Using temperature and humidity sensors in vineyards is becoming a well-established practice. They provide valuable data for precision viticulture that help farmers monitor and optimize growing conditions. Up until a few years ago, deploying multiple sensors in a vineyard was quite costly and complicated. Mainly because the communication technology to backhaul the data required specialized equipment. However, with ubiquitous access to WiFi and LoRa communication technology, the cost per sensor has dramatically dropped, enabling better applications for viticulture management.

One practical application of sensors, especially in Central Virginia this time of the year, is frost mitigation. Frost is particularly damaging to grapevines during bud break. This is when the young tissue is most vulnerable. Temperature sensors deployed throughout a vineyard can alert farmers to frost risk zones, allowing for targeted protection measures. Since vineyards often have varied terrain with different sun exposure, elevation, and airflow, utilizing multiple sensors enables farmers to map, visualize, and understand microclimates accurately. For a great example of a frost mitigation technique, click this link for a recent article on using anti-frost candles to protect the buds.

                                                  [ one weather station]

                                           [multiple temperature sensors]

With the sensors in place early this year at Fidelis Farm, I witnessed first-hand one of the risks of growing grapes in central Virginia. A few days ago, the temperature in Crozet, VA, plummeted below freezing. As a farmer, which graph is most helpful? The first part describes a vineyard block as a whole (the forest) using one trend line, which doesn’t show the different microclimates. The second part provides details about these microclimates within the vineyard block (the trees). Each line represents different zones within the block, with some sensors showing temperatures that differ by up to 4 degrees from the central weather station. While 4 degrees might not seem like much, this insight can significantly help reduce the risk of losing all of the crop. The geolocation of each sensor made it easy to identify the colder zones in the vineyard.

                                [ numbered sensors with geolocation]

To facilitate sensible frost mitigation, it seems that one weather sensor per vineyard block is not enough. It is better to have at least one sensor per acre. For next year, I am tempted to attach a strobe light to each sensor. This way, when the temperature drops below freezing, it triggers the strobe light to alert the farmers where to concentrate their efforts. Wouldn’t that be cool?

As always, all the VineIQ code is open source. Feel free to clone it:

https://github.com/Fidelis-Farm-Technologies/VineIQ

Assuming elevation, terrain, bodies of water, wind, temperature, and humidity are some key traits that characterize a vineyard's micro-climate, how would you determine where to strategically place one sensor? But, what if you could avoid this dilemma by deploying a dozen sensors instead? Better yet, how about attaching the sensors directly to the trellis system within the fruit zone inside the canopy?

For approximately $35 per unit (without hub) on Amazon, the YoLink Smart Outdoor Temperature & Humidity Sensor is one of the most cost-effective ways to deploy multiple sensors in a vineyard. At Fidelis Farm, I recently installed 10 sensors plus a hub (one is required) for less than $400. That is an incredibly low price when compared to the price of purpose-built sensors offered by traditional vendors serving agricultural market.

Specification. YoLink sensors include a long-range radio (LoRa) to transmit small amounts of data at a long range using minimal power. Therefore, the batteries should last over 2 years—enough for 2 growing seasons. The sensor offers commercial-grade accuracy for temperature and humidity. Plus, it is IP67 certified (waterproof). You will need at least one hub to deploy sensors. A hub is a piece of hardware that functions as a gateway between the sensors' radio communication and Internet connectivity. Be aware that when ordering the sensors on Amazon, you can buy the sensor bundled with or without a hub. One hub will serve many sensors; therefore, typical deployments will require buying the first sensor bundled with the hub. Then, the subsequent sensors without a hub.

Installation. Installing the hub and sensors is a fairly easy process. First, install the YoLink app on your phone. Both iPhone and Android are supported. Then, add the hub to your app by selecting "Add Device" and scanning the QR label on the hub. After connecting the hub to the Internet, use the app to add and bind the sensors to the hub by scanning the QR code of each sensor. When you complete the process, your app should look similar to the screen shot below.

Note: although the sensors come with nylon strings to use for attachment, I prefer to use zip ties to secure them on the trellis system.

Collection. The YoLink app will display measurements within minutes. After the sensors transmit data over radio waves to the hub, the hub forwards the data over the Internet to the YoLink cloud service for storage and visualization. As part of the cost, YoLink includes basic cloud storage with each sensor.

Visualization. The YoLink app includes basic data time series visualization. Since I have a few weeks of measurements, I can view daily, weekly, and monthly temperature and humidity measurements. Although the simple graphs are adequate for personal and home use, more features are required for this project. Over the course of this growing season, VineIQ will add more functionality and analysis to support vineyard-specific use cases.

