Frequently Asked Questions

Physics The simulation uses the Logarithmic Wind Profile, a standard atmospheric boundary layer model widely adopted in wind engineering and by the World Meteorological Organization (WMO).

Where z₀ is the surface roughness length (selectable in the left panel), z_ref = 10 m is the standard meteorological reference height, and U_ref is the mean wind speed at that height. This means wind is slower near the ground (where friction dominates) and faster at higher altitudes — which is critical for modeling how snow particles behave at different heights.

You can see this profile visualized in real-time in the right panel's Log Wind Profile chart.

Physics Real wind is never steady — it fluctuates constantly. The simulation models this using a Turbulence Intensity (TI) parameter, which represents the ratio of wind speed fluctuation to the mean speed.

The gust factor is generated by superimposing two sine waves at different frequencies, creating a pseudo-random fluctuation that mimics the gusty nature of real wind. While this is a simplification compared to full spectral analysis (e.g., Kaimal spectrum used in advanced wind engineering), it effectively captures the essential behavior: periods of stronger and weaker wind that cause snow drift intensity to vary over time.

Each particle also receives an additional local random turbulence component, so not all particles move in perfect unison — just like real snowflakes.

Barrier When a snow particle enters a barrier zone (defined by its position, width, and height), its horizontal velocity is reduced by a factor of (1 − density). This models the Wind Shadow Effect — vegetation disrupts airflow, causing wind-borne snow particles to decelerate and settle near the barrier rather than being carried onto the road.

Additionally, particles passing through dense barriers fall more slowly (their downward velocity is capped), simulating how foliage and branches create air resistance that slows particles in all directions.

This mechanism is consistent with the principles used in designing snow fences and shelterbelts in cold-climate transport infrastructure.

Barrier Density is the inverse of optical porosity. A density of 0.8 means the barrier blocks 80% of the wind passing through it, allowing only 20% to continue at full speed. This directly corresponds to how dense the foliage is.

• Low density (0.1–0.3) — Open structure, like a sparse shrub row or picket fence. Reduces wind speed moderately.
• Medium density (0.4–0.6) — Semi-porous, like a well-maintained shelterbelt. Often considered the optimal range for snow control, as it reduces wind without creating large leeward vortices.
• High density (0.7–1.0) — Nearly solid, like a dense conifer hedge. Blocks most wind but can cause snow to accumulate immediately behind the barrier rather than dispersing.

You can adjust this slider and observe how the snow drift pattern changes in real-time.

Barrier Tree height determines the vertical extent of the wind shadow. Only snow particles below the barrier height are affected by the velocity reduction. Particles above the treeline continue at full wind speed.

This is why short shrubs (2–3 m) have limited effect on snow driven aloft during strong winds, while 8–10 m conifers can intercept snow across a much larger cross-section of the wind flow. In the simulation, increasing tree height from 3 m to 10 m produces a visible increase in the protected zone.

In practice, the protected downwind distance of a shelterbelt is approximately 10–15× the barrier height, which is why the "Distance from road" parameter also matters — the barrier must be positioned so that its wind shadow covers the road surface.

Metric The simulation defines two driver-perspective view frustums:

• Forward Visibility — Right lane, looking ahead (−Z direction), 5 m wide × 3 m tall × 40 m deep
• Reverse Visibility — Left lane, looking behind (+Z direction), same dimensions

Each animation frame, the system counts how many snow particles are inside each frustum. More particles = more visual obstruction = lower visibility.

Where max_blockers is scaled to the total particle count. This approach is conceptually analogous to optical extinction — in real-world meteorology, visibility is measured by how much suspended particles scatter and attenuate light. Our particle-density method is a 3D discrete approximation of this principle.

Metric The monitor panel uses several intuitive visual elements designed for non-technical audiences:

• SVG Gauge Meters — Semi-circle arcs showing forward/reverse visibility as a percentage, color-coded green (GOOD, ≥60%), yellow (MODERATE, 30–60%), or red (POOR, <30%).
• Barrier Protection Score — A prominent card showing the overall visibility improvement percentage (e.g., "+52%") compared to a no-barrier baseline.
• Shield Comparison Bars — Green bars (🛡️ With Barriers) side-by-side with red bars (🚫 Without Barriers), showing how much safer each lane is.

