Doppler radar can report folded radial wind velocities when true motion exceeds the Nyquist interval set by its sampling. A colour reversal may be an ambiguity, not an abrupt physical wind reversal.
The gold standard records raw velocity, pulse metadata, quality masks and an evidence-supported unfolding alongside independent verification.
eduKateSG · Weather Radar Velocity Dealiasing
Find your next learning step
Work Nyquist folding, compare PRF and range, and examine what makes a reliable unfolding.
Full chapter index · Practice and transfer · Education root · How X Works
Full chapter index
Understand Doppler velocity · Chapters 1–2
Calculate aliasing · Chapters 3–4
Evaluate correction models · Chapters 5–7
Verify the interpretation · Chapters 8–10
Read connected science · Chapters 11–12
CHAPTER 1 OF 12 · Understand Doppler velocity
1. Radial motion is the Doppler radar measurement
Doppler weather radar measures a component of target motion along the radar beam. The observable is radial velocity, which follows the path toward or away from the antenna according to the adopted sign convention. It is not automatically a complete three-dimensional wind vector.
Precipitation, insects or other scatterers can return a signal. Interpreting that signal as wind requires understanding the type of scatterer and how closely its motion represents the surrounding air.
The radar samples pulses at a finite repetition rate. The phase change used to infer motion is ambiguous when it falls outside the measurable interval; a fast inbound or outbound signal can appear with the opposite sign.
NOAA’s Doppler-pattern guide explains this wrapping as velocity aliasing. A dramatic colour shift in a radar product can therefore be a sampling artefact rather than a physical discontinuity in the atmosphere.
The gold standard preserves the raw velocity, sign convention, location, time, Nyquist interval and the separate corrected estimate so that meteorological interpretations can be challenged later.
CHAPTER 2 OF 12 · Understand Doppler velocity
2. Nyquist velocity depends on wavelength and PRF
For a simple monostatic Doppler radar using a uniform pulse train, the Nyquist radial-velocity magnitude is V_N = λ·PRF/4, where λ is wavelength and PRF is pulse repetition frequency.
The observed velocity generally lies within an interval approximately −V_N to +V_N, while true speeds outside it can be folded by integer multiples of 2V_N. A folded measurement is therefore one of several possible true radial velocities.
Raising PRF increases the unambiguous velocity interval under this model, but can reduce the unambiguous distance because a new pulse might be emitted before a distant echo returns. The two constraints form the Doppler range–velocity tradeoff.
More complicated operational radar pulse sequences and processing can address parts of this tradeoff, but their acquisition and dealiasing assumptions must be stated.
Do not reuse the simple formula without checking whether a product uses multiple pulse repetition frequencies, a nonuniform Nyquist interval or other special processing. A file labeled ‘velocity’ does not supply all this information by itself.
CHAPTER 3 OF 12 · Calculate aliasing
3. Worked example: +36 metres per second folds negative
Assume a hypothetical radar with Nyquist interval ±30 m/s. The interval’s full width is 2V_N = 60 m/s. A true radial velocity of +36 m/s lies above the +30 m/s boundary.
Subtract one full interval: 36 − 60 = −24 m/s. The radar may therefore report −24 m/s even though the target’s real radial motion was +36 m/s.
The NSSL Doppler guide uses this particular numerical illustration. It shows that the sign of an uncorrected velocity display can be misleading when aliasing occurs.
Conversely a measured −24 m/s could represent a true −24 m/s or +36 m/s, and in principle other values separated by integer multiples of 60 m/s. A single gate cannot uniquely identify which is correct.
A more general expression is true radial velocity = observed value + 2kV_N for integer k. Choosing k is an inference requiring other information, not a property visible from the observed number alone.
CHAPTER 4 OF 12 · Calculate aliasing
4. Worked example: the opposite fold and a continuity guess
Within the same ±30 m/s interval, a true inbound velocity of −36 m/s may appear as +24 m/s because −36 + 60 = +24. Positive and negative wraps are symmetrical under the chosen model.
Suppose neighbouring valid range gates appear near +28 m/s while one measured gate shows −24 m/s. A correction method assuming local continuity might favor +36 m/s for the isolated gate, since the difference from the neighbours is then 8 rather than 52 m/s.
That preference is plausible but not guaranteed. A genuine front, strong convergence, shear or nearby opposing air mass could create a real large change. A smoothing assumption cannot be used as a universal physical law.
Always inspect a broader spatial and temporal context, including neighbouring beams and any independent wind information. A pretty smooth map made by forcing all gates together may be wrong.
