Crop Registration

Manual Crop Registration Today

The greenhouse tomato industry continues to expand, with experienced growers overseeing more facilities and plants, prompting questions about how growers manage this expansion and ensure industry success. Plants are pivotal to greenhouse operations, driving labor, forecasting, and livelihoods while sustaining food systems. Despite advancements in technology and selective breeding, challenges persist in prioritizing plant-centric approaches and achieving crop uniformity. To address these challenges, we've collaborated with growers globally, revealing that traditional crop registration methods fall short in quantifying overall plant performance. As we navigate tighter margins and the need for efficiency, empowering growers with better insights and control becomes paramount in leveraging the knowledge of master growers and sustaining industry progress.

The way we grow tomatoes in greenhouses is constantly improving, yet the methods for understanding how plants respond to their environment haven't kept up. This makes it difficult to consistently forecast production, optimize plant balance through climate and irrigation, allocate labor efficiently, and meet operational targets. Traditionally, growers have relied on a combination of quantitative and qualitative methods to assess crop performance and achieve production goals aimed at maximizing yield while minimizing costs.

Every industry relies on measuring its production output. In CEA, our "product" is “manufactured” or grown by a population of living organisms. Despite the fact that we refer to our greenhouses as “controlled,” our products develop under dynamic conditions and we are learning through research that the crop is not as homogenous as we expect. Think of plants as gears in a machine. When some gears slow down, it's a sign the whole system needs attention. How much production must be compromised before we can see which gears need attention? This inherent unpredictability makes measuring more than just a small registration field crucial to fully understanding the crop, yet we as an industry settle. Accurately understanding the crop creates a new norm allowing growers to:

• Optimize growing conditions: Precise data on plant development and crop response across the far reaches of a greenhouse facility allows for adjustments to irrigation, fertilization, climate control, and labor allocation, leading to improved yields and resource efficiency.
• Identify problems early: Early detection of issues like nutrient deficiencies, labor quality, or pest and diseases (like ToBRFV) allows for timely intervention, minimizing losses and ensuring better crop quality.
• Gain insights to drive decisions and increase sell-thru: Data and analysis can help growers anticipate trends and adjust their production strategies or forecasts accordingly. The friction between growers and sales is palpable in today's industry. Reduce this pain by helping match supply with demand.

For decades, crop registration has supported growers in measuring plant development and production and gauging crop response to climate. However, as greenhouses expand and time for crop walks to verify the data becomes harder, this method is demonstrating gaps. With profit margins narrowing, growers seek more comprehensive crop insights to inform their decisions, driving the need for innovative approaches to crop monitoring

‍

Moving Beyond The Status Quo: A Critique of the Crop Registration Model

Exploring the practical implementation of manual crop registration with many growers has consistently revealed a discrepancy between its real-world application, and the intended purpose as a traditional resource. In fact, we’ve documented over a dozen different approaches all attempting to achieve the same goal – to find a true understanding of plant balance to inform many decisions.

Plant balance is derived from Crop Registration which is a manual practice of first measuring and then recording the vegetative and generative characteristics of a very small sample of plants. While intended to be a scientific method, the data shows it suffers from limitations in accuracy and confidence regarding its representation of the entire population. Rarely do samples account for the areas of the greenhouse that challenge growers the most such as gables or other microclimates where production losses are known to occur.

Typically, registration fields are positioned centrally within rows to avoid these problematic areas, yet the resulting data often fails to incorporate them. Looking at the data, the greenhouse map below illustrates the areas where crop registration fields would not typically be set up in order to avoid influences like gables and post rows throughout the duration of the crop; this accounts for 52% of the greenhouse area. Without considering the time it takes to measure a plant by hand, it is difficult to understand why we would avoid more than half of the greenhouse area for sampling. Then subsequently make assumptions about the performance and status of the crops in those areas.

After examining the crop registration processes among many medium to large-scale growing operations indicates 2 major issues with traditional practices - sampling and methodology. While many ag-tech companies have advocated for the importance of this quantitative process and attempted to support improvements in methodology in the form of SOP development and digital recordkeeping, they only address half the issue due to the statistical significance problems with sampling. Doing this ignores the possibilities beyond traditional manual data collection.

‍

Methodology in Crop Data Collection

After reviewing the methods used by multiple growers, we've identified shared concerns arising from the inherent inconsistency in the manual and human aspects of data collection. Examining each stage of the process, from measurement and recording to transcription, analysis, and reporting, we've pinpointed the following flaws:

• Labor intensity of measuring
• Time spent by highly skilled workers usually technicians and growers
• Human bias for anticipated or desired outcomes
• Human error potential in each step of the process
• Plant treatment differs from neighboring plants - crop registration plants get the best crop care and excess handling compared to the other plants
• Lack of consistency over time and among teammates
• Data management, usage, and interpretation are lacking

“When we rely on people to take this data it is not accurate and includes bias - it's hard to trust.”- Felix Tarrats, Horticultural Consultant

Given these shortcomings, growers find it challenging to rely on the data and always make good decisions. Ironically, while crop registration is intended to reinforce instinct gained from experience, growers not only lack confidence in the methodology but also lack statistical confidence in their sampling.

