Satellite data is information collected by sensors orbiting Earth that measure reflected or emitted energy from the planet’s surface and atmosphere. These measurements capture everything from visible light to heat signatures to radar echoes, producing digital images and datasets that reveal patterns invisible from the ground. The global satellite data services market was valued at roughly $12 billion in 2024 and is projected to reach $55 billion by 2034, reflecting how central this information has become to industries ranging from farming to disaster response.
How Satellites Collect Data
Satellites carry instruments that detect energy across the electromagnetic spectrum, far beyond what human eyes can see. A single satellite like Landsat 9 collects data in 11 separate bands, each tuned to a different slice of the spectrum: visible blue, green, and red light, near-infrared, shortwave infrared, and thermal infrared. Each band reveals different characteristics of the surface below. Visible bands show what a landscape looks like. Near-infrared bands highlight vegetation health because healthy plants reflect strongly in that range. Thermal bands measure surface temperature.
The sensors fall into two categories. Passive sensors measure energy that the Earth naturally emits or reflects from the sun. They work across all weather conditions, day and night, and are the backbone of most environmental monitoring. Active sensors, by contrast, send out their own signal (typically radar pulses) and then measure what bounces back. Precipitation radars, for example, send a signal through the atmosphere and read the echo from rainfall to calculate how much rain is falling. Cloud-profiling radars build three-dimensional maps of cloud structure the same way.
Four Types of Resolution
Not all satellite data is created equal. The usefulness of any dataset depends on four types of resolution, each describing a different dimension of detail.
- Spatial resolution is the size of each pixel on the ground. Landsat 9’s main bands have a spatial resolution of 30 meters, meaning each pixel represents a 30-by-30-meter patch of Earth. Its panchromatic band sharpens that to 15 meters. Some commercial satellites achieve sub-meter resolution, while weather satellites may use pixels a full kilometer wide.
- Spectral resolution describes how finely a sensor can distinguish different wavelengths. A sensor with three to ten broad bands is called multispectral. Hyperspectral sensors can capture over 200 narrow channels, letting scientists identify specific minerals, crop species, or water pollutants.
- Temporal resolution is how often a satellite revisits the same spot. Some orbits produce a revisit every one to two days, while Landsat’s orbit returns to the same location every 16 days. Higher temporal resolution is critical for tracking fast-changing events like floods or wildfires.
- Radiometric resolution is how many shades of intensity a sensor can record per pixel. An 8-bit sensor distinguishes 256 levels of brightness, while higher-bit sensors capture thousands or millions of levels, revealing subtler differences in surface features.
These four resolutions involve trade-offs. A satellite that covers the whole planet every day typically does so at coarser spatial detail, while a high-resolution satellite may only revisit a location once a week or less.
From Raw Signal to Usable Information
The data that comes down from a satellite isn’t a ready-made map. It goes through several stages of processing, standardized by NASA into levels. Level 0 is the rawest form: unprocessed instrument readings with communication artifacts stripped out. Level 1 adds time stamps, location references, and calibration so the data can be placed on a map. By Level 2, scientists have converted those calibrated signals into meaningful physical measurements like surface temperature, vegetation density, or ocean color, still tied to the satellite’s original pixel grid.
Level 3 takes those measurements and maps them onto a uniform grid over consistent time periods (weekly, monthly), filling in gaps where clouds or other obstructions blocked the view. Level 4 goes further still, combining satellite observations with computer models to produce derived products like global carbon flux estimates or weather forecasts. Most researchers and businesses work with Level 2 or Level 3 data, which balances scientific rigor with practical usability.
The finished products are typically stored in specialized file formats. NetCDF is one of the most common for scientific satellite data because it’s self-describing: the file itself contains metadata explaining what each variable means, what units it uses, and how the data is structured. This makes it far easier to share data between research groups without losing critical context. GeoTIFF is another widely used format, especially for image-based products, because standard mapping software can open it directly.
Environmental and Climate Monitoring
Satellite data’s longest-running application is tracking environmental change. The Landsat program has been imaging Earth continuously since 1972, creating a record spanning more than 50 years. That archive allows scientists to measure deforestation rates, glacier retreat, urban sprawl, and coastline erosion with a consistency no ground-based network could match.
Thermal satellite data is particularly valuable for studying urban heat islands, areas where cities run significantly hotter than surrounding countryside due to concrete, asphalt, and reduced tree cover. Satellites map land surface temperature across entire metropolitan areas in a single pass, identifying which neighborhoods absorb and radiate the most heat. When those thermal maps are combined with census and health data, city planners can build heat vulnerability indexes that pinpoint where cooling interventions (tree planting, reflective roofing, public cooling centers) will save the most lives.
On a larger scale, satellite observations feed directly into global climate models. Instruments measuring sea surface temperature, ice sheet thickness, atmospheric gas concentrations, and ocean height provide the continuous, planet-wide measurements that ground stations simply cannot.
Agriculture and Land Management
Farmers and agricultural agencies use satellite data to monitor crop health across vast areas without setting foot in the field. Near-infrared and red-light bands are combined into vegetation indexes that quantify how vigorously plants are photosynthesizing. A sudden drop in that index over a region can signal drought stress, pest damage, or disease weeks before it becomes visible to the eye.
Satellite-derived soil moisture measurements help optimize irrigation timing, and thermal data can flag waterlogged or compacted zones in a field. At a national scale, governments use satellite crop monitoring to forecast harvests, manage food reserves, and respond to agricultural emergencies. The combination of frequent revisit times and broad coverage makes satellites uniquely suited to this kind of large-area surveillance.
Who Provides Satellite Data
Several government programs distribute satellite data at no cost. NASA’s Earthdata platform hosts petabytes of freely accessible datasets from dozens of missions. The U.S. Geological Survey manages the Landsat archive and makes every image available for free download. The European Space Agency’s Copernicus program operates the Sentinel satellite constellation and follows the same open-data philosophy.
On the commercial side, companies operate constellations of smaller satellites offering higher spatial resolution, faster revisit times, or specialized sensors not available from government missions. These commercial providers sell data to industries like insurance (assessing disaster damage), finance (estimating retail activity from parking lot imagery), energy (monitoring pipeline infrastructure), and defense. The market’s projected growth to over $55 billion by 2034 is driven largely by this expansion of commercial applications, fueled by cheaper launch costs putting more satellites into orbit every year.
Limitations Worth Knowing
Satellite data has real constraints. Optical sensors cannot see through clouds, which means tropical regions and high-latitude areas often have persistent gaps. Radar sensors solve this for some applications, but radar data requires more specialized processing and interpretation. Spatial resolution limits what you can identify: a 30-meter pixel can distinguish a forest from a parking lot, but it cannot count individual trees. And while temporal resolution has improved dramatically with multi-satellite constellations, there’s still a lag between when an event happens and when a satellite happens to pass overhead.
Data volume is another practical challenge. A single day of observations from one satellite constellation can produce terabytes of raw data. Storing, transferring, and processing that volume requires significant computing infrastructure, which is why cloud-based platforms for satellite data analysis have grown rapidly alongside the data itself.