Fog is the most underspecified atmospheric element in AI video prompting. Most fog prompts say "foggy morning" or "misty atmosphere" and receive a generic grey haze that has no depth, no light interaction, and no physical logic. The difference between that and cinematically useful fog is always the same set of physical facts: how far objects are from the camera determines how much the fog obscures them; what light source is present determines whether the fog scatters that light as god rays or wraps it as a soft diffuser; and what kind of fog it is — ground fog, valley fog, sea fog, jungle mist — determines the height of the layer and how it moves.
Fog is not a single look. It is a physics system that behaves differently with every light source, every surface, and every distance. A sunrise over a misty lake eliminates all hard shadows by converting the sun into an enormous soft diffuser. A lighthouse beam sweeping through maritime fog makes the beam itself visible as a solid rotating column — because the fog particles scatter the light at 90° to the beam direction. Morning mist across the Great Wall unifies fifteen rapid aerial cuts by applying a consistent atmospheric haze to every shot's colour temperature. Jungle mist combined with low-angle sunlight produces the crepuscular rays that no other atmospheric element can. And neon in urban fog creates the depth-gradient colour separation that aerial landscape photographers use in mountains — except here the depth markers are buildings and their signs rather than mountain ridges.
These five Seedance fog prompts each demonstrate a different structural use of fog physics as a cinematic tool. None of them says "foggy atmosphere" and hopes the model gets it right. Each specifies the fog type, the light interaction, and the spatial relationship between camera, fog, and subject — and that specificity is why each one produces footage that reads as atmospheric and directed rather than hazy and generic.
1. The Monet lake — morning mist as omnidirectional soft diffuser
See the full prompt on scenic.sh →
"Monet impressionist oil painting style, thick and loose brushwork, soft blending, hazy edges. First-person perspective hands rowing through a lake surrounded by weeping willows and water lilies, transitioning from enclosed garden to open water at sunrise."
Why this works: At 95 likes, this Monet impressionist lake prompt demonstrates what happens when the style reference and the atmospheric physics align. Impressionism is not primarily a brush technique — it is a way of seeing diffuse, scattered light. Monet spent years painting the same subjects at different times of day precisely because diffuse morning mist changes how every surface reads: shadows disappear, colour saturation drops by 40–60%, and edges soften to the point that near-field subjects separate from background only by tonal value, not by contrast or edge sharpness. That is also a physical description of what morning mist over a still lake actually does to light.
The prompt uses the Monet style reference to compress an entire light-and-atmosphere specification into one phrase. The model's training data associates "Monet impressionist style" with: low-contrast diffuse light, saturated colour applied in loose strokes that soften edges, a warm-neutral palette at dawn, and water surfaces that read as colour planes rather than reflective mirrors. This style reference loads the fog physics automatically. The "hazy edges" and "soft blending" are the descriptive equivalents of the physical process — the fog's fine water droplets scatter each light ray slightly, producing the soft edge where the weeping willow meets the water.
The first-person rowing perspective is what makes this atmospherically functional rather than purely decorative. At water level, morning mist is densest just above the water surface (cold air drains to the lowest point), so the near-field is hazy while the weeping willows above the waterline separate more clearly from the sky. This height stratification — denser fog at the waterline, lighter mist above — is what gives lake morning fog its characteristic two-register look: the surface feels enclosed and private while the canopy above is open to the lightening sky. The prompt encodes this by placing the camera at water level and naming the willows as the overhead element.
The transition from "enclosed garden to open water" is a fog revelation device. In the enclosed garden section, the weeping willow branches overhead create a leafy tunnel that limits the visible sky and concentrates the mist between the foliage. As the rower exits into open water, the visible sky expands and the mist thins — the fog horizon moves to the far bank rather than the next lily pad. This spatial transition mirrors the emotional opening that the scene conveys: the fog-physics change (from enclosed diffuse to open hazy) and the feeling of the scene (intimate to spacious) are the same event described at two levels.
Takeaway: Style references for impressionist painters activate fog-and-diffuse-light physics automatically — the model's training data associates those styles with specific atmospheric qualities (soft edges, low contrast, diffuse directionless light). At water level, name the stratification of morning mist by height: denser at the water surface, lighter above the waterline, clearing toward the sky. A spatial transition from enclosed to open (garden tunnel to open lake) is a fog-density transition: the same fog changes behaviour when the overhead canopy disappears.
2. The lighthouse keeper — fog as a medium that makes light visible as an object
See the full prompt on scenic.sh →
"It is the end of the lighthouse keeper's shift. An older man in a thick wool sweater slowly climbs the spiral staircase toward the lantern room carrying a steaming mug and a weathered logbook."
