The deep preparation of vineyard soils has a long history in South Africa. Abraham Perold (1926) had already discussed the pros and cons of “delving vs. ploughing” and based his recommendations on soil type. Deep soil preparation has however, only been scientifically grounded with the visit of H. Schulte-Karring (1976) and publishing of the results of soil preparation trials at Robertson and Stellenbosch (Claassen, Van Zyl & Kleynhans, 1973; Saayman & Van Huyssteen, 1980; Saayman, 1982). Since then deep soil preparation on wine farms has become an essential practice aiming to break up compact soil layers, as well as abrupt transitions between soil horizons down to a depth of 1 000 mm. Soil preparation also provides an ideal (and often the only) opportunity to rectify unfavourable soil chemical conditions such as high acidity and low phosphate content.


Soil layers with a high density limit root penetration forces roots to only utilise the overlying soil (Schulte-Karring, 1976; Saayman, 1982). The inability of roots to penetrate into compact subsoil layers is related to the lack of macro pores in these layers (Van Huyssteen, 1989). Roots cannot penetrate a pore which has a smaller diameter than the root tip (Wiersum, 1957). A root cannot decrease its diameter, but will rather thicken when the growing tip experiences resistance. The presence of a good distribution of macro pores is consequently a prerequisite for good root distribution in soil. Various researchers also found a positive linear relationship between rooting depth and aboveground grapevine performance (Saayman & Van Huyssteen, 1980; Van Huyssteen, 1982, Myburgh et al., 1996).


In the past the results of different cultivation practices have been characterised by measuring bulk density, soil strength, shear strength, infiltration and soil hydraulic conductivity. Impediment of grapevine root growth was successfully characterised by Leopoldt van Huyssteen (1983) in terms of bulk density and soil strength. Soil strength was measured by a constant-speed penetrometer. In a wide variety of soil types, a critical, albeit poorly defined, penetrometer resistance of 2 000 – 2 500 kPa was found above which root penetration is drastically impeded (Zimmerman & Kardos, 1961; Taylor & Gardner, 1963; Taylor & Burnett, 1964; Greacen et al., 1969; Bar-Yosef & Lambert, 1981).


The current study was conducted at three locations in the Western Cape in order to determine i) whether the positive effect of deep soil preparation on the soil physical condition is sustainable in the long term and ii) which methods best indicate the physical differences between different soil preparation practices.


Material and methods


Measurements to determine the degree of re-compaction were conducted in an old soil-volume experiment where soil preparation was done in 1984, i.e. 26 years previously. In this 1984-experiment a series of soil-volumes were created using an excavator on a Cartref soil of granitic origin. After mixing the excavated soil with calcitic lime and superphosphate, the soil was backfilled into the open trenches (Myburgh et al., 1996). The treatments thus created, appear in Table 1. In the current study only eight of the 10 treatments were included; the 400 mm irrigated and 1 200 mm irrigated treatments were excluded.


TABLE 1. Soil preparation treatments (1984) applied in the soil-volume trial at Nietvoorbij (Myburgh et al., 1996).

Soil preparation treatments
Dryland Irrigated
Control, ploughed to 200 mm
Ridges, 400 mm high
Trenches, 400 mm deep Trenches, 400 mm deep
Trenches, 600 mm deep
Trenches, 800 mm deep Trenches, 800 mm deep
Trenches, 1 000 mm deep
Trenches, 1 200 mm deep Trenches, 1 200 mm deep


This sandy clay loam soil contained 50% gravel and had a high bulk density in its natural state. After soil preparation, the soil was left to consolidate for a year and then it was planted to Pinot noir/99 Richter. At that stage soil bulk density was a favourable 1 550 kg/m3 compared to 1 750 kg/m3 of the control. Penetrometer studies at that stage, i.e. a year after soil preparation, also showed that the soil was effectively loosened to the pre-determined depths. No vehicle traffic was allowed in the vineyard after planting.



The study at Broodkraal near Piketberg investigated selected physical properties of a Tukulu soil on three plots that represented the following soil treatments: i) control (natural state), i.e. undisturbed soil next to the other two treatments on the same soil type, ii) newly prepared soil (soil prepared in 2010 and measurements taken in 2010), and iii) soil one year after soil preparation (soil prepared in 2009, but measurements taken in 2010). The last-mentioned treatment was prepared and planted to Thompson Seedless/ Ramsey in 2009. Soil preparation consisted of cross ripping to a depth of 1 000 – 1 200 mm followed by construction of ridges.



The result of two soil preparation methods were compared in 2010, i.e. six years after execution in 2004. The soil was classified as Tukulu and Oakleaf forms. The first method involved 600 mm deep ripping followed by delve ploughing, while the second treatment consisted of ripping (600 mm depth) followed by tilling with a soil mix implement (Fig. 1).



FIGURE 1. Soil mix implement used for soil preparation at Kanonkop.


