Step through real satellite observations, release an air parcel into a measured atmosphere, and discover why moisture, lifting and the surrounding air all matter.
Enable JavaScript to change the conditions and run the interactive experiment.
Becoming cloudy does not guarantee continued ascent. What the parcel meets above—and how far it is lifted—can change its path.
A parcel is cloudy and still moving upward, but the surrounding layer makes its buoyancy negative. What follows?
Velocity and acceleration differ. The first turning point depends on remaining motion and work along the connected path; cloud formation alone does not establish a thunderstorm.
Three satellite observations show a featured cloud developing over Western Australia. Move between the frames and identify the broad change. Coverage gaps and different viewing geometries mean these are not identical cameras tracking the same individual droplets.
An air parcel is a selected amount of air we follow conceptually. Fronts, terrain and other processes can provide upward displacement. Our experiment supplies the initial lift explicitly, then investigates what can happen after release.
The parcel stays at surrounding pressure. As pressure falls during ascent, it expands and follows an approximate adiabatic cooling path. This cooling begins before the first visible cloud cue.
The parcel eventually reaches its lifting condensation level. Some vapor can condense into droplets. The pseudoadiabatic branch represents how condensation changes the cooling path, while removing condensate from the thermodynamic parcel.
Buoyancy depends on the parcel relative to surrounding air. Water vapor affects that density comparison too. Virtual temperature is a useful way to account for vapor; the model includes it even when the detailed values are hidden.
In the authored warmer-layer case, condensation occurs near 830 m, but the first upward turning point is near 920 m after release at 250 m. Those are outputs of this particular calculation, not universal cloud heights.
A parcel can still move upward while negative buoyancy slows it. It might cross a thin unfavorable interval using motion gained earlier. If its remaining upward-motion energy reaches zero first, this ascent ends there.
A full thunderstorm also involves processes such as precipitation, downdrafts, ice, winds and electrical activity. This lesson isolates part of the updraft story. Neither saturation nor the top of our displayed column establishes lightning or a storm forecast.
The pinned source is NOAA IGRA station USM00072357, nominal May 18, 2026, 00 UTC. Heights are relative to its first valid 345 m geopotential surface. The separate 2300 release-time field is preserved without inventing a release date. A rising, drifting radiosonde is approximated as a frozen vertical profile.
Temperature and dew point are interpolated linearly with source height; logarithmic pressure is interpolated linearly. The final source sample brackets the 5 km model limit. Missing and quality-removed required values were excluded rather than changed to zero.
Before saturation, the approximate parcel temperature is T₀(p/p₀)^(Rd/cp), with constant initial vapor mixing ratio. The condensation intersection is solved numerically. Above it, a pressure-coordinate liquid-water pseudoadiabatic equation is integrated with RK4 steps no larger than 50 Pa.
dT/dp = (Rd T + Lv rs) / {p[cp + Lv² rs ε/(Rd T²)]}. Here rs is saturation mixing ratio, ε = 0.622, Rd = 287.05 J/(kg K), cp = 1004 J/(kg K), and Lv = 2.5 × 10⁶ J/kg. Constants and liquid-water treatment are approximations, not a complete mixed-phase cloud model.
r measures kilograms of vapor per kilogram of dry air. For the vapor/dry-air mixture without condensate, Tv = T(1 + r/ε)/(1 + r). Buoyancy is B = g(Tv,parcel − Tv,environment)/Tv,environment. The units, moisture basis and omission of condensed-water loading matter.
Signed buoyancy work A is the height integral of B, using a 10 m grid with linear buoyancy between points. After release at zᵣ, E(z) = 2 + A(z) − A(zᵣ), in J/kg. Two J/kg corresponds to the specified 2 m/s upward start. The first inaccessible zero ends the path; favorable conditions farther above cannot be reached by skipping that barrier.
The warm-layer calculation has positive signed net work up to 5 km, yet its 250 m release stops much lower. Summing a favorable region aloft does not show that this parcel can reach it. The displayed truncated signed integral is not full CAPE or CIN.
The counterfactual adds a triangular temperature increment from 500 to 2,500 m, peaking at +6 K at 1,500 m. It preserves measured pressure, height and surrounding vapor ratio. This is a sensitivity experiment, not a newly balanced atmosphere, measured cap or climate scenario.
The drier case lowers the initial parcel dew point by 6 K and holds the parcel temperature and surrounding profile fixed. Its vapor ratio and density change. Its condensation level is higher, around 1,585 m, and its 250 m release stops below that level in this calculation.
Replay progress is parameterized by height along the connected upward path. It is not elapsed atmospheric time. We stop at the first turning point, because reversing a condensate-removing ascent would not provide a valid descent model.
A warm afternoon alone does not reveal the full vertical profile. Upper-air observations help establish what a lifted parcel may encounter, alongside the many additional measurements used in actual forecasting.