Why this works: At 58 likes, this psychological horror lighthouse prompt demonstrates the single most cinematically powerful function of fog: converting a light beam from an invisible illumination plane into a visible, tangible object. In clear air, a lighthouse beam illuminates whatever it strikes — but the beam itself is invisible from the side. In fog, every particle of water vapour in the beam's path scatters the light at 90° to the beam direction, which is exactly toward the camera. The beam becomes visible as a solid rotating column of light sweeping through the darkness.
This is the physics the prompt builds its horror around. The lighthouse beam is not described in the prompt text excerpt — it is implied by the setting. But the night-scene analysis of this prompt in the Seedance gallery identifies it precisely: "a practical light — the rotating lighthouse beam — combined with fog as a visible diffusion medium. The beam sweeps through fog as a moving shape rather than a flat illumination plane." The key insight is that the beam's regular mechanical sweep gives the fog a rhythm. The fog column doesn't glow steadily — it pulses with each sweep of the beam. And in the prompt's horror context, "the second light's appearance" (a non-mechanical, irregular light source) is disturbing precisely because it breaks that mechanical rhythm: something in the fog is generating light at an irregular frequency.
Maritime fog is the densest naturally occurring fog type, generated by warm, moist ocean air flowing over cold coastal water. The density that makes maritime fog visually distinctive is what makes the lighthouse beam most visible — denser fog = more scattering particles per cubic metre = brighter, more defined beam column. The windswept lighthouse on black cliffs is a spatial specification that positions the fog source (the ocean) below and to one side of the structure, so fog rolls in horizontally at structure-height rather than rising from below. This is what makes lighthouse fog dramatic rather than just grey: it arrives in rolling horizontal waves that interact with the structure rather than filling it from the ground up.
The keeper's details — thick wool sweater, steaming mug, weathered logbook — are atmospheric props that establish the human register before the fog physics take over. They mark the setting as lived-in and physically real, which makes the fog's subsequent behaviour more unsettling. A character who is credibly cold and tired in a credibly damp environment makes the lighthouse beam's behaviour more alarming than the same events in a stylized setting.
Takeaway: Fog makes light beams visible as objects — name the light source and the fog together so Seedance renders the beam as a solid sweeping column rather than flat illumination. Maritime fog is horizontally rolling (moves in from the ocean at structure-height, not rising from below) — position the camera to see the waves arriving rather than the fog sitting still. The rhythm of a mechanical light source through fog (regular sweeps) makes any irregular light source disturbing by contrast — the fog's mechanical pulse is what you're violating when you introduce the horror element.
3. The Great Wall aerial — morning mist as a multi-shot colour coherence system
See the full prompt on scenic.sh →
"Cinematic 15-second historical landmark tour of the Great Wall of China at sunrise composed of 15 rapid 1-second shots, each cut cleanly with smooth visual continuity, ultra-realistic ancient stone wall stretching across mountainous terrain..."
Why this works: At 35 likes, this Great Wall aerial demonstrates one of the most useful structural functions of morning mist in multi-shot sequences: it is the only naturally occurring atmospheric element that applies a consistent light quality, colour temperature, and contrast level to every shot in a sequence regardless of camera angle. The aerial drone analysis of this prompt identifies the mechanism precisely: "The consistent 'soft sunrise lighting with warm golden tones and morning mist' instruction, applied globally, keeps the light coherent across what would otherwise look like 15 separate clips."
The coherence problem in rapid multi-shot sequences is severe: fifteen 1-second shots cut together from different camera angles and altitudes each capture a slightly different specular environment. An aerial wide shot and a close-up orbit shot of the same wall section at the same moment would show different reflections, different shadow directions, and different colour temperatures from slightly different sun angles. Morning mist eliminates this problem by imposing a diffusion layer over every shot: the sun's colour temperature is shifted to warm grey-blue by the mist, hard shadows are eliminated (so shadow direction doesn't vary by angle), and the distant mountain ranges are softened to tonal planes rather than hard forms. The fifteen shots become fifteen framings of a single atmospheric state rather than fifteen disconnected captures.
The mist also encodes height in these aerial shots. At high altitude, the Great Wall sections below the camera are partially obscured by the mist layer — a peak emerges above the mist while the valley is filled with the fog bank. This is the topographic function of valley fog: it inverts the visual priority of the landscape by hiding the valleys and revealing only the elevated structures. The Great Wall's ridgeline route, which follows the mountain crests, is thus rendered by the fog: sections on ridges emerge from the mist, sections in valleys disappear. The fog is drawing the Great Wall's route on the landscape for the camera.
