Heights: Meteorological, Probe, and Pedestrian
Overview
Three distinct heights govern the ArchiWind workflow. They serve different roles and must not be confused with one another:
| Symbol | Name | Role |
|---|---|---|
| $z_{\text{met}}$ | Meteorological reference height | Height at which the wind data source’s statistics were recorded |
| $h_{\text{probe}}$ | Upstream CFD probe height | Height of the upstream reference probe placed inside the CFD domain |
| $z_{\text{proj}}$ | Pedestrian projection height | Height at which comfort and safety results are evaluated |
This page defines each height, how it is determined, and how they combine to produce the per-direction velocity scaling factor used across ArchiWind’s climate-weighted outputs. For the underlying log-law wind profile and per-direction roughness determination, see Inlet Boundary Conditions and Reference Point Handling.
1. Meteorological Reference Height ($z_{\text{met}}$)
$z_{\text{met}}$ is the height AGL at which the wind statistics were recorded at the meteorological station. It is a property of the wind data source, not of the CFD model.
| Wind data source | How $z_{\text{met}}$ is determined |
|---|---|
| GovMap CSV | User-specified at project setup from the available 30 m, 60 m, 100 m, or 150 m parameter blocks (default 150 m). The chosen value selects the correct block and becomes $z_{\text{met}}$ for the simulation. |
.tim and .tab files | Must be provided explicitly as a separate simulation parameter, since neither format carries height metadata. |
2. Upstream CFD Probe Height ($h_{\text{probe}}$)
$h_{\text{probe}}$ is the height AGL at which an upstream reference probe set is placed inside the CFD domain. It is a property of the simulation geometry and is entirely independent of $z_{\text{met}}$.
This probe-based scaling is applied specifically to IS 5281 simulations (see IS 5281 (Israeli Standard)). It is a second correction, layered on top of the per-direction roughness correction ($\alpha_\theta$) described in Inlet Boundary Conditions and Reference Point Handling: where $\alpha_\theta$ transfers the meteorological wind climate from the recording location to the target site using land-cover roughness alone, the probe-set scaling additionally checks that transfer against the CFD-resolved flow field itself, catching effects, such as terrain-induced distortion or domain-edge artefacts, that a purely analytical roughness ratio cannot capture.
Its purpose is to anchor the post-processing velocity scaling: by sampling the CFD-computed speed at points that lie in undisturbed upstream flow, the workflow establishes a per-direction ratio between the expected real-world wind speed and the normalised CFD speed.
Placement
A set of five probes is generated per wind direction, arranged across the inflow perpendicular to the wind direction at cross-wind offsets of $-45, -22.5, 0, +22.5, +45$ m from the direction’s centreline. Sampling five points instead of one guards against a single unrepresentative sample, for example a probe that happens to sit in a local wake or acceleration zone, skewing the scaling factor for that direction.
Each probe’s height above local terrain is determined as follows:
- Automated (default): $h_{\text{probe}} = 1.5 \times H_{\text{tallest}}$, where $H_{\text{tallest}}$ is the height of the bounding box of the built models (building or surroundings). The result is clamped to $[10\text{ m}, 150\text{ m}]$ AGL.
- User override: the operator may supply an explicit value, bypassing the $1.5\times$ rule, for example when a more precise height is known. The same $[10\text{ m}, 150\text{ m}]$ clamp still applies.
Each probe’s absolute elevation also varies per wind direction, placed at $z_{\text{terrain}}(x, y) + h_{\text{probe}}$, the local terrain elevation at that probe’s upstream location. This keeps every probe at a consistent height above ground regardless of topographic variation around the site.
Per-direction velocity scaling factor
For each direction $d$, the velocity scaling factor is computed by comparing the theoretical ABL log-law speed at $h_{\text{probe}}$, referenced to $z_{\text{met}}$, against the median streamwise CFD-simulated speed across the direction’s five probes:
$$ SF[d] = \frac{u_{\text{ABL}}(h_{\text{probe}}, z_{0,d})}{\text{median}\left(u_{\text{streamwise}}^{(1)}, \dots, u_{\text{streamwise}}^{(5)}\right)_d} $$
where
$$ u_{\text{ABL}}(h, z_0) = u_{\text{ref}} \cdot \frac{\ln\left(h_{\text{probe}} / z_0\right)}{\ln\left(z_{\text{met}} / z_0\right)} $$
| Symbol | Description |
|---|---|
| $u_{\text{streamwise}}^{(i)}$ | Component of probe $i$’s CFD-simulated velocity aligned with the incoming wind direction $d$ |
| $z_{0,d}$ | Upstream roughness length for direction $d$ (from land cover) |
| $z_{\text{met}}$ | Meteorological reference height (wind data source) |
Using the streamwise component rather than the raw velocity magnitude matters because probes placed near complex terrain or building massing can otherwise see appreciable cross-flow; taking the component aligned with $d$ keeps the scaling factor tied to the wind direction it is meant to represent.
