Composition of effluents from wineries in the Western and Northern Cape provinces Part 2

A Technical Guide
for Wine Producers

RECENT ARTICLES | WYNBOER HOME

Composition of effluents from wineries in the Western and Northern Cape provinces (Part 2): Impacts on soil and the environment

Reckson Mulidzi

Reckson Mulidzi¹, Giel Laker² & John Wooldridge¹
¹ ARC Infruitec-Nietvoorbij, Stellenbosch
² Department of Plant Production & Soil Science, Pretoria University, Pretoria
Keywords: Exchangeable sodium percentage, phosphorus, potassium, soil, texture, trace elements.
Introduction
This is the second of two articles concerning winery waste waters (effluents). The first article (Mulidzi et al., 2008) showed that winery effluent composition differs between wineries, and also with time during the annual wine making cycle. The potential of winery effluents to pollute soils and the general environment therefore varies between wineries and over time, depending on the nature of the processes that led to the generation of the effluent. Many of the less desirable inorganic constituents of the effluents were introduced during cleaning and other winery operations. This being the case, it is clear that the first stage in managing winery effluent involves control over the use of products which have undesirable components, and the substitution of more environmentally friendly products wherever possible (Van Schoor, 2004). There is, however, a second stage, or aspect, of effluent management; that which concerns the effluent disposal process itself (Van Schoor, 2000, 2001, 2004).
Regardless of any pre-treatment that may take place, the ultimate repository of winery effluents is the soil. How the effluent reaches the soil differs. Either the effluent is discharged onto a (usually) grass-covered soil surface through an overhead sprinkler irrigation system, or it is pumped into a dam or pond. In both cases the desired outcome is that the soil will remove contaminants from effluents passing through it, so that the water that reaches streams, or which joins the body of ground water, will be clean and pollutant-free. This is not always the case (Mulidzi (2001). Much depends on the products used in the winery, and on the nature of the disposal site. Factors which must be considered when selecting a disposal site, and which have been detailed by Van Schoor (2004), include position in the landscape, proximity to water courses and the properties, characteristics and qualities of the soil, all of which will affect not only the on-site impact of the effluent, but also the potential for off-site pollution.
This article concerns the effects of the effluents described in the companion article (Mulidzi et al., 2008) on the soils of the disposal sites attached to the participating wineries, and the associated environmental hazards. For further details of these investigations refer to the M.Sc. thesis of Mulidzi (2001) and the Ph.D. dissertation of Van Schoor (2003).
Materials and methods
At six of the ten participating wineries (Paarl 1 and 2, Stellenbosch, Robertson 1, Worcester and Berg River, effluent was distributed through overhead irrigation (sprinkler) systems onto kikuyu grass pasture. At the remaining four wineries (Paarl 3, Robertson 2 and the Olifants and Orange rivers), the effluents were pumped into ponds from which evaporation and seepage could occur. At each site, soil samples were collected from the 0-30 cm, 30-60 cm and, where possible, 60-90 cm depth intervals on the same days that the effluent samples were taken. Soil samples from deeper in the profile were also collected, where soil depth permitted (Peters, 1981). At Robertson 2, where the soil beneath the pond was stony and shallow, samples could only be obtained from the 0-30 cm depth interval. Control samples were taken from locations at each disposal site where the soils were similar to, but apparently unaffected by, the disposal process. At each disposal site the soil profiles were investigated and classified in accordance with the Taxonomic System (Soil Classification Working Group, 1991) during May 2000. Preparation and analysis of the soil samples was carried out using standard laboratory procedures (Mulidzi, 2001). Extractable P was determined by the method of Bray No. 1. In view of the serial sampling procedure employed, the data were considered to reflect changes over time at the localities from which they were derived. No statistical treatments were therefore employed.
Results and discussion
Soil characteristics
Soils at the disposal sites varied in form, texture and stone content (Table 1). At three of the sites the soil cover had been largely removed during excavation of the ponds, locally revealing weathered, fractured rock (Paarl 3, Robertson 2) or fragmented dorbank (Oliphants River) beneath a disturbed and rocky cover of variable depth. In contrast, the fourth pond (Orange River) was underlain by sandy soil to the full sampled depth, with an increased stone content below 60 cm. Of the irrigated sites, that at Worcester was situated on stratified sandy alluvium. Robertson 1 was also sandy and low in clay and silt at all sampled depths, without stratification. The Berg River and Paarl 1 sites were characterized by high (relative to the rest of the soil population) clay and silt contents in the 0-30 and 30-60 cm horizons, whereas at Paarl 2 the clay contents of the 0-30 and 30-60 cm horizons were appreciably lower than in the underlying material. At Stellenbosch, the clay content of the 30-60 cm horizon was double that of the 0-30 cm layer.
Table 1: (click table to enlarge in separate window)

