Exploring NASA AVIRIS3 imagery collected over FLX vineyards
Katie Gold, August 4, 2025
Written with assistance from generative AI
On Monday July 28th, 2025, AVIRIS-3 successfully captured high resolution (0.5-1.25m) hyperspectral imagery over Cornell research farms and commercial grapevine, apple, and onion production in the Finger Lakes region. These flights were funded by the NASA Agriculture program office in support of Drs. Katie Gold & Yu Jiang’s NASA ACRES consortium (#80NSSC23M0034) funded research project on early disease detection. Cornell University is the largest sub-award in the NASA Acres consortium led by Dr. Alyssa Whitcraft at the University of Maryland-College Park.
The aircraft’s hyperspectral sensor recorded hundreds of narrow “colors” of light for each half-meter pixel, creating maps that let us flag subtle differences in canopy health, water status, nitrogen, and leaf chemistry. Below, we walk you through the two early-release map layers now live in the online viewer and share potential insights you may be able to derive from the imagery at this time. The AVIRIS team is posted on the east coast until August 15th and we hope to capture imagery over the Lake Erie grape belt before then!
Open the online map viewer
- Click the link below (it works on desktop or mobile):
Interactive viewer — https://popo.jpl.nasa.gov/mmgis-aviris/?mission=Acres - In the upper-left corner, use Layers ▸ Flight Lines to zoom to the strip that covers your vineyard.
| PCA Product
“AVIRIS-3 Realtime PCA 1” West Canandaigua and West Seneca lines only |
RGB quick-look
“AVIRIS-3 Realtime Quicklook” All lines |
 |
 |
How to understand the two map layers you’ll see
| Product | What it is | What the colors could be flagging | When to trust it |
| PCA experimental layer (W-Cdga & W-Seneca) | A “principal-components” composite built from 430 hyperspectral bands. Brightness is removed so color only shows differences in canopy chemistry and structure. | • Areas with denser, greener foliage; zones with thinner canopies, stress, bare soil, cover crop gaps | Early scouting for “hot-spots.” Great for relative canopy differences within the same line, but DO NOT compare color between different flight lines! Colors are randomly selected for each individual line |
| RGB quick-look layer (all other lines) | False-color image made from 1660 nm, 850 nm & 560 nm wavelengths. | • Bright green: healthy foliage
• Red-orange: senescing foliage, gaps, soil, roads, etc. • Dark/black: water or shaded areas |
General plant health. Requires a healthy amount of prior knowledge of the vineyard for best use! |
Why two products? AVIRIS processed the PCA layer in-flight only for the two “demo” strips above. All other lines were downlinked as a fast preview. Higher-level foliar-trait maps (N, water, NSC, phenolics) will follow lab calibration later this season.
Basic navigation tips
| Action | How |
| Turn layers on/off | Click the eye icon next to a layer name. |
| Adjust color or brightness | Click the three-line “Layer Settings” icon; slide Opacity or Contrast as needed. |
| See what time the plane passed | The Time-Slider at the bottom lets you move along the flight track minute-by-minute. |
| Locate your block fast | Toggle the Context → Finger Lakes Boundaries layer, then zoom with your mouse wheel or the “+” button. |
Practical uses for growers right now
| Question | What to look for in the map |
| Which panels are lagging in canopy growth? | In the PCA strip, look for big color differences (e.g. magenta or yellow pixels next to blue-green neighbors) |
| Is there spatial variability in my vineyard this year? | Compare color changes within the same block; sudden dull patches might align with damage or differences in vigor. Target areas that look chemically different (different color in PCA) or chlorotic (orange in RGB). |
Caveats & next steps
- Early-look only. These images were captured July 28, 2025 and have no direct plant health indicators built in—they’re snapshots of canopy condition and variability. Right now, this imagery can do a great job of indicating differences in canopy condition, but not what caused them.
- Pixel size ≈ 0.5–1.2 m. One pixel ≈ at most one panel. Very small features (less than a panel) may blur. Young vines and closely cropped VSP trained vines will be most difficult to assess in this imagery.
- Lighting matters. Shaded end-rows and headlands can look stressed even when vines are fine—compare with your field notes.
- Trait maps coming. Once foliar samples are analyzed, we’ll post nitrogen, water, NSC and phenolics layers here and announce via the Finger Lakes Grape Program list.
See something cool?
We would love to hear your thoughts on the imagery and to what extent it aligns with what you are seeing on the ground. Send Katie and Dana an email (kg557@cornell.edu and dana.chadwick@jpl.nasa.gov ) and let us know what you notice!