Effect of Different Environmental Factors on the Performance of Sauvignon Blanc Grapevines in the Stellenbosch/Durbanville Districts of South Africa

Effect of Different Environmental Factors on the Performance of Sauvignon blanc Grapevines in the Stellenbosch/Durbanville Districts of South Africa. I. Geology, Soil, Climate, Phenology and Grape Composition W.J. Conradie1, V.A. Carey1*, V. Bonnardot2, D. Saayman3 and L. H. van Schoor1 1) ARC Infruitec-Nietvoorbij, Private Bag X5026, 7599 Stellenbosch, South Africa 2) ARC Institute for Soil, Climate and Water, Private Bag X5026, 7599 Stellenbosch, South Africa 3) Distell, Box 184, 7599 Stellenbosch, South Africa Submitted for publication: June 2002 Accepted for publication: August 2002 Key words: Geology, soil, climate, terroir, grapevine, phenology, growth, yield, must analysis

A study in five commercial Sauvignon blanc vineyards, grown at different localities in the Western Cape of South Africa, was carried out over a period of seven years. These localities (four within the district of Stellenbosch and one in Durbanville) were within a radius of 15 km and underlain by different geological formations. Two experimental plots, representing different soil forms, were identified at each locality. In some cases soils could be related to the s Grapevine . materials l geologica f o s mixture m fro d develope s soil e th s case r othe n i t bu , materials t paren g underlyin were not irrigated and root distribution was mostly affected by factors such as soil moisture, compacted layers and percentage stone, and not necessarily by geological parent material. Relatively good root distribution was obtained for at least one soil from each parent material (granite, hornfels, shale). Management practices, such as liming and fertilisation, changed the chemical properties of the soil. However, the lowest levels of potassium were found in soils originating from phyllitic shales. Despite their geographic proximity, meso-climates differed between the five localities, especially during summer. This was mainly due to their various landscape positions and distance from the ocean. Maximum temperatures for February differed by 3.1°C between the warmest and coolest locality. The lowest temperature variability index was experienced at the localities closest to the sea, due to moderate maximum and minimum temperatures. Time of budburst was not necessarily earliest at the warmest localities, but at the warmer sites grapes were generally harvested one to three weeks earlier than at the cooler localities. Ripening was also affected by soil type, resulting in differences of up to 12 days for time of harvest for grapes from different soil types, at the same locality. Acidity was lowest, and pH highest, for grapes from the warmest locality, but low acidity and moderately high pH values were also found for grapes from the coolest locality, probably on account of differences between clones. Within a specific locality, however, acidity tended to be lowest, and/or pH highest where water stress was the highest. Investigation of the effect of soil and climate on wine quality is part of an ongoing study.

In accordance with the Wine of Origin legislation, vineyards in terroir concept has been done in South Africa. The current state South Africa are currently demarcated by technical experts of knowledge, especially as far as the South-Western Cape is con- (Saayman, 1998). Such demarcation is theoretically based, due to cerned, is summarised below. insufficient historical information on grapevine performance/ Climate wine quality being available in South Africa, which is a relatively young wine-producing country. Areas are allowed to express The Western Cape Province is situated between approximately their specific wine style and character after demarcation, instead 33°S and 34°S, and experiences a warm, temperate climate of proving originality beforehand. Demarcation is not necessari- (Kendrew, 1961; Schulze, 1972). In an earlier study (Le Roux, ly specific enough to meet the demands of consumers (Van Zyl, 1974) temperature was identified as the single most important cli- 2000). Further delimitation must therefore be applied. The terroir matic factor whichthat affects grape quality and character in the concept provides a rational basis for such delimitation. A terroir Western Cape. Using data from the few weather stations available can be defined in different ways, but Carey (2001) described it as at that stage, the Western Cape was divided into five climatic "a complex of natural environmental factors, which cannot easily classes according to the growing degree day model of Amerine & be modified by the producer. This complex will be expressed in Winkler (1944). In a later study (De Villiers et al, 1996) the the final product, with the aid of various management decisions, South-Western Cape was subdivided into different climatic resulting in distinctive wines with an identifiable origin. regions, largely according to the index of Smart and Dry (1980), Therefore the terroir cannot be viewed in isolation from manage- which is based on mean January temperatures, because January is ment and cultivation practices, although such practices do not the warmest month for regions with the same continentality. form part of the intrinsic definition". In view of the complexity of However, De Villiers et al. (1996) used mean February tempera- soil/climate/cultivar relationships, little experimental work on the tures for their classification, because February was the warmest

Africa South Matieland XI,7602 Bag (Stellenbosch), Private University, Oenology,and Stellenbosch Viticulture of Department *Presentaddress: Acknowledgements: Technical assistance by the project team at Nietvoorbij, as well as financial support by Winetech, is appreciated.

S. Afr. J. Enol. Vitic., Vol. 23, No. 2, 2002

78 EnvironmentalDifferent of Factors Effect 79

month at 80 percenand/oy rqualit t oe f thwin e , weathegrowth r e stations grapevin an o t d becausy directl e l most materia t en n I . March d an y Februar g durin n ripe e Cap n Wester e th n i s grape n o y geolog f o t effec e th n o s studie r othe o N . character e win - differ h wit s region n i n grow blann s cwa Sauvigno e wher l tria a grapevine performance have been reported from South Africa. ent macro-climates (Stellenbosch, Elgin, Robertson), a "cool cli- Soil type mate" and a "warm climate" wine style could be identified f o l shortfal l principa e Th . 1999) , Haasbroek & (Marais r - Hunte , vine d classifie t tha d showe ) 1983 , (Seguin x Bordeau n i h Researc ) al,el 1996 s Villier e D ; 1974 , Roux e (L s map c climati y earl e th e th e regulat o t l soi f o y abilit e th o t y superiorit r thei e ow s yard was that t thbu r e distancewate e s betweeexcessiv g n weathedrainin y r stationrapidl , s wervines e eth too t o r greawate t fof o r y suppl t pas e th r Ove . them n betwee d interpolate y reliabl e b o t a dat , that e rat a h suc t a t i g deliverin d an , water t sufficien g retainin - pro n bee s ha stations r weathe f o k networ e th , however , decade although the vines experience some stress towards ripening, this gressivelo t ye expandedguid r poo . a Intensiv e b o t d e recordinfoun s wa ge oftyp climati l Soi . c data, sucexcessive t h no a s i s stres that which is now available for the Bottelary-Simonsberg- vine performance and wine quality, unless considered in con- s area n withi n eve , that n show s ha , area g wine-growin g Helderber junction with climate, cultural practices and the specific require- with a e uniforclimat m d macroan s -soil climatef o t , certaiaccoun n O n . aspects cultivars o fl climate maindividua f o y s ment vary considerably (Carey, 2001). In order to gain further infor- being vastly different, models developed in Europe cannot be mation on genotype/climat o Rankint g eal. interactionset e Accordin . , anAfrica d thh e effectSout o t sy of sucdirectl h d applie y proximit e clos n i s plot e referenc , character e win n o s interaction s it f o s term n i e climat o t e rol subordinate a s play l soi , (1971) to the - weathehowev r , stationsAfrica h werSout e n monitoreI . regions d (Carem war yn i et y al., 2001)qualit e .win Thn o e s effect effect of season and site climates on the aroma profile of the er, soil may have a pronounced effect, as was shown by Saayman wines - wavar s s cleawine r t for alCinsau lf o thre s e cultivarcharacteristic s e (Sauvignoth t tha d n foun blanco wh ,, (1977) Chardonnay and Cabernet Sauvignon). However, the effect of soil ied according to the soil type on which the vines were grown, water status on aroma profile was not taken into account in this even when meso-climatic conditions were the same. Since 1977 . al.,et 2001) y (Care y stud e grapevin n o e typ l soi f o t effec e th n o s studie c scientifi r furthe o n . Africa h Sout n i d reporte n bee e hav e performanc Geology l soi f o s statu t Nutrien Apart from the indisputable effect of climate on wine character, n o t effec g predeterminin l potentia a e hav o t d considere s i y geolog e hav y ma h whic s element o tw e th s soil d vineyar n Africa h Sout n I - pri e th s wa y Geolog . 1998) , (Wilson y qualit d an r characte e win a definite effect on wine quality are nitrogen and potassium, assum- mary ke- yele tl o identifessentia r y thothe e e "Uniteth f o s s Terroir deficiencie d es Baseseriou " o in e thar e e Midther -t tha g in Loire Valle, y inripening Franc s e delay (Morlat n , 1996)nitroge e . ThereExcessiv is . , however1992) , , no indi(Saayman - s ment y qualit h hig f o e win n i s result t tha n formatio l geologica l vidua d unwante s stimulate d an t ro h bunc f o e occurrenc e th s enhance (Seguin, 1983) and grapevines are found on many geological for- growth, resulting in unbalanced and herbaceous wines. Insufficient mationsd . Than s e effectyield d s of geologdecrease n i t y onresul winy ema , qualithand yr areothe noe tth directn o , , nitrogen e determin t tha l soi a f o s propertie l physica e th y mainl s i t i e sinc musts with too little assimilable nitrogen, thus contributing to quality through the effects of such variables as drainage, soil tem- retarded or stuck fermentation (Gladstones, 1992 and references perature and water availability on vine growth patterns. The indi- therein). The extent to which a specific soil will be able to supply rect effectc s oorgani f pass it t n geologicao t l processedependen y s largel o ne b vineyar l wil s d performancegrapevine o t n nitroge s material e th d create s processe h suc e sinc , profound e b y ma material content (Conradie, 1986). As the organic material content which weatheriny largel , g ha1986) s, sculpted t(Stevenson o e form thclimat eo t moderd nrelate landscapey normall .s i l soi a f o different nitrogen-supplying capacities can be expected for differ- s soil e Stellenbosch/Durbanvill a th n I Afric h Sout f o n regio e e hav y ma t soil-s tha Kt level fac d terroirst well-recognise a en s i t I . t oldes e Th . materials l geologica f o y variet a m fro d develope a considerable effect on the acid balance in grape juice and on the rocks are eth e sedimentarwhil , y1989) , formationSaayman s & oe f the Malmesbur(Conradi e win g y Groupresultin e .th f o H p These includy b d e shalesaffecte e b , schisto als l , phyllitwil m e anpotassiu d l greywacksoi o t s e (Therograpevine f no ete respons s sediment y pre-Cambriae lat Malmesbur n e I . th s al.,1992) ntime - satura m potassiu , soil e th f o t conten y cla e th s a h suc , factors r othe were folded into a mountail soi e n th chai n i n d ans K/Man d intruderatio gx dcomple b ye graniteexchang s oe f thth ef o n tio Cape Granite Suite. This landscape eventually subsided and was (Etourneaud & Loue, 1984). Here geology does play a role, in the p u e mak h whic s shale d an s sand f o e sequenc k thic e th y b d covere e Cap n Wester e th n i s soil d vineyar f o e percentag e larg a t tha e sens e Cap e th d strippe y subsequentl n Erosio . Supergroup e Cap e th - potassium-con n i h ric y relativel s i h whic , granite m fro d derive e ar e th f o n plai l coasta e th w no s i t wha f o h muc m fro p Supergrou Wooldridge; 1964 . Kaolinic, - 1988) , (Visser gran ,s mineral g tainin Westernd Capean m , once potassiu agair fo ny exposincapacit g g the pre-Capbufferin w lo ea flooe hav r s witsoil h d its ite-derive low granite hills and occasional, steep-sided remnants of the have been shown to promote luxurious consumption of potassium - intermedi r o w lo y occup s vineyard t Mos . Group n Mountai e Tabl in grasses (Wooldridge, 1988). Grapevines may also consume - 10 m fro e rang h whic s altitude t a , landscape e th n i s position e at excessive amounts of potassium from chemically weathered 500 m. Few of the vineyard soils in the Coastal Wine-growing granitic soils. However, luxurious consumption of potassium will k roc g underlyin e th f o g weatherin situ in y b d derive e wer n Regio also be induced by excessive fertilisation. formations. Long continued transport and reworking at these low altitudes led to the formation of soil parent materialy s stud whic t h concurren f -o e Scop tain mineral material derived from a number of rock types. In a Because scientific data on the demarcation of terroir units is lim- recent s study (Valocalitie d n Schoorselecte t a , 2001n ) it waundertake ss nowa t possibly stud a e , to relatAfrica eh par-Sout n i d ite

2 200 , 2 Enol. . J . No , . Afr Vitic. 23 . . S ,Vol 80 Factors Environmental Different of Effect

representing different landscape positions and underlain by con- present article is limited in scope to phenology, growth, yield and trasting rock types in the Stellenbosch/Durbanville districts. chemical composition of grapes. If localities and soils can be sep- s vineyard s Africa' h Sout f o % 17 n contai y currentl s district e Thes arated in terms of the above-mentioned parameters, it should be cult - importan t mos e th blancn g bein , (SAWISSauvigno . 2000) , possible to produce, within the Stellenbosch/Durbanville dis- tivar for production of high-quality white wines in these districts, tricts, a range of Sauvignon blanc wines with region-specific was used as test material. The hypothesis was that soil type may characteristics. Differences in actual wine quality will then play an important role in determining wine characteristics, result- become the subject of a separate article. ing in different wine styles for different soils, even under identi- S METHOD D AN S MATERIAL y identif o t s wa l goa e ultimat e l meso-climatiTh ca . cconditions ) etc. , soil , altitude , position e landscap , (climate s condition f o s set o t 5 (1994/199 s season n seve r ove t ou d carrie s wa y stud e Th that will l leaAl . d to winevineyards s c withblan specifin c stylesSauvigno l . As a preliminarcommercia e fiv n yi ) step 2000/2001 o t s wa y stud t curren e th f o e objectiv t firs e th l goa s thi s toward were within a radius of 15 km, four situated in the district of . localities d meso-climatn i s selecte e th t a edifference y quantif Stellenbosch and one in Durbanville (Fig. 1). Their co-ordinates, The second objective was to identify two different soil forms at altitudes, aspects, slopes and distances from the sea are sum- each localityn . Physica(Thero n l and chemicaunderlai e ar r l characteristicRive s Kuil t a s s osoil fe soil 1e .Th s werTabl n i ed marise to be determined, while soil and River-Helderbers planKuil t e wateth y rb statu) gs2001 , were toSchoor bn e Va ; al.,et 1992 monitored on a regular basis. Thirdlyd , the interactiosituate e ar y nPapegaaiberh Valle betwee n Bot . Devo n d clipluton g an - e Granit mate, landscape position,g geology and River-Helderber soi s l characteristicsKuil e th n whe d , as forme s hornfelsn wa o h whic , e Th . determined e b o t s wa , performance e grapevin n i d manifeste e th n i s MalmesburA o . int d ysediments intrude n pluto e Granit

Hi! Towns • Automatic Weather stations /S/ Rivers

FIGURE 1 Study area and locations of the experimental sites.

S. Afr. J. Enol. Vitic., Vol. 23, No. 2, 2002 Effect of Different Environmental Factors 81

TABLE 1 Characteristics of the five experimental localities, all planted to Sauvignon blanc/99 Richter. l Geologica e Distanc e Distanc e Slop t Aspec e Altitud s Coordinate y Localit Soil Description n formatio m ) fro (m m fro (%) form False Bay Table Bay (km) (km)

3 1 5 14. E ES r Kuil33°57'18"sRive 0 25 S 24 Kuils River - Tukulu Light textured, gravelly topsoil, mottled, high- 18°44'05"E Helderberg ly weathered subsoil, with signs of wetness n pluto e Granit Vilafontes Medium textured, gravelly topsoil, medium l subsoi d texture

Papegaaiberg 33°55'46" S 148 NW 5.7 18 33 Hornfels Avalon Medium textured, yellow-brown, weakly 18°50'06" E l subsoi d mottle , structured Tukulu Medium textured, yellow-brown, weakly l subsoi n i s wetnes f o s sign , structured

Devon Valley 33°54'44" S 200 WNW 6.4 20 32 y Homfelweakl , s yellow-brown , Oakleatextured m f Mediu 18°49'55" E structured h hig , yellow-brown , Glenrostextured m a Mediu percentage stone

Helshoogte 33°54'12"S 413 5.2 24 41u PorphyritiTukul e cbiotit Medium textured, yellow-brown, weakly 18°55'15"E Stelf o e - granit structured, slight signs of wetness with depth lenbosch pluton Hmton Medium textured, reddish-brown, very weakly structured, well-drained

Durbanville 33°50'00" S 230 SE 8.6 27 12 Phyllitic shales Westleigh Medium textured, dark colour, signs of water 18°36'47" E (Tygerberg forma- table in winter y tioweakl n, of the Malmescoloured k dar , - textured m Mediu ) group y bur structured, signs of wetness in subsoil

, Helshoogte t a l materia t paren e th s wa e granit , River s Kuil f o e cas (one wire for the cordon arms and two to four wires for the but its, origirows nt (Stellenboscadjacen o tw h f o pluton g ) waconsistin s , differentplots l . The soilsExperimenta . at foliage) m for h Malmesburwhic m , fro d ysediments originate e Durbanvill . type l soi h eac n o d selecte e wer , each s vine t adjacen 0 1 h wit the oldes. t geologicacordon r lmete formatior pe s bud n i6 n1 tho t e Westery nannuall Caped . spur-prune e wer s Vine Suckering (removal of shoots not located on spurs) was done Within each vineyard two contrasting soil forms (not more than before bloom. Apart from this, growers applied normal viticultur- n Africa h Sout e th g usin , 1) e (Tabl d identifie e wer ) apart m 0 6 p to e th e abov m c 0 (3 g Toppin . site l individua h eac t a s alpractice Soil Classification System (Soil Classification Working Group, wire) was usually done once or twice during the growing season. - cat t firs a s a d classifie e ar s soil , system s thi o t g Accordin . 1991) e zon h bunc e th n i s leave e th f o e som e Durbanvill f o e cas e th n I s propertie d an e presenc e th f o s basi e th n o , forms l soi o int y egor were also removed after veraison. According to the general prac- n o d base s i s familie o int n classificatio r Furthe . horizons r maste f o a o t d ploughe e delv n bee d ha s soil l al , districts e thes r fo e tic A, B an. d E horizoplanted e wer n s propertiesgrapevine , e degrebefor e om fm leaching0 80 y , clay moveapproximatel f -o h dept ment an- d wetnessincreas f o . m Fairlai e yth h largwit e , differenceadded o als ss mawa ye thereforlim s e proces occus thi rg Durin within 0 a3 soid l formapplie .y normall s Producer . 5.5 t leas t a o t ) (KC1 H p l soi g in In view of the study being conducted in commercial vineyards, kg N/ha/yr (15 kg in spring and 15 kg post-harvest), while P and with the prerequisite being differen. t meso-climates/geologicaproduction o t g accordin d applie le wer K formations/soil types at the different localities, availability of Automatic weather stations (MC Systems, Cape Town) were n blan9 9 / c (Sauvigno l materia g Plantin . limited s wa s site d an y localit h eac t a s plot o tw e th n betwee y halfwa d erecte Richterd ) wawin , s obtainesunshine f do s fromhour , differenradiation t , commerciarainfall , l nurseriestemperature d , recorde implyinr go d that cloneaverage e s (botwer s h rootstocvalue e k Thes an . d scionminute )y werever en not necesdirectio d -an d spee - deci l Manageria . localities d e fivth selecte e t a l identica y saril summed for the period of an hour and the hourly data set was sions a- t the studietempera n d sitemea d s an (e.g m . canopmaximu y , managementminimum y ) alsdail e o createcalculat d o t d use - experi s thi f o e outcom e th d affecte e hav y ma h whic , limitations tures, number of hours with temperatures above 30°C and below ment. Planting density varied from 4000/h as (Amerin (2.5 d mWinkleran & ex ) degree-day 1. 0g 1944 m, ) atgrowin , 12°C Kuils River, to 2778/ha (3 m x 1.2 m) at Durbanville. Vineyards Huglin index (Huglin, 1978). Temperature variability indices for , investigation e th f o t star e th t a d ol s year 0 1 y approximatel e wer February were calculated according to Gladstones (1992). Plant s system s trelli l vertica n o d traine e wer d an , irrigated t no e wer n a s a d (PAWh estimate s Marc e )f wa , o d en e th t a e availabl , water

S. Afr. J. Enol. Vitic., Vol. 23, No. 2,2002 82 Effect of Different Environmental Factors

index of water stress- . In thiscompar approachn i d san e , similacoars e r tomor thd ean ony e cla formu s les d - Tukulcontaine m ufor lated by Riou in 1994 and reported by Tonietto (1999), the fol- ison to its Vilafontes counterpart. Normally coarser fragments are lowing- equatioorig e n thos foo t r seasonan l watecompariso n i r e balancgranit e m wafro s d used tderive o s calculatsoil n i d e expecte

PAWe: g Malmesburm fro g (Theros suggestin , yrock inatin al.,net. 1992) PAWe = PAWb + P - ET Eq 1 that the Tukulu may have been of purer granitic origin than the Vilafontes. The other granitic soils (Helshoogte) contained less where PAWe bparticl it s ththa eg amount oconfirmin fs watethu , r availablsoils r eRive s in thKuil e e 0-1.th n 2 mtha d soisan l e coars layer at the paren beginnin g g ounderlyin fe thth eo t seaso d n relate (September)y necessaril t , fono rs i eacn h soil distributio e siz form. Soils were assumed to be at field capacity at the beginning material. Greater mixing of parent materials or further advanced d an h Marc l unti r Septembe m fro l rainfal l tota s i P . season e th f o weathering must have occurred at Helshoogte than at Kuils River. - calculat s wa T E . period e sam e th r ove n evapotranspiratio s i T E For the hornfelsic soils (Papegaaiberg and Devon Valley), a higher - mea , (ETo) n evapotranspiratio e referenc y monthl n mea g usin d e proportion of finer sand, in comparison to coarser sand, was sured at each locality over the seven seasons, and crop coeffi- expected (Van Schoor, 2001). This was actually the case for the f o n consumptio r wate e th m fro d calculate e wer r latte e Th . cients ) (29.4% d san e fin + d san e fin y ver h wit , Valley n OakleaDevo t fa dry land grapevines in Stellenbosch and mean monthly ET0 mea- being more abundant than medium + coarse sand (11.6%). For the sured at Nietvoorbij in the Stellenbosch district (Myburgh, 1998). other three hornfelsic soils, however, very fine sand + fine sand was - (nor s month c specifi r fo T E d exceede l rainfal y monthl e Wher 2)e . (Tabl d san e coars + m mediu o t n proportio r simila a n i d foun (rainfall-ET" - regard "excess s e th ) wa ) September/October y mall This could be ascribed to contributions from surrounding granitic ed as drainagee fin e . mor y significantl d containe e Durbanvill m fro s Soil . material Leaf water potentials (pressure chamber technique of and very fine sand (43.5%) in comparison to the other localities. Scholander et al. (1965)) were measured once per week from This was expected, because the Durbanville soils are underlain by e wer s leave t sunli e matur y full , Uncovered . March o t r Novembe phyllitic shales, which weather into predominantly fine particles l experimenta r pe s leave e thre n o e mad e wer s Measurement . used (Theron et al., 1992). Particle size analyses suggested that four of plot betweeh bot nd 12:0an y 0 anValle d n 15:00Devo .t Changea f sOaklea i , n soilRiver wates rKuil content a u t wer(Tukul es soil e th measured weekly by means of a neutron probe at 300 mm depth at Durbanville) were related to underlying parent materials. The 150f . o h 0mm dept a o t n dow , intervals . materials f o s admixture m fro d develope e hav t mus s soil x si r othe g diggin y b 3 199 n i t ou d carrie e wer n distributio t roo n o s Studie This is a common phenomenon in the Western Cape. profile pits parallel to experimental rows. A root plot was done for Bulk densities were relatively high at Kuils River, Papegaaiberg . Bohf o d m(1979) metho l wal e profil e th g usin , form l soi h eac and Devon Valley, compared to lower values at Helshoogte and For they purposma ) 5 ee of thi(Tabl sl study materia onl c y twoorgani r root classeHighe . 2) s e were (Tabl identi e - Durbanvill - diame 1.m > ( 0 m m mediu r o ) diameter . finm i.e m e , 0 (<1. fied have resulted in less compact soils at the latter two localities, but ter). Soilsd , samplecoul y d accordindensit k bul d g an t ol the differenmateria t t paren horizon n s betwee identified n patter r , clea o n : extract 2 . No y (Bra P , KC1) (1:2.M H 1 p r 5n i fo d analyse e wer be detected. 0.03 M NHtF in 0.01 M HC1), K, Ca, Mg and Na (all extracted Water-holding capacities were low for both soils at Kuils River with 1. M NH4C1)fractions d , CEsan e C (witcoars h h NILChig d an ls at the content p Hy ocla f thw elo o soilt e ) du an , d2) e (Tabl n Classificatio e Walkley-Blac l th y (Soi b d C c kmetho organi e th d an g Papegaaiber t a u Tukul e th y b n retentio r wate w Lo s soil e th f o s capacitie g water-holdin e Th . 1991) , Group g Workin - con l grave h hig a f o t accoun n o s wa , Valley Glenrosn Devo t aa , cores l soi d undisturbe n o d kP5 10o at (2. measure 0 e kPawer ) tent. In spite of a low clay content, water-holding capacity was s densitie k Bul . equipment e pressure-plat d standar g employin e b d coul s Thi . Durbanville t a u Tukul e th r fo h hig y relativel were also determined on undisturbed soil cores. ascribed to high percentages of very fine sand + fine sand. Dates of budburst, flowering and harvest were recorded annu- Root distribution s wa t Budburs . 2000/2001) o t 5 (1994/9 s season n seve r fo y all c specifi a n withi , 3) e (Tabl n distributio t roo t tha d indicate s Result t a g flowerin d an 4 e stag d reache d ha s bud e th f o % 50 n whe d note parent material, can be largely affected by factors such as soil , mass e Can . 1995) , (Coombe m syste L E- d modifie e th f o 3 2 e stag - per d an s layer d compacte , dry) o to r o t we o to r (eithe e moistur t A . determined o als e wer s mas y berr d an s mas h bunc , yield e grap y b n light-texturea r underlai , Rive s dKuil topsoil t A . stone e centag harvest the musts were analyzed for soluble solid content (°B), a light-textured subsoil with signs of wetness, could be identified s g/a L d tartari. cacid) l titratabl(expresse y tota d eacidit an H p for the Tukulu, in comparison to a medium-textured topsoil and a d considere s wa h whic , ripeness l optima t a d harveste e wer s Grape 0 (0-30 l topsoi e Th . 1) e (Tabl s Vilafonte e th r fo l subsoi t we s les to be at approximately 23 °B, at a titratable acidity of 8 g/L and a mm) of the Tukulu contained less than 7% of the total number of pH of betweee th g n 3.durin 0w anlo dg 3.2bein . e moistur l soi f o t accoun n o y probabl , roots Viticultural and climatic data were analysed statistically, using largest part of the growing season. Medium roots were largely - cal e wer s value D LS t s Student' . replicates s a s season n seve e th concentrated in the 300-600 mm layer, but fine root development culated to facilitate comparison between localities. was obviously enhanced by the slightly wet subsoil, resulting in N DISCUSSIO D AN S RESULT fine roots being evenly distributed from 300 mm to 900 mm. Total r numbe e th e doubl n tha e mor e wer s Vilafonte e th r fo s root e fin s soil f o s analyse l Physica found for the Tukulu, while an appreciable number (20% of total) Particle size distribution (homfelsg s (dept) Papegaaiber h t weighteA . d topsoil mea y n valuesclaye e )mor foe rth thn i e difd - foun s wa e th ) soils c (graniti Kuilr t A . s2 Rive e Tabl n i n show e ar s soil t feren the Tukulu soil again showed inferior root distribution in compar-

