Yield Estimation Funded by the GWRDC Regional Grassroots Solutions program

ACCURACY OF YIELD ESTIMATION PRIMARY INFLUENCES ON YIELD Yield estimation is a very important component of successful wine • Environmental conditions including mean grape production but is difficult to perfect consistently. Generally, temperatures below 15°C, greater than 35°C yield estimates vary by 33% from actual tonnes harvested1. or wet, cold conditions, Inaccurate yield estimation can have dire consequences for: • management practices, • Vineyard logistics, including harvesting and transport, • Pests and disease, and • Growers with fruit in excess of contracted tonnes, and • Vine genetics2. • Allocation of winery staff, crushers, fermenters and wine Unfavourable conditions, especially around storage. flowering, can result in bunch shatter or hen and Poor yield estimation can also have longer term impacts on chicken3 (Figure 2). budget, supply and marketing for and wineries.

IDENTIFYING VARIABILITY “live green ovaries” Large variation in vineyard blocks makes yield estimation (bunch shatter) especially difficult. Estimating yield within similar ‘production (do not develop into berries) zones’ rather than homogenously across a block can capture variability within the block better. ‘Production zones’ can be identified using detailed mapping of : “hen berries” (seeded, full development) • Plant cell density (Figure 1), • Yield maps, and • Topography maps. “chicken berries” (small, seedless, still ripen) “Yield varies from season to season, Photograph: C. Collins, Adel. Uni. between regions, across vineyards, between blocks, within individual vines and between berries on an Figure 2: A bunch displaying mixed fruit set individual bunch” 2

FRUIT SET CHARACTERISTICS BETWEEN VARIETIES Recent research3 demonstrated reduced flower number rather than poor fruit set was a significant contributor to lower yield in some varieties. Understanding the differences that exist between varieties can assist yield estimates by influencing the berry number and bunch weight used in calculations.

Table 1: Variation in flowering and fruit set between 10 varieties grown in Adelaide Hills, Adelaide Plains, McLaren Vale and Padthaway. , • Lower bunch number with higher flower number , • Higher bunch weight, greater number of berries Zinfandel. • Less bunch shatter

Cabernet Sauvignon, • Higher number of flowers and live green ovaries . • More hen and chicken • Less fruit set

Chardonnay, • Lower number of flowers, seeded berries and live , Shiraz, green ovaries • Lower total berry number Figure 1: Plant cell density map Tempranillo, • Less hen and chicken showing vigour variation across a block4 Sauvignon Blanc. Wine Grape Yield Estimation Funded by the GWRDC Regional Grassroots Solutions program

HOW TO ESTIMATE YIELD There are several key measurements and periods for yield estimation during the growing season (Table 2). Prior to any yield estimation, ascertain the exact area of the block by counting the vines. Any missing vines should be excluded. “Growers have a good feel for 1. BUNCH COUNTING & CALCULATIONS5 average production but consistently • Count six weeks after bud burst. overestimate production in low- • Count all bunches on 30 vines per block (In blocks larger cropping years and underestimate in than 4 hectares, count an additional 10 vines/ha). 1 • Calculation assumptions include: high-cropping years” ⇒ Bunch weight (g) - refer to relevant historical data, Table 2: Yield estimation timing and measurement5 ⇒ Bunch gain/loss factor - compensates for any variation in bunch number between bunch counting and , and Dormancy Bud dissection ⇒ Harvest efficiency factor (HEF) - see opposite. 6 weeks post budburst Yield (t/ha) = (onward) (1) Bunch counts vines/ha x bunches/vine x bunch weight (g) x bunch gain/loss x HEF 1,000,000 Flower counts Flowering 2. BERRY COUNTING & CALCULATIONS5 4 weeks post fruit set (2) Berry counts • Sample approximately one month after flowering. (onward) • Remove all bunches from 3 - 5 vines throughout the (3) Bunch weighing block, sampling at least 60 bunches. • Remove and count ‘normal’ berries from each bunch Pre-harvest (3) Bunch weighing (do not include chicken or live green ovaries). Post-harvest Analysis and • Calculation assumptions include: evaluation ⇒ Berry loss factor - compensates for any berry loss between berry counting time and harvest, HARVEST EFFICIENCY FACTOR (HEF)5 ⇒ Berry weight (g) - refer to relevant historical data, and HEF (used in calculations opposite) reflects the ⇒ Harvest efficiency factor (HEF) - see opposite.

percentage of in the vineyard that are actually Bunch weight (g) = harvested. This excludes any dropped, missed and berries/bunch x berry loss factor x berry weight (g)

unharvested fruit. Yield (t/ha) = vines/ha x bunches/vine x bunch weight (g) x HEF ⇒ Meticulous hand harvesting with small transport 1,000,000 losses. HEF = 1.00 (100% harvested), 5 ⇒ Hand harvesting with small transport losses. 3. BUNCH WEIGHING & CALCULATIONS HEF = 0.95 (95% harvested), • Sample at onset of veraison (1%) and again pre-harvest. • ⇒ Very efficient machine harvesting with small Remove all bunches from 5 - 10 vines throughout the transport losses. HEF = 0.90 (90% harvested), block, sampling at least 200 bunches, and ⇒ Inefficient machine harvesting with small transport • Weigh and calculate average bunch weight. • losses. HEF = 0.85 (85% harvested). Onset of veraison calculation assumptions include: ⇒ Bunch weight gain factor (BWGF) - 1.7 - 2.0 depending FURTHER INFORMATION on cultivar, season and site (refer reference 5 ) ‘How to forecast wine grape deliveries using Grape Forecaster ⇒ Harvest efficiency factor (HEF) - see opposite.

Excel workbook version 7’ is an excellent resource describing Yield (t/ha) = vines/ha x bunches/vine x bunch weight (g) x BWGF x HEF the methodology, timing and assumptions required for accurate 1,000,000 yield estimates using the ‘Forecaster Version 7’ model. This • Pre-harvest calculation assumptions include: excel spreadsheet based model was developed by DPI Victoria ⇒ Harvest efficiency factor (HEF) - see opposite. and GWRDC and is available on the GWRDC website located Yield (t/ha) = vines/ha x bunches/vine x bunch weight (g) x HEF at: www.gwrdc.com.au/site/page.cfm?u=98 1,000,000 REFERENCES 1 Dunn, G.M. and Martin, S.R. (2003) The current status of crop forecasting in industry. 4Proffitt, T. and Winter, E. (2008) Adoption of Precision on the rise. The Australian and New In: Bell, S.M. de Garis, K.A., Dundon, C.G. Hamilton, R.P., Partridge, S.J. and Wall, G.S. (Eds.). ASVO Zealand Grapegrower and Winemaker. 539, 36-38. proceedings, Tanunda. Aust. Soc. Vitic Oenol., Adelaide, SA. 5Martin, S., Dunstone, R., Dunn, G. (2003) How to forecast wine grape deliveries using Grape Forecaster 2Dunn, G.M. and Martin, S.R. (1998) Optimising vineyard sampling to estimate yield components. The Australian Grapegrower and Winemaker. 414a, 102-107. Produced by SA Central (Adelaide Hills, McLaren Vale, Langhorne 3Dry, P.R., Longbottom, M.L., McLoughlin, S., Johnson, T.E., Collins, C. (2010) Classification of Creek, Kangaroo Island, Currency Creek, South Fleurieu). reproductive performance of ten winegrape varieties. Australian Journal of Grape and Wine Research. Prepared by Scholefield Robinson Horticultural Services. 16, 47-55. August 2010©