‍

Challenges in Crop Sampling: Unveiling the Foundational Flaws

Getting to the root of the issue, a sample is defined by the Oxford dictionary as “a small part or quantity intended to show what the whole is like”, and many different acceptable methods for sampling have been defined. However, because of the time-intensive nature of measuring only a handful of plants (a registration field), the ability to measure a sample of statistical significance has always been out of reach for growers. Therefore, sampling methods are driven by convenience and cost, not science and the data lacks confidence which is reflected in the errors throughout the crop cycle. The fewer the errors, the better the yield per crop.

The rationale? The assumption that all plants are genetically identical underestimates the potential that environmental factors have on variable crop response or phenology among plants. While many growers say they don't need to see all of their plants to understand their crop, most will agree that they wish they could see more and the reason for this is apparent in the data.

Manual data collection, whether on paper or via an app, remains inefficient. Growers typically measure only 5-10 plants per population, consuming 10-20% of their weekly time. This limited sampling compromises confidence in population representation and increases the pressure on growers to accommodate more crop walks in their busy schedules. As a result, many populations, including trials, lack adequate crop registration, hindering growers' ability to assess crop performance and balance comprehensively, and objectively select new varieties.

“What you cannot measure, you cannot improve.” - Greenhouse Grower

Relying on these traditional methods has been important for weekly crop management and planning. They're also crucial for evaluating trial varieties, which affects long-term strategy. However, fully understanding and observing crops through crop registration requires better coverage.

The variability within a population of plants indicates the need for greater sample sizes and well-distributed coverage. A study by the Horticultural Team at IUNU in some commercial greenhouses, looking at an entire row side of plants, shows this variability. It also highlights how small crop registration fields aren't sensitive enough to spot trends in crops.

‍

‍

Understanding Plant Variability: Insights from Crop Registration, A Case Study

There are several ways to measure and assess plant balance effectively. The VeGe calculator, developed by Godfrey Dol, is one of the most widely recognized methods that involves using crop registration metrics to assign a score to a population. Horticulturists at IUNU used this scoring method on a more granular level to first measure and then assign individual scores to each plant on one row side. The scores were charted to illustrate variability among each plant in sequence starting at the walkway. The analysis revealed significant variability among the plants in just one side of a row (195 - 209 plants).

According to Dol's philosophy, a score of 100 generally indicates balanced growth for most tomato varieties; scores above 100 suggest more vegetative growth and those below 100 indicate generative growth. Drawing from their experience with a particular variety of tomatoes (TOV), the growers opted to adjust the threshold, considering a score of 120 as balanced for this crop.

Using this adjusted threshold, the same half row of plants underwent evaluation twice, thirteen weeks apart and at different growth stages—first during the vegetative phase in week 36 (209 plants) and then during the generative phase in week 49 (195 plants). The overall assessment of the crop's status initially relied on LUNA AI, confirmed by grower observations throughout the crop cycle.

The graphs below depict the lack of uniformity observed in both instances of vegetative and generative growth across all plants from the walkway to the gable, based on their individual scores. Horizontal lines indicate the overall average score, alongside the original and updated thresholds.

In the same sample where the study occurred, there was a small area where the grower checked 5 plants every week as their crop registration field (≈ 2.5% of the half row sample). We compared the average VeGe score of these 5 plants to the average score of all the plants in the half row sample.

The average VeGe score of the whole row and the crop registration consistently differed by nearly 10%. In the last check, the half row's average score was below the grower's balance threshold, meaning the plants were growing more generatively. Conversely, the average score of the crop registration plants remained above the threshold, still indicating a vegetative status. This is an indication that a data-driven comprehensive approach will show a more accurate understanding of the plant population.

Within one row side we surfaced challenges inherent with small samples; most importantly it illustrates a lack of sensitivity when the crop is changing course. Growers often say they reflect on the data as a way to simply visualize crop trends, but these small samples don't show the real picture. If this grower had only looked at the crop registration field, they might have made the wrong decisions, they could have been steering the crop inappropriately.

‍

‍

‍

‍

The Impact of Crop Steering – Avoid Taking a Wrong Turn

Crop steering is a method of directing the crop towards desired growth patterns and production outcomes using key inputs such as:

1. Crop Care Labor (e.g., de-leafing, pruning, clipping/twisting, lowering, and harvesting)
2. Utilities & Environmental Controls (e.g., light, temperature, CO2, humidity)
3. Fertilizer & Irrigation (e.g., nutrients, saturation and timing)

Understanding the current status of their crops is crucial for growers, especially for long-growing crops like tomatoes. Growers must continuously monitor their crops, adjust plans accordingly, and observe how the crops react, creating a constant feedback loop between the grower, the climate, and the crop.

While many growers use remote technology to monitor climate conditions and adjust their settings once or twice a day, crop registration occurs only once per week. Assessing the crop's response to the environment requires spending more time in the greenhouse. Unfortunately, due to the size of some greenhouses and busy schedules plants are left out of the critical feedback loop. With both short and long-term crop health and production at stake, steering in the wrong direction can have consequences.