2 200 , 2 . No , 23 . Vol , Enol. J . .Vitic. Afr . S Effect of Different Environmental Factors 83

ison to- its Avalooriginat e n counterpart(granit e , mostHelshoogt likelt A . y on accounGlenrosa e th t of fo the e highcas e th - n i n tio r Fo . 2) e (Tabl y capacit e water-storag r lowe d an t conten l grave r e ing from the Stellenbosch pluton) root distribution was better for s wa n distributio t roo ) Valley n hornfelsid (Devo secon l e soi cth the wetter Tukulu soil than for the well-drained Hutton soil. The - frac e ston h hig GlenrosaOakleae e a th th o r n t fo e ftha r du , bette growe th f o - s part r latte e th g durin y dr y overl e b o t Huttod ntende

TABLE2 Soil particle size distribution (%), gravel content, bulk density and water-holding capacity for soils at five localities in Stellenbosch/ . mm) 0 100 f o h dept a o t s mean d weighte h Durbanvill(dept s edistrict

Locality Soil form Clayd) Fine silt (D Coarse sile t (fin D y Ver Fine Medium Coarse % Gravel Bulk Water holding (< 0.002 mm) (0.02- (0.05- sand(') sand») sandd) sand(') (> 2.0 mm) density capacity 0.002 mm) ) 0.02mm (0.10- (0.25- (0.50- (2.0- (kg/m3) (mm/m) ) 0.05mm ) 0.10mm 0.25 mm) 0.50 mm)

Kuils River Tukulu 8.0 11.3 6.7 8.0 14.9 12.5 37.1 37.2 1520 110 Vilafontes 20.4 11.7 5.5 7.2 15.0 11.7 27.1 30.6 1580 111

Papegaaiberg Avalon 32.4 8.0 5.3 10.2 16.0 13.6 11.5 5.0 1460 149 Tukulu 33.4 7.3 9.8 8.8 16.2 12.8 12.4 36.1 1670 118 y Valle n Devo Oakleaf 29.5 12.0 14.4 14.3 15.1 5.8 5.8 25.9 1594 147 Glenrosa 32.7 20.5 9.8 7.2 7.3 4.3 15.5 57.0 (2) 101 <3)

Helshoogte Tukulu 30.6 14.5 10.1 8.4 11.5 9.2 14.0 9.8 1250 136 Hutton 31.8 12.8 8.4 8.7 14.1 10.7 11.7 8.9 1360 120

Durbanville Westleigh 19.4 13.1 13.7 19.1 17.3 7.1 3.9 3.4 1290 143 Tukulu 8.9 12.0 10.3 24.7 25.9 8.9 5.3 3.7 1320 120

) Particle size analyses for soil fraction < 2mm, excluding gravel. > Not determined. . Oakleaf e th r fo d obtaine e valu m fro d Estimate . fraction e ston h hig o t e du d determine t no y capacit g Water-holdin >

TABLE 3 Stellenboschn i s localitie t / t fiva differen r e Richte 9 blanc/9 n Sauvigno f o y densit t roo d an n distributio t roo n o m for l soi f o t Effec . districts e Durbanvill

Root density in 0-900 mm soil Root distribution (%) (roots/mr laye 2)

Locality Soil form s root e Fin Medium roots Medium roots (> 1.0 mm) ) mm 0 (s <1. root e Fin (< 1.0) mmmm )0 1. > (

0-300 mm 300-600 mm 600-900 mm 0-300 mm 300-600 mm 600-900 mm • r Rive s Kuil Tukulu 6.5 44.8 48.7 6.7 74.1 19.0 232 58 Vilafontes 20.0 38.1 41.9 32.2 29.0 38.7 544 62

Papegaaiberg Avalon 25.3 39.3 35.4 26.4 31.5 42.1 863 140 Tukulu 26.8 31.8 41.4 15.7 48.7 25.6 261 160 y Valle n Devo Oakleaf 19.8 32.8 47.4 16.5 47.8 35.6 344 115 Glenrosa 30.1 32.5 37.4 21.0 37.1 41.9 246 105

Helshoogte Tukulu 38.0 25.5 36.5 28.2 42.3 29.4 561 170 Hutton 31.2 34.1 34.7 31.5 50.8 17.7 369 124

Durbanville Westleigh 22.6 38.7 38.7 21.1 25.0 53.8 106 52 Tukulu 36.7 34.1 29.3 33.0 38.7 28.3 581 106

2 200 , 2 . No , 23 . Vol , Enol. J . .Vitic. Afr . S 84 Effect of Different Environmental Factors

ing season. At Durbanvill interstratified an ee (phyllitiillit . d (e.g c s shalesvalue ) rooC tCE distributio r highe h n wawit s s mineral relativeld y an poo e r for theHelshoogt r Westleigfo d hfoun s soilwa , t probablconten yC ot n accounhighes e t Th o . f a minerals) water table during the first part of the growing season. In general, Durbanville, suggesting higher levels of organic material in soils - illustrat s a , form l soi o t d relate y directl t no s wa n distributio t roo h Stellenbosc e th m fro d an s shale y Malmesbur m fro d derive ed by the fact tha hornfelm t roofro g ts density originatin (tota e l roots/mthos o t n 2) withincompariso thn i e n Tukulpluto u e Granit form ranged from 731 (Helshoogte) to 290 (Kuils River). and from the Kuils River-Helderberg Granite pluton. However, Furthermore, , relativelmaterial t paren y gooo t d d roorelate t t distributiono y n wanormall s i s s obtainesoil f o dt for atconten C e th least one soil from each individual parent material, suggesting that but rather to prevailing climatic conditions during the process of . material t paren y b d affecte y indirectl y onl s i n distributio t roo soil formation (Stevenson, 1986), thus implying that soils at s soil f o s analysi l Chemica Helshoogte and Durbanville were formed under cooler condi- e b d shoul n regio h Stellenbosc e th m fro s soil d Vineyar . tions s soil n betwee d occurre s analysi l chemica n i s difference r majo o N f o d deman e th y satisf o t N t sufficien g supplyin f o e capabl - depth y onl y Consequentl . shown) t no a (dat y localit e sam e th t a grapevines, where the organic material content exceeds 1% weighted mean values for the two soils at each locality are shown n fertilisatio N d standar e th t tha d suggeste s Thi . 1986) , (Conradie in Table 4. A relatively low pH value was found for the soils at programme for this region (30 kg N/ha/yr) could be waived at Kuils River (Table 4), mainly on account of a value of 4.13 in the Helshoogte and at Durbanville. subsoil (650-1000 mm) (data not shown). At this locality, in contrast to the other four, lime was probably not adequately mixed Potassium in subsoil 5 5. o t 0 5. f o s value H p t A . preparation l soi g durin l subsoi e th o int The lowest value (18 mg/kg) for K in the subsoil (600 - 1000 d impede y seriousl e b t no d shoul e performanc e grapevin mm) was found at Durbanville (Table 5), thus being in agreement (Conradien , 1983)contai s . Nonshale t e tha o t f thfac e e soilth sh showewit ) d 1988 an , y signs of sodicityWooldridge ; 1964 , (Visser n tha r highe r neve s wa n saturatio a N t tha t fac e th y b d evidence s a less total potassium than granite (Kuils River and Helshoogte). - 2 mg/k16. (Papegaai gm fro d range t conten s Phosphorou . 3% However, since hornfels contains even less K than phyllitic berg) to 48.3 mg/kg (Durbanville). In view of the fact that the shales, high potassium levels at Papegaaiberg and Devon Valley P contene t ohav f y parenma s t materiallevel high se fromThes . the Westermaterial t n Capparen h e tendwit e s to bcorrelat e t no d di low (Visser, 1964), these relatively high values must have been been derived from K fertilisers that leached into the subsoil or the result of P fertilisation. Grapevine performance should not be they may have been the result of mixing of parent materials. , toxicity P r o y deficienc P f o t accoun n o r eithe , impeded y undul Relatively high clay contents at both these localities (Table 2) at any og f the concentrationpointin s thu , extent s e showsom no t i n K Tablf o g e 4 (Saaymanleachin d , impede 1981 e ;hav d coul Conradie e &wher Saayman , River s , 1989)Kuil t A . . materials t paren f o g mixin s toward e mor The higheso int d t Ca leache leve e l hav wa y s ma foun m d for potassiu Durbanville , 2) e (Tabl , w thulo s beinwa t g in conten y cla agreement with a relatively high pH value. Potassium was high- the subsoil, thus explaining a relatively high value in comparison est for Papegaaiberg and lowest for Durbanville. Fertilisation to the other granitic soil (Helshoogte). Basically the above results practices may have had a significant effect on the K level of soils. were in agreement with the suggestion that management practices - affect least e b l wil l subsoi e th f o t conten K t tha d accepte s i t i f I can change the chemical composition of soils to such an extent ed by fertilisation, (Seguin s , K levels incharacteristic th l e deepeorigina rr soilthei layere hav s r may blonge e o relatn y - the t tha n i d discusse e b l wil t aspec s Thi . material t paren e th f o t tha o t d e 1986). In general, with Durbanville being the only exception, . section t nex e th g underlyin o t d relate e b t no d coul s subsoil f o s level m potassiu t canno r paramete s thi t tha g suggestin s thu , formations l geologica m fro d derive s soil r fo t highes e wer C CE d an e valu S h Bot terroirc .specifi a f o s merit e th n o e decid o t d use e b Malmesbury shale (Durbanville). Since kaolinite, with a low CEC (White, 1987), is usually the predominant clay mineral in soils Climate s result e abov e th , 2001) , Schoor n (Va e Cap n Wester e th m fro The macro-climate in the Stellenbosch/Durbanville districts can indicate that shale-derived soils contain a higher fraction of clay be classified as Mediterranean, i.e. mild, rainy winters and warm,

TABLE 4 s soil o tw r fo s mean Stellenbosch/Durbanvilln d i s (0-100s fivt a weighte e) localitie soil h f o 0 mm s (dept s edistrict characteristic l Chemica at each locality).