The soil mix implement was specifically developed to work between vine rows. A rotating cylinder of 800 mm diameter equipped with tungsten tines that give it extra reach, creates the loosening action of the implement. Both soil preparation methods aimed to loosen the soil down to a depth of 1 000 mm. Land prepared by these two methods was planted to Cabernet Sauvignon/99 Richter in 2004. Due to the nature of the treatments, they could not be replicated, but comprised large areas of land adjacent to each other. Undisturbed soil of the same soil family in an adjacent vineyard served as a control.


Soil physical measurements

Soil bulk density was determined in triplicate at various depths of the different treatment plots. A pocket vane tester was used for determining shear strength (Fig. 2). This instrument was placed against the profile wall, the vanes pushed into the soil and the instrument slowly turned clockwise until the soil fails. Since soil strength is a highly variable property sensitive to changes in soil water content, all shear strength measurements were executed at field capacity after rain on the same day in a specific trial.



FIGURE 2. Measurement of soil shear strength using a pocket vane tester.


Soil strength was measured by a constant speed penetrometer equipped with a rod that can reach a depth of 800 mm, and a conic tip (30° angle) with a base area of 1.3 cm2. This particular instrument can take measurements at 10 mm intervals and to a maximum resistance of 5 000 kPa. Soil water content during penetrometer measurements was close to field capacity. Penetrometer measurements were taken at 10 randomly positions on each treatment plot of the soil-volume experiment at Nietvoorbij. Due to equipment breakdown, none of the other two trial sites could be included in the penetrometer study.


A double-ring infiltrometer (Fig. 3) was used for measurement of saturated hydraulic conductivity from the soil surface. The apparatus consists of two vertical cylinders filled with water and two double rings which are placed on the soil. The infiltrometer is mounted in such a way that one cylinder supplies water to the outer ring and the other cylinder supplies water to the inner ring. Water from the outer ring cancels lateral water flow from the inner ring and measurements are only taken of the latter. Measurements were done in duplicate on all treatments of the three trial sites, namely Nietvoorbij, Broodkraal and Kanonkop.



FIGURE 3. Double-ring infiltrometer for determination of saturated hydraulic conductivity.


A mini disc infiltrometer was used to measure unsaturated flow in the soil. Measurements were taken at 150 mm increments from the soil surface down steps in the profile wall. Infiltration rate was determined at a tension of 0.1 kPa on loose soils, while 0.05 kPa was used on compact soil to make provision for the slow water infiltration under those conditions. Measurements on each treatment plot at both Nietvoorbij and Kanonkop were replicated at two sites, but Broodkraal was not included in these measurements.


Root studies were conducted at one grapevine per treatment at Nietvoorbij and Kanonkop, using the profile wall method. In addition to mapping their positions, roots were also classified in different diameter classes of fine (<2 mm), medium (2 – 5 mm), coarse (5 – 10 mm) and thick (>10 mm).


Results and discussion

Twenty six years after soil preparation, the beneficial effect of this action could still be detected on most of the treatment plots at Nietvoorbij by the majority of soil physical determination methods. Penetrometer readings were 1 969 kPa (2 000 – 2 500 kPa is accepted as the upper boundary above which no root penetration can occur) on average in the 200 – 600 mm soil layer in comparison with 3 538 kPa at the same depth in the control. The abrupt increase in soil strength below the working depth of the excavator could also still be measured. Despite the lasting difference between loosened and undisturbed soil, serious re-compaction had taken place and it was clear that deep soil preparation would be necessary before a new vineyard could be planted.


The penetrometer results also corresponded well with the compaction pattern indicated by the bulk density values. Current results confirmed the findings of previous studies, namely that penetrometer resistance is a sensitive, accurate and quick method to measure soil strength. Automated constant speed penetrometers are however expensive, but as an alternative a steel rod can be pushed manually into a wet soil to discover compacted layers. Although such a crude method will not give quantitative values, together with experience it can be used to give a fair indication of soil compaction.


Measurement of shear strength on the three test sites using the pocket vane tester gave variable results. This instrument did not perform well on the gravelly Cartref soils at Nietvoorbij, but on the fine textured soils of Kanonkop and Broodkraal it could successfully distinguish between the undisturbed controls and their matching deep-tilled treatments. Shear strength could even detect re-compaction in the deep-tilled soil at Broodkraal after one year (Fig. 4).


FIGURE 4. Shear strength on treatment plots at Broodkraal.


Saturated, as well as unsaturated, hydraulic conductivity showed limited differences among soil preparation treatments at Nietvoorbij. High values of unsaturated hydraulic conductivity were caused by preferential flow since the apparatus is particularly sensitive to changes in soil macro porosity. Results obtained with saturated hydraulic conductivity and bulk density measurements did not agree consistently. Bulk density and penetrometer resistance were more sensitive to soil preparation effects than saturated hydraulic conductivity. More differences between treatments would probably have been found if infiltration time had been increased to three hours and even longer.


