Accurate crop forecast needed
Technology that will help apple growers assess the fruit on their trees is being developed.
New technology might help growers avoid sending low-value fruit to the packing house.
An accurate crop estimate can have a major impact on the profitability of the Washington apple industry, says Bruce Grim, manager of the Washington Apple Growers Marketing Association. "It's hugely important." That's because growers tend to make their picking decisions in the orchard based on what they expect the state's crop size to be and how their fruit will fit the profile of what can be picked, packed, and shipped profitably, he said.
Currently, Washington compiles its crop estimate in two ways. Members of the Wenatchee Valley Traffic Association compile an estimate for the northern district based on what volume each shipper thinks they will handle. The Yakima Valley Growers-Shippers Association asks its members to estimate what the entire southern district's crop will be. The two associations then combine their data to produce a statewide crop estimate.
Keith Mathews, executive director of the Yakima Valley Growers-Shippers Association, believes the two traffic associations adopted different estimating methods because of the different structure of the industry in their districts. Whereas shippers in the Yakima district have always been primarily independent companies, fruit cooperatives used to predominate in north central Washington. The grower-owners of cooperatives tended to stay with the same packer over the years, making it relatively easy for the cooperatives to predict the volume of fruit they would handle from year to year, whereas in the Yakima area, growers were more likely to move around between independent fruit packers, making a districtwide estimate more practical.
Since 1980, the industry's August 1 estimate has, on average, been 3 percent lower than the volume actually shipped. But in 2008, the August 1 estimate was 99.6 million boxes, which will be about 10 percent lower than the actual volume. The greatest discrepancy was in 2004, when the actual crop came in at 105 million boxes, 19 percent above the initial estimate of 88 million. During the last five years, the estimate has only once been within 5 percent of the actual crop.
Mathews said crop estimating became trickier when the industry began growing varieties other than Red and Golden Delicious. The harvest timeframe used to be much shorter, and there was less potential for changes in the crop after the estimate was made. Nowadays, the industry is growing late-maturing varieties such as Pink Lady, Braeburn, and Fuji, which stretch harvest into late October and November. While the estimated volume has a significant impact on the marketing of the crop, Mathews said he did not know what level of accuracy was attainable with the current system.
Four years ago, the Washington Growers Clearing House Association and the Washington wine industry received a grant from the U.S. Department of Agriculture's Risk Management Agency to explore ways to generate data for decision making in tree fruits and wine grapes. As part of that project, the Clearing House contracted with Eileen Perry at Washington State University's Center for Precision Agriculture to work with remote sensing and satellite imagery to generate timely and accurate maps of acreages. Dan Kelly, assistant manager at the Clearing House, said the ultimate intent was to be able to distinguish the trees by variety and develop crop estimates. Estimates that could be compiled earlier in the season would provide marketers with better information when they're setting up promotions and negotiating prices in advance of harvest. However, funding for the project ended after two years.
Kelly still believes it's important for the industry to have accurate crop estimates and thinks that a robotic scout being developed by Vision Robotics Corporation of San Diego, California, could be the most realistic way to achieve this in the near future. The scout should be able to go through the orchard before harvest to count fruit and record its size. It likely would be used in sample blocks to extrapolate the size of the crop.
The system has been in development since 2006, funded in part by the Washington Tree Fruit Research Commission. This season, the company's scientists have been focusing on ensuring that the scout can capture images fast enough to move down the row at a speed of two miles per hour. They've also been refining the scout's lighting system so that it can work in a wider range of light conditions and can more easily detect the fruit, said Dr. Derek Morikawa, chief executive officer of Vision Robotics.
The scout must be able to sense fruit both in strong sunlight and in deep shadow in the interior of the tree. The vision system has also been improved. Cameras are now oriented upwards to reduce the likelihood that the leaves will obscure the fruit.
The scout will need to be towed through the orchard, and Morikawa planned to bring the latest version to Washington this summer to be tested in conjunction with an autonomous vehicle being developed by Carnegie Mellon University in Pennsylvania. Both the scout and the Carnegie Mellon projects are now supported with funding from a federal Specialty Crop Research Initiative grant for development of comprehensive automation for specialty crops.
Morikawa expects that the company will have a commercial prototype ready by 2010 and that it will be ready to commercialize in 2011.
Grim said the scout could be of great help to the industry. "If we can utilize that technology to go out and count apples and determine the size before we pick, and come up with a more accurate means of estimating the crop, then we'll make decisions in the orchard that will impact the volume that we take to the warehouse, and that will directly impact our bottom line."