Big Data and the Productivity Challenge for Wine Grapes Nick Dokoozlian Agricultural Outlook Forum February 20160 Big Data and the Productivity Challenge for Wine Grapes Outline • Current production challenges • Lessons learned from annual crops • How will we utilize Big Data to meet our challenges? – Measure – Model – Manage • Summary 1 The Productivity Challenge for Wine Grapes • Suitable land, labor and water for agriculture are becoming more scarce and expensive • Need to increase grape supply without increasing production area and environmental impact • Must increase both yield and quality simultaneously • Similar challenges are faced by nearly all agricultural commodities worldwide 2 How are annual crops addressing these challenges? Dramatic increases in the productivity of agronomic crops have been achieved during the past century via: • Genetics – traditional breeding and genomics • Improved agronomic practices and resource management • Application of remote sensing and other technologies 3 How are perennial crops different in their approach? Progress has been much slower in wine grapes and other perennial crops: • Critical mass – limited acres = limited attention despite farm-gate value • Genetics – research, breeding cycle and market tradition • Production cycle and innovation adoption • Yield – quality relationships 4 IntegratedObjective systems of Today’s are required Meeting for improving productivity and quality Germplasm Systems Precision Improvement Biology Agriculture • Clonal • Elucidate the • Characterize the selection regulation of parameters • Cultivar and key yield and regulating vine rootstock fruit quality productivity and improvement pathways quality via traditional • Functional • Model key breeding genomics – relationships • Pest/disease linking genes • Variable rate resistance to key traits management 5 IntegratedObjective systems of Today’s are required Meeting for improving productivity and quality Germplasm Systems Precision Improvement Biology Agriculture • Clonal • Elucidate the • Characterize the selection regulation of parameters • Cultivar and key yield and regulating vine rootstock fruit quality productivity and improvement pathways quality via traditional • Functional • Model key breeding genomics – relationships • Pest/disease linking genes • Variable rate resistance to key traits management 6 The Future of Grape Growing MEASURE Measures Automated sensors used to measuring intra- construct field variability – geospatial crop load, canopy maps of key size, irrigation relationships requirements MANAGE MODEL Information used to spatially alter cultural practices 7 Historical sensors are site specific Historical soil measures X X X 9 Sensors provide high density soil information 10 High Resolution Maps - EM Sensor Characterizing Yield Variability 12 13 Why does variability matter? Cabernet Sauvignon 9.2 tons/acre 22.7 tons/ha 14 Block acreage (%) 10 15 20 0 5 0‐1 Why doesvariabilitymatter? Colony 1‐2 2‐3 Mean yield= 9.2 tonsperacre 3‐4 2A 4‐5 Cabernet 5‐6 Yield 6‐7 7‐8 Sauvignon (tons/acre) 8‐9 9‐10 10‐11 / 11‐12 32.1 12‐13 13‐14 acres 14‐15 15‐16 16‐17 17‐18 15 18‐19 19‐20 Block acreage (%) 10 15 20 0 5 = 30%yield increase Block improvement 0‐1 What isthesizeofprize? Why doesvariabilitymatter? producing below Colony mean blockyield Why doesvariabilitymatter? 