Concord Fruit Thinning: Using Vine Biology and Mechanized Management to Address Market Demands in New York

Transcription

1 Concord Fruit Thinning: Using Vine Biology and Mechanized Management to Address Market Demands in New York

2

3 Energy In: Energy Out: Energy Balance:

4 Energy In: Energy Out: Vine Size Maximize sunlight interception per unit land area (and minimize internal canopy shading) Crop Size Match crop size with growing season potential, variety, location, market Energy Balance: Juice Soluble Solids (brix) % Crop 75% Crop 100% Crop Days After Veraison Crop Load Vegetative : Reproductive Leaf Area : Fruit Weight Yield : Pruning Weight

5 Canopy Characteristics: Vine size, leaf area, canopy fill Soil water and nutrient availability AND the capacity of the root system to acquire soil resources directly affects vine growth and productivity

6 Kimball and Shaulis, 1958 Quantity and Quality of Retained Nodes

7 Pool, Controlling Crop Solely By Non- Selective Machine Pruning (Hedging)

8 Pool,

9 Morris-Oldridge System Approach Lower 180 pruning Shoot Thinning Shoot Positioning Fruit Thinning

10 Vineyard Crop Load: Total Soluble Solids ( o brix) with leaf removal overcropped 120 Node Pruning with thinning Balanced Pruning at undercropped Y = 20 - (45/x) r = Leaf Area : Fruit (cm 2 /g) Kliewer and Dokoozlian, AJEV, 2005 Thompson Seedless - CA Bates, Wine East, 200 Concord, NY Exposed Leaf Area : Fresh Fruit Weight

11 Crop Load: Where vine size and crop size come together. Juice Soluble Solids ( o brix) Small Medium Large Yield (tons/acre) Leaf Area:Fruit (cm 2 /g)

12 Concord Crop Estimation and Adjustment in the Field

13 Lake Erie GDD Accumulation May 1, 2014 Lake Erie = 143 heat units Concord Predictions Bloom - /20 Veraison - 8/27 00 /1/ heat units Big jump should help 200 move up predicted 0 bloom date Day of Year Concord Bloom Date y = x R² = Lake GDD Accumulation on May 1

14 Date 9/10 8/27 8/13 7/30 7/1 7/2 /18 /4 5/21 5/7 4/23 4/9 Veraison Bloom Bud Break SEASON 3 Days in 40 Years Days in 40 Years 8 Days in 30 Years Year Concord phenology dates in the Lake Erie Belt collected from the Cornell research vineyard since 195.

15 Seasonal Variation

16 Concord Berry Curve Fresh Berry Weight (g) Fresh Berry Weight (g) cell division lag phase veraison ripening typical harvest dehydration Days After Bloom Year Mean Days After Bloom Fresh Berry Weight (g) Concord Berry Curve (Lake Erie) Year Mean Days After Bloom R-Square (berry wt and max berry wt) Days After Bloom

17 Dr. Terry Bates: Crop Estimation and Thinning Table: 7/1/2003 Time of Season 20DAB 25DAB 30DAB 40DAB 50DAB Veraison % of Final Berry Weight Pounds of Fruit Removed in 1/100th of an Acre Row Spacing determines length of 1/100th of an acre 10.0 feet row spacing = 43.5 feet = 1/100th of an acre 9.5 feet = 45.9 feet = 1/100th of an acre 9.0 feet = 48.4 feet = 1/100th of an acre 8.5 feet = 51.2 feet = 1/100th of an acre 8.0 feet = feet = 1/100th of an acre 7.5 feet = 58.1 feet = 1/100th of an acre Calculation 43, 50 square feet per acre Divide by row spacing and then divide by 100 to get 1/100th of an acre Example: A grower has 9 foot row spacing and clean picks 48.4 feet at 25 days after bloom. The fruit weighs 80 pounds and the grower estimates that the berries are between 35% and 40% of final berry weight. According to the table, the crop estimate is between 10.0 and 11.4 tons per acre. Disclaimer: This table gives the relationship between time of season and % final berry weight on an average year. Year to year variability in weather related berry growth adds error this table. Information on current year berry growth can be obtained from the Fredonia Vineyard Lab (or) it is strongly suggested that individual growers start collecting berry weight information from their own individual vineyard blocks.