Integration. Fortunately, YoLink allows functional control and access to data generated by the sensors through an openly published application programming interface (API). This is quite promising because it allows for third-parties, like VineIQ, to develop custom applications. For example, if you do a Google search on "yolink api github," you will see a few open-source projects that already integrate with the YoLink API.

In my next post, I will discuss the VineIQ YoLink logger software (published on GitHub here) and its integration with QuestDB to store time-series data and application-specific visualizations with Grafana.

Sense. Use cost-effective, commercial-off-the-self hardware sensors to measure temperature, humidity, solar radiation, rain accumulation, soil moisture, and wind speed/direction, etc.

Collect. Integrate with sensor hardware vendors by connecting to third-party APIs to collect and store all the data in the cloud using an instance of QuestDB (database).

Visualize. Use Grafana to build dashboards with gauges and time series charts of various historical data points. Calculate second-order metrics such as vapor pressure deficit (VPD), growing degree days (GDD), etc.

Predict. Implement machine and deep learning models using PyTorch to make predictions such as leaf wetness duration, risk index for frost, risk index for powdery mildew, irrigation management, and spray drift management. Eventually (but not immediately), expand capabilities to include nutrient management and crop yield estimation using hyper-spectral imaging cameras.

Share VineIQ

There is a myth that Lord Kelvin said, "To measure is to know. If you cannot measure it, you cannot improve it." Although these are not his exact words, in principle, this was his mantra1 and the inspiration for Project Vine IQ. 

Soon, dormant buds in my vineyard will emerge and produce shoots, marking the beginning of a new growing season in Central Virginia. This year, I will be ready to measure and collect basic telemetry from my vineyard (something I've wanted to do for years). I’m excited to get to know my vineyard and hope to contribute to viticulture practices in Virginia. 

Growing grapes in Virginia…

Growing grapes in Virginia is challenging2. Even Thomas Jefferson struggled to grow European grapes successfully in Central Virginia. However, with the modern advances in science, technologies, and analytics, viticulture managers are able to overcome many of the challenges by making data-driven decisions, provided they measure and collect relevant data from their vineyard.

Why collect data?  

Collecting data facilitates data science, which ultimately enables the extraction of meaningful insights. Although our entire economy is data-driven, the viticulture industry needs to catch up in adopting modern remote sensing and data collection practices. One reason for this is that the existing technology is expensive. Another is that it needs to be simpler to deploy and manage. Finally, many farmers must see the value of capturing and analyzing the data before using it. In their minds, "why fix what isn't broken?"

What can we do with the data?

Data-driven analytics can give farmers insights and enable them to optimize frost mitigation, nutrient input, fungicide application, pesticide management, and much more. Furthermore, data-driven visualization of charts, graphs, maps, enable intuitive understanding of trends and outliers in a vineyard. And, lastly, data-driven machine learning algorithms detect hidden patterns, regularities, and anomalies that are not easily perceived by farmers at first-glance. Good analytics requires lots of historical data to provide useful insights. Although it takes additional time and effort to collect, data enables valuable insights not only at a local level, but regionally and globally.

How do we collect the data?

The goal is to collect data with technology that is easy to use and cost-effective. Ultimately, the farmer needs to see the value in collecting data. A great approach is to use commodity off-the-shelf hardware with accessible application programming interfaces (API) and community-driven open-source software. First, hardware vendors benefit from the wide adoption of their products. Secondly, the development of open-source software applications efforts typically operate at a faster velocity and incorporate useful features. Nothing kills an open-source project like useless features. Thirdly, a community-driven effort benefits from the contributions of domain experts: researchers, developers, technologist, and, most importantly, farmers.

Subscribe now

Why should I subscribe?

First, this project is an open-source project. In this publication, I will share details about the project. Plus, I plan to publish the hardware specs, code, and documentation on GitHub along the way that you may use and provide feedback. Secondly, this project is community-driven. If you have domain expertise, as a farmer, scientist, or developer, and would like to make a contribution, the first step is to subscribe. Then send me a message and let me know how you would like to contribute.

What are the benefits?

As a subscriber, you will help promote the project.

• As a free subscriber, you will receive regular updates describing the progress of the project.

• As a paid subscriber, you will get access to the implementation details and will be able to participate in threads, chats, and post feedback.

• As a sponsoring subscriber, in addition to the benefits of a paid subscriber, I will publish your organization’s logo and will include access to VineIQ’s software-as-service (later this year). For this, you are responsible for purchasing and setting up the hardware — with support from me via email. Then, VineIQ will do the rest… collect, analyze, and visualize your data on a private instance in the cloud, managed by Fidelis Farm & Technologies. This subscription offer will be limited to the first five (5) sponsors.

Subscribe now