More than 99% of energy emitted by the sun to the Earth's atmosphere and surface e in the wavelength region of less than 4 microns, called shortwave radiation or solar radiation. The temperature of Earth's atmosphere and the ground is generally within the range of 180 K-330 K, and thus more than 99% of the thermal radiation energy emitted by Earth is in the region of wavelengths longer than 4 microns, called long-wave radiation. The radiation are absorbed, scattered end emitted when it travel through and interacts with the atmosphere and ground in both shortwave and long-wave spectral region. The imbalances in radiant charge resulting in temperature change in the ground and the atmosphere which varies a lot in different places, forming the heat and cold sources, promoting movement and changes of the atmosphere. In the few decades, the research and application of atmospheric remote sensing technology promote the development of various radiative transfer models. These models is widely used in the climate study, atmospheric correction and other fields. In practice, the radiative transfer models adopt various methods to simplify the computation. That’s the reason that we divide the models into two categories: the mid- and low- resolution radiative transfer models and the high-resolution radiative transfer models.
The radiation detected by a satellite instrument has passed from the source, perhaps the sun, to the earth’s surface or a cloud top, through the atmosphere. Usually, there are three necessary steps in the radiative transfer simulation: first, computing the absorption coefficients using the spectrum data; then calculating the scattering parameters of clouds, aerosols and molecules; at last, the radiative flux of the whole atmosphere are derived by adding the contributions of sublayers. The 6S model simply the first step by assuming the absorption lines obey the random exponent distribution; MODTRAN and 1DMWRTM use two-flow approximation in the simulation of multi-scattering, and RT3 adopt double-adding method to resolve the multi-scattering problem. Regression model and Pre-computed LUT model are also called fast model, such as RTTOV (regression).
Most atmospheric constituents absorb and re-emit infrared photons, but the strongest spectral features are due to rotational-vibrational transitions allowed by the quantum mechanical selection rules. The emission spectrum of the Earth’s atmosphere contains many spectral features due to the different molecules that absorb and emit radiation; furthermore, due to the highly variable vertical temperature profile in the atmosphere, the observed spectrum is a superposition of all these processes at different altitudes, so that radiative transfer models are required to calculate infrared atmospheric spectra. The molecular line strengths (due to temperature-dependent population of the molecular energy levels), and this leads to vertical information than can be extracted form atmospheric spectra in the thermal infrared – even when using nadir geometry from space. The pressure-dependences of the molecular line-widths provide additional vertical information, although a high spectral resolution is required to observe this effect. LBLRTM (Line-By-Line Radiative Transfer Model) is an accurate line-by-line model that is efficient and highly flexible with the Voigt line shape used at all atmospheric levels.