Mismanagement and misuse of resources throughout the tomato crop's growth can be costly and detrimental to production. Steering the crop too generatively can lead to issues like crop stress, excess fruit load, small fruits, weak plants, and crop loss. The result is early crop termination and major production gaps that critically impact budgets, labor retention, and customer relationships. Conversely, steering too vegetatively can result in less fruit with lower quality and excessive use of resources such as labor, heating, and lights.

According to baseline work with customers, labor and utilities have shown to make up more than 75% of input costs in medium to large scale operations thus indicating that efficient crop steering can critically impact the bottom line.

The level of control that growers have on manipulating climate, irrigation, and labor is not possible on a per-plant basis. However, having an appropriate sample size that provides enough sensitivity, enables growers to effectively control on a compartment, valve, or row level respectively.

With this kind of efficiency, a grower achieves a significant reduction in production costs. A 5% reduction in annual labor and utility expenses in a 10 hectare greenhouse with supplemental lighting translates to an estimated savings of $270K and $180K, respectively.

‍

Bridging the Gap: The Role of Crop Registration in Planning and Execution

Manual Crop Registration falls short of its potential to lead. The results lack sensitivity due to inadequate sampling and infrequent measurements, occurring only once a week.

CEOs and GMs need to quantify current and future outputs for planning purposes. Metrics like yield and forecast data serve as lagging indicators of success. In contrast, growers focus on plant balance to manage inputs like climate, irrigation, and labor planning, using leading indicators to drive outcomes at least 8 weeks ahead.

Despite its flaws, manual crop registration still serves as a critical resource for quantifying these leading indicators, which are then used to provide lagging insights or trends. This foundational measuring device impacts stakeholders across an entire organization.

In an industry driven by technology, reliance on manual methods like measuring plants by hand with a tape measure has persisted. This manual approach, prone to error, remains the most widely practiced method of quantifying leading indicators in the industry.

‍

Optimizing Greenhouse Coverage: Achieving Statistical Significance

Determining greenhouse coverage is crucial for representation and practicality. Different objectives drive different coverage needs; for instance, crop registration requires less coverage than issue detection. We found that at least 10% coverage within a population, evenly distributed throughout greenhouse rows, achieves statistical significance for crop registration.

To validate our recommendation, we simulated data capture across an entire greenhouse. The purpose was to establish the true mean of a variety of plant metrics. We tested various sample sizes, starting with small sets of 5 plants, and progressed to percentage-based sampling when the rate of error did not improve. Each population size was analyzed across 1000 sampling scenarios to compare with the true mean of the population. The chart below plots the average error of the 1000 sample sets for each of the plant metrics to ensure we are accounting for the dynamic nature of plants and the challenges of sampling the various plant characteristics. Because we understand that some metrics perform better or worse, it's important to note that 10% is where all of these metrics converge to less than 5% error. This is the beginning of the convergence of statistical significance in the plant population.

While selecting the 'right 5 plants' to represent the average of the total population is theoretically possible, it's highly improbable. Our studies show that biased sampling leads to incorrect results the majority of the time, sometimes even doubling the variability

Uniform random sampling is statistically ideal but impractical in greenhouse rows. To overcome this, we've set a 10% threshold to address infrastructure limitations and the impracticality of manual plant sampling.

By automating the sampling process, we can ensure consistency and efficiency. Why manually sample, when you can automate? Why measure once a week, when you can daily or as frequently as you want? Why be limited by the number of plants you can measure by hand when you can measure thousands? These are questions we ask our industry.

‍

This is why

We envision a future where every grower is empowered by their greenhouse production potential. The best growers continue to push the upper limits of yield and quality. Our industry holds vast opportunities waiting to be unlocked, yet we hinder our progress with inadequate sample sizes and misguided decision-making. This belief drives our mission. We are committed to transparency, sharing the data and rationale behind our development of an automated crop registration system with LUNA. You no longer need to settle, a brighter path to achieving the insights you seek has arrived.

If you have any questions or would like to talk to one of our experts, schedule some time with us here or reach out to us at Info@IUNU.com

‍

‍

‍

About the Authors

‍

Erika Verrier is a Vine Crop Specialist at IUNU on the Horticultural Team. Erika has over a decade of practical experience in the greenhouse supporting large-scale tomato greenhouse operators and growers. Her background also includes work in education, leadership and food-systems. Erika’s work in CEA has focused on risk management programs including Integrated Pest & Disease Management (IPDM), Food Safety, Biosecurity and many years of manual crop registration for growers. Erika joined IUNU in the fall of 2021 and has been on the forefront of developing the vine products here ever since. Additionally, she established the Baseline and Risk Assessment programs at IUNU to enhance our ability to drive value for growers of all crops we work with. Recently, Erika authored a report with IUNU on “Learning to Live with ToBRFV,” which has been read by more than 500 growers worldwide and is now excited to release the study titled “Crop Registration: A Compass for Crop Management. Erika thinks that the greatest part of working at IUNU is the impact we drive through innovation, and collaborating with a talented team to solve major challenges for growers in the industry. “If we know better, we can grow better.”