Article summary: A 2025 Remote Sensing study looked at why common vegetation indices (like NDVI) can “saturate” once pasture gets dense, and how newer approaches improve pasture biomass estimates by using the full Sentinel-2 spectral signal, adding weather context, and filling gaps when images are missing or misaligned. The practical takeaway for grazing managers: satellite-backed pasture numbers are getting less noisy, especially when combined with consistent on-farm routines, so your rotation decisions can be calmer, faster, and less reactive.
You already know the feeling: you want confidence in the next grazing call, but a full pasture walk takes time, weather gets in the way, and the “gut feel” decision can creep in when things get busy.
The good news is that satellite pasture estimates are improving year by year, not because satellites magically got perfect overnight, but because the science and modelling underneath them is getting smarter.
A 2025 study in Remote Sensing put plain language around why earlier satellite approaches sometimes struggled in heavy covers, and what’s now being done to work around it. Here’s the science-to-practice version.
The old shortcut: a simple “greenness score” that can top out
Many satellite pasture tools started with vegetation indices like NDVI. Think of NDVI as a quick “how green is it?” score based on how plants reflect different light.
It’s useful, but it has a known limitation: it can saturate in dense pasture.
An analogy: the bathroom scale that tops out
Imagine a bathroom scale that reads up to 120 kg, then stops. Once you’re over that point, the number no longer tells you what’s really happening.
NDVI can behave a bit like that in thick pasture. Once the canopy is dense enough, the signal can plateau, even if the pasture keeps piling on. Practically, that means a simple greenness score can struggle to separate “good cover” from “very high cover”.
The 2025 study described this effect in pasture systems and noted the saturation point can occur once pasture mass gets into the heavier range.
The upgrade: stop relying on one score, use the full “signature”
Instead of compressing the satellite signal into one index, newer approaches use full-band reflectance, particularly from satellites like Sentinel-2.
Another analogy: black-and-white vs full colour
If NDVI is like judging a paddock from a single black-and-white photo, full-band reflectance is like seeing the whole colour image plus extra information your eyes cannot see.
Sentinel-2 doesn’t just capture “visible” light. It also measures bands such as:
-
near-infrared
-
red-edge
-
short-wave infrared
Those bands are often more informative when pasture is dense, when the canopy structure matters, or when moisture and plant condition start to influence what the satellite sees.
The study found that models using the full set of Sentinel-2 bands outperformed index-only approaches, especially where NDVI would otherwise flatten out.
Add the missing context: weather explains the story between satellite passes
Satellites take snapshots. Pasture growth is a moving picture.
That’s where weather data helps. Temperature, radiation, humidity and rainfall all influence growth, and the effect often shows up with a lag.
When a model blends satellite information with weather context, it’s less likely to overreact to a single noisy image and more likely to track the underlying growth pattern.
In plain terms: the model gets better at understanding why a paddock is changing, not just whether it looks greener today.
The real-world headache: cloud and “missing days”
If you farm in a cloudy region, you don’t need a scientist to tell you the main weakness of satellites: sometimes they simply cannot see the ground.
There are two common “gaps” that create noise:
-
cloud cover blocks the view
-
timing mismatches (your on-ground reads and satellite images don’t line up neatly)
The 2025 study tackled this by using interpolation and progressive updating to reduce the penalty of missing or misaligned observations.
A simple way to think about interpolation: joining the dots
If you’ve got two reliable points, interpolation is a sensible way of joining the dots between them.
In practice, it can help models learn from paddock changes even when:
-
a cloudy spell wipes out images for a week or two, or
-
your plate meter walk is weekly but the best satellite image is on a different day.
The point is not to “make up data”. It’s to reduce the blind spots so the model has a smoother, more realistic training signal.
So what does this mean for grazing decisions?
It does not mean you should stop walking paddocks or blindly trust a single number.
It does mean that satellite-backed estimates are getting better at the things that matter for decision-making:
1) Better separation at the top end
Where older greenness-only methods could struggle in heavy covers, full-band approaches can do a better job of distinguishing “high” from “very high”. That matters when you’re trying to keep quality in front of the herd, avoid heading, or decide what to graze first.
2) More stable trends through messy weeks
Blending satellite with weather context, and smoothing around missing images, generally improves the usefulness of the trend line. That helps you:
-
hold your nerve on a rotation,
-
spot a genuine slowdown in growth,
-
and avoid reacting to one odd reading.
3) More confidence in paddock ranking
Even when absolute values vary, paddock ranking is often where the practical value is highest:
-
Which paddocks are running away?
-
Which are falling behind?
-
Which should be prioritised to hit residual targets?
Where satellites still struggle (and what to do about it)
Even with better models, there are times when satellite estimates can be challenged. The best approach is to know the common trouble spots and build a routine around them.
Satellites can struggle when:
-
Cloud persists for long stretches
-
Pasture is very dense or rank (the “top end” problem)
-
Tree shelter, steep terrain or mixed land cover confuses the signal
-
Rapid management changes happen across a few days (strip grazing, break feeding, partial grazes)
-
Senescence or dead material increases (green signal and total feed on offer can diverge)
What good operators do instead:
-
Trust the trend, then verify the outliers
-
Record grazing events consistently so the model is anchored to what actually happened
-
Use quick visual checks when the system flags something unexpected
-
Keep one consistent measurement routine as a reality check (even if it’s on a subset of paddocks)
The quiet superpower: consistency beats perfection
Here’s the soft truth most grazing teams learn the hard way:
You don’t need a perfect number to make great grazing decisions.
You need a consistent yardstick.
When your measurement approach is consistent (satellite-backed, plate meter, pasture walk, or a mix), your rotation decisions become:
-
less emotional,
-
less reactive,
-
and easier to communicate to the team.
A practical “science-to-practice” routine you can use
If you want to turn improving satellite estimates into better grazing outcomes, focus on the basics:
-
Pick a measurement cadence you can stick to (weekly works well for many)
-
Record grazing dates and areas cleanly (so changes in cover make sense)
-
Use pasture numbers to support rotation rules (pre- and post-grazing targets, round length, supplement triggers)
-
Ground-check the extremes (very high, very low, or “doesn’t look right”)
-
Review the wedge and growth trend as a team (make decisions explicit, not assumed)
If you do that, you’ll find the biggest benefit isn’t “better satellite science”. It’s that your rotation calls get calmer, faster, and easier to repeat.
- The Dedicated Team of Pasture.io, 2025-12-30