Wind profiling describes how wind changes with height and time. A radar wind profiler does not directly photograph a column of arrows: it records returned signals, estimates radial velocities and combines those observations through geometry.
The gold standard keeps that chain visible. State the beam directions, sign convention, sampled heights and averaging period before interpreting the profile. This guide focuses on Doppler radar profiling, while recognising that other instruments can also measure winds aloft.
eduKateSG · Wind Profiling
Find your next learning step
Follow a beam measurement into a wind vector, then test height, timing, shear and data quality.
Full chapter index · Practice and transfer · Education root · How X Works
Full chapter index
Define the observation · Chapters 1–2
Recover the motion · Chapters 3–4
Read height and time · Chapters 5–7
Check and transfer · Chapters 8–10
Choose further reading · Chapters 11–12
CHAPTER 1 OF 12 · Define the observation
1. A profile has a vertical and a time coordinate
A wind profile is a set of wind estimates at different heights for a stated place and period. It may show speed and direction, or horizontal components such as eastward u and northward v.
Before comparing two profiles, inspect the height reference and reporting interval. A level reported above ground is not the same as the same number above sea level. An hourly estimate is not an instantaneous gust.
Begin a learner’s sketch with a vertical axis, a time label and a coordinate key. These small additions prevent a detailed-looking diagram from being interpreted without its physical frame.
A Doppler radar estimates motion along its line of sight. A wind vector perpendicular to that beam can have a zero radial component while its full speed is nonzero. Multiple suitably directed observations are needed to recover more components.
Radar profilers commonly use vertical and tilted beams. Combining them assumes that their sampled motions can represent a common wind field during the observation. The beams sample different volumes, so rapidly varying or horizontally uneven flow can challenge the reconstruction.
Keep three stages separate: the returned signal, the radial-velocity estimate and the transformed wind vector. A problem in an earlier stage does not disappear when the final display has arrows.
CHAPTER 3 OF 12 · Recover the motion
3. Worked example: recover one horizontal component
Use a two-dimensional teaching model. Let θ be beam tilt from vertical, u positive eastward and w positive upward. Define radial velocity positive along the outgoing beam. For an eastward-tilted beam, vr = u sinθ + w cosθ.
Take θ = 30°, vr = 5 m/s and w = 0. Then u = 5/sin30° = 10 m/s. The radial measurement is only half the horizontal component because of the geometry.
Now let w = 1 m/s with the same measured vr. Then u = [5 − cos30°]/sin30° ≈ 8.27 m/s. Assuming zero vertical motion would overestimate u in this example. These invented values illustrate projection, not a recommended instrument configuration.
For a westward-tilted beam at the same angle, vr,W = −u sinθ + w cosθ under the same convention. Subtracting the two measurements removes the common vertical term: u = (vr,E − vr,W)/(2sinθ).
Let vr,E = 6 m/s and vr,W = −4 m/s at 30°. Then u = 10 m/s. Their sum gives w = (6 − 4)/(2cos30°) ≈ 1.15 m/s in this ideal common-flow model.
Write the signs before inserting values. An instrument product may define positive velocity differently. Reversing the convention reverses the signs in the equations; it does not mean the physical wind reversed. Also retain the common-flow assumption when interpreting the recovered components.
CHAPTER 5 OF 12 · Read height and time
5. Range gates are not automatically height gates
A radar estimates range from round-trip travel time: r = cΔt/2. Under straight-beam, flat local geometry, the height increment above the instrument is z = r cosθ for tilt θ measured from vertical.
For a hypothetical slant range of 1,000 m and θ = 30°, z ≈ 866 m. Treating slant range as height would misplace this sample by about 134 m. More complete processing may account for additional geometry.
A gate represents a sampled volume, not a dimensionless point. Closely spaced plotted levels do not alone establish independent vertical resolution. Check the actual instrument and processing documentation.
| Display quantity | Check first | Why it matters |
|---|---|---|
| Height | Reference and beam geometry | Slant range is not vertical height |
| Wind arrow | Components and direction convention | Motion toward differs from wind from |
| Time label | Start, end and averaging interval | A mean is not a peak gust |
CHAPTER 6 OF 12 · Read height and time
6. Vector shear is more than a speed difference
Wind shear describes change of the wind vector with position. In a simple vertical comparison, use Δu/Δz and Δv/Δz, or the magnitude of the vector difference divided by height separation.
