Four Types of Satellite Image Resolution That Actually Matter
Quick Answer: Satellite imagery has four types of resolution: spatial (pixel size on the ground), spectral (number and width of wavelength bands), temporal (how often the same area is revisited), and radiometric (how many brightness levels the sensor distinguishes). For most Earth observation tasks, temporal and spectral resolution matter more than raw spatial resolution.
Resolution Is Not Just About Pixel Size
When people say "high-resolution satellite imagery," they almost always mean spatial resolution — the size of a single pixel on the ground. A 10-meter pixel versus a 0.3-meter pixel is an obvious, intuitive difference.
But spatial resolution is only one of four resolution types, and in many practical applications, it's not the most important one. I've seen projects fail because teams chased the highest spatial resolution available while ignoring that their analysis actually needed better temporal coverage or more spectral bands.
1. Spatial Resolution
What it means: The area on the ground that one pixel represents.
| Satellite | Spatial Resolution |
|---|---|
| Sentinel-2 | 10m (visible/NIR), 20m (red edge/SWIR) |
| Sentinel-1 SAR | ~20m (IW mode) |
| Landsat 8/9 | 30m (multispectral), 15m (pan) |
| Commercial VHR | 0.3–3m |
At 10m resolution, you can distinguish field boundaries, forest edges, and major land cover types. At 0.3m, you can identify individual cars and building footprints.
When spatial resolution is critical: Urban mapping at building level, infrastructure monitoring, damage assessment of individual structures, vehicle detection.
When it doesn't matter much: Regional vegetation trends, ocean color analysis, atmospheric studies, large-area land cover classification. A 10m pixel is more than sufficient for these tasks, and the extra data volume of higher resolution just slows you down.
2. Spectral Resolution
What it means: The number, width, and placement of wavelength bands the sensor captures.
This is where satellite imagery differs fundamentally from a photograph. A camera captures three broad bands (red, green, blue). Sentinel-2 captures 13 bands from visible through shortwave infrared, each tuned to reveal specific surface properties.
| Band Region | What It Reveals | Example |
|---|---|---|
| Blue (490 nm) | Water penetration, coastal mapping | Shallow bathymetry |
| Red (665 nm) | Chlorophyll absorption | Vegetation health via NDVI |
| Red Edge (705–783 nm) | Vegetation stress | Early crop disease detection |
| NIR (842 nm) | Vegetation structure | Biomass estimation, water boundaries |
| SWIR (1610, 2190 nm) | Moisture content | Soil moisture, burn severity, mineral mapping |
Why this matters more than you think: Two satellites could have identical spatial resolution, but if one has SWIR bands and the other doesn't, their analytical capabilities are fundamentally different. SWIR bands enable soil moisture estimation, geological mapping, and burn severity assessment that visible/NIR bands simply cannot provide.
Sentinel-2's spectral resolution is a key reason it dominates environmental monitoring despite having "only" 10m spatial resolution. Those 13 carefully placed bands enable a range of analysis that many commercial 0.5m satellites with only 4 bands cannot match.
3. Temporal Resolution
What it means: How often the sensor captures the same location on Earth.
| Satellite | Revisit Time |
|---|---|
| Sentinel-2 (A+B combined) | ~5 days |
| Sentinel-1 | ~6 days |
| Landsat 8 + 9 combined | ~8 days |
| Commercial constellations | 1–3 days |
For many applications, temporal resolution is the most limiting factor:
Agriculture: Crop growth stages change on weekly timescales. Miss a critical observation window (flowering, early stress), and the analysis loses value regardless of spatial resolution.
Disaster response: After a flood or earthquake, you need the first available image — hours and days matter more than pixel size.
Change detection: Detecting deforestation, urban expansion, or seasonal patterns requires consistent repeat observations. A single very-high-resolution image tells you what an area looks like today, but not how it's changing.
Cloud interference: Over tropical regions, any given optical observation has a ~70% chance of being cloud-contaminated. With 5-day revisit, you might get one clear view per month. With 16-day revisit, you might wait months. This is why SAR (which penetrates clouds) is so valuable in tropical monitoring.
4. Radiometric Resolution
What it means: The number of brightness levels the sensor can distinguish — essentially, the precision of each measurement.
| Sensor | Radiometric Resolution |
|---|---|
| Sentinel-2 | 12-bit (4,096 levels) |
| Landsat 8/9 | 12-bit (4,096 levels) |
| Older Landsat (5, 7) | 8-bit (256 levels) |
Higher radiometric resolution means the sensor can distinguish subtle differences in reflected or emitted energy. This matters for:
- Water quality: Detecting slight differences in chlorophyll concentration across a lake
- Shadow analysis: Distinguishing between deep shadow and dark surfaces
- Quantitative analysis: Any application that requires precise reflectance values rather than just visual interpretation
In practice, modern sensors (12-bit or better) have sufficient radiometric resolution for almost all applications. This was more of a limitation with older 8-bit sensors where dark areas like shadows and deep water all collapsed into the same few pixel values.
The Tradeoff Triangle
Satellite engineers face a fundamental tradeoff: you cannot maximize all four resolutions simultaneously. Higher spatial resolution means a narrower swath (less area covered per pass), which reduces temporal resolution. More spectral bands require longer integration times, which can limit spatial resolution or coverage rate.
This is why different satellites exist for different purposes:
| Priority | Best Fit |
|---|---|
| High spatial, few bands | Commercial VHR for defense/mapping |
| Many bands, moderate spatial | Sentinel-2 for environmental monitoring |
| All-weather, moderate spatial | Sentinel-1 SAR for flood/disaster/ice |
| Very high temporal | Geostationary satellites for weather |
Practical Decision Framework
When starting a new satellite monitoring project, ask these questions in order:
- What change timescale am I observing? This determines temporal resolution needs
- What surface property do I need to measure? This determines spectral resolution (and whether you need SAR or optical)
- What is the smallest feature I need to detect? This determines spatial resolution
- Do I need to work in cloudy regions? If yes, SAR becomes essential regardless of other considerations
Most beginners start with question 3 (spatial resolution) and work backwards. Experienced analysts start with questions 1 and 2.
Off-Nadir Delta provides access to Sentinel-1 SAR and Sentinel-2 optical data — a combination that covers most environmental monitoring needs with strong temporal, spectral, and spatial resolution across both radar and optical domains.
