What does this equation mean?

Atmospheric composition is inferred the same way interiors are — from a signal, not from the substance itself — but the physics is radiative transfer rather than mechanics. The workhorse technique is transmission spectroscopy : as a planet with an atmosphere passes in front of its star (for an exoplanet) or as sunlight grazes a planet’s limb (for a solar-system body observed from orbit), starlight is filtered through the atmosphere’s outer layers before reaching the detector. Each gas absorbs a characteristic set of wavelengths, governed to first approximation by the Beer–Lambert relation, $I(\lambda) = I_0(\lambda)\, e^{-\tau(\lambda)}, \qquad \tau(\lambda) = \int n(z)\,\sigma(\lambda, z)\, dz$. where $I_0(\lambda)$ is the incident spectrum, $\tau(\lambda)$ the wavelength-dependent optical depth, n(z) the number density of an absorbing species at altitude z , and $\sigma(\lambda, z)$ its absorption cross-section, which itself depends on temperature and pressure. The instrument records I($\lambda$) ; the abundance profile n(z) is recovered by inverting this relation against a model atmosphere, which means every composition result carries the assumptions of that model — cloud and haze opacity, temperature structure, and the line-list data used for $\sigma(\lambda,z)$ — bundled in with the raw spectral measurement.