Locality pH Resistance P Ca Mg K Na S value CEC Saturation C ) (ohm ) (KC1 (mg/kg) (meq/lOOg) (meq/lOOg) (meq/lOOg) (meq/lOOg) (meq/lOOg) (meq/lOOg) (%) (%)

Kuils River 4.63 a «) 3299 b 39.2 be 1.74 a 0.87 ab 0.26 ab 0.008 a 2.88 a 4.69 a 61. 4 a 0.49 a Papegaaiberg 5.03 b 2566 ab 16.2 a 1.75 a 0.71 ab 0.50 c 0.079 ab 3.04 a 4.99 a 60.9 a 0.58 a y Valle n Devo 5.33 be 1934 a 33.9 be 3.12ab 1.78 b 0.32 b 0.041 a 5.26 ab 6.51 ab 80.8 ab 0.34 a Helshoogte 5.29 be 3399 b 23.7 ab 3.81 be 0.54 a 0.27 ab 0.002 a 4.62 ab 6.19 ab 74.6 ab 1.15b Durbanville 5.42 c 1693 a 48.3 c 5.28 c 1.39ab 0.21 a 0.243 b 7.12 b 7.96 b 89.4 b 1.29b

S. Afr. J. Enol. Vitic., Vol. 23, No. 2, 2002 Effect of Different Environmental Factors

TABLES and the Santiago region in Chile. During the maturation period of Potassium content of subsoil horizons (600-1000 mm) at five Sauvignon blanc (February), day and night temperatures were localitiey s in Stellenbosch/DurbanvillFebruar r fo s temperature m e districtsmaximu d .an m minimu , Mean . warm were 21.4°C, 15.6°C and 28.3°C, respectively. However, despite Locality K CEC K saturation all of the stations being within a radius of 15 km, significant dif- . 6) e (meq/lOOg(Tabl d ) detecte e b d (meq/lOOgcoul s ) aspect c climati (%r ) fine n i s ference

Kuils River 0.179 4.64 3.86 Summer Papegaaiberg 0.256 g s >30°Cgrowin , temperature 3.0h 4 wit ) (Dec-Feb s 8.42 hour f o r Numbe y Valle n Devo s toward d 0.23pointe 1 x inde n 6.5Hugli e 9 th d an ) 3.50(Dec-Feb s degree-day Helshoogte 0.151 4.24 3.56 Kuils River, Helshoogte and Durbanville being cooler than Durbanville 0.046 6.76 0.68 Papegaaiberg and Devon Valley. Mean and maximum tempera- , d 3.1°Can respectively C , 1.8° y b d differe y Februar r fo e tur between the warmest (Papegaaiberg) and the coolest locality x inde y variabilit e temperatur y Februar e Th . (Durbanville) dry summers- dif s . Mealocalitie n t data cooles fo re ththre e e fivth e t weathetha d r stationshowe ) s 1992 fo , r the (Gladstones seven-yeax r inde perio s thi dr (1994-2001fo s value r ) are(Lowe show. n in amplitude Fig l . 2. Ththerma f eo s coldesterm tn i d fere month (July) had a mean temperature of 12.1°C, while 530 mm of suggest a more temperate climate, while higher values point o t l Apri m fro d recorde s wa ) total l annua e th f o % (77 n rai - mini e th y Februar g Durin ) climate. l continenta e mor a s toward Septembere b .y Dryfirstl conditiond coul s sThi . wereRiver experiences Kuil t a t d frohighes ms Decembewa e r to temperatur m mu , mm) 9 (5 l rainfal ) (December-February r summe l tota h wit , March ascribed to the proximity of the sea and secondly to its location being less than 9% of the annual total. Temperature during the on an east-facing slope, thus warming up earlier in the morning. growin- g seasoclimat nn i place e rol dt the studysignifican a are y a pla in e regionaltitud Hd I an o fa thse ee Winkleth o t y r Proximit f o x al,inde et n r a 1974 h wit ) (Winkle n classificatio y degree-da y da h hig s prevent e breez a se e th f o r ai t mois l coo e Th . patterns c i Huglie th r nfo ) degree-days 3 (222 e valu e Th . degree-days 3 178 temperatures, especially in February (Bonnardot, 1999; Bonnar- heliothero n e b d - shoul e ther t tha d indicate ) 1978 , (Huglin x Inde - clos s location e th , Durbanville d an r Rive s Kuil . 2001) al, et t do mic constraint for ripening of most cultivars. According to these est to the sea, therefore experienced a maritime climate (lowest r Montpellie s a r cluste e sam e th n i s fall a are y stud e th s indice o tw d an m maximu e moderat o t e du x inde y variabilit e temperatur d locate l al , Italy n i a Veron d an e Franc f o h sout e th n i e Orang d an minimum temperatures) and, as a result, recorded the fewest in the Mediterranean climatic zone, as well as the coastal regions hours with temperatures above 30°C. The cooling effect of the a Australi n Wester f o t par n souther e th , USA e th n i a Californi f o sea breeze decreases rapidly with distance from the sea

30 -

§ is 20 2.

10 -•

y JiMa m g r Au Ap Jul r Ma b Fe n Ja c De v No t Oc p Se Months

) Rainfall - • »- Min . •Max. TempTemp . . •• •+ --Mean Temp| .

FIGURE 2 Mean monthly rainfall and temperature (minimum, mean and maximum) values for . (1994-2001) a are y stud e th n i s station r weathe e fiv e th

S. Afr. J. Enol. Vitic., Vol. 23, No. 2, 2002 86 Factors Environmental Different of Effect

TABLE 6 Climatic parameters as measured at five localities in the Stellenbosch /Durbanville districts (mean for the period 03/1994 to 03/2001).

Locality February February February Growing Huglin Temp. Temp. February Rainfall July July July Number meanm maximinimu - degree- index >30°C <12°C temp. (Oct- mean maxi- mini- of days temp. mum temp. days (°C) (Dec-Feb) (Dec-Feb) varia- Mar) temp. mum mum with (°C) temp. (°C) (Dec-Feb) (hours) (hours) bility (mm) (°C) temp. temp. mean (°C) (°C) index «> (°C) (°C) temp. <10°C (May- June) r Rive s Kuil 21.1c<2 ) 27.6c 16.0 a 942 b 2167 b 107 b 23 a 34.5 a 155 b 12.4 ab 15.8 c 9.7 a 3a Papegaaiberg 22.2 a 29.8 a 15.5 c 1039 d 2420 d 195 e 39 b 41.4 b 153 b 12.5 a 17.3 a 8.6 b 2a Devon Valley 21.8 b 29.1 b 15.8 ab 1000 c 2325 c 164 d 42 b 39.0 c 141 be 12.2 be 16.6 b 8.8 b 4ab Helshoogte 21.3 c 27.9 c 15.6 be 948 b 2146 ab 125 c 86 c 37.8 c 214 a 11.3d 15.6 c 7.7 c 9c Durbanville 20.4 d 26.7 d 15.2 d 903 a 2059 a 71 a 37 b 35.6 a 131 c 11.9 c 16.0 be 8.6 b 7 be

d TMd xan an m minimu d an m maximu y dail Z[(TD= e x TDmin- inde averag (TMy e + )TDmiTMmin)]- d th an TDe e nar wher variabilit , e CTemperatur > ma x x ma x ma ma TDmin the average highest maximum and average lowest minimum temperatures for February (Gladstones, 1992). .