Suppose the lower wind is (u,v) = (5,0) m/s and the upper wind is (0,5) m/s, 500 m higher. Both speeds are 5 m/s, so their scalar speed difference is zero. But the vector difference is (−5,5) m/s, with magnitude √50 ≈ 7.07 m/s.
The finite-difference shear magnitude is about 0.0141 s⁻¹. Equal speeds therefore do not imply no shear. This is a mathematical example, not a flight-safety assessment or an operational threshold.
CHAPTER 7 OF 12 · Read height and time
7. Average vectors before interpreting the period
Consider two equal-duration winds: (10,0) m/s and (−10,0) m/s. Their vector mean is (0,0), while their mean speed is 10 m/s. The two summaries answer different questions.
A small mean vector can hide substantial changing motion. Likewise, an hourly profile can conceal shorter changes. Always identify what was averaged and which variability is still available.
Meteorological wind direction normally names where wind comes from, unlike a vector arrow that may indicate where air moves. Label the convention rather than asking readers to infer it from an unlabeled compass symbol.
CHAPTER 8 OF 12 · Check and transfer
8. Quality control starts with the returned signal
Atmospheric echoes can be mixed with returns from birds, insects, aircraft, nearby objects or interference. A prominent spectral peak is not automatically the desired air-motion signal. NOAA’s instrument description explicitly treats signal selection as part of profiling.
Useful checks compare neighbouring times and heights, beam consistency and independent observations. They should also preserve genuine atmospheric changes rather than rejecting every sharp feature as an error.
For a teaching audit, flag a suspicious point and write two hypotheses: a real change in wind, or a non-atmospheric contribution. Then state which additional evidence would distinguish them. Smoothness is a clue, not proof.
CHAPTER 9 OF 12 · Check and transfer
9. A local profile is not a complete weather forecast
A profiler adds observations aloft near its site. Meteorological interpretation places those observations alongside other measurements and atmospheric models. One column does not establish conditions everywhere nearby or at every future time.
Comparisons need matched sampling: a model grid value, an aircraft observation and a radar-derived average may cover different volumes and periods. Disagreement can reflect representativeness as well as instrument error.
When the question concerns building comfort or turbine loading, connect the profile with the relevant engineering model. A wind estimate is an input to that model, not a complete decision on its own.
First, set w = 0, θ = 20° and u = 12 m/s in the eastward-beam model. Predict vr. Second, compare winds (6,0) and (0,8) m/s across 200 m. Calculate the finite-difference vector shear magnitude.
Checks: vr = 12sin20° ≈ 4.10 m/s. The vector difference is (−6,8), magnitude 10 m/s, giving 10/200 = 0.050 s⁻¹. Carry units through both calculations.
Finally, explain what changes if the tilted beams sample different winds. The equations can still produce numbers, but the recovered vector may not represent one common air motion. This is the important transfer test: correct algebra does not validate an unsuitable observation model.
CHAPTER 11 OF 12 · Choose further reading
11. Connected radar and atmospheric reading
Follow the measurement
- Radar — rebuild range, Doppler, clutter and component geometry.
- Weather radar and rainfall estimation — compare a wind retrieval with a different radar-derived product.
Place the profile in context
- Meteorology — connect observations, forecasts, uncertainty and verification.
The broader Radar root also links to specialist Doppler-folding and two-projection guides. Choose those when the difficulty is the measurement equation; choose Meteorology when it is the interpretation across space and time.
- NOAA PSL: wind profiling instruments — beams, range gates, spectral processing and the transformation into winds.
- NOAA CAP profiler documentation — reporting intervals, time stamps and quality-controlled products.
- NOAA profiler training manual 2 — quality control of wind-profiler data.
All numerical examples are idealised teaching cases. They do not provide current Singapore weather, aviation guidance or equipment settings. Source-specific network descriptions are not universal specifications for every profiler.
The learning goal is a short, checkable explanation from radial observation to wind vector, including the assumption that makes the conversion possible. Return to the Education root to practise carrying that explanation into an unfamiliar sensor problem.
Follow the optical route to wind profiles
The Gold Standard of Wind Lidar. Use Doppler shifts in laser returns, scan geometry and platform corrections to understand wind lidar and its measurement limits.