How Meteorologists Measure Weather

By WeatherRecall editorial project. Date reviewed: 2026-10-06. Last updated: 2026-10-06.

Meteorologists observe an atmosphere that changes in three dimensions and through time. No single instrument measures all of it. Surface equipment describes local conditions, balloons sample vertical profiles, radar follows precipitation structures, and satellites observe radiation across wide areas. These methods overlap but do not produce identical quantities. Understanding their differences makes weather charts and historical summaries easier to interpret.

Surface measurements need a defined exposure

Air temperature is measured in a ventilated shelter rather than by leaving a sensor in direct sunlight. Humidity describes moisture relative to temperature-dependent saturation; relative humidity can therefore change when temperature changes even without adding water vapour. Pressure instruments measure atmospheric pressure at their elevation. Wind readings need both speed and direction, along with height and surrounding obstacles.

Precipitation gauges accumulate water at a small collecting surface. Their readings should distinguish liquid-equivalent precipitation from snowfall depth. Snow depth also depends on settling and melting after snow falls. These are separate variables, as the NOAA daily station overview makes clear in its list of temperature, precipitation, snowfall, and snow-depth elements.

Station pressure and mean-sea-level pressure are not interchangeable. Sea-level reduction estimates what pressure would be at a common reference elevation using assumptions about the atmosphere. This supports spatial comparison, but it is a derived value rather than a barometer physically located at sea level beneath every station.

Sampling the atmosphere above the ground

A radiosonde carried by a balloon measures conditions as it ascends. Temperature, moisture, pressure, and position information describe a vertical profile, while the balloon's movement provides wind information. The package drifts horizontally, so the profile is not a perfectly vertical column over its launch point.

Aircraft observations can add temperature and wind information along flight routes. Their coverage depends on where and when aircraft fly. Ocean buoys and ships extend monitoring over water but also have differing exposures and reporting practices. Combining these networks requires source-specific quality checks and accurate timing.

A profile helps identify layers that a surface thermometer cannot reveal, such as a temperature inversion. It remains a sample through a moving atmosphere. Conditions between launches or away from the balloon's path require additional observations or estimation.

Radar and satellites measure signals before weather quantities

Weather radar sends electromagnetic pulses and measures returned signals. Those returns provide information about precipitation particles; Doppler measurements reveal motion along the radar beam. Estimating rainfall at the ground requires assumptions and processing, because the radar samples a volume aloft rather than collecting water at the surface.

Beam height increases with distance, terrain can block coverage, and particle type affects the relationship between radar signal and rainfall. Radar and gauge totals may therefore differ without either being meaningless. Their comparison needs matching time intervals and an understanding of whether a point gauge represents the surrounding radar area.

Satellites measure radiation in selected wavelength bands. Cloud imagery, temperature sounding, and moisture retrievals depend on how radiation interacts with the atmosphere and surface. A satellite image's apparent surface temperature is not automatically shaded two-metre air temperature. Remote sensing broadens coverage, but retrieved variables retain algorithm and sampling limitations.

Distinguish measured, derived, and modelled variables

Dew point can be calculated from temperature and humidity; apparent temperature combines several influences on thermal sensation. Neither should be silently substituted for ordinary air temperature. The Open-Meteo historical variable definitions identify heights, units, pressure types, and apparent-temperature calculations, showing why variable names matter.

ERA5 reanalysis combines observing information with a numerical model to reconstruct conditions. A model's two-metre temperature estimates a specified atmospheric quantity at grid scale; it is not another physical thermometer observation. Reports should identify this distinction even when the displayed units are identical.

Illustrative example: directions need circular arithmetic

Illustrative measurements only; these are invented readings, not real observed data. Imagine two equal-speed winds of 10 m/s, one from 350° and one from 10°. A simple arithmetic average of direction gives (350 + 10) ÷ 2 = 180°, falsely suggesting a southerly wind.

Both directions are near north. Averaging their east-west and north-south components makes the opposing sideways components cancel. Their resulting vector points from north, with magnitude 10 × cos(10°), approximately 9.85 m/s. The scalar average of the two speeds remains 10 m/s. Vector-mean speed and mean speed are different statistics.

For wind from opposite directions at equal strength, the mean vector can be near zero even though neither observation was calm. A direction assigned to that near-zero vector is unstable and should not be interpreted as a reliable prevailing wind. Check the averaging method before using a daily direction summary for an exposure decision.

Make comparisons physically consistent

Illustrative pressure comparison, not real observations: suppose a hill station reports surface pressure of 950 hPa while a coastal station reports 1010 hPa. That difference alone does not prove a strong weather-system pressure gradient; elevation contributes to it. Compare compatible sea-level-reduced fields or examine the elevation and reduction method first.

State sensor height, site elevation, and exposure where known.
Match instantaneous values with instantaneous values and averages with averages.
Retain units, quality flags, and observation timestamps.
Identify retrieved or derived quantities rather than labelling everything measured.

WMO's climate overview describes international standards that support comparable, quality-controlled observations. Consistency supports long-term analysis; it does not remove all local exposure differences.

Use the date tool for a historical starting point, the comparison tool for context, and methodology to check what the displayed variables actually represent.

References

About the author

WeatherRecall editorial project compiles educational summaries from listed public documentation. It is not a credentialed meteorologist and has not performed original field research, balloon launches, or instrument tests for this article. The hypothetical calculations explain measurement interpretation, not actual atmospheric conditions.