Phosphorus
Few plants require soil P concentrations above about 50 mg/kg (Bingham, 1966). Higher P concentrations may be associated with such problems as reduced nodulation in legumes, zinc (Zn) and copper (Cu) deficiencies and interference with sugar metabolism (Rossiter, 1955; Silber et al., 2002). It was therefore of concern that the P concentrations in the topsoil and subsoil, but rarely the deeper soil horizons, commonly exceeded 100 mg P/kg before wine making commenced in December or January (Table 2). These high soil P concentrations could have been residues from the previous season, or recent deposits from preparatory washing and sterilizing. Since the effluents were not analysed for P, this matter could not be investigated further. Mid- (February, March) and late-season (April, May) soil P concentrations were appreciably lower than the December/January values at most wineries, commonly approaching the control values. Where high end-season P concentrations were observed, as at Stellenbosch in April and May, and the Berg, Olifants and Orange River sites in April, these were speculatively attributed to the use of high-P washing materials. If this proves to be the case, then it will be evident that such products constitute far more of a P hazard than does the winemaking process itself.
At four of the six irrigated disposal sites, seasonal average soil P concentrations and control P concentrations decreased progressively with increasing horizon depth. At each sampled horizon the seasonal average P concentrations were, with one exception (Berg River, 60-90 cm), greater than the control. Lack of consistency in the P concentrations in successively deeper horizons at individual sites over time suggests that the rate of downward movement of P varied between profiles.
Where the effluent was pumped into dams, seasonal average P concentrations in the 60-90 cm horizon were not lower than in the 30-60 cm horizon at Paarl 3 and Orange River, perhaps implying that ponding alters the pattern of P retention in the lower soil material, perhaps as a consequence of permanent saturation. At the Orange River site, seasonal average P concentrations at all three sampled depths exceeded those in the controls.
The highest P concentrations were observed in the soil material underlying the pond at Olifants River. As at the irrigated sites the P concentrations varied from month to month in all horizons, low values following high values as in the extreme case of the shift from 3173 mg p/kg in February to 138 mg P/kg in March in the 30-60 cm horizon.
That soil P concentrations in the individual horizons commonly fluctuated between high and low values as the season progressed, and did not show a progressive increase over time, suggests that little P absorption took place in the profiles at either the irrigated or ponded sites. This implies that a large proportion of the P in the effluent remained in solution where it could potentially reach free water, perhaps contributing to eutrophication.
Table 2: (click table to enlarge in separate window)

Potassium
Potassium (K) is both an enzyme co-factor and the agent responsible for maintaining osmotic relations. Absorption of K is rapid and, where K is abundant, tissue K concentrations may greatly exceed requirements. Although the plant may show no apparent ill effects, luxury K consumption may lead to cation imbalances, which can detrimentally affect quality in deciduous fruit, and the vinification of grapes. Situations where exchangeable K exceeds about 4% of the sum of exchangeable base cations are usually undesirable, as are soil solution K concentrations high enough to promote K concentrations above 4%. Since K is usually more mobile in soils than P, the likelihood of K leaching into, and polluting, ground or surface water is relatively great. Seasonal average soil K levels in all horizons tended to be appreciably higher than the average P levels. However, seasonal peaks in K were not necessarily coincident with peaks in P. At the Worcester site, for example, high P values were occasionally observed despite the fact that K concentrations were low in all seasons and at all depths (Mulidzi, 2001; Van Schoor, 2003). This indicates that not all of the P and K derives from the same source, and highlights the fact that not all wineries have the same effluent-related problems. As was the case for P, seasonal average soil K concentrations at the irrigated sites decreased progressively with increasing profile depth, but did so at only four of the six sites. That these sites did not all coincide with those that showed progressive decreases in P probably reflects differences in the soil factors that affect the movement of P and K through soil profiles. At only one site (Paarl 2) was there an increase in K content with depth. Also like P, soil K concentrations varied strongly with season in all horizons. That low values followed high values implies that little K was retained in extractable form. At those sampling dates where low K values were observed these values were, on occasion, lower than the controls. Where several such low values occurred during a season the seasonal average soil K concentrations were also lower than the controls. In most cases, however, the seasonal averages were higher than the controls. Much the same K distribution patterns were observed at the ponded as at the irrigated disposal sites. As for P, the Olifants River site was characterized by exceptionally high K levels at most sampling dates.
Comparison between the effluent K values cited in the companion article (Mulidzi et al., 2008) and in Table 3 show that the incidences of high and low K concentration in the effluent do not consistently coincide with those in the soil, as at Stellenbosch, where high effluent K concentrations in February and March were associated with low soil K levels throughout the depth of the profile at those sampling dates. That the soil K concentrations at all depths were relatively high in April and May suggests that a measure of lag time may be involved.
Table 3: (click table to enlarge in separate window)