The "rapid 1-second shots" structure creates a kinetic pace that would fragment badly without the mist's unifying effect. At 15 cuts per 15 seconds, each shot has no time to establish its own atmosphere. Mist solves this by making atmosphere a global constant rather than a per-shot decision — the viewer reads the atmospheric state immediately because it's the same in every frame, and the rapid cuts create rhythm without disorientation.
Takeaway: Specify morning mist as a global atmospheric constant at the top of multi-shot prompts — it imposes colour temperature, contrast level, and haze density consistently across all angles so the model doesn't generate different atmospheric states per cut. Valley fog inverts landscape priority: it fills the low points and reveals the ridgelines, which makes elevated structures (walls, peaks, towers) emerge from the fog naturally. At high pace (15+ cuts in 15 seconds), atmospheric coherence prevents visual disorientation — mist is the most reliable single-phrase solution.
4. The BBC Earth jungle — tropical mist as a scattering medium for god rays
See the full prompt on scenic.sh →
"Ultra-realistic cinematic jungle sequence, filmed like a premium wildlife documentary. Dense tropical rainforest at golden hour, sun rays piercing through mist, exotic birds flying overhead, elephants, deer, monkeys, and tigers moving naturally through the environment."
Why this works: At 24 likes, this BBC Earth jungle sequence demonstrates the god-ray technique — crepuscular rays made visible by the mist medium they travel through. In clear air, a shaft of sunlight entering a dark forest is invisible: the light illuminates everything it strikes, but the beam itself cannot be seen from the side. In humid tropical mist, every fine water droplet in the beam's path scatters the light at 90° toward the camera, making the beam visible as a solid column from first leaf gap to forest floor. "Sun rays piercing through mist" is a three-element specification: sun (directional light source at sufficient intensity), rays (multiple parallel shafts, implying gaps in the canopy overhead), and mist (the scattering medium that makes the shafts visible). All three are required; any two will not produce the effect.
The golden-hour timing is structurally essential, not decorative. At golden hour, the sun is at 5–15° above the horizon, which means its rays enter the canopy at a shallow angle. A shallow entry angle through vertical canopy gaps produces long horizontal shafts — the beam travels nearly parallel to the forest floor before striking it, giving the longest possible visible column through the mist. At midday, the sun's 70–80° elevation angle produces near-vertical shafts that are short in horizontal extent and visible only briefly before hitting the ground. The prompt's "golden hour" instruction encodes the long-shaft geometry without stating it explicitly.
The wildlife in this prompt — birds overhead, elephants, deer, monkeys, tigers — are not just scene dressing. They move through the god-ray shafts, which creates the key cinematographic event in a god-ray sequence: an animal passing through a shaft of mist-lit light and becoming briefly, brilliantly illuminated against the dark forest behind them. The mist scatters the light sideways onto the animal's coat or feathers at the moment of passing, producing an instant of rim-lit luminosity that the rest of the dark forest floor cannot replicate. The documentary register ("filmed like a premium wildlife documentary") instructs the model to place the camera far enough back that wildlife moves through the foreground while the shafts are visible in the mid-ground — the wildlife and the light shafts coexist as separate visual elements rather than one overwhelming the other.
The "dense tropical rainforest" specification matters for mist density. Tropical humidity saturates the air at near-100% relative humidity by dawn, and the canopy holds moisture from overnight rain. This gives jungle mist its fine-droplet character — smaller and more densely suspended than valley fog — which produces the cleanest scattering for god rays. Coarser fog particles scatter light too broadly, making the beam edges soft. Fine humidity-saturated mist produces the sharp-edged, defined shafts visible in BBC Earth footage.
Takeaway: God rays require three named elements — light source (directional sun), scattering medium (mist or fine humidity), and canopy gaps that break the beam into shafts. Golden hour at 5–15° elevation produces the longest horizontal god-ray geometry; midday produces short near-vertical shafts. Position the camera to show wildlife moving through the shafts — the moment of illumination (animal in the beam column against dark forest) is the visual event, not the shaft itself. "Dense tropical rainforest mist" specifies fine-droplet humidity that produces sharper, better-defined shafts than coarser valley fog.
5. The Gothic vigilante — urban neon fog as depth-gradient colour separation
See the full prompt on scenic.sh →
"A stormy midnight city covered in heavy rain. Towering skyscrapers disappear into thick fog while neon reflections shimmer across wet streets."
