Infrared (IR) detectors can be used as airborne limb-viewing surveillance systems for missile detection. These systems'
performances are impacted by the atmospheric inhomogeneous background. In fact, the probability of target detection
can be heavily affected. Consequently, the knowledge of these radiance small-scale fluctuations and their statistical
properties is required to assess these systems' detection capability. A model of two-dimensional radiance spatial
fluctuations autocorrelation function (ACF) is developed. This model is dedicated to airborne limb-viewing conditions in
the thermal IR. In the stratosphere and in clear-sky conditions, the structured background is mainly due to
internal-gravity-wave-induced temperature and density spatial fluctuations. Moreover, in the particular case of water vapour
absorption bands, the mass fraction fluctuations play a non negligible role on the radiative field. Thereby, considering
the temperature field and the water vapour field as stochastic processes, the radiance ACF can be expressed as a function
of the temperature ACF and the water vapor mass fraction ACF. A local thermodynamic equilibrium model is sufficient
for stratospheric conditions and sunlight scattering is neglected in the thermal IR. In addition, determination of the
radiance fluctuations ACF requires the knowledge of the absorption coefficient and its first derivatives with respect to
the temperature and water vapour mass fraction. Thus, a line-by-line model specific to water vapor absorption bands has
been developed. This model is used to precalculate the absorption coefficients and their derivatives. This look-up table
method allows circumventing the computational cost of a line-by-line calculation. A detailed description of the radiance
fluctuations ACF model is presented and first results are discussed.