The aim of dealiasing is an evidence-supported unfolding, not simply the removal of inconvenient colour boundaries.
CHAPTER 5 OF 12 · Evaluate correction models
5. Worked example: pulse repetition versus unambiguous range
A simple pulse radar’s maximum unambiguous range is Rmax = c/(2PRF), where c is approximately 3×10⁸ m/s. Combine that with V_N = λPRF/4 for a fictional uniform pulse train.
Choose wavelength λ = 0.10 m and PRF = 1,000 s⁻¹. Then V_N = 0.10×1,000/4 = 25 m/s. The range is Rmax = 3×10⁸/(2,000) = 150,000 m, or 150 km.
Increase PRF to 1,500 s⁻¹. The Nyquist magnitude becomes 37.5 m/s, while the unambiguous range drops to 3×10⁸/(3,000) = 100 km.
The product of V_N and Rmax stays λc/8 in this idealized model. Thus increasing one limit reduces the other for an unchanged wavelength and uniform sampling approach.
Actual WSR-88D acquisition schemes can use pulse techniques and corrections beyond this simple calculation. The numbers are invented and must not be treated as configuration advice for a weather radar.
CHAPTER 6 OF 12 · Evaluate correction models
6. Region-based dealiasing needs more than one gate
Region algorithms identify connected radar areas where measured radial velocities are relatively consistent, then evaluate possible shifts by whole Nyquist intervals to join them into a wider unfolded field.
The open-source Py-ART implementation of region-based dealiasing documents the use of velocity regions, merging and an optional reference velocity field that anchors the otherwise ambiguous solutions.
The method relies on data quality and some smoothness of relevant physical motion. A region separated by missing or clutter-contaminated gates may not have enough evidence to determine its correct integer fold.
A radar scan with nonuniform Nyquist velocities can require different treatment from one uniform pulse train. Proper metadata about each ray or gate must be available to the correction procedure.
Store the original data, excluded gates, reference wind profile, method parameters and software version. An unlabelled corrected product is hard to verify when algorithms are later updated.
An external atmospheric profile can help anchor an interpretation but must be time- and location-matched to the beam and radial component.
CHAPTER 7 OF 12 · Evaluate correction models
7. True wind shear can resemble a velocity discontinuity
Wind changes with height and position, and radar beams usually sample higher air as range increases. The spatial velocity field can therefore vary for real atmospheric reasons even when the target motion is correctly measured.
NOAA’s Doppler-pattern guide illustrates radial signatures of wind changing with height and shows that aliased data can create abrupt changes across the colour scale. Both situations can be present in one weather system.
Convective boundaries, gust fronts, shear zones and storm circulations may contain genuine strong gradients. A dealiaser that forces all such gradients to be small may infer incorrect wind structures.
Different radar elevations and nearby azimuths can provide context about whether a surprising sign change is a plausible physical feature or a sampling wrap. The evidence still has limits where scans are incomplete.
An algorithm may make opposite errors: wrongly unfolding a true gradient or failing to unfold a genuinely aliased region. Both should count in a serious performance assessment.
Do not use one unexamined dealiased gate as proof of a dangerous circulation or direction reversal.
CHAPTER 8 OF 12 · Verify the interpretation
8. Clutter, range folding and weak signals deserve separate masks
Velocity aliasing differs from range folding. Range folding incorrectly assigns a target’s distance when a return is received after subsequent pulses. Changing PRF may affect both velocity and range ambiguity.
Ground clutter, insects, birds and other nonmeteorological returns can contribute radial velocities that have little relation to the wind being estimated. Some targets may have near-zero velocity even when the wind nearby is strong.
Low signal-to-noise, beam geometry and incomplete sampling can make individual gate velocities unreliable. A correction method should mark poor data instead of manufacturing a plausible value solely from the adjacent field.
Modern radar processing can include clutter filtering, multi-PRF approaches, gap handling and velocity unfolding. The NOAA Radar Operations Center documents a two-dimensional velocity-dealiasing method among its radar techniques.
Quality control and dealiasing should be evaluated as linked but distinct operations. A corrected field that quietly excluded difficult observations may look more successful than its complete acquisition actually was.
CHAPTER 9 OF 12 · Verify the interpretation
9. Multiple pulse repetition intervals add information
A radar can use distinct pulse repetition intervals to observe different ambiguous representations of a target’s motion. Comparing those measurements can help resolve the true velocity within an extended feasible range.
This does not eliminate noise, missing measurements, clutter, atmospheric shear or range ambiguity. An algorithm may still need a spatial or temporal reference and may be unable to resolve some gates confidently.
