Error analysis for mesospheric temperature profiling by absorptive occultation sensors

An error analysis for mesospheric profiles retrieved from absorptive occult ation data has been performed, starting with realistic error assumptions as would apply to intensity data collected by available high-precision UV pho todiode sensors. Propagation of statistical errors was investigated through the complete retrieval chain from measured intensity profiles to atmospher ic density, pressure, and temperature profiles. We assumed unbiased errors as the occultation method is essentially self-calibrating and straight-line propagation of occulted signals as we focus on heights of 50-100 km, where refractive bending of the sensed radiation is negligible. Throughout the a nalysis the errors were characterized at each retrieval step by their mean profile, their covariance matrix and their probability density function (pd f). This furnishes, compared to a variance-only estimation, a much improved insight into the error propagation mechanism. We applied the procedure to a baseline analysis of the performance of a recently proposed solar UV occu ltation sensor (SMAS - Sun Monitor and Atmospheric Sounder) and provide, us ing a reasonable exponential atmospheric model as background, results on er ror standard deviations and error correlation functions of density, pressur e, and temperature profiles. Two different sensor photodiode assumptions ar e discussed, respectively, diamond diodes (DD) with 0.03% and silicon diode s (SD) with 0.1% (unattenuated intensity) measurement noise at 10 Hz sampli ng rate. A factor-of-2 margin was applied to these noise values in order to roughly account for unmodeled cross section uncertainties. Within the enti re height domain (50-100 km) we find temperature to be retrieved to better than 0.3 K (DD) / 1 K (SD) accuracy, respectively, at 2 km height resolutio n. The results indicate that absorptive occultations acquired by a SMAS-typ e sensor could provide mesospheric profiles of fundamental variables such a s temperature with unprecedented accuracy and vertical-resolution. A major part of the error analysis also applies to refractive (e.g., Global Navigat ion Satellite System based) occultations as well as to any temperature prof ile retrieval based on air density or major species density measurements (e .g., from Rayleigh lidar or falling sphere techniques).