18

19

20 Early Season NDVI Sampling Map Predicted Yield Actual Yield

21 Crop Estimation by Stratified Sampling Fresh Berry Weight (g) lag 49 phase 41 cell division veraison ripening typical harvest dehydration Days After Bloom

22 2014 Concord Crop Estimation Single Mean 3 Vigor Classes Continuous by NDVI Actual Yield Y Y Y Y Pred Yield NDVI Class Linear NDVI TA_Harvest Single Low Medium at 1.4 factor High X X X X Estimated Yield (tons/acre) Single Mean Estimated Yield (tons/acre) Low Medium High Vigor Class Predicted Yield (tons/acre) y = 25.51x R² = NDVI (20 Days After Bloom) Yield (tons/acre) y = 17.7x x x R² = NDVI at 20 Days After Bloom 1.4 (70% of final berry weight) 1.5 (5% of final berry weight) Tons (% from actual) (3.3%) (10.%) (0.1%) (7.2%) (2.5%) 1.9 (9.8%) 133.8

23 CMU s High-throughput Vineyard Imaging System Collect and convert raw image data into measurements of yield components Berry Count Berry Size Berry Quality 23

24 Y Y Linear NDVI TA_Harvest at 1.4 factor Non-destructive Cluster Counting Predicted of BarnBlockBerryCounts_KRIG X X Yield (tons/acre) y = 17.7x x x R² = NDVI at 20 Days After Bloom Canopy size and light interception are relatively high and uniform in this block; therefore, there is a poor NDVI to Yield relationship. Yield (tons/acre) y = x R² = Berry Counts

25 Shoots/Vine Spatial Crop Estimation: Progress and Hurdles Clusters/Shoot Berries/Cluster NDVI Stratified Sampling Yield (tons/acre) Yield (tons/acre) y = 17.7x x x R² = NDVI at 20 Days After Bloom Yield Monitor y = x R² = Berry Counts Berry Imaging Estimated Pattern Fresh Berry Weight Fresh Berry Weight (g) Days After Bloom Year Mean Actual Fresh Berry Curve Sampling? Modeling? Imaging?

26

27 Methods of Mechanized Crop Control in New York Concord Viney Potential for Variable Rate Management and Improved Production Efficiency Dormant Mechanical Cane Pruning Earlyseason Shoot Thinning Mid-Season Fruit Thinning

28 Management Classification Map: What does it mean? Spatial Map built from multiple data layers. (i.e. Soil, Canopy, Crop) Each color area is statistically different based on the layers in the analysis But are they VITICULTURALLY different? Each color area has attributes which can be further investigated. (i.e. Red area has high soil EC, low vine growth, and low yield) Stratified manual samples can be taken to ground truth spatial data The GROWER can then add his/her decision process to MANAGEMENT

29 Turning the management decision into ACTION No thinning Some thinning Heavy thinning Data Driven Decision Maker

30 Making the Implement Variable Rate: Response Curve 2015: The effect of paddle speed on the level of shoot thinning and yield reduction in Concord was tested. Example: 40 RPM paddle speed (2 MPH ground speed) led to 15-20% shoot and yield reduction.

31 How do we go from spatial DECISION to variable rate MANAGEMENT ON-THE-FLY? Some thinning 40 RPMs No thinning 0 RPMs Heavy thinning 50 RPMs

32 Currently, a research technician processes the spatial data into a management map. (The SCRI project aims to make this semi-automated) GPS Synched The management shapefile is imported into AgLeader SMS software and target flow rates are assigned to the zones. The PWM valve controls the flow of hydraulic fluid to the slapper motor. Trimble and Raven have similar products The AgLeader 1200 In-Command unit and Direct Command Liquid Module interface to regulate a hydraulic PWM valve. JUST DRIVE!!!

33

34 Concord Fruit Thinning in NY Prune for high crop potential Procedure for Crop estimate at 30 DAB Match crop size with season potential Use spatial vineyard pattern to direct samples and make weighted estimate Developing Variable Rate Crop Adjustment

35