The simple +36-to−24 example assumes one uniform Nyquist cointerval. A multi-PRF strategy has different sampling details, so blindly applying 2V_N using an assumed single number can produce incorrect results.
NOAA ROC documentation describes pulse repetition frequency selection and operational dealiasing techniques; their method-specific documentation is needed to understand a real radar product.
Scientific reporting should include the acquisition mode alongside Nyquist and correction metadata. Otherwise two velocity analyses made by different modes may not be comparable.
A robust result also has an explicit uncertain or uncorrected status when observations are insufficient for a unique solution.
CHAPTER 10 OF 12 · Verify the interpretation
10. Worked example: score a corrected radial field
Take four fictional independent radial-velocity references in m/s: [20,35,−10,−38]. Suppose a dealiased product, correctly matched to the same beams and times, estimates [21,32,−11,−35].
With error defined as estimate minus reference, the differences are [+1,−3,−1,+3] m/s. Their mean is (1−3−1+3)/4 = 0 m/s, despite every individual estimate having some error.
Mean absolute error is (1+3+1+3)/4 = 2 m/s. RMSE = sqrt[(1²+3²+1²+3²)/4] = sqrt5 ≈ 2.24 m/s.
The arithmetic is only an illustration. A valid field test must use genuinely independent wind observations projected onto each radar line of sight, with appropriate spatial, temporal and vertical matching.
Large mistakes by whole Nyquist interval widths may be rare and easily hidden by average scores. Evaluate their frequency, the fraction of rejected gates and performance in different weather regimes.
An algorithm’s smoothness is not an independent validation metric: a real storm may contain sharp changes.
CHAPTER 11 OF 12 · Read connected science
11. Practice, FAQs and changed-weather transfer
Practice A: With ±30 m/s Nyquist, confirm that true +36 can look like −24 m/s while true −36 can look like +24.
Practice B: At λ = 0.10 m and PRF = 1,000 Hz, calculate Nyquist magnitude 25 m/s and unambiguous range 150 km. Repeat at 1,500 Hz.
Practice C: A gate reads −24 m/s and neighbours read +28. Explain why a +36 correction is plausible under continuity but still needs independent checks.
FAQ — Does radar directly measure total wind speed? No. It observes the radial velocity of scatterers along the radar beam.
FAQ — Does a sign change always imply a gust front? No. Folding or invalid returns may create misleading displays.
FAQ — Are all dealiased gates guaranteed correct? No. Ambiguity can remain when independent information is absent.
Transfer the example to a radar scan with variable PRF, missing gates and strong real wind shear. Explain which original assumptions break down.
Verified eduKateSG routes for related weather-radar topics:
- The Gold Standard of Weather Radar Rainfall Estimation — reflectivity-derived rain rather than Doppler motion
- The Gold Standard of Weather Radar Convective Cell Tracking — track motion of reflectivity objects
- The Gold Standard of Weather Radar Dual-Polarization Hydrometeor Classification — polarimetric class inference
- The Gold Standard of Meteorology — real winds and convective gradients
- The Gold Standard of Atmospheric Science — vertical wind structure
- The Gold Standard of Statistics — error assessment and uncertainty
- The Gold Standard of Weather Forecast Verification — comparison design and scoring
CHAPTER 12 OF 12 · Read connected science
12. Official radar science and next learning step
NOAA’s Doppler guide explains sampling ambiguity, Nyquist intervals and the velocity–range tradeoff. Its chapters include worked radar examples, while Py-ART documents how region-based dealiasing handles connected observations and optional reference fields.
All radar velocities, pulse rates and comparison scores above are fictional or textbook-style teaching numbers, not live atmospheric conditions or forecasts of hazardous weather.
For a reproducible project retain raw velocity, per-ray Nyquist interval, scan timing, validity masks, dealiasing settings, corrected velocity, reference observations and independently evaluated errors.
Original NOAA radar references and implementation guidance:
- NOAA NSSL Doppler Velocity Patterns: Chapter 1 — Nyquist aliasing and the ±30-m/s worked example
- NOAA NSSL Doppler Velocity Patterns: Chapter 2 — radial patterns and wind shear
- NOAA Radar Operations Center: Radar Techniques — advanced dealiasing and PRF selection
- Py-ART Region Dealiasing Documentation — region joining and optional external velocity reference
- Py-ART API: dealias_region_based — method description and parameters
The immutable Clementi Mathematics teaching reference is retained as an instructional model, not a radar operations authority.