What Is Atmospheric Science?

Atmospheric science is the study of Earth's gaseous envelope: its composition, vertical structure, motion, energy budget, and chemical transformations, together with the methods used to observe and predict them. It covers the atmosphere across every timescale, from the seconds of a turbulent eddy to the centuries of a changing climate, and it is conventionally divided into meteorology, which treats weather and short-term motion; climatology, which treats long-term statistics and variability; atmospheric chemistry, which treats trace constituents and their reactions; and aeronomy, which treats the ionized upper layers. The discipline draws its theoretical base from fluid dynamics, thermodynamics, radiative transfer, and physical chemistry, and its practical base from instrumentation and large-scale computing.

The atmosphere is stratified by temperature into troposphere, stratosphere, mesosphere, and thermosphere, with most weather confined to the lowest ten to fifteen kilometers. Roughly 99 percent of the mass is nitrogen and oxygen, but the trace constituents, water vapor, carbon dioxide, ozone, methane, and aerosols, control the radiative and chemical behavior that atmospheric science is largely about.

Atmospheric Dynamics and Numerical Weather Prediction

Dynamics treats the atmosphere as a thin, rotating, stratified fluid. Its governing equations are the Navier-Stokes momentum equations in a rotating frame, mass continuity, the ideal gas law, and the first law of thermodynamics, together known as the primitive equations after the hydrostatic and shallow-atmosphere approximations are applied. Because those equations have no analytic solution for realistic conditions, forecasting is done numerically: the atmosphere is discretized onto a grid or spectral basis and integrated forward from an initial state. That initial state comes from data assimilation, which blends a short prior forecast with new observations using variational or ensemble Kalman methods. Reference material from NOAA on numerical weather prediction describes the operational model suite, and a textbook treatment of numerical weather prediction basics covers the numerical methods and error growth involved. Sensitivity to initial conditions, established by Edward Lorenz in 1963, sets a practical predictability horizon of roughly two weeks and is the reason forecasts are issued as probabilistic ensembles.

Radiation, Chemistry, and Aerosols

The energy budget determines the climate state: incoming solar shortwave radiation is partly reflected by clouds, aerosols, and surface albedo, and the remainder is absorbed and re-emitted as longwave radiation that greenhouse gases intercept. Radiative transfer calculations resolve this across thousands of spectral bands. Atmospheric chemistry addresses the reactions that set trace gas concentrations, including stratospheric ozone depletion by halogenated compounds, tropospheric photochemical smog produced from nitrogen oxides and volatile organics, and the oxidation chemistry governed by the hydroxyl radical. Aerosols link the two branches, scattering and absorbing radiation directly while also acting as cloud condensation and ice nuclei, which makes aerosol-cloud interaction one of the largest remaining uncertainties in climate projection.

Climatology and Variability

Climatology studies the statistical description of atmospheric states and how those statistics change. Its subjects include the general circulation, the Hadley, Ferrel, and polar cells that transport heat poleward, and modes of internal variability such as the El Nino Southern Oscillation, the North Atlantic Oscillation, and the Madden-Julian Oscillation. Reanalysis datasets, produced by running a fixed assimilation system over decades of archived observations, provide the gridded record on which most of this work depends, and general circulation models developed at centers including NCAR provide the projections. The use of a community climate model as a weather prediction model illustrates how closely the weather and climate modeling traditions have converged. Machine learning emulators trained on reanalysis now compete with physical models at medium range, and a study of inductive biases in deep learning weather models examines what physical structure those systems retain.

Applications

Atmospheric science has applications in a range of fields, including:

• Operational weather forecasting and severe storm warning
• Aviation, marine, and space launch operations planning
• Air quality regulation and pollution exposure assessment
• Renewable energy resource assessment and grid forecasting
• Agriculture, hydrology, and water resource management
• Satellite remote sensing and radio propagation modeling