Exchangeable sodium percentage (ESP)
As its name suggests ESP is the percentage of the cation exchange capacity of the soil that is occupied by sodium. At high ESP’s, clays readily disperse on wetting, forming a puddled layer. This dries to form a hard, impermeable crust which reduces infiltration and gas exchange (Mulidzi, 2001). In effluents, high ESP’s have the added disadvantage of contributing to dispersion of the organic fraction, facilitating leaching into the deeper subsoil (Mulidzi et al., 2002). Whereas it was once believed that problems are only experienced at ESP’s of 15 and higher, some dispersion-susceptible soils in South Africa, Zimbabwe and Australia are now known to begin breaking down at ESP values just above 2. On the other hand, some red soils on the South African Highveld may remain stable at ESP values over 40. The effect of ESP is therefore strongly dependent on soil type and its inherent stability.
At the disposal sites, ESP’s ranged from <1 to >60 (Table 4). As with P and K, the higher ESP’s tended to be observed in December or January, sometimes with a secondary peak later in the season, presumably reflecting activities in the winery at the time. There was no consistent relationship between ESP and sodium absorption ratio (SAR) or electrical conductivity in the effluent, although the high December and January ESP’s observed at Robertson 1 were probably directly linked to the high SAR’s (29.0-33.1) of the effluents passing through the profile at that stage. Normally the main source of sodium in winery effluents is sodium hydroxide (caustic soda), which is diluted to 5% and used for tank cleaning purposes (Van Schoor, 2000).
The seasonal average ESP’s were not always greater than the controls. At Paarl 2, ESP’s were high, relative to the controls, at all sampling dates and at all soil depths. The SAR’s, electrical conductivities and chloride concentrations in the Paarl 2 effluents were also consistently high, and were ascribed to the use of poor quality borehole water (Mulidzi et al., 2002; Mulidzi et al., 2008). Although the kikuyu grass cover crop at Paarl 2 was apparently unaffected by the high electrolyte concentrations, the potential off-site pollution hazard due to lateral leaching is great. That leaching of salts was in fact taking place was confirmed when examination of a drainage ditch below the disposal site in early May, at the end of the dry period, revealed the abundant presence of white sodium chloride deposits on the ditch wall closest to the disposal site. At Paarl 3, subsoil ESP’s were also high throughout the study period, especially in December, despite the low SAR (0.5-3.2) values in the effluents throughout the entire study period (according to Van Schoor (2004) the maximum permitted SAR is 5). At Worcester, ESP’s were high at all soil depths during December, despite a very low effluent SAR (0.7) during that month. Conceivably, the high early-season ESP’s at these sites may have been residual deposits from tank washing at the end of the previous season.
Indications of clay dispersion and resultant soil physical problems, such as crusting and low infiltration rates were absent at most sites, mainly because the soils were reasonably sandy and open. The exception was Paarl 1, where, although the soil ESP and effluent SAR levels were not exceptionally high, the irrigated kikuyu pasture was established on a Sterkspruit soil with its diagnostic, strongly structured prismacutanic subsoil. Such soils are highly dispersive, particularly in waters which have low electrolyte contents. Ponding might have been a better disposal option than irrigation at this particular site. In general, sandy soils are unsuitable for ponding because of their high infiltration rates, low water storage capacities and excessive leaching. Very sandy soils may be too permeable to be used as disposal sites even where they are grass covered and overhead irrigation is employed.
Table 4: (click table to enlarge in separate window)

