Remote Sensing for Farm Irrigation: How Satellite NDVI Data Tells You Where to Water (And What It Costs Per Acre)

What satellite NDVI actually tells you that soil sensors miss

Soil moisture sensors tell you what’s happening at one point. A tensiometer buried 12 inches deep in row 47 tells you about row 47. Nothing about row 52, the low spot where water pools at the far end of the field, or the sandy patch by the road that dries out two days before everything else.

That’s the issue with point-based monitoring: decisions about 40 acres based on three or four data points. It works, sort of. Farmers have been doing it for decades. But it leaves money on the table, water in the pump you didn’t need to run, and yield in under-watered parts of the field you don’t know about.

Satellite NDVI flips the problem around. NDVI stands for Normalized Difference Vegetation Index, a number calculated from how plants reflect red and near-infrared light. Healthy vegetation with lots of chlorophyll absorbs red light and reflects near-infrared. Stressed vegetation does the opposite. The math spits out a value between -1 and 1, and for actively growing crops, anything above 0.5 means “doing fine,” while a patch reading 0.35 next to an area reading 0.65 tells you something is wrong in that patch.

What makes this useful for irrigation: water stress shows up in NDVI before you can see it. By the time leaves wilt or turn yellow, you’ve already lost yield. NDVI picks up the photosynthetic drop when plants close stomata to conserve water. That happens days before visible wilting.

The difference from soil sensors is coverage. A single Sentinel-2 image covers roughly 100 square kilometers at 10-meter resolution. Every pixel in your field gets a reading. Overlay NDVI maps on your field boundaries and you immediately see which zones are behind. Not because a sensor told you one spot is dry, but because the crop itself is showing you across every square meter.

An almond grower in California’s Central Valley runs 12 soil moisture stations across 200 acres. His sensors said everything was fine. Sentinel-2 NDVI maps showed a strip along the southern edge running 0.1 to 0.15 lower. Pressure drop at the end of his longest laterals was the culprit. He’d been losing roughly $9,000 a year in yield on that strip. Twelve sensors didn’t catch it. One satellite image did.

Where the data actually comes from (and what “free” really means)

There are three practical satellite sources for farm NDVI data right now.

Sentinel-2 is the workhorse. Two satellites (2A and 2B), operated by the European Space Agency, each revisiting any given point on Earth every 5 days. Combined, you get a new image of your field roughly every 2 to 3 days. Resolution is 10 meters per pixel. Not fine enough to see individual plants, but good enough to map zones within a field. The data is completely free. You can download it from the Copernicus Open Access Hub or use one of several platforms that process it for you.

Landsat 8 and 9 are NASA/USGS satellites with 30-meter resolution and a 16-day revisit cycle (8 days combined). The spatial resolution is coarser. A 30-meter pixel covers 900 square meters, so it’s better suited to large fields and regional monitoring than zone-level irrigation management. Also free.

Planet Labs runs a constellation of 200-plus small satellites with 3-meter resolution and daily revisits. This is the premium option. You can spot individual tree rows and stress patterns within a single irrigation block. It starts around $2 to $5 per acre per year. For a 500-acre farm, that’s $1,000 to $2,500 a year. Not trivial, but not insane compared to what under-watering costs.

The “free” in free satellite data has an asterisk. Raw images need atmospheric correction, cloud masking, and NDVI calculation. If you’re comfortable with QGIS and Python, you can do it yourself. Otherwise use a platform that handles it, and that’s where the actual cost lives.

How to turn satellite images into an irrigation schedule

The practical workflow comes down to comparing zones.

Step one: get a baseline NDVI map of your field during a period when you know irrigation is uniform. Right after a heavy rain, or early in the season when soil moisture is consistent. This is your reference. It shows you what the crop looks like when water isn’t the limiting factor.

Step two: pull new NDVI images every week during the growing season. Most platforms will do this automatically and notify you when a cloud-free image is available. Compare each new image to your baseline. Zones that are trending downward relative to the field average are the ones that need more water or less time between irrigations.

Step three: ground-truth. When a zone consistently runs 0.1 below the field average, go look at it. Soil probe. Check emitters. The satellite tells you where to look; your boots tell you why.

Step four: adjust your irrigation zones accordingly. If you’ve already got a system with solenoid valves and a controller, you can split irrigation blocks so the stressed zones get an extra pulse or longer runtime. If you don’t have zone control, the NDVI data builds the business case for the upgrade.

Satellite NDVI is a relative tool, not an absolute one. A value of 0.72 doesn’t mean “irrigate now.” It means “this zone is 0.08 below the field average, and that gap has been widening for two weeks.” The comparison across space and time is what matters.

What this actually costs per acre, and when it beats soil sensors

Let’s put numbers on it.

The free route (Sentinel-2 data processed with QGIS or Google Earth Engine) costs nothing in cash. The labor is a few afternoons of learning if you’re comfortable with software. For a small farm, the learning curve might not be worth it. For 200 acres, a few afternoons against $9,000 in recovered yield makes the math obvious.

If you use a paid platform, companies like CropX, SWIMM, Ceres Imaging, and Taranis bundle satellite NDVI with soil sensor integration, weather data, and irrigation recommendations, expect to pay $5 to $15 per acre per year. Many of these platforms also sell or integrate soil moisture probes, so you’re not choosing between satellite and sensors; you’re getting both. At $10 per acre on 200 acres, that’s $2,000 a year. If the satellite data helps you catch one pressure-regulation problem or one under-watered zone per season, it pays for itself several times over.

A full soil sensor network with enough density to give you the same spatial coverage runs $3,000 to $8,000 in hardware for a 200-acre farm, plus installation, plus annual maintenance and calibration. The satellite approach doesn’t replace soil sensors. You still want a few probes for depth information and real-time readings. But it tells you where to put them and fills in the gaps between them.

The sweet spot is three to five soil moisture stations placed based on NDVI zone maps, plus weekly satellite monitoring. Spatial coverage from above, depth and real-time data from the probes. Total cost: roughly $2,500-$4,500 a year including platform fees for a mid-sized farm. The water savings alone often cover it. The yield gains from catching stress early are the real return.

The catch: clouds, resolution, and when satellite data fails

Satellite NDVI has blind spots, and you need to know them before you rely on it.

Cloud cover is the biggest one. In regions with persistent cloud during the growing season (Pacific Northwest, Southeast Asia during monsoon, the UK) you might go two or three weeks without a usable image. That’s too long. In those conditions, satellite NDVI supplements soil sensors. It doesn’t replace them.

The 10-meter resolution from Sentinel-2 works for field-scale zone mapping but not small plots. A one-acre market garden with 30 different crops in tight beds has management units smaller than the pixel size. That scale needs drone imagery, not satellite.

NDVI also saturates. Once a crop canopy is full and dense. Corn in August, mature alfalfa, values top out around 0.8 to 0.9 and stop changing. The index can’t distinguish “well-watered” from “over-watered” once the canopy closes. For full-canopy crops, you need additional indices like NDWI (water index) or thermal infrared.

There’s also a timing lag. NDVI responds to stress after the plant has already started conserving water. It’s earlier than visual symptoms, but still days after the stress began. A soil moisture sensor buried in the root zone registers the drop in real time. That’s why the recommend approach is satellite plus probes, not satellite alone.

None of this makes satellite NDVI a bad tool. It’s a tool with a specific job: showing you where the crop isn’t happy so you can go investigate with a probe and your boots. For a few hundred to a few thousand dollars a year, that’s a job worth automating.