 21 sec after.

The dry sand flows easily from under the moist surface layer - 10 min 50 sec.
The penetration is deeper beneath depressions and absent between ridges.
Some of the sand right at the bottom is still fairly dry. After 12 min 6 sec.

Testing the effect of litter and lime in a soil depression

I also did another small comparison with four rainwater, four rainwater with lime, two of each with litter in the depression.
Immediately after adding water, and after 2 min 37 sec.
Lime is on the left, plain rainwater on the right, litter in top two and bottom two depressions.
Note that water from the top two depressions came into contact. 

4 min 14 sec and 5 min 38 sec.

7 min 40 sec and 9 min 42 sec, all water has gone into the sand.

Sample 1 Sample 2 Sample 3 Sample 4
Rainwater 04:14:00 00:05:38 00:11:14 04:14:00
Lime + rainwater 00:02:37 04:14:00 04:14:00 04:14:00
Litter 00:02:37 00:02:37 00:02:37 04:14:00

Second application of water 

Soon after second application and after 5 minutes. 
The water is not obvious, so I turned on the flash to reflect off the surface for later shots.

7 min 10 sec and 8 min 9 sec.

17 min 6 sec and the sand below the surface after 44 minutes of first application and about 28 minutes following second application. Note how the water does not spread within the body of sand.

Sample 1 Sample 2 Sample 3 Sample 4
Rainwater 00:05:06 00:08:09 00:08:09 00:05:06
Lime + rainwater 00:05:06 00:08:09 00:08:09 00:05:06
Litter 00:05:06 00:05:06 00:05:06 00:05:06

Soil texture

The sand is fairly pale greyish beige, with blacking dust from organic matter. This is a winter rainfall area and so the summers are hot and dry. The plant matter is derived from the local natural vegetation that is called Cape Flats Sand Fynbos, although as the soil appears to be alkaline it would probably more correctly be Cape Flats Dune Strandveld. The sand is normally dry with a very fine black dust that seems to contribute to the hydrophobic nature of the soil. This dust blows readily when sifting the sand, mowing the lawn, or in the regular strong winds on the flats.

Determining the soil texture using the sediment method https://www.thoughtco.com/quick-sediment-testing-particle-size-1441198. See also the mason jar soil test https://preparednessmama.com/jar-soil-test/,

The sediment in the jar after 24 hours is 110 mm, with a portion of 13 mm floating at the top of the body of water.

When inverting the bottle, the sediment simply remained in place and did not become suspended again without several shakes of the bottle. After shaking the bottle again and measuring the sediment after 40 seconds, there is about 95 mm, but the water and sediment are all so black that it is really difficult to see where the interface was between sediment and suspended matter. Light does not shine through the water to help determining the solid and liquid interface.

After 30 minutes the sediment is 108 mm and the floating portion 8 mm. The water is turbid.

After 24 hours, there is still about 10 mm floating, the line between greyish sand and blackish humus at about 85 mm, and the humus layer about 25 mm thick. The water is brown although clear enough to see some things floating in it although one cannot see through it. Knocking the bottle results in some floating material sinking and some sedimented material rising. Some bubbles also rise if the lower portion is disturbed by squeezing the bottle.
After shaking, after 24 hours and after 48 hours.

24 hours

Water penetration in fractions

The soil was sieved using mesh of 2 mm, 0.5 mm and 0.2 mm. The litter is referred to as that which did not pass through the 2 mm, coarse fraction passed through 2 mm, medium passed through 0.5 mm and fine through 0.2 mm. The recognised soil particle sizes are: Coarse gravel > 5 mm > Fine gravel > 2 mm > Coarse sand > 0.2 mm > Fine sand > 0.02 mm > Silt > 0.002 mm > Clay.
See and Daubenmire, R.F., 1948. Plants and environment. A textbook of plant autecology. Wiley International 

I used sieves with mesh as shown: 2 mm, 0.5 mm and 0.2 mm

Fractions were put into syringes to see the different rates of penetration of water. The fractions had stood for about 15 weeks and had good chance to dry out. Each syringe was shaken up after 10 minutes to mix the water and soil and then photographed again after 5 minutes and 10 minutes.

Litter fraction: The water went in immediately and 15 ml fraction and 15 ml water made up 20 ml volume. This did not change after 15 minutes, and remained after shaking the mixture.
Coarse fraction: The water went in within a few seconds and 15 ml of sand and water made up about 20.5 ml. The volume remained 20.5 mm after 15 minutes and reduced to 20 ml after shaking.
Medium fraction: The water went in within a few seconds and 15 ml of sand and water made up about 23 ml. This remained for 15 minutes, and reduced to 22 ml after shaking with about 1 ml of dark organic sediment on top of the sand. 
Fine fraction: The water went in about 5 mm within a few seconds and 15 ml of sand and water made up about 25 ml in the first half minute. The water had gone about 95% of the way through the soil within 5 minute and the volume was 20 ml. The volume remained 20 ml for 15 minutes, and reduced to 19.5 ml after shaking with about 2 ml dark organic sediment on top of the sand.

I know that this is not all a randomised design with replicates and does not lend itself to statistical analysis, but it might give someone an idea or two for undertaking a proper experiment that might help gardeners like me who have to deal with soil that is similar to the Cape Flats sand. I think that our greatest challenge is the hydrophobic nature of the soil. Even with the addition of compost, the water does not enter the soil to feed the roots, but sits at the top, or might go into a very shallow top layer, and then be more able to evaporate. Even when wet, the soil is low in nutrients.