NDVI vs NDRE: What They Actually Show a Greenkeeper

Multispectral drone surveys produce two numbers greenkeepers see most often: NDVI and NDRE. Both are vegetation indices — a way of turning a camera's raw light readings into a single number for how a patch of turf is doing. They are related, often shown side by side, and routinely confused for the same thing. They are not. This guide explains what each one actually measures, in plain terms, and when the difference matters.

What a vegetation index is

A plant's leaves absorb most visible light (that's why they look green) but reflect a lot of near-infrared light, which the human eye can't see at all. A healthy, dense canopy reflects a lot of near-infrared. A stressed, thinning or bare patch reflects much less. A vegetation index is just a ratio of how much infrared light comes back compared to some other band — expressed as a single number, usually between -1 and 1, so a whole course can be mapped in one consistent colour scale.

A multispectral camera like the DJI Mavic 3 Multispectral captures several separate bands per photo — green, red, red edge and near-infrared — which is what makes both indices possible from the same flight.

NDVI: the standard index

NDVI (Normalized Difference Vegetation Index) compares near-infrared reflectance to red light reflectance. It's the index most people mean when they say "NDVI map," and it's a good general read on vigour: deep green on the map means vigorous turf, yellow means weaker, brown means bare, sand or heavily stressed ground. Open water reads below zero.

NDVI's limitation shows up on dense, healthy canopy. Red light is absorbed so strongly by chlorophyll that once a canopy is thick enough, almost all of it is already being absorbed — adding more healthy leaf tissue doesn't change the red reading much. NDVI then saturates: a very healthy green and an extremely healthy green can look almost identical in the red band, even though the plant's actual chlorophyll content still differs.

NDRE: the red-edge index

NDRE (Normalized Difference Red Edge) uses the same near-infrared band, but compares it to red edge light instead of red — a narrow band that sits right at the boundary where chlorophyll absorption falls off. Red edge reflectance keeps responding to changes in chlorophyll content even on dense, already-healthy canopy, where red light has already maxed out.

That makes NDRE more sensitive on exactly the turf where NDVI has the least room left to move: closely mown greens and tees with thick, healthy cover. It's a better early-warning signal there — on sparse or bare ground (bunker faces, paths, thin rough) NDVI is usually the clearer read, because there's little canopy left for the red-edge comparison to work with.

Why they can disagree — and why that's useful

A patch of turf that reads strongly healthy on NDVI but is drifting down on NDRE is a real pattern worth attention: it suggests early stress inside a canopy that still looks fine from the red band's point of view. Flying both indices from the same flight (they come from the same four bands, so there's no extra cost) and watching for that kind of disagreement is more informative than either number alone.

Reading the numbers on a real course

As a rough guide from imagery processed so far: healthy autumn greens typically read NDVI around 0.7 to 0.85; fairways somewhat lower; bunkers near 0; open water below 0. MachAIr's own engine flags turf as "stressed but alive" in the NDVI 0.2–0.5 range, and the equivalent NDRE band for the same condition sits lower, around 0.05–0.2 — NDRE's scale runs narrower than NDVI's for the same ground truth, which is normal and expected given what the red-edge band is measuring.

These are working figures from processing real and reference multispectral imagery, not a peer-reviewed benchmark — treat them as a starting point for your own course, and watch how your own numbers move over a season rather than anchoring hard to any single reading.

Getting imagery you can trust

Both indices are only as good as the imagery behind them:

• Radiometric calibration matters more than the index you pick. A sunlight sensor (or a reflectance calibration panel) ties each flight's readings to the actual light conditions at the time, which is what makes one flight comparable to the next. Without it, a cloudier day can look like stressed turf that isn't.
• Fly within about two hours of solar noon, on dry turf (dew and frost reflect light and make healthy grass look artificially stressed), and at the same time of day each visit — so two maps actually compare fairly.
• Consistency beats precision. The same course, flown the same way each time, gives a trend you can act on even if the absolute numbers are approximate.

Honest limits

NDVI and NDRE describe light reflectance, not a diagnosis. Both can flag a problem; neither tells you whether it's drought, disease, compaction, nutrient deficiency or wear — that still takes a greenkeeper's eyes on the ground. Band alignment and camera calibration issues are real and documented failure modes of multispectral processing pipelines generally, not something specific to any one platform, so a map showing coloured fringing along sharp edges (bunkers, paths) is a processing artefact to double-check, not a finding to act on.

Multispectral imagery is a way to see where to look sooner and more precisely across a whole course — not a replacement for walking it.