$SF[d]$ is applied uniformly to all surface cell velocities in direction $d$ during the exceedance calculation.
The two heights appear together in the formula: $z_{\text{met}}$ sets the normalisation of the ABL profile, while $h_{\text{probe}}$ determines the specific elevation at which that profile is evaluated and compared to the CFD output.
Guardrails
Before $SF[d]$ is computed, ArchiWind validates the probe set for each direction and blocks the calculation if either check fails:
- Invalid probe location. Every probe must resolve to a valid point inside the fluid domain, not inside solid geometry or outside the mesh. A probe that cannot be sampled invalidates the direction’s scaling factor.
- Near-zero sampled velocity. Because the sampled CFD speed appears in the denominator of $SF[d]$, a probe velocity too close to zero, for example from a probe placed in a stagnation region, would produce an unphysically large or unstable scaling factor. Directions failing this check are rejected rather than silently propagating a bad factor into the exceedance calculation.
Diagnostics (quality control)
Independently of the guardrails above, ArchiWind reports the following non-blocking diagnostics per direction so users can judge how representative the upstream reference conditions are:
Streamwise spread (CV): coefficient of variation of the five streamwise probe speeds, computed as standard deviation divided by the median. Lower values indicate a more spatially uniform upstream flow across the probe line.
Minimum alignment: lowest ratio u_streamwise / |U| among the five probes.
Median alignment: median of the five u_streamwise / |U| ratios. Values closer to 100% indicate that the local flow is well aligned with the intended inflow direction; the median is less sensitive than the minimum to a single outlier probe.
These diagnostics do not stop the run. Values shown in orange in the generated report exceed the recommended upstream probe-set diagnostic limits and should be reviewed.
Example
The table below is taken from a real IS 5281 report and shows the scaling factor and diagnostics computed for each of the 12 wind directions:
| Wind direction | Median u_streamwise (m/s) | Scaling factor | Streamwise spread CV | Minimum alignment | Median alignment |
|---|---|---|---|---|---|
| N | 7.303 | 1.0799 | 0.4% | 100.0% | 100.0% |
| NNE | 6.954 | 1.1340 | 0.8% | 100.0% | 100.0% |
| ENE | 6.219 | 1.6409 | 0.3% | 99.9% | 99.9% |
| E | 6.282 | 1.8038 | 0.1% | 99.9% | 99.9% |
| ESE | 7.238 | 1.0895 | 0.1% | 99.9% | 99.9% |
| SSE | 7.562 | 1.0429 | 0.1% | 99.9% | 99.9% |
| S | 7.800 | 1.0110 | 0.1% | 99.9% | 99.9% |
| SSW | 7.938 | 0.9935 | 0.1% | 100.0% | 100.0% |
| WSW | 7.686 | 1.0261 | 0.4% | 99.9% | 99.9% |
| W | 7.461 | 1.0570 | 0.3% | 99.8% | 99.8% |
| WNW | 7.423 | 1.0623 | 0.9% | 99.8% | 99.9% |
| NNW | 7.578 | 1.0406 | 0.3% | 100.0% | 100.0% |
All twelve directions fall within recommended limits here (spread CV well under 1%, alignment above 99.8%), so none are flagged in the report. The scaling factor itself varies by direction because it combines two direction-dependent inputs, the CFD-sampled probe speed and the ABL profile evaluated with that direction’s roughness length $z_{0,d}$, so directions with similar probe speeds can still end up with noticeably different $SF[d]$ values.
3. Pedestrian Projection Height ($z_{\text{proj}}$)
$z_{\text{proj}} = 1.5$ m AGL is the height at which comfort and safety results are evaluated on all exposed surfaces, in accordance with IS 5281. This is where CFD velocity and turbulence fields are sampled to produce the exceedance maps. See Pedestrian Wind Comfort and Safety for how these fields are classified.