A balloon samples the air along its path; a satellite observes a broad area; an astronaut photograph can reveal side structure. Combining evidence requires their dates, positions and limitations, rather than treating every picture as the same event.
The selected ISS image dates to 2008, the satellite sequence to 2020 and the sounding to 2026. Those are observation dates, not the discovery dates of convection. IGRA’s evolving integration and quality-control work is described in its 2018 methods paper.
Compare the three supplied images before checking their times. Locate the featured cloud; note one observed change and one thing an image cannot tell you. Missing coverage is not dark weather.
Draw a vertical strip. Mark 250 m release, about 830 m condensation, about 920 m warm-case stop, and 2,500 m higher release. These are declared model quantities, not observed cloud heights.
Use one marker for the measured environment and another for the warmer-layer trial. An upward-moving parcel can slow before it stops. Compare the supplied outcomes rather than treating every downward buoyancy arrow as immediate descent.
Move only the warm-case release to 2,500 m. Explain how the imposed lift changes access to the air above. Then compare the drier parcel separately and name which part changed.
Write a conclusion about conditions that affect continued ascent. Do not turn that result into a prediction about tomorrow, rain amount, lightning or an exact storm height.
How can a parcel become cloudy yet fail to keep rising?
Original indoor paper activity using supplied observations and model results. No outdoor storm watching, smoke, hot water or pressure vessels. The model is not a weather warning system.
Conditions allow some vapor to condense. Saturation and continued ascent are different questions.
It expands as surrounding pressure falls. The approximate adiabatic path starts before condensation.
It loses enough upward-motion energy to reach a turning point. The changed surroundings alter buoyancy; the tinted layer is still air.
It slows, and can cross the layer or stop. Motion and acceleration have different roles.
Vapor ratio and density change; condensation occurs higher here. The parcel intervention leaves surrounding measurements unchanged.
The imposed lift places the parcel beyond the earlier barrier. The forcing that supplied this lift is external to the model.
Real cloud development illustrates a broader story; the calculation uses a different observation. Different places, dates and instruments must remain distinguishable.
It continues through this modeled range under the assumptions. A chosen numerical boundary is not a physical forecast.
Ingredient framework and density comparisons. The lesson does not adopt simplified wording as a quantitative forecast or use the page’s erroneous thermodynamics-law wording to derive lift.
National Weather Service · thunderstorm ingredientsDevelopment, precipitation, updrafts and downdrafts. Context for the processes omitted from our numerical calculation.
National Weather Service · thunderstorm developmentDurre et al. (2016), Integrated Global Radiosonde Archive. Norman USM00072357, nominal 2026-05-18 00 UTC, retrieved 2026-09-09. Selected raw record, flags, bracketing rows and hashes are preserved locally.
NOAA NCEI · IGRA version 2IGRA source integration, processing and quality control; quality checks do not make observations exact or remove all instrument and timing differences.
Durre et al. (2018) · original methods paperEquation 10 saturation-vapor-pressure approximation. Our declared constant-coefficient dry path and numerical condensation intersection are not the complete Bolton treatment.
Bolton (1980) · original researchmoist_lapse documents and implements the equation attributed to Bakhshaii and Stull (2013). Brytalearn independently integrates the displayed equation with specified constants; it does not claim the reference software’s validation for the complete scene.
Unidata MetPy v1.7.1 · primary implementationWeather and Forecasting 9:625–629. The effect of neglecting virtual-temperature correction. Supports keeping vapor’s density effect in the calculation; not validation of this selected 2026 trial.
Doswell & Rasmussen (1994) · original paperWestern Australia, January 14, 2020. Terra/MODIS around 11 a.m., Aqua/MODIS around 1 p.m., Suomi NPP/VIIRS around 2 p.m. NASA/Lauren Dauphin; original coverage gaps retained, optional crop disclosed.
NASA Earth Observatory · growth of a summer stormISS Expedition 16 photograph captured February 5, 2008; NASA/JSC Image Science & Analysis Laboratory and ISS Crew Earth Observations. It provides visible structure, not a geometric reconstruction or altitude measurement.
NASA · ISS016-E-27426Educational/informational use with acknowledgment; no endorsement implied. The selected media are identified as NASA-origin assets. This permission is not labeled as a Creative Commons license.
NASA · image and media-use guidelinesUse the supplied materials indoors. The activity does not require outdoor observation or photography. The lesson is not a warning system.
National Weather Service · lightning safetyIndependent subject review is pending.
Read the sources and model assumptions
Interactive explorationFollow water above and below the surface. Pave part of the land, change evaporation, and watch the water budget rearrange.
Interactive explorationFreeze two model temperatures, change infrared transfer and release time. Follow every energy route, move the accounting boundary and compare the model with real atmospheric evidence.
Interactive explorationGive water an energy budget. Watch its temperature and phase masses, catch a melting plateau, compare two starting states and inspect an open ice structure.