The no-barrier baseline is estimated by tracking particles intercepted by barriers each frame. Without barriers, these particles would have continued at full wind speed into the driver's visibility zone. The formula:

This approach lets you adjust any barrier parameter and immediately see how the protection score changes — far more intuitive than reading raw numbers.

Physics The default settings are chosen to represent a typical winter blizzard in the Saguenay–Lac-Saint-Jean region:

• Wind speed: 15 m/s (54 km/h) — Environment Canada historical data shows Chicoutimi winter storms frequently reach 40–70 km/h sustained winds
• Terrain roughness: 0.03 m (Open Terrain) — Appropriate for agricultural areas and open highway corridors typical of the Route 175 / Boulevard Talbot corridor
• Turbulence Intensity: 15% — Standard for open terrain in moderate-to-strong winds
• Barrier: Conifer, 8 m height, density 0.8 — Representative of mature spruce (épinette) commonly planted as shelterbelts in Québec

Users can adjust all parameters to simulate different scenarios — from calm snowfall to extreme blizzard conditions.

Physics The roughness length (z₀) characterizes how much the ground surface slows down wind. The simulation offers five presets:

• Smooth (0.01 m) — Frozen lakes, ice, very flat snow-covered fields
• Open Terrain (0.03 m) — Open farmland, highway corridors with minimal obstacles
• Agricultural (0.1 m) — Farmland with hedges, scattered buildings, low crops
• Suburban (0.3 m) — Residential areas, small towns, mixed land use
• Forest / Urban (1.0 m) — Dense forest edges, urban centers with tall buildings

For a highway in the Saguenay region, Open Terrain (0.03 m) or Agricultural (0.1 m) are the most appropriate choices. The lower the roughness, the faster the wind near the ground, and the more critical vegetation barriers become.

Metric Snow Intensity controls the total number of particles simulated (from 10,000 to 100,000). Higher values produce denser snowfall and more statistically stable visibility readings, but require more GPU/CPU resources.

• 10k–30k — Light snowfall, runs smoothly on most devices
• 50k — Moderate blizzard (default), good balance of realism and performance
• 80k–100k — Heavy blizzard, may cause frame drops on older devices

If the simulation feels sluggish, reducing the particle count is the most effective way to improve performance. The physics and barrier effects remain identical regardless of particle count — only the visual density and metric precision change.

Barrier Based on established shelterbelt design principles and what the simulation demonstrates:

• Porosity: 40–50% (density 0.5–0.6) — Allows enough wind through to prevent snow accumulation directly behind the barrier, while still significantly reducing wind speed over the road
• Height: 6–10 m — Tall enough to protect the full driver visibility zone (0–3 m height)
• Distance: 10–20× tree height from road edge — The wind shadow maximum protection occurs at roughly 5–10H downwind; placing the barrier too close can cause snow to deposit on the road itself
• Both sides — Enable barriers on both the upwind (primary protection) and downwind sides for variable wind directions

Try toggling between different configurations in the simulation to see how the "No Barriers" baseline comparison changes — this is the most intuitive way to understand the trade-offs.

Physics Wind direction controls which side's barriers are most effective. At 0° (crosswind, West to East), the left-side (West) barrier directly intercepts wind-blown snow before it reaches the road — this is the scenario where barriers make the biggest difference.

As the wind shifts toward parallel to the road (±90°), snow blows along the road rather than across it, reducing the barrier effectiveness but also reducing the cross-road drift problem itself.

The "No Barriers" baseline estimation accounts for wind direction — the wind_angle_factor ensures the comparison is larger for crosswind conditions and smaller for parallel winds. This matches real-world observations that crosswinds are the most dangerous for snow drift on highways.

Scope CFD tools like OpenFOAM or ANSYS Fluent solve the full Navier-Stokes equations and can produce highly accurate flow fields — but they require hours of computation on powerful hardware for a single configuration, and their results are not interactive.