Root distribution correlated well with depth of soil preparation at the two sites where root studies were conducted. Despite serious re-compaction at Nietvoorbij, grapevine root distribution in the soil-volume trial still reflected the positive effect of deep soil preparation 26 years after it had been carried out (Fig. 5 and 6). An explanation for this long-term benefit of soil preparation is the fact that roots can explore the full potential rooting depth within two years after planting thus establishing the root frame very early, i.e. before re-compaction normally takes place. Thereafter rooting depth remains constant, but rooting density increases with time. Although a very reliable indicator of soil conditions, root mapping is tedious and time consuming, but well worth the effort. In fact, root distribution remains the ultimate indicator of soil conditions for grapevine growth.


FIGURE 5. Root distribution on a control plot at Nietvoorbij.


FIGURE 6. Root distribution on a deep-tilled (120 cm) plot at Nietvoorbij.


In the case study at Broodkraal, bulk density not only showed the positive effect of soil preparation, but also quantified soil re-compaction that took place within one year after deep tilling (Fig. 7). Such re-compaction is a natural process due to, among other factors, soil wetting and the weight of the soil mass itself. Despite re-compaction, the soil physical condition (bulk density values <1 650 kg/m3) would still allow grapevine root penetration.


FIGURE 7. Bulk density on treatment plots at Broodkraal.


At Kanonkop, five years after the soil was deep-tilled, all the soil physical determination methods, including saturated and unsaturated hydraulic conductivity, clearly indicated the positive effect of two soil preparation methods compared to an undisturbed control. The delve plough and the soil mix implement both gave similar results (Fig. 8) and this result was confirmed by the excellent root distribution on the respective treatment plots. The first abrupt decrease in root penetration occurred at 1.0 m in the rip/soil mix combination and at 1.1 m in the rip/delve plough treatment.


FIGURE 8. Bulk density on treatment plots at Kanonkop.


In summary it can be concluded that significant re-compaction can already occur one year after soil preparation and it increases with time. Deep soil preparation will therefore normally be necessary when an existing vineyard has to be replaced after 20 – 25 years. There are various instruments that can be used successfully to determine soil compaction. Determination of soil strength using a penetrometer is probably the easiest, quickest and most accurate. Due to the high cost and unavailability of constant-speed penetrometers, the pocket vane tester or a simple steel rod with handle are alternatives that can be used to detect compact soil layers.



The authors are grateful to Winetech for providing funding for the study.


Literature cited

Bar-Yosef, B. & Lambert, J.R., 1981. Corn and cotton root growth in response to soil impedance and water potential. Soil Science Society of America Journal 45, 930 – 935.

Myburgh, P.A., Van Zyl, J.L. & Conradie, W.J., 1996. Effect of soil depth on growth and water consumption of young Vitis vinifera L. cv. Pinot noir. South African Journal of Enology and Viticulture 17, 53 – 62.

Richards, D., 1983. The grape root system. Horticultural Reviews 5, 127 – 168.

Saayman, D., 1982. Soil preparation studies: I. The effect of depth and method of soil preparation and of organic material on the performance of Vitis vinifera (var. Colombar) on Clovelly/Hutton soil. South African Journal of Enology and Viticulture 3, 61 – 74.

Saayman, D. & Van Huyssteen, L., 1980. Soil preparation studies: I. The effect of depth and method of soil preparation and of organic material on the performance of Vitis vinifera (var. Chenin blanc) on Hutton/Sterkspruit soil. South African Journal of Enology and Viticulture 1, 107 – 121.

Schulte-Karring, H., 1976. Bodenschäden und massnahmen zu ihrer behebung. Aufgezeigt an beispielen aus dem Südafrikanischen weinbau. Der Deutsche Weinbau 31, 941 – 943.

Taylor, H.M. & Burnett, E., 1964. Influence of soil strength on the root growth habits of plants. Soil Science 97 – 98, 174 – 180.

Taylor, H.M. & Gardner, H.R., 1963. Penetration of cotton seedling taproot as influenced by bulk density, moisture content and strength of soil. Soil Science 96, 153 – 156.

Van Huyssteen, L., 1983. Interpretation and use of penetrometer data to describe soil compaction in vineyards. South African Journal of Enology and Viticulture 4, 59 – 65.

Van Huyssteen, L., 1989. Quantification of the compaction problem of selected vineyard soils and a critical assessment of methods to predict soil bulk density from soil texture. PhD. (Agriculture)-thesis, Stellenbosch University, Stellenbosch, March 1989.

Zimmerman, R.P. & Kardos, L.T., 1961. Effect of bulk density on root growth. Soil Science 91 – 92, 280 – 288.



Eduard Hoffman

You may like to read these:

Go Back