Why this works: At 14 likes, this Gothic vigilante sequence demonstrates the urban application of atmospheric perspective — the depth-gradient technique that landscape photographers use to separate mountain ridges by fog density, applied to city buildings and neon signs. The core physics is identical to landscape aerial perspective: objects at different distances from the camera receive different amounts of atmospheric haze between them and the lens. Near-field subjects (street level, the vigilante, neon signs on the ground floor of the closest building) are sharp and colour-saturated. Mid-ground subjects (buildings at 2–5 blocks distance) are partially obscured, with their neon signs reduced to soft coloured glows rather than legible text. Far-field subjects (towers several blocks away) disappear into the fog entirely, readable only as darker shapes against the lighter fog background.
But urban fog adds a dimension that landscape fog does not have: neon colour temperature variations by distance. The near-field neon (red, blue, green, amber shop signs at street level) is fully saturated and hue-accurate. At mid-ground distance, the fog scatters those colours sideways, so a red neon sign becomes a soft pink glow that bleeds beyond the sign's physical edges. At far-field, the neon reduces to a monochromatic warm or cool glow with no discernible hue — you can see a building has lit signs but not what colour they are. The fog doesn't just obscure the signs; it desaturates and broadens their colour halos by distance, which creates a colour-temperature gradient from the foreground (high saturation, multiple hues) to the background (low saturation, monochromatic warm glow).
The "skyscrapers disappear into thick fog" instruction is doing specific altitude work. Urban fog in a stormy midnight context sits as a dense layer from street level to somewhere between 100–300 metres, with the upper sections of skyscrapers emerging above the fog band or disappearing into it depending on their height. The instruction to have skyscrapers "disappear" specifies that the fog layer extends high enough to cut off the upper floors — this is the visual effect that makes a city look infinite: the towers have no visible tops, so the eye cannot establish their scale. "Disappear into thick fog" rather than "are partially obscured by fog" encodes the specific outcome: the buildings have no visible top edge.
The wet street surface multiplies every neon source in the scene. At street level, neon signs are visible directly and in reflection (the wet pavement mirror). The reflection adds a second copy of every sign, inverted, at a slightly different colour temperature (reflections shift toward the darker, cooler end of the sign's palette). The combination of heavy rain (creating the reflection surface), fog (diffusing and colour-grading by distance), and neon (the light sources being diffused and reflected) is the three-element night system this prompt encodes.
Takeaway: Urban atmospheric perspective works identically to landscape perspective — near-field subjects are sharp and colour-saturated, mid-ground neon reduces to soft coloured glows, far-field signs become monochromatic warm halos. Specify "disappear into thick fog" for tall buildings rather than "partially obscured" — disappearance removes the visible top edge and makes height unknowable, which makes the city read as infinite. Wet streets multiply every neon source via reflection; name both the neon colour and the reflective surface together so the model renders both the sign and its street-level double.
Fog & mist prompt cheat sheet
What these five prompts share across five different environments:
- Fog is a depth-stratified medium, not a uniform filter — objects at different distances from the camera receive different levels of haze, saturation, and edge sharpness. Name the near-field, mid-ground, and far-field treatments separately to produce a three-dimensional atmospheric scene rather than a flat grey overlay.
- Every fog type has a physical origin — morning lake mist (cold ground air), maritime fog (warm air over cold ocean water), valley fog (cold air draining into low terrain), jungle humidity mist (near-100% RH in closed canopy), urban storm fog (precipitation + cold surface). The origin determines the fog layer's height, density, movement, and how it interacts with available light.
- Fog and light source together produce the effect, never fog alone — diffuse sunrise light + morning mist = omnidirectional soft diffuser; lighthouse beam + maritime fog = visible rotating column; directional sunlight + jungle mist = god rays; neon signs + urban fog = depth-gradient colour separation. Specify the light source and its angle alongside the fog type.
- Fog is the most reliable multi-shot coherence tool — applied as a global atmospheric instruction at the top of a multi-shot prompt, morning mist imposes consistent colour temperature, contrast, and haze density across all angles and distances, preventing the atmospheric discontinuity that rapid cutting otherwise creates.
- Physical verbs matter more than aesthetic adjectives — "skyscrapers disappear into thick fog" produces a specific visual outcome (buildings with no visible top edge); "fog creates a moody atmosphere" delegates the outcome to the model. "Sun rays piercing through mist" specifies the physical event (beam entering scattering medium); "ethereal foggy light" does not.
→ Browse the Fog & Mist AI video gallery on scenic.sh for more prompts
→ For atmospheric depth in night scenes, see 5 Seedance Night Scene Prompts
→ For rain as a visual and light-interaction element, see the Rain & Wet Weather gallery
→ For the complete Seedance prompt technique guide: How to Write Seedance 2 Prompts