f o s station d inlan e th n i g (Bonnardoresultin , - al,et t2001) expect e wer s deficit r wate , Consequently . period s thi r ove y cantl Papegaaiberg and Devoh nhig Valle , y experiencin l formssoi e Theoretically . th f o ge a morsom n eo continentas grapevine l r fo d e climate (highest temperature variability index). Higher tempera- degrees of soil-water depletion (Table 7) and water stress were tures ate thesth , e stationcontents) s couly cla d alsw o(lo havr e Rive bee s n partlKuil t a y s duesoil toh thbot e r lowefo d r expecte - experi o als e and/ot (Ne WNW)d Wan raspec altitud Helshoogt . , content) l grave h hig d an e climat m (war g TukulPapegaaiber t ua enced a more continental climate, but was cooler than Pape- the Glenrosa at Devon Valley (high gravel content) and the Tukulu . altitude r highe a t a n locatio s it o t e du , Valley n Devo d an g gaaiber at Durbanville (low rainfall during summer). Lowest water stress This localit- Leaf y . also experienceHelshoogte t a s d soil th eh highesbot n o s t rainfall grapevine durinr fo d g the expecte s wa active growind an y g seasoJanuar nr fo (Oct-Mar ) noon )d and aroun th d e highest(measure numbes r opotential f r wate hours witr h Rive temperature s Kuil r fo s s belostres r w 12°wate t C (duhighes e e toth cooled r nights)showe ) 7 e . (Tabl y Februar Winter (both soils) and the Glenrosa at Devon Valley in comparison to lower stress at Helshoogte. At Papegaaiberg (warmest location), g durin d reduce e wer s localitie n betwee s difference e Temperatur however, water stress appeared to be low for the Tukulu, as well as winter. Mean and maximum temperatures for July differed by e closur a stomat e induc d coul s stres r wate e Excessiv . Avalon e th 1.2°C and 1.7°C, respectively, between the warmest (Papegaai- d an 3 198 , Coombe & t (Smar y da e th f o t par t warmes e th g durin - situat , locality r latte e Th . locality ) (Helshoogte t cooles d an ) berg ed at the highest altitude and facing south, also experienced the s day f o r numbe highest e th d an s temperature m minimu t lowes f o e releas e th r fo s Condition . 10°C < f o s temperature n mea h wit grapevine buds from dormancy (9 consecutive days with mean TABLE 7 temperatures below 10°C during May h and JuneMarc f o ) should en 2 mt da ) b ee (0-1. soil-wate best t atavailabl conten rt Plan this locality. t a ) February d an y (Januar l potentia r leaf-wate n (PAWemea d an ) a se e th f o y proximit e th o t e du , that d showe s result e abov e Th Stellenbosch/Durbanvilln i s localitie e fiv (1994-2001)s edistrict . and the complexity of the topography, finer climatic aspects can Locality m for l Soi PAW r Stellenboschfo e pictur " / "global e th n withi y considerabl y var e Leaf-water potential (mm) (-MPa) Durbanville. Similar indications were obtained in a study on nat- e Bottelary-Simonsbergth n i e viticultur r fo l terroi- s ura unit r Kuils River Tukulu 13 1.53 (Careya . are g 2001) , wine-growin g Helderber Vilafontes 15 1.56 s statu r wate l Soi Papegaaiberg Avalon 31 1.34 Due to relatively high rainfall up to the end of September (Fig. 2), Tukulu -6 1.35

soils were at field capacity on 1 October. From October to t

December monthly rainfall averageds 38 mm for the five localities Devon Valley Oakleaf 34 1.50 Glenrosa -21 1.59 e th g durin d replenishe y partl g bein r wate l soi n i g resultin , 2) . (Fig first part of the growing season. Soil water was therefore sufficient Helshoogte Tukulu 57 1.38 to meet the requirements of the grapevines up to the end of Hutton 38 1.43 December for most soil forms (data not shown). From January to March monthly rainfall averaged only 15 mm for the five locali- Durbanville Westleigh 38 1.27 Tukulu 10 1.28 - signifi t conten r wate l soi e increas t no d di t i t tha g implyin , ties

2 200 , 2 . No , 23 . Vol , Enol. J . .Vitic. Afr . S Effect of Different Environmental Factors 87

references therein). At Papegaaiberg e thiscan mag y hav eAvaloe resulteth Papegaaiber d t a nan dy in Valle n OakleaDevo t a f leaf-water potentials not giving a true reflection of water stress. masses were low in comparison to predicted values. In view of - con y relativel d remaine g Papegaaiber t a s potential t tha t fac e Th e (Tabl s localitie t warmes e th g bein g Papegaaiber d an y Valle n Devo stant over January and February, while all the other localities 6), excessively high temperatures may have suppressed CO2 assim- n i n tha y Februar n i ) stress r wate r (highe s value r lowe d showe ilation, responsible for vegetative growth (Williams et al., 1994 and - con r Unde . assumption s thi d supporte , shown) t no a (dat y Januar , counterparts e respectiv r thei o t n compariso n I . therein) s reference r leaf-wate n pre-daw r o - stem , stress r wate e excessiv f o s dition t a f Oaklea e th r fo h hig y relativel l stil e wer s masse e can , however potentiale sth shoulg d ratheillustratin s r thu b e, measured Papegaaiberg (Chon t a n e etAvalo al.,e th d 2001)an y . AValle t n Devo Durbanvill. e grapevinelocalities m swar experiencet a n eve , d thegrowth lowese t watevegetativ n r o stress m for , l mainsoi f -o t effec ly on account of the cool climate (Table 6). Apart from the differ- Grapevine performance d differe o als , s leaf-watelocalities potential r n betwee s ence n I . locality e sam e th t a s form l soi t differen n o s grapevine n betwee Phenology s stres r wate r highe , counterparts e respectiv r thei o t n compariso Budburst (Table 8) was earliest at Kuils River (both soils), Devon Glenrose th , t Kuila t s a sRiver Vilafonte e th r fo d measure s wa Valley (both soils) and Durbanville (Tukulu only), while it was Devon Valley and the Hutton at Helshoogte. significantly later for Papegaaiberg (both soils), Helshoogte (both n a h wit e increas o t d tende mass/) e (can mh cordon growt e Vegetativ g Papegaaiber f o w vie n I . only) h (Westleig e Durbanvill d an ) soils increase in expected PAW (Fig. 3). Cane masses were high for the being the warmest locality, in contrast to Durbanville being the n compariso n i e Helshoogt t a n Hutto e th d an r Rive s TukulKuil t ua e y wholl e b o t r budbreaf appea o t e no dat d k, di 6) e (Tabl t cooles l soi f o t movemen y PAWm fro Capillar . d predicte s value e th o t temperature dependent. Soil temperatures during early spring e water , from the Hofa'cker deepe & t r layers (Alleweld coul t d havebudburs f resulteo t donse ine th PAW d (calcuaffecte e - hav y ma

lated to, 1.2 m) beinDurbanville t a g d under-estimatefoun s wa n d for thesassumptio s e thi soilsr fo .e e For theEvidenc . 1975)

1.0

Hels 5 0.682 + X 0.0043 = y Hels2 R2 = 0.6073 0.9

o • Kuils 1 "E 8 0.8

0.7 • Devon 1 0 Pap1 Pap 2 0.6 Devon 2

0.5 -50 -25 0 25 50 75

PAWe (mm)

FIGURES

Plant available soil-water content (0-1.2 m) at the end of March (PAWC) and cane mass for the five localities in Stellenbosch/Durbanville . Durbanville) = r Du , Helshoogte = s Hel , Valley n Devo = n Devo , Papegaaiberg = p Pa , River s Kuil = s (Kuil . (1994-2001) s district

S. Afr. J. Enol. Vitic., Vol. 23, No. 2,2002 $g Effect of Different Environmental Factors

TABLE 8 . 2001) o t 4 blanc/9n 199 m 9 fro RichteSauvigno n r fo e r(mea performanc t plan d an y phenolog n o m for l soi d an y localit f o t Effec

Locality m For l Soi Date of Date of Date of Cane mass Yield Bunch Berry Sugar Titratable pH budburst flowering harvest (kg/vine) (kg/vine) mass (g) mass (g) (°B) acidity (g/L) Kuils River Tukulu 11/0) 9ad 04/1 labc 11/02 a 0.84 abc 2.21 d 1 19.65 b 1.95 b c ab 8 22. 8.86 ab 3.19 a Vilafontes 12/09 ab 06/1 led 12/02 a 0.73 ab 2.10cd 1 17.36 b 2.09 c 23.4 be 8.04 ab 3.2b 7a

Papegaaiberg Avalon 16/09 cd 01/11 a 10/02 a 0.69 ab 1.79 be 118.24 b 1.96 b 23.4 be 8.33 ab 3.27 ab Tukulu 16/09 cd 01/11 a 12/02 a 0.64 a 1.40 ab 101.92 ab 2.07 c 22.8 abc 7.74 a 3.35 b

Devon Valley Oakleaf 12/09ab 02/11 ab 09/02a 0.84 abc 1.71 abc 84.82 a 1.69a 23.3 be 9.17b 3.20a Glenrosa 12/09 ab 02/1 labc 09/02 a 0.70 ab 1.99 cd 109.04 b 1.72 a 23.9 c 7.98 ab 3.22 ab

Helshoogte Tukulu 17/09d 11/11 f 01/03 d 0.94cd 1.70abc 104.18ab 1.77 a 23.1 be 8.79 ab 3.21 a Button 16/09 cd 09/11 ef 24/02 c 0.95 cd 1.84 cd 102.48 ab 2.13 cd 23.3 be 8.14 ab 3.20 a

Durbanville Westleigh 14/09 be 07/1 Ide 03/03 d 1.07 d 1.38 a 116.76 b 2.22 d 22.0 a 8.40 ab 3.20 a Tukulu 10/09 a 05/11 be 19/02 b 0.91 bed 1.83 cd 108.45 b 2.21 d 22.6 ab 7.73 a 3.30 ab

TABLE 9 Number of days and number of heat units experienced from budburst to harvest at five localities in Stellenbosch/Durbanville districts (mean from 1994 to 2001).

