Wait, what? A satellite brochure says, “Spectral resolution: 10 nm.” A student looks at the number and says, “That satellite can see objects only 10 nanometres apart on the ground.” The unit is tiny, the word is “resolution,” and the conclusion feels natural. But two different kinds of scientific detail have just been mixed together.
This Reality Lab teaches one exact evidence-transfer job: when a remote-sensing specification gives spectral resolution, keep it attached to wavelength detail—not ground spatial detail. A sensor can distinguish narrow wavelength bands while still having large ground pixels. Another sensor can show smaller ground features while measuring broader wavelength bands.
Quick Answer
Spectral resolution describes how finely a sensor separates wavelength information—often through the number, width and placement of spectral bands. Spatial resolution describes the ground area represented by a pixel or the spatial detail that can be distinguished in an image. They are different properties.
USGS remote-sensing guidance distinguishes spatial, spectral, temporal and radiometric resolution. Therefore, “10 nm spectral resolution” does not mean “10 nm ground pixels,” and “5 m spatial resolution” does not mean the sensor separates wavelengths 5 m apart. The same word—resolution—needs its scientific qualifier.
The Owned Learner Job
When a product page, infographic, news graphic or data catalogue says a sensor has “high resolution,” ask: high resolution in what sense?
- Name the resolution type before reading the number.
- Read the unit and ask what physical quantity that unit belongs to.
- Separate wavelength detail from ground detail, time frequency and signal-level sensitivity.
- Match the claimed advantage to the resolution type that could actually support it.
- Check the sensor band definitions and product metadata before comparing instruments.
This page does not own the physics of electromagnetic radiation, satellite orbits or image processing. It applies PSLE Science evidence reasoning to a real-world technical specification.
Build an Original Sensor Comparison
Imagine two fictional Earth-observation sensors. These values are constructed for learning and are not copied from a commercial specification.
| Sensor | Spectral bands | Typical band width | Spatial pixel size | Claim |
| A | Many narrow bands | 10 nm | 30 m | “More spectral detail” |
| B | Fewer broad bands | 60 nm | 5 m | “More ground detail” |
Sensor A has the finer wavelength sampling in this simplified comparison. Sensor B has the finer spatial sampling. If someone says “A has better resolution” without naming the kind of resolution, the statement is incomplete.
The important scientific habit is not to hunt for the smallest number. It is to ask what is being resolved?
Observed, Claimed and Inferred
| Layer | Statement | Evidence status |
| Reported | Spectral resolution or band width = 10 nm | Wavelength information is divided on roughly that spectral scale under the stated specification |
| Reported | Spatial resolution = 30 m | A spatial scale associated with the image product or sensor |
| Supported claim | Sensor A can separate finer wavelength intervals than a sensor with much broader bands, if definitions are comparable | A spectral comparison |
| Unsupported leap | Sensor A can distinguish two ground objects 10 nm apart | Confuses wavelength interval with ground distance |
| Unsupported leap | The sensor with the narrowest bands always makes the sharpest-looking map | Image sharpness and ground detail are not determined by spectral resolution alone |
Use the general canonical skill How to Read Units, Scales and Measurement Resolution Before Using PSLE Science Data whenever a familiar word carries different scientific meanings.
Unit Check: Nanometres of What?
A nanometre is a unit of length. That fact alone does not tell you whether the length is on the ground, inside an instrument, or along the electromagnetic spectrum. In a spectral-resolution specification, nanometres refer to wavelength spacing or band width. The value belongs to the wavelength axis, not automatically to the Earth-surface distance axis.
This is a classic evidence trap: the same unit dimension can appear in different scientific roles. A learner must preserve the quantity name, not only the unit.
Spatial Resolution: A Different Question
Spatial resolution is about how the sensor samples or represents space on the ground. A 30 m product does not mean every object inside a 30 m pixel is identical, nor does it mean the sensor pinpoints one exact 30 m object. The existing Reality Lab Vol.062 owns the real-world job of interpreting a 30 m spatial pixel.
Here we make a different comparison: spectral detail can become finer while spatial pixels remain the same size. Conversely, spatial pixels can become smaller while the sensor keeps broad spectral bands.
