Using NDVI to make vineyard decisions
The use of NDVI in agriculture is not a new idea. With numerous studies demonstrating NDVI’s ability to help winegrowers understand vine vigor among other use cases, the number of NDVI providers has exploded in recent years.
With the vast number of choices available, winegrowers may have trouble evaluating different services. Images provided by different providers and on different days are also difficult to compare and reconcile with other data sources. The Fruition Sciences support team occasionally receive questions on how to correlate NDVI results with plant-based measurements obtained from the field.
One of the most frequent issues we have seen is the confusion around NDVI calibration with ground reflectance. Even though NDVI is useful in capturing variability within a vineyard, the results may be affected by a variety of factors. According to Johnson et al (2003), these factors include soil brightness, atmospheric turbidity, and canopy structure among others. Two NDVI images taken on different days for the same block can look entirely different because of different levels of cloudiness. This results in the difficulty in comparing NDVI results of two different dates.
According to Iowa State’s Integrated Crop Management News, the vast majority of images provided by NDVI image providers are uncalibrated. This may be less of an issue when no perennial structures are left in the field from one season to the next (typical for annual crops like lettuce). However, this presents a challenge for winegrowers who seek to understand what constitutes a high-quality image and how vineyard spatial structure changes from one season to the next.
Vineyard NDVI map (source)
It’s important to use ground-based measurements to correct the biases of NDVI results. In another study, Johnson and Scholasch (2005) used a handheld spectroradiometer to measure the reflectance at four types of ground targets (an unlined asphalt road, a gravel parking lot and two concrete surfaces). These measurements were used to calibrate the NDVI readings.
What to ask your NDVI provider
Fruition Sciences has compiled a list of questions for you to evaluate your NDVI provider:
- What is your experience with the use case we need (e.g. field monitoring, leaf area calculation, pest and disease detection, etc.)?
- What is the spatial resolution required for this use case?
- What are potential sources of bias? How do you do ground calibration?
- How do you correct for these sources of bias?
Fruition Analytics integrates vineyard data from a variety of sources including plant-based measurements from Sap Flow and Physiocap as well as NDVI.