SnowSim takes a fundamentally different approach: it prioritizes real-time interactivity over numerical precision. The particle-based approach uses simplified but physically grounded rules (logarithmic wind profile, barrier porosity reduction) to produce qualitatively correct results that respond instantly when you adjust a slider.

The two approaches are complementary:

• CFD — For detailed engineering analysis, regulatory compliance, and academic research
• SnowSim — For rapid exploration, stakeholder communication, educational demonstrations, and intuitive understanding of barrier effects

Scope Being transparent about limitations is important for proper interpretation:

• Flat terrain only — No hills, embankments, or road cuts are modeled. Real terrain features significantly affect wind flow patterns.
• Simplified turbulence — Uses sinusoidal gust approximation rather than stochastic spectral models (Kaimal, von Kármán).
• No temperature or moisture effects — Snow crystal type, humidity, and temperature affect real drift behavior but are not modeled.
• Baseline estimation is approximate — The barrier protection score is estimated from intercepted particle statistics, not from a parallel barrier-free simulation.
• No vehicle aerodynamics — Passing vehicles create turbulence and snow re-suspension, which is not included.
• Capped snow accumulation — Settled snow is limited to 2,000 particles for performance; long-duration progressive drift buildup is not modeled.

Despite these simplifications, the qualitative behavior — that vegetation barriers reduce wind-driven snow reaching the road and improve driver visibility — is consistent with established research and field observations.

Scope SnowSim is designed as an interactive demonstration and communication tool, useful in several contexts:

• Stakeholder Communication — Show municipal officials, transportation departments, or community members how different tree-planting strategies affect road safety. A live demo is far more persuasive than static charts.
• Educational Use — Teach students about atmospheric boundary layers, wind engineering, and infrastructure resilience in a hands-on, visual way.
• Rapid Scenario Comparison — Quickly explore "what-if" questions: What if we plant taller trees? What if we add barriers on both sides? What if the wind is from a different direction?
• Pre-study Scoping — Before commissioning expensive CFD studies or field measurements, use the simulation to identify the most promising barrier configurations to investigate further.
• Community Engagement — Let residents interact with the tool to understand why road-side tree planting is being proposed and how it benefits winter driving safety.

Barrier The simulation models Black Spruce (Picea mariana), the most common boreal conifer in the Chicoutimi / Saguenay–Lac-Saint-Jean region. Key characteristics:

• Brown cylindrical trunk — approximately 20% of total height, 12–16 cm radius
• Dark green conical crown — approximately 85% of total height, starting at ~15% above ground
• Narrow, tapered shape — the crown narrows toward the top, which is physically accurate for boreal spruce

The two-part geometry (trunk + crown) means that changing tree height only stretches the tree vertically — the width stays constant, so density is not artificially exaggerated when you increase height.

The cone-shaped collision zone also follows this geometry: the effective blocking width is wider at the crown base and narrows toward the tip, and the foliage density decreases toward the top. This matches real conifer structure — lower branches are denser, upper branches are sparser.

Physics Yes. When snow particles are slowed by a vegetation barrier and then reach the ground, they settle in place as persistent white dots instead of being recycled. This models the real-world phenomenon of snow deposition on the lee side of shelterbelts.

• Only particles that were actually intercepted by a barrier accumulate — free-falling snow simply recycles
• The accumulation pool is capped at 2,000 particles to maintain smooth performance
• Accumulated snow appears as small white dots near the base of barriers
• More snow accumulates on the windward (upwind) side, consistent with shelterbelt research

While this is a simplified representation (no drift shapes or progressive mounding), it provides an important visual cue that barriers are working — you can literally see the snow being caught before it reaches the road.

Scope Yes, SnowSim can be installed and used completely offline in two ways:

• PWA (Progressive Web App) — Visit the live simulation in Chrome, Edge, or Safari, then click "Install" or "Add to Home Screen". This works on Windows, macOS, Android, and iOS — no app store needed.
• Desktop Installer — Download pre-built installers from the GitHub Releases page:
  • .dmg for macOS
  • .exe for Windows

Both options include all simulation code and assets — no internet connection required after installation. This is especially useful for field demonstrations, conference presentations, or stakeholder meetings where network access may be limited.