1‐2 vines 40% of opportunity 2‐3 Mean yield= 9.2 tonsperacre 3‐4 2A 4‐5 Cabernet 5‐6 Yield 6‐7 7‐8 Sauvignon (tons/acre) 8‐9 9‐10 10‐11 / 11‐12 32.1 12‐13 13‐14 acres 14‐15 15‐16 16‐17 17‐18 16 18‐19 19‐20 Block acreage (%) 10 15 20 0 5 = 30%yield increase Block improvement 0‐1 What isthesizeofprize? Why doesvariabilitymatter? producing below Colony mean blockyield 1‐2 vines 40% of opportunity 2‐3 3‐4 2A Mean yieldperacre=9.2tons 4‐5 Cabernet 5‐6 Yield 6‐7 7‐8 Sauvignon (tons/acre) 8‐9 9‐10 • • • 10Establishment -‐ $35,000 11 Capital avoidance/acre $100/acre Estimated cost= $900/acre revenue = Annual increase in additional acreage without planting increase 30% yield / 11‐12 32.1 12‐13Land - $50,000 13‐14 acres 14‐15 15‐16 16‐17 17‐18 17 18‐19 19‐20 What isthesizeofprizefor BigData? Block acreage (%) 10 15 20 0 5 0‐1 Colony 1‐2 2‐3 Mean yield= 9.2 tonsperacre 3‐4 2A 4‐5 Cabernet 5‐6 Yield 6‐7 7‐8 Sauvignon (tons/acre) 8‐9 9‐10 10‐11 below districtaverage / 11‐12 32.1 12‐13 produce quality 20% of vines 20% of 13‐14 acres 14‐15 15‐16 16‐17 17‐18 18 18‐19 19‐20 What isthesizeofprizefor BigData? Additional revenue basedon Additional Additional cost=$500/acre Block acreage (%) = $2,200/acre at farmgate fruit qualityimprovement 10 15 20 0 5 0‐1 Colony 1‐2 2‐3 Mean yield= 9.2 tonsperacre 3‐4 2A 4‐5 Cabernet 5‐6 Yield 6‐7 7‐8 Sauvignon (tons/acre) 8‐9 9‐10 10‐11 below districtaverage / 11‐12 32.1 12‐13 produce quality 20% of vines 20% of 13‐14 acres 14‐15 15‐16 16‐17 17‐18 19 18‐19 19‐20 Integrated systems - analytics - Proximal Remote Proximal Proximal Sensor Sensor Sensor Sensor Plant Vegetation Fruit available Yield Index Quality water in soil (NDVI) 20 Modeling Yield and Fruit Quality Data with Soil Parameters Significant Correlations with Fruit Yield Parameter Correlation (r2) Subsurface K+ 0.903 Soil rooting depth 0.774 Subsurface pH – 0.805 Subsurface P – 0.805 Subsurface organic matter – 0.882 Subsurface K/Mg ratio – 0.890 Significant Correlations with Fruit Quality Parameter Correlation (r2) Soil rooting depth – 0.673 Surface CA – 0.506 Subsurface CA / Mg ratio – 0.510 Surface CEC – 0.554 21 Variable rate management 22 Relative importance of soil parameters to block yieldwhc variability 100% 5% Anions 11% Soil compaction 80% variance 13% Textural class 60% 16% pH explained to 40% 0.46 0.46 0.46 0.46 0.46 0.46 0.46 0.46 46% 20% 46% Water holding capacity Contribution Contribution to total variance (%) 0% 23 Vine water use is variable based on canopy size Sap Flow - Colony 2A - 2012 High Vigor 120 Low Vigor 28 gallons per 100 vine per week O/hour) 2 80 60 17 gallons per Sap Flow (g H (g Flow Sap 40 vine per week Row direction 20 0 00:00 04:00 08:00 12:00 16:00 20:00 00:00 Time 24 Variable Rate Drip Irrigation Changes in canopy vigor (NDVI) Each square = 30m x 30m LANDSAT Pixel Before After variable rate variable rate irrigation irrigation Pixel level management based on canopy size 25 Impact of Precision Irrigation 2012 Block Yield 2015 Block Yield 8.1 tons/ac 10.2 tons/ac Yield improved 20%; Water use efficiency improved 30% Summary and future challenges • Sensor technology has advanced real-time, high density data collection – Geospatial analytics for characterizing vineyard variables – environment, growth, yield and quality • Our ability to measure exceeds our ability to interpret – Understand what is important and actionable • Large gaps exist in variable rate application technologies for geospatial management – Example: Variable rate drip irrigation • Research collaboration (USDA - ARS) and industry partnership are essential to advance 27.