WSU Viticulture and Enology

Research and Extension

Maximize Your Trellis Investment: Continuously Monitor Your Vines and Estimate Yields


Major funding provided by the USDA-Agriculture Research Service and a grant from the Washington Wine Industry Foundation in Cooperation with USDA-Risk Management Agency. Other funding provided by the Viticulture Consortium-West, a USDA-CSREES special grants program.


Electronic devices and wireless communications are part of today’s farm, and used to their potential, they can provide a grower with rapidly updated information on the status of the vineyard. To maintain competitiveness in the global market, the grape and wine industry must maximize the utility of all its inputs and farming tools, including the trellis system.

We have devised a way to give a grower a nearly continuous stream of information related to vine growth and crop development by measuring the tension in the trellis wire and plotting its increase during the season. Surf this page and its links for a description of our system, explanatory diagrams, and plots of actual trellis wire tension across the season in 12 different vineyards around the Yakima Valley of Washington State.


Imagine booting up your computer and watching your vines and fruit grow week by week. Rather than “guesstimating” the progress of the crop, your computer screen would display this season’s progressing growth curve and the complete growth curves from key previous years.

Wineries and fruit processors want accurate and early crop estimates to manage harvest scheduling, processing, storage, and marketing of the final product.

Historically, crop yield in grapes has been estimated by collecting and weighing clusters or specified numbers of berries, a labor-intensive and thus expensive operation. These estimates are snapshots—fixed at one time of a dynamic process.

What if this practice could be automated, and what if a grower could use the same equipment for multiple purposes, thereby increasing the return on investment?

In 2001 our research team began to investigate an innovative method of monitoring a trellised crop and automating yield estimation. In 2005, we were awarded a U.S. patent for the apparatus and method of doing so (see Tarara et al. 2005, U.S. Patent no. 6,854,337). Our approach exploits the fact that trellised crops, of which grapes are a prime example, are physically supported by a system of vertical posts and horizontal wires. A row of grapevines can be thought of as a “load” that hangs from the support wire, in our case a cordon wire. As the load increases so does the tension in the wire. Measuring the tension in the wire continuously allows us to monitor the increasing load throughout the growing season, and from the increase in tension, monitor shoot growth and overall crop development.

How it Works

To understand automated yield estimation technology, explore these three components.

The Trellis Tension Measurement System

Figure 1. Schematic diagram depicting components of the continuous monitoring and yield estimation system in a trellised crop. A single cordon wire is shown for simplicity. Yellow arrows depict forces or loads on the trellis wire that influence wire tension

Figure 1. Schematic diagram depicting components of the continuous monitoring and yield estimation system in a trellised crop. A single cordon wire is shown for simplicity. Yellow arrows depict forces or loads on the trellis wire that influence wire tension

We insert an electronic device, a load cell, into the trellis wire to measure tension. Many forces affect tension in the wire including vine growth, temperature, and wind. Our goal is to develop a predictive model and computer program that account for these variables to improve yield estimates.

The load cell is connected to a data logger or computer that records increases in tension during the season. Tension can be used to follow crop development and to make estimates of final yield. The physical concept is straightforward but there are many environmental and biological factors to be considered for successful application.

Brick hang

For example:

  • daily temperature cycles
  • wind
  • management practices
  • variation in support among posts and vine trunks
  • variation in a biological system from year-to-year

After four seasons of working with our system, we have learned how to account for most of these factors to obtain more accurate yield estimates.

To improve our ability to estimate yield from wire tension it is useful at any given time to know how much of the mass hanging on the wire is due to fruit or vegetative growth. Before budbreak, a standard weight is hung on the cordon wire. The weight is correlated with the change in tension that we measure.

During the growing season, samples of shoots and fruit are collected and weighed so that we can develop a system that will automatically account for the weight of the shoots upon the trellis wire.

Details of the method can be found in Transactions of the American Society for Agricultural Engineering (2004, volume 47, pages 647-657; or by request).

The Bridge

The bridge

We are continually testing and improving our system so that ultimately, there will be an affordable and viable solution for grape growers. Our latest research focuses on solutions for implementing the system in commercial vineyards:

  • accounting for season-to-season variability
  • making the system easy to maintain
  • applying the system to different trellis types

Initially we used load cells free to sense tension along entire vineyard rows, dubbed the “Open” trellis monitor. For commercial vineyards, we designed a robust steel brace to isolate and immobilize a section of cordon wire between two trellis posts, dubbed the “Bridge”. The section of trellis wire under the bridge is the sensing region for a load cell. We are comparing the “bridge” trellis monitor to the “open” trellis monitor in research plots and commercial vineyards.

The “bridge” provides a sensing region short enough (ca. 20 to 25 feet) to:

  • set initial wire tensions uniformly each year
  • permit periodic calibration with known weights
  • buffer effects of disturbance to the trellis wire by some vineyard operations.

Our goals:

  • Make the system easier for growers to manage
  • Allow straightforward comparisons among bridges and between years.

The trellis tension monitor was developed on experimental blocks of Concord juice grapes at WSU-IAREC (Washington State University-Irrigated Agriculture Research & Extension Center) in Prosser, Washington. Later we installed both “open” and “bridge” trellis monitors in 10 commercial Concord grape vineyards.

Experiments in the commercial vineyards are designed to:

  • allow us to compare juice processors’ yield estimates via hand sampling to ours via the trellis tension monitor
  • test the practicality of the “bridge” version of the trellis monitor

Processing Data

Vinifera datalogger

We would like to develop a predictive model and computer program to account for variables influencing tension in the trellis wire so that the trellis tension monitor can become a viable tool for the grape industry to monitor vineyards and to estimate yields.

Data Collection:

In our experiments we use research-grade dataloggers to scan the signals from the load cells every 10 seconds. The datalogger computes and stores an average value every 15 minutes. Because 90 of the 10-second readings are combined for the average, environmental influences like wind gusts do not affect the final value. However, other effects must be accounted for after we download the 15-minute values. We call this effort “post-processing.”

Data Post-Processing Sequence:

  • Account for the effect of wire temperature on wire tension.
    – steel wire stretches and contracts as temperature rises and falls every day
  • Apply calibration factor
    – each spring we hang known weight from the wire to calculate the change in wire tension for every unit of weight on the wire. This pre-budbreak “baseline” tension is subtracted from the current value to determine the change in trellis tension created by the vine this season.

Yield Estimation Sequence:

  • Compare the “post-processed” or final, corrected values of wire tension (or change in wire tension during the season) with those from previous years.
  • Adjust the ratio between today’s tension value and the previous years’ tension value by the previous years’ actual yields to estimate this year’s harvest weight (meaning this year’s yield estimate).

Advantage of the Trellis-Tension Monitor for estimating yield:

Because the tension data are nearly continuous, the grower, processor or winery can make frequent adjustments to the crop estimate as the season progresses. This flexibility would have been beneficial in 2005, particulary in California, where the berries “sized up” (enlarged) rapidly during the ripening period, resulting in a serious underestimation of the crop that was ultimately delivered to the processors’ and wineries’ doors.

For further information or suggestions please contact us by email, phone or mail.


Dr. Julie Tarara, Research Horticulturist, USDA-ARS
Email: Phone: 509-786-9392

24106 North Bunn Road
Prosser, Washington 99350

Research and Extension, 2710 University Drive, Richland, WA 99354-7224, 509-372-7224, Contact Us
© 2017 Washington State University | Accessibility | Policies | Copyright | Log in