Trace elements
Many South African soils are deficient in trace elements. It is nevertheless not always possible to accurately define the limits of those concentration ranges that lead to deficiencies or toxicities. For the present purpose it will be assumed that Cu, Zn, manganese (Mn) and boron (B) concentrations of 2.5, 15, 10 (EDTA) and 1.5 (hot water) mg/kg, respectively, are adequate for deciduous fruit and vines. Within individual seasons and sites, the observed concentrations of these elements tended to be strongly skewed toward low values with infrequent high values (full data not shown), giving a wide min.-max. range (Table 5). That the range of trace element concentrations varied from site to site (Cu and Zn ranged particularly widely at Stellenbosch, Mn and B at Olifants River; Cu, Zn and Mn ranges were least at Paarl 2) emphasises that effluent management problems, and their solutions, vary from winery to winery. Averaged over the season, (Table 5), Cu and Zn concentrations in the disposal site soils consistently tended to diminish with increasing depth as, with less consistency, did the Cu and Zn concentrations in the control soils. Seasonal average Cu and Zn concentrations in the control soil horizons were mostly, but not always, higher than in the corresponding horizons in the controls. Differences were greatest in the topsoils, implying that both Cu and Zn move through the soil at a slow rate.
At most wineries the seasonal average concentrations of Mn and B also tended to decrease with depth at the irrigated and control sites. There were however, more exceptions to this generality for Mn and B than for Cu and Zn, perhaps indicating greater mobility under the locally-prevailing conditions. The Orange River control site subsoil Mn concentration was appreciably greater than in the overlying horizon. This was attributed to subsurface transfer from the disposal site.
Table 5: (click table to enlarge in separate window)

Conclusions
Collectively, the results obtained from this project, as reported by Mulidzi (2001), Mulidzi et al. (2002, 2008), Van Schoor (2003), and in the present paper, enable the following conclusions to be drawn.

Wineries differ greatly in terms of the potential environmental hazards posed by their effluents. These differences reflect both the compositions of the effluents, the effectiveness of the effluent disposal practices followed and the suitability of the disposal site for the purpose.
At all the participating wineries the composition of the effluent exceeded acceptable limits, at least for some parameter (e.g. chemical oxygen demand; COD) at some stages of the season. The extent of the problem posed by these excesses differed between wineries.
Variation in effluent parameters tended to be seasonal, presumably reflecting changing winery activities over time. The occurrence of regular trends based on patterns of activities and product usage greatly facilitates the implementation of corrective measures.
High COD levels may be the greatest overall effluent-related problem faced by wineries, and therefore the highest priority for rectification.
High effluent acidity (low pH) seems to be well managed through pre-disposal intervention measures at the majority of wineries, though not all.
Some wineries appear to manage their effluents well yet still have problems. As in the case of Paarl 2, the problem (high sodicity and salinity levels) apparently arises from the use of very poor quality borehole water. In these cases, the only way to avoid environmental damage is to locate an acceptable alternative water source.
In general, effluent disposal is poorly planned and managed, and disposal sites rarely seem to have been selected because their soil properties are appropriate for effluent disposal. In particular, deep sandy soils, as at Orange River, are unsuitable for disposal by ponding, mainly because of their high infiltration rates, high permeabilities and low water storage capacities. In this case, the actual ponding takes place at the water table around two metres below the soil surface.
Many disposal sites are too limited in area to permit the large volumes of effluent to be absorbed without surface runoff. This problem invariably persists despite the presence of a kikuyu sward and a sandy subsoil (E horizon).

