Search

We present a neural network algorithm for spectroscopic retrievals of concentrations of trace gases. Using synthetic data we demonstrate that a neural network is well suited for filