Wait, What? A Cloud Is Usually Full of Water That Is Too Small to Fall as Rain
A typical cloud contains enormous numbers of tiny droplets, yet many clouds do not rain. A cloud droplet can be only tens of micrometres across, while a raindrop is much larger.
How do microscopic droplets grow through the size gap into particles that fall faster than the surrounding air can support them?
That size-gap problem is cloud microphysics.
aerosol seed → activated droplet or ice crystal → growth → collision/riming/aggregation → precipitation-sized particle → fall through changing air below
The One-Sentence Answer
Learn cloud microphysics by separating cloud formation from precipitation formation: first explain how aerosols activate into droplets or ice, then follow the particle-size distribution as collision and ice processes create rain, snow, graupel or hail.
Stage 1: Saturation Is a Thermodynamic Condition
Cloud formation becomes possible when air reaches sufficient relative humidity and supersaturation. Cooling rising air is a common route, but saturation alone does not specify how many droplets form, how large they are or whether precipitation develops.
Stage 2: Homogeneous Condensation Is Hard in the Real Atmosphere
Water vapour does not easily create large droplets from nothing. Atmospheric aerosol particles provide surfaces that lower the barrier to droplet formation. Particles capable of activating under cloud supersaturation are called cloud condensation nuclei, or CCN.
Stage 3: Aerosol Size and Chemistry Both Matter
A larger soluble particle generally activates more easily than a tiny insoluble one. Important aerosol types include sea salt, sulfate, organics and dust. The cloud therefore inherits information from the aerosol population below it.
Stage 4: Köhler Theory Combines Curvature and Solute Effects
A tiny curved droplet has elevated equilibrium vapour pressure because of curvature. Dissolved solute lowers vapour pressure. Köhler theory combines these effects and predicts a critical supersaturation for activation.
Stage 5: Activation Is a Threshold Process
Below critical supersaturation, a wet aerosol tends toward an equilibrium size. Above it, continued condensation can produce a much larger cloud droplet. Small changes in updraft or aerosol population can therefore change droplet number strongly.
Stage 6: Updraft Controls Supersaturation
Rising air expands and cools, tending to create supersaturation. Growing droplets remove vapour and consume that supersaturation. Cloud state is a competition between cooling and condensation.
Stage 7: More Aerosol Can Produce More but Smaller Droplets
If similar liquid water is distributed among more activated CCN, average droplet size can fall. This is central to the Twomey effect: more numerous, smaller droplets can reflect more sunlight.
Stage 8: Aerosol Effects Are Not One Universal Cloud-Brightening Rule
A 2025 multimodel study found substantial differences among models in aerosol–cloud responses. Higher aerosol loading often suppressed initial warm-rain formation, but cloud adjustments also changed humidity, stability and convection.
Stage 9: Condensation Alone Is Too Slow to Make Large Raindrops Efficiently
Condensational growth is rapid for very small droplets but becomes relatively slow as drops grow. Clouds therefore need mechanisms that accelerate the transition from ~10 μm droplets toward millimetre raindrops.
Stage 10: Collision–Coalescence Builds Warm Rain
Larger droplets fall slightly faster, overtake smaller ones and sometimes merge. The larger droplet then falls faster still. This positive feedback can make rain in warm tropical clouds without any ice.
Stage 11: Not Every Collision Produces Coalescence
Collision efficiency depends on droplet size, relative velocity and airflow. Coalescence efficiency depends on surface properties and collision dynamics. Two droplets can collide and separate.
Stage 12: Turbulence Can Alter Collision Rates
Turbulent eddies create velocity fluctuations, clustering and relative motions that can alter collision probabilities. The exact importance depends on droplet size and cloud regime.
Stage 13: Raindrops Cannot Grow Without Limit
Large drops become aerodynamically distorted and can break apart. Real rain reaches a dynamic size distribution determined by coalescence, breakup and evaporation. Small drops are nearly spherical; large falling drops flatten rather than becoming cartoon teardrops.
Stage 14: Mixed-Phase Clouds Contain Supercooled Liquid and Ice Together
Liquid water can remain unfrozen below 0°C. Mixed-phase clouds contain both supercooled droplets and ice crystals. Their phase partition is a major microphysical uncertainty in climate modelling.
Stage 15: Freezing Does Not Automatically Occur at 0°C
Tiny pure droplets can remain liquid far below 0°C. Ice formation can require sufficiently low temperature or suitable ice-nucleating particles.
Stage 16: Ice-Nucleating Particles Are Rare but Important
Selected mineral dust and biological particles can promote heterogeneous ice nucleation. Their ability depends strongly on composition and temperature.
Stage 17: The Bergeron–Findeisen Process Exploits Different Vapour Pressures
At subfreezing temperatures, equilibrium vapour pressure over ice is lower than over liquid water. Vapour can therefore deposit preferentially on ice while nearby droplets evaporate. Ice crystals grow at the expense of supercooled liquid.
Stage 18: Ice Crystal Shape Depends on Temperature and Supersaturation
Snow crystals can grow as plates, columns or dendrites. The habit depends strongly on growth conditions. One chemical compound, H₂O, produces many structures because growth kinetics change.
Stage 19: Aggregation Builds Snowflakes
Ice crystals collide and stick. Many snowflakes are aggregates of multiple crystals rather than one giant perfect crystal.
Stage 20: Riming Builds Graupel
An ice particle falls through supercooled droplets. Droplets freeze onto it. Heavy riming can produce graupel, which is physically different from a pristine snow crystal.
Stage 21: Hail Requires Strong Convective Cycling
In powerful thunderstorms, strong updrafts can keep ice particles aloft while they collide with supercooled droplets. Repeated movement through regions of different water content and temperature can produce layered hailstones.
Stage 22: Cloud Electrification Is Connected to Ice Collisions
Collisions among graupel, ice crystals and supercooled droplets can transfer charge. Updrafts separate particles by size and charge until strong electric fields can develop. Lightning is therefore partly an ice-microphysics problem.
Stage 23: Precipitation Can Evaporate Before Reaching the Ground
Rain or snow falling through dry air can evaporate or sublimate. Radar can detect precipitation aloft while the surface remains dry. This is virga.
Stage 24: Falling Hydrometeors Have Terminal Speeds
A falling particle accelerates until drag balances effective weight. Terminal speed depends on size, shape and density, so rain, snow and hail fall differently.
Stage 25: Radar Reflectivity Depends Strongly on Particle Size
Weather radar sends electromagnetic waves and measures returned energy. In the Rayleigh regime, large particles contribute disproportionately.
radar reflectivity ≠ direct rainfall rate without assumptions
Stage 26: Dual-Polarisation Radar Adds Shape Information
Horizontally and vertically polarised signals respond differently to non-spherical particles, helping distinguish rain, snow, hail and mixed precipitation.
Stage 27: Singapore Uses Radar to See Storm Structure
Meteorological Service Singapore uses Doppler weather radar to monitor precipitation, storm movement, vertical structure and wind shear. Radar provides an operational bridge from microphysics to local severe-weather monitoring.
Stage 28: Satellites Measure Different Cloud Layers
Passive imagers infer optical thickness, effective droplet radius and cloud-top temperature. Active radar and lidar probe vertical structure. No single satellite instrument measures the entire cloud state.
Stage 29: EarthCARE Combines Radar, Lidar, Imager and Radiation
ESA’s EarthCARE mission observes liquid and ice cloud profiles, aerosols, vertical motion and radiative effects, connecting microphysics directly to climate radiation.
Stage 30: NASA’s CloudCube Pushes Toward Multifrequency Microphysics
On 10 June 2026, NASA described CloudCube, a compact radar concept spanning multiple frequencies. Different wavelengths respond differently to droplets and ice, improving constraints on particle size and phase.
Stage 31: Aircraft Cloud Measurements Are Invaluable but Difficult
Aircraft carry droplet probes, imaging probes and ice-particle instruments. But particles can shatter on inlets, be undersampled or be misclassified. In-situ measurement can disturb the object being measured.
Stage 32: Bulk Microphysics Schemes Compress Particle Distributions
Weather models cannot track every droplet. Bulk schemes predict selected moments of particle populations, such as mass and number. A one-moment and a two-moment scheme therefore preserve different amounts of information.
Stage 33: Spectral-Bin Microphysics Resolves More of the Size Distribution
Bin schemes divide hydrometeors into many size categories and represent microphysical evolution in greater detail, at much higher computational cost.
Stage 34: Cloud Microphysics Is a Major Forecast Uncertainty
Small errors in droplet number, ice fraction or rain conversion can influence storm lifetime, latent heating and radiation. A microscopic assumption can alter a kilometre-scale weather system.
Stage 35: Cloud Seeding Is Not a Rain-Making Switch
Cloud seeding attempts to alter microphysical pathways in suitable existing clouds. The World Meteorological Organization’s 2025 statement emphasises careful research and evaluation. Seeding cannot create a cloud from clear dry air.
Stage 36: Professional Cloud Microphysics Is Population Dynamics
What droplet/ice size distribution exists, which growth process is dominant, and which observation can distinguish condensation, collision, riming or aggregation?
Evidence: How Do We Know Aerosols Change Cloud Droplets?
Evidence comes from parcel theory, aircraft measurements, ship-track observations, satellite retrievals, cloud chambers and model experiments. More CCN often correspond to more numerous, smaller droplets under comparable cloud-water and updraft conditions.
Misconceptions Worth Hunting
- Cloud droplets simply fall and become rain.
- Condensation alone makes raindrops.
- Water freezes at exactly 0°C in clouds.
- Every aerosol particle becomes a cloud droplet.
- More aerosol always means more rain.
- Radar directly measures rainfall rate.
- Cloud seeding creates rain from clear sky.
- A snowflake is always one perfect ice crystal.
Transfer Check
Take two clouds with identical liquid-water content. Cloud A has fewer CCN; Cloud B has many more. Which likely begins with larger average droplets? Cloud A. Which may initiate collision–coalescence sooner? Often Cloud A, all else equal.
Cool both below 0°C. Must every droplet freeze? No.
Radar shows strong echoes aloft but no rain-gauge signal below. Evaporation below cloud can explain the mismatch.
How We Know the Learning Has Held
A learner should be able to explain CCN; Köhler activation qualitatively; updraft and supersaturation; warm-rain collision–coalescence; supercooled water; ice nucleation; Bergeron–Findeisen growth; aggregation, riming, graupel and hail; radar particle-size sensitivity; radar versus satellite versus aircraft measurements; bulk versus bin microphysics; and why aerosol–cloud effects remain uncertain.
Model Limits
Köhler theory idealises droplet chemistry. Bulk schemes compress size distributions. Radar retrievals assume particle shape and scattering models. Satellite effective radius is an optical weighted property. Aircraft probes sample tiny volumes. Professional cloud microphysics keeps aerosol + supersaturation + phase + particle-size distribution + vertical motion + measurement method visible.
Teaching Guide
Teach in this order: vapour → CCN → supersaturation → activation → droplet growth → warm-rain collisions → supercooled water → ice nucleation → Bergeron growth → snow/graupel/hail → radar/satellite → parameterisation.
Begin with: “Why can a dark cloud contain enormous amounts of water and still produce no rain?”
At advanced level, compare aerosol concentration, cloud-droplet number and radar reflectivity. Ask which describes the seed population, activated droplets and large falling particles.
Connect This to the eduKate Learning Estate
Research Foundations and Further Learning
- ESA EarthCARE mission.
- NASA Global Precipitation Measurement mission.
- WMO Statement on Weather Modification, 14 June 2025.
- NASA CloudCube, 10 June 2026.
- 2025 mixed-phase-cloud and aerosol–cloud-interaction literature.
The Quiet Ending
The beginner asks, “How does a cloud make rain?” The developing atmospheric scientist asks, “Which particles activated, and how did they grow?” The advanced learner asks, “Was warm-rain collision, ice growth, riming or aggregation dominant?”
Which evolving particle-size distribution and phase pathway explains the observed precipitation, and which instrument actually constrains that process?
eduKate Sengkang 