Budburst to flowering Flowering to harvest

Locality Soil form s day f o r Numbe Heat Units (« Number of days Heat Units «> r Rive s Kuil Tukulu »b >a 4 5 298 a 99 ab 963 a Vilafontes 55 b 302 a 98 a 973 a

Papegaaiberg Avalon 46 a 306 a 101 ab 1080 bed Tukulu 46 a 301 a 103 abc 1090d c

Devon valley Oakleaf Slab 312 a 99 ab 1023 abc Glenrosa Slab 317 a 99 ab 1017b a

Helshoogte Tukulu 55 b 294 a HOcd 1101 cd Hutton 54 ab 289 a 107 be 1061 be

Durbanville Westleigh 54 ab 292 a 116d 1143d Tukulu 56 b 290 a 106 be 1006 ab

C> Growing degree-days above 10°C (Amerine & Winkler, 1944). .

where budburst occurred four days later for the Westleigh (high between localities, all localities experienced a comparable num- water table during winter and therefore lower soil temperatures) ber of heat units (ranging from 289 to 317) over this period (Table than for the Tukulu (drier and warmer than the Westleigh). Time 9). These results were in agreement with the suggestion (Huglin of floweringn o t , howeverdependen y , did nolargel s i t diffeg r withiflowerin f no e a specifidat t tha c) localit1998 , y and Schneider & was earliest for the two warmest localities (Papegaaiberg and temperature, especially on maximum temperatures. Devon Valley), while it was latest for one of the cooler localities Grapes from the two warmer localities (Papegaaiberg and e th r fo d foun e wer s date g flowerin e Intermediat . (Helshoogte) r earlie s week e thre o t k wee e on d harveste e wer ) Valley n Devo other cool localities (Kuils River and Durbanville). Even though than those from two of the cooler localities (Helshoogte and d differe g flowerin d an t budburs n betwee s day f o r numbe e th s unit t hea f o r numbe e th s localitie r fou e thes r Fo . Durbanville)

S. Afr. J. Enol. Vitic., Vol. 23, No. 2,2002 Environmental Factors Environmental Different of Effect 89 experienced between flowering and harvest ranged from 1006 to centage of bunches being exposed (determined by means of point 1143 (Tabl- rela e, 9). Iacidity n h the (hig cas e e ofrevers th e e Th othe . r relativelshown) t no ya coodat - ld localitmetho yt quadra (Kuils River) grapes ripened at the same time as those from the tively low pH) was true for the Oakleaf at Devon Valley, despite warmer localities (Table 8). This was probably on account of the high maximum temperatures (Table 6) and high soil K. At the same low water-holding capacities of soils at Kuils River (Table 2), locality, however, acidity tended to be lower and pH higher for the resulting in relatively high water stress during the latter parts of Glenrosa (higher water stress). Any factor, such as water stress, that the season (Table 7). In spite of relatively cool conditions, ripen- reduces the photosynthetic activity of leaves will result in increased ing was thus enhanced, resulting in a low number of heat units potassium accumulation in berries and higher must pH values (mean value of 968) between flowering and harvest (Table 9). At (Freeman et al, 1982). It must be noted that the trial was carried Durbanville grapes from the wetter soil (Westleigh) were har- out in commercial vineyards with different clones, which may have vested 12 days later than those from the drier soil (Tukulu). had an effect on sugar/ acid/pH balances. Within a specific locali- Grapevines on the Westleigh grew vigorously (Table 8) and active ty, however, acidity tended to be lowest and/or pH highest for the growth could still be detected during the ripening phase, resulting soil with the lowest water content. Apart from the effect of con- in sugar accumulation being slowed down. Different soil water trasting associations of landscape/aspect/climate on the perfor- regimes could also have been the reason why grapes from the mance of Sauvignon blanc (phenology, vigour, yield, sugar accu- wetter soil (Tukulu) at Helshoogte were harvested 5 days later mulation, acidity), soil type may therefore also have a pronounced than those from the drier one (Hutton). Date of ripening was, effect. Wine style should differ between different localities and, in therefore, not affected by climate only, but also by the soil. some cases, also between soils at the same locality. Growth and yield CONCLUSIONS In general, the highest cane masses (Fig. 3, Table 8) were obtained It is often difficult to relate soils to specific geological parent at the localities (Helshoogte and Durbanville), where soils con- materials. Management practices (especially loosening and tained relatively high amounts of organic C (Table 4), usually in chemical amelioration of soils before grapevines are planted) combination with a high soil water status (Table 7). Yield was low- may change the chemical properties of soils to such an extent that est for the Westleigh at Durbanville, probably on account of infer- inherited characteristics are no longer distinguishable. Potassium tility, induced during spring by too vigorous growth at floral initia- levels nevertheless appear to be lower for soils originating from tion and therefore lower sunlight interception in the vine canopy. In phyllitic shales in comparison to granitic soils. Geology also contrast, yield was also low for the Tukulu at the warmest locality plays a role in that it influences the physical properties (especial- (Papegaaiberg). In this case low yield could be ascribed to a low ly particle size distribution and water-holding capacity) of soils. number of bunches per vine and fewer berries per bunch (Table 8). Geological maps may therefore be valuable tools where decisions At this locality a combination of factors (high temperature/soil with concerning cultivars and rootstocks must be made for specific a high gravel fraction and a low water-holding capacity/poor root localities. Such maps must, however, be considerably more distribution) resulted in a terroir with low potential for Sauvignon detailed than those which are currently available. blanc. Yield was highest at the relatively cool locality (Kuils River), Within the Western Cape of South Africa appreciable climatic where grapes ripened earliest. Comparable yields (1.70 - 2.09 differences can occur between localities in close proximity. This kg/vine) were obtained at the other sites. Berries were largest at the is largely on account of different aspects, altitudes and distances coolest locality (Durbanville) and smallest at one of the warmer from the sea. These subtle climatic differences do have an effect localitien s (DevonSauvigno Valley)f o n . At threcompositio e l of the localitiechemica e th n so (Kuild an sy River, phenolog n o Papegaaiberg and Helshoogte) berry size was also significantly blanc grapes. Ripening can either be enhanced or delayed, while . form l soi y b d affecte sugar/acid/pH balances are also affected. These differences Must composition should be reflected in wine quality. In spite of grapes from the Westleigh at Durbanville being har- Even though climate appears to be the most important driving vested last, sugar content was still the lowest (Table 8). In con- force affecting grapevine performance, the effect of soil cannot be trast, a significantly higher sugar content was obtained for grapes discounted. Especially in the case of soils with low water-holding from the eb Tukuls ualway t at Helshoogtecanno e ,climat whic" h "good wer a f eo alss o harvesteadvantage e d th rela s - capacitie tively late. Slow ripening in the case of the Westleigh at fully exploited. High water stress may result in grapes with low Durbanville could obviously be ascribed to problematic soil con- acidity/high pH. On the other hand, adequate ripening may also be ditions resulting in excessive vegetative growth and not to cli- problematic on soils that are too wet. Even though the use of dif- matic factors. In general, grapes from all other sites could be har- ferent clones created limitations in the current study, it may still be vested fairly close to the desired sugar content (23°B). argued that the correct scion/rootstock combination will help to ensure that each situation is exploited to its fullest advantage. Titratable acidity was lowest and pH highest for grapes from the Tukulu at Papegaaiberg (Table 8). This may have been on account In the past it has been argued that application of the terroir con- of the warm climate (Table 6), leading to metabolism of malic acid, cept is of doubtful value in relatively warm wine-growing coun- and the high K content of the soil (Table 5). However, low titratable tries such as South Africa. Results from this study, however, sug- acidity and a moderately high pH value was also found for the gested that Sauvignon blanc wines with region-specific charac- Tukulu at Durbanville, despite the cool climate and low soil K. The teristics can be produced within specific districts. Differences in lower acidity can possibly be ascribed to canopy-management wine style, as induced by different terroirs, should be the subject practices followed at this locality, which resulted in a large per- of further research.

Enol2,200. . J . No , . Vitic.2Afr 23 . . S ,Vol 90 Effect of Different Environmental Factors

LITERATURE CITED Morlat, R., 1996. Elements importants d'une me'thodologie de caracterisation des facteurs naturels du terroir, en relation avec la reponse de la vigne & travers le vin. cr , austrieb f au n umweltfaktore n vo s Einflus . , W.1975 , Hofacker & . G , Alleweldt In : Proc. l Colloque International "les Terroirs Viticoles", 17-18 July 1996, bliite, fruchtbarkei , 103-11514 s triebwachstumd Viti . . un t rebe r de i be , France . 17-31pp ,Angers .