The Four Resolution Families
| Resolution family | Learner question | Example unit or description |
| Spatial | How finely is the ground divided or distinguished? | metres per pixel / ground sampling scale |
| Spectral | How finely is wavelength information separated? | nanometres; number and width of bands |
| Temporal | How often is the same place observed or a product updated? | minutes, days, revisit interval |
| Radiometric | How finely can differences in recorded signal be represented? | bit depth or radiometric sensitivity description |
USGS uses these categories because “resolution” is not one universal score. A sensor can be strong for one job and less suitable for another.
Comparison Check: More Bands Is Not Automatically Better
Suppose Sensor C has 200 narrow bands and Sensor D has 10 broad bands. It is tempting to say C is simply “better.” But useful evidence depends on the task. Narrow bands may help distinguish subtle spectral features. Broad bands may provide strong signal for another purpose. Data volume, signal-to-noise, calibration, atmospheric effects, coverage, spatial scale and processing can also matter.
The existing Reality Lab Vol.452 owns the specific “more bands = automatically better sensor” trap. This page routes to it rather than re-teaching that job.
Band Centre Is Not the Whole Band
A sensor catalogue may label a band by a centre wavelength such as 665 nm. That does not necessarily mean the sensor measured only one exact wavelength. Real bands cover a range described by a response curve or bandwidth. Reality Lab Vol.497 owns that distinction.
Why does this matter here? Because a spectral-resolution claim should be interpreted from the actual band width and response, not merely from a single number printed beside the band name.
Worked Case 1: Finding a Small Pond
Sensor A has 10 nm spectral bands but 30 m spatial pixels. Sensor B has 60 nm bands but 5 m spatial pixels. A tiny pond is only 8 m across. Which sensor is automatically better for finding its outline?
The answer is not “A, because 10 is smaller than 60.” Those numbers refer to different quantities. For the pond’s spatial outline, Sensor B’s finer ground sampling may be more useful. Spectral characteristics still matter for separating water from other surfaces, but the spectral-width number does not become a ground-size number.
Worked Case 2: Separating Similar Vegetation Signals
Now imagine two plant conditions that have subtle differences in reflectance over a narrow wavelength region, while the fields themselves are large. Fine spectral information may matter more than very small ground pixels. Sensor A could be useful for the evidence job even though its spatial pixels are larger.
This is the deeper lesson: the scientific question decides which resolution matters.
Worked Case 3: The Advertisement
A brochure says: “10 nm resolution—ten times more detailed than a 100 nm sensor.” That may be a reasonable spectral statement if the definitions and wavelength ranges are comparable. But if the brochure shows a sharper map and implies the ground image is ten times sharper, the visual has crossed from a spectral claim into a spatial claim without supplying spatial evidence.
Representation Check: A Colour Image Can Hide the Spectrum
Many satellite images shown to the public use three displayed colours even when the sensor collected more bands. Conversely, a colourful false-colour image can combine selected wavelength bands into red, green and blue display channels. The displayed colours are a representation; they are not a direct inventory of how many spectral bands were measured.
For false-colour interpretation, route to Reality Lab Vol.335.
Method and Metadata Check
- What wavelength range does the sensor cover?
- What exactly does the source mean by spectral resolution: band width, sampling interval, resolving power or another definition?
- What are the centre wavelengths and response curves?
- What is the spatial resolution of the relevant band or product?
- Are some bands collected at different spatial resolutions?
- Has the image been resampled or pan-sharpened?
- What calibration and quality information accompanies the data?
USGS Sentinel-2 documentation is a useful real example because different bands have specified centre wavelengths, bandwidths and spatial resolutions. The table structure itself demonstrates why wavelength and ground sampling must remain separate columns.
Processed Images Can Add Another Layer
A product can be resampled to smaller output pixels without creating new native spatial information, and pan-sharpening can combine higher-spatial-resolution information with multispectral data without meaning every colour band was originally measured at the sharpened pixel size. Those are separate Reality Lab jobs:
- Vol.395 — resampled pixel size is not new native detail
- Vol.398 — pan-sharpened output does not mean every band was measured at that spatial resolution
Alternative Explanations for a ‘Sharper’ Scientific Result
If one image or classification looks better than another, do not immediately attribute the improvement to spectral resolution. Other explanations may include:
- smaller spatial pixels
- better signal-to-noise
- better atmospheric correction
- different illumination or season
- more suitable wavelengths for the target
- better calibration
- a different algorithm
- less cloud or haze
- better reference data
- display sharpening or resampling
A scientific claim should identify which change actually produced the improvement, not simply point to one impressive specification.