Recommendations

Because practices and usages differ widely between cellars, every cellar should urgently initiate an environmental monitoring/auditing program extending over at least one, and preferably two annual cycles of activity. The information that is generated during this process is vital if a cost-effective effluent management system is to be devised for any specific winery.
Monitoring/auditing should not be restricted to effluent composition and effluent management in the cellar, but should also include on-site soil monitoring at disposal sites, as well as off-site monitoring in drainage ditches, streams and water bodies that are or could be affected by effluent disposal. Routine effluent sampling should be carried out at roughly 4-week intervals, commencing before the season begins and ending after the completion of all major winery activities. Soil samples, representative of the 0-30 cm, 30-60 cm and 60-90 cm depths intervals should also be taken at 4-week intervals. Where the soils at the disposal site are deep, it is desirable that samples should also be taken from even greater depths, possibly to depths of three metres. This should be done once every six months during the initial intensive monitoring period.
On-site and off-site monitoring at and around effluent disposal sites enables information to be collected regarding the suitability and, even more importantly, non-suitability of different types of terrain and of soil, for specific types of effluent disposal. Such information is urgently required so that land suitability criteria may be identified which will, in turn, allow soils and terrains to be classified in terms of their suitability for effluent disposal, whether by irrigation or ponding, wherever new disposal sites are being evaluated.
Many of the cellars which currently use excessively small or otherwise unsuitable disposal sites do not have viable alternative sites, or lack access to the land necessary to expand their existing effluent disposal operations. In these cases alternative methods of disposal methods must be identified. Alternatively, the pre-disposal quality of the effluent must be improved, to the point where any run-off will not create an environmental hazard.
Where larger or more suitable disposal sites are not available, the only feasible option may be to utilize reed beds or other types of wetland to purify effluent before discharge (Van Schoor, 2002). Wetland technologies are currently being intensively investigated at ARC Infruitec-Nietvoorbij, the objective being to develop improved, more cost effective, and more environmentally friendly methods of disposal for winery effluents.
For further information contact Reckson Mulidzi. E-mail: mulidzir@arc.agric.za
References
Bingham, F.T., 1966. Phosphorus. In: H.D. Chapman (ed.). Diagnostic criteria for plants and soils. University of California, Riverside, 324-361.
Mulidzi, A.R. 2001. Environmental impact of winery effluents in the Western and Northern Cape provinces. MInstAgrar dissertation, University of Pretoria.
Mulidzi, R., Laker, G., Van Schoor, L. & Louw, P.J.E. 2002. Fate of organic components of winery effluents in soils. Wynboer Tegnies 154, 82-83.
Mulidzi, R., Laker, G. & Wooldridge, J., 2008. Composition of effluents from wineries in the Western and Northern Cape provinces I: seasonal variation and differences between wineries. Wynboer Tegnies. (in press)
Peters, W.B. 1981. The role of water and land resource information in World Bank programs for agricultural development. SMSS Techn. Mon. 1, 115-133. SCS, USDA, Washington.
Rossiter, R.C., 1955. The influence of soil type on phosphorus toxicity in subterranean clover (Trifolium subterraneum). Australian J. Agric. Res. 6, 1-8.
Silber, A., Ben-Jaacov, J., Ackerman, A., Bar-Tal, A., Levkovitch, I., Matsevitz-Yosef, T., Swartzberg, D., Riov, J. & Granot, D., 2002. Interrelationship between phosphorus toxicity and sugar metabolism in Verticordia plumose L. Plant and Soil 245, 249-260.
Soil Classification Working Group, 1991. Soil classification. A taxonomic system for South Africa. Memoirs on the agricultural natural resources of South Africa No. 15. Department of Agricultural Development, Pretoria.
Van Schoor, L.H. 2000. Bestuursopsies om negatiewe omgewingsimpakte by wynkelders te minimaliseer. Wynboer 132, 11-14.
Van Schoor, L.H., 2001: A formula for the quantification and prioritisation of negative environmental impacts. Wynboer 142, 15-17.
Van Schoor, L.H., 2002. Die gebruik van kunsmatige vleilande in die suiwering van kelderafvalwater. Wynboer 152, 9-11.
Van Schoor, L.H., 2003. The development of a prototype ISO 14001 environmental management system for wine cellars. Ph.D. dissertation, University of Stellenbosch, unpublished.
Van Schoor, L.H., 2004. Guidelines for winery wastewater and solid waste management at existing wineries. Winetech, South Africa. Downloadable from www.ipw.co.za

Wynboer is incorporated in WineLand, magazine of the SA wine producers.
Subscribe to WineLand

Visit our sister sites:

South African wine farmers' representative organisation

Facts, figures, contact details and much more in the 2009/10 Directory