Amerine, M.A. & Winkler, A.J., 1944. Composition and quality of musts and Myburgh, P.A., 1998. Water consumption of South African vineyards: A model- wines of Californian viticulturald grapesselecte . Hilgardif o ,s aeffect 15d , 493-673combine d . quantifie e th n o d base h approac g lin , University h Stellenbosc , Dissertation D Ph . parameters l meteorologica d an l soi Private Bag XI, 7602 Matieland (Stellenbosch), South Africa. . 33 . Vol , studies l Ecologica . systems t roo g studyin f o s BohmMethod . , W.1979 , Springer-Verlag, Berlin. J.C.M., e , RankineBoehmCellier& Fomachon . , ,,E.W Influenc K.M.,B.C. . 1971 , of grape variety, climate and soil on grape composition and on the composition preliminaire Etud period a . l t brises e1999 e de e med e , sd V r pendan , Bonnardot . 33-50 , 10 s Viti . wines e tabl f o y qualit d an Stellenbosch-Kleie d e regioa l s nviticol dan n n Drakensteimaturatio Afriqun ne e 3 26-3 , 12 n . AICPublicatio Sud u d . Int . Proc : In . quality e win n o e climat d an l soi f o t effec e Th . , D.1977 , Saayman Symp. Quality of the Vintage, 14-21 February 1977, Cape Town, South Africa, pp. Bonnardot, V, Carey, V, Planchon, O. & Cautenet, S., 2001. Sea breeze mecha- 197-208. . area g e Stellenboscth n i producin e hwin effects s it f o s observation d an m nis . Wynboe10-14 , 147 r Saayman, D., 1981. Wingerdvoeding. In: Burger, J. & Deist, J. (eds). Wingerdbou in Suid-Afrika. Carey, V., 2001. Spatial characterization of natural terroir units for viticulture in the Bottelaryberg-Simonsberg-Helderberg winegrowing area. M.Sc. Thesis, h Sout , Matielan 2 XIg 760 ,Ba e d(Stellenbosch) Privat , University h Stellenbosc Nietvoorbij, Stellenbosch, South Africa, pp. 343-383. Africa. Saayman, D., 1992. Natural influences and wine quality. Part 2: The role of soil. Carey, V.A., Bonnardot, V.M.F., Schmidt, A. & Theron, J.C.D.. 49-51 , , 20011992 . t The interAugus , - Wynboer action between vintage, vineyard site (mesoclimate) and wine aroma of Vitis vinifera L. cvs. Sauvignon blanc, Chardonnay and Cabernet Sauvignon in the Saayman, D., 1998. The development of vineyard zonation and demarcation in Stellenbosch-Klein Drakenstein wine growing area, South Africa (1996-2000). In: South Africa. In: Proc. Simposio Internazionale Territorio & Vino, 19-24 May , Adelaide , 2001 r Octobe , Congress e Win & e Vin Procd O.I.V26 . Worl . 1998, Siena, Italy, pp. 35-42. Australiah t , pp. 139-152. SAWIS, 2000. South African Wine Industry Statistics No. 24, SAWIS, Paarl, Gaudillere& r . D , wate 2001, m ,J.P. .Ste C., ,Dubourdieu Chon6Leeuwen , n X.,Va South Africa. . 477-483 , 87 . Bot . Ann . status r wate e grapevin f o r indicato e sensitiv a s i l potentia Scholander, PR, Hammel, H.T., Bradstreet, E.D. & Hemmingsen, E.A., 1965. Sap Conradie, W.J., 1983. Liming and choice of rootstocks as cultural techniques for pressure in vascular plants. Science 148, 339-346. . Afr S . Enol . J . . .soils Viticd 39-44 , aci 4 . n i s vine r Elsevie . Climatology f o y Surve d Worl , Africa h Sout . 1972 , B.R. , Schulze Conradie, W.J., 1986. Utilisation of nitrogen by th. e grapevin501-586 , e a(15) s affecte0 1 , d by timCompany g e Publishin of application and soil type. S. Afr. J. Enol. Vitic. 7, 76-82. Seguin, G., 1983. Influence des terroirs viticoles sur la constitution et la qualit£ s nitrogenm phosphoru ,long-ter f o Saaymans & . W.J Effect , . , D.1989 , Conradie des vendanges. Bull. OIV 56 (627), 8-12. e grap d an s analyse f Lea . II . vines Chenin o c n nblan fertilizatio m potassiu d an composition. Am. J. Enol. Vitic. 40, 91-98. Seguin, G., 1986. "Terroirs" and pedology of wine growing. Experientia 42, 861- 873. Coombe, E.G., 1995. Adoption of a system for identifying grapevine growth stages. Austr. J. Grape Wine Res. 1, 104-110. Smart, R.E. & Coombe, B.G., 1983. Water relations of grapevines. In: Kozlowski T.T. (ed). Water deficits and plant growth VII. Academic Press, New York. pp. De Villiers, F.S., Schmidt, A., Theron, J.C.D. & Taljaard, R., 1996. 137-196. Onderverdelin e Wes-Kaapsdi n va g e wynbougebied s bestaandvolgen e e kli- maatskriteria. Wynboer Tegnies 78, 10-12. Smart, R.E. & Dry, PR., 1980. A climatic classification for Australian viticultural regions. Austr. Grapegrower and Winemaker 196, 8-16. Etourneaud, F. & Loue, A., 1984. Le diagnostic petiolaire de la vigne en relation : In . t l emagnesium e 1'interpretatioc m ave r lepotassiu 1'analyse npou d l so e ed Proc. 6th International taxonomiColloquiuA - n - csys m foclassificatio r l theSoi . optimisatio1991 , nGroup g of planWorkin t n nutrition, Classificatio l Soi Septembe- r nat 1984 n o s , MontpellierMemoir : , France, Development pp l . 189-198.Agricultura f o t Departmen . Africa h Sout r fo m te ural agricultural resources of South Africa No. 15, Department of Agricultural Development, Pretoria, South Africa. Freeman, B.M., Kliewer, W.M. & Stern, P., 1982. Influence of wind breaks and r potentialwate f ,lea stomata- f o n conduc l fluctuatio l diurna n o n regio c climati tance, and leaf temperature of grapevines. Am. J. Enol. Vitic. 33, 233-236. Stevenson, F.J., 1986. Cycles of soil - Carbon, Nitrogen, Phosphorus, Sulfur, . USA , MicronutrientsYork w Ne , Wilen Sons Joh .y &

. Winetitles. Adelaide , Environment d an e Viticultur . , J.1992 , Gladstones e th f o y Geolog e Th . , J.1992 , Rogers & . H.P , , P.O.Siegfried , Gresse , J.N. , Theron Cape Town Area. Explanation of Sheet 3318. Department of Mineral and Energy devaluatioe mod u possibilits nde Nouvea . y heliothermique1978 , P. , Huglin s . Africa h Sout , Pretoria , Affairs . Agric.Acad , R 1117-1126C . viticole . u milie n d'u t ['influence x mesou ed mondiau -s viticole s macroclimat s Le . , J.1999 , Tonietto Huglin, P. & Schneider, C., 1998. Biologie et Ecologie de la vigne (2nd ed). climat sur la typicite de la Syrah et du Muscat de Hambourg dans le Sud de la . Paris , Lavoisier France - Me'thodologie de caracterisation. ENSA-Montpellier. Doctoral thesis.

Kendrew, W.G., 1961. Th- e climatemin n s ofractio f y thecla Continentsd an n . Oxfordistributio e dsiz Universite particl , y PressGeology , . Schoorn 2001 , Va L.H. , London. eralogy of selected vineyard soils in South Africa and the possible relationship with grapevine performance. M.Sc. Thesis, Stellenbosch University, Private Bag Le Roux, E.G., 1974. 'n Klimaatsindeling van die Suidwes-Kaaplandse XI, 7602 Matieland (Stellenbosch), South Africa. Wingerdbougebiede. M.Sc. Thesis, Stellenbosch University, Private Bag XI, 7602 . Africa h Sout , (Stellenbosch) d Matielan Van Zyl, P., 2000. Stellenbosch Offspring. In: Stellenbosch Appellations and Flagships. Supplement to Wine, June 2000. Marais, J., Hunter, J.J. & Haasbroek, P.D., 1999. Effect of canopy microclimate, . Afr . S . quality e win d an n compositio e grap c blan n Sauvigno n o n regio d an n seaso Visser, J.N.J., 1964. Analyses of rocks, minerals and ores. Handbook 5. . 19-30 , 20 . Vitic . Enol . J Geological Society of South Africa, Pretoria, South Africa.

2 200 , 2 . No , Vitic.. 23 . Enol ,Vol . J . Afr . S Factors Environmental Different of Effect gj

White, R.E., 1987. Introduction to the principles and practice of soil science. 2 Winkler, A.J., Cook, J.A., Kliewer, W.M. & Lider, L.A., 1974. General Edition, Blackwell Scientific Publications, Oxford, UK. nd Viticulture. University of California Press, Berkeley, California.

Williams, L.E., Dokoozlian, N.K. & Wample, R., 1994. In: Schaffer, B. & Wooldridge, J., 1988. The potassium supplying power of certain virgin upland XIg ,Ba e Privat , University h Stellenbosc , Thesis . M.Sc . Cape n Wester e th f o s soil l Vo , crops t frui f o y physiolog l environmenta r fo k Handboo . (eds) . P.C , Anderson I, Temperate crops. CRC Press, Boca Raton. pp. 85-133. 7602 Matieland (Stellenbosch), South Africa.

Wilson, J.E., 1998. Terroir. The role of geology, climate and culture in the mak- ing of French Wines. University of California Press, Berkeley, California.

2,200. No , 223 . Vol , Enol. J . .Vitic. Afr . S

fe'