What Evidence Strengthens or Weakens the Claim?
| Claim | Evidence that helps | Evidence that does not finish the job |
| Sensor separates finer wavelength features | Comparable band widths/response functions and validation | Small ground pixels alone |
| Sensor maps smaller ground objects | Spatial resolution plus object contrast and validation | Narrow spectral bands alone |
| Sensor classifies a target better | Task-specific validation on representative data | One resolution number |
| Image contains native 5 m information | Native sensor/product documentation | A 5 m resampled display grid |
Tempting but Invalid Reasoning
- “10 nm resolution means 10 nm ground detail.” The nanometres describe wavelength detail here.
- “Smaller number always means better sensor.” Only after the resolution type and task are matched.
- “More spectral bands means a sharper image.” Spectral richness and spatial sharpness are separate.
- “5 m pixels mean 5 m spectral resolution.” Metres and nanometres here refer to different axes.
- “A colour image shows only three measured bands.” The display may be built from a larger multispectral dataset.
How Far Can the Conclusion Travel?
A careful conclusion is: “This sensor has finer wavelength-band detail under the stated spectral specification.” To claim it resolves smaller ground objects, inspect spatial resolution. To claim it revisits more often, inspect temporal resolution. To claim it distinguishes smaller signal differences, inspect radiometric performance.
The broader owner How to Learn Earth Observation and Remote Sensing carries the deeper remote-sensing mechanisms. Reality Lab stays with the evidence-transfer problem.
PSLE-Style Transfer Case
A fictional sensor catalogue says: “Band width = 15 nm; spatial resolution = 20 m.” A student writes, “Objects must be at least 15 nm apart for the satellite to see them.”
Repair: The 15 nm value belongs to wavelength separation. The 20 m value is the listed spatial scale. The student must not move a number from one scientific axis onto another.
Now change the catalogue: “Band width = 15 nm; spatial resolution = 5 m.” The spectral-resolution statement is unchanged, but the spatial evidence is different. A correct answer tracks which part of the claim changed.
Delayed Independent Return
After a short break, explain this without looking back: Sensor X has 5 nm spectral bands and 30 m pixels. Sensor Y has 50 nm bands and 3 m pixels. Which sensor has finer spectral resolution? Which has finer spatial resolution? Why is it scientifically wrong to call one simply ‘ten times more detailed’ without naming the evidence job?
Explained Practice
- “Spectral resolution = 20 nm, so each pixel is 20 nm wide.” Wrong. Wavelength interval and ground pixel size are different quantities.
- “Sensor A has more bands, so it must see smaller houses.” Not established. Check spatial resolution and target contrast.
- “A 10 m product was resampled to 2 m, so native spatial resolution improved fivefold.” Not necessarily. Resampling does not create new measured spatial detail.
- “Two sensors both say high resolution.” Incomplete. Identify whether the source means spectral, spatial, temporal or radiometric resolution.
Routes to Existing Canonical Owners
- How to Read Units, Scales and Measurement Resolution Before Using PSLE Science Data
- How to Identify What Evidence a PSLE Science Question Actually Gives You
- Reality Lab Vol.062 — interpreting spatial resolution
- Reality Lab Vol.451 — swath width is not pixel resolution
- Reality Lab Vol.452 — band count is not a universal quality score
- Reality Lab Vol.497 — central wavelength is not one exact measured wavelength
Parent and Tutor Teaching Guide
Draw four boxes labelled space, wavelength, time and signal levels. Give the child specification cards such as “10 m,” “10 nm,” “5-day revisit,” and “12-bit.” Ask the learner to place each card in the correct box before discussing which is ‘better.’
Then give two fictional sensors with crossed strengths: one better spectrally, one better spatially. Ask which sensor is better for mapping tiny ponds and which might be better for detecting a narrow spectral feature in large fields. The child must justify the choice by matching evidence to task, not by choosing the smallest printed number.
Authoritative Sources
- Singapore Examinations and Assessment Board — 2026 PSLE Science syllabus
- Ministry of Education Singapore — 2023 Primary Science Teaching and Learning Syllabus
- U.S. Geological Survey — Introduction to Aquatic Remote Sensing
- U.S. Geological Survey EROS — Sentinel-2 archive and band specifications
The Quiet Habit
Scientific words often come in families. “Resolution” is one of them. Do not let the shared word erase the qualifier. Ask: resolution of where, what wavelength, what time, or what signal level? Once the evidence keeps its correct axis, the claim becomes much easier to judge. That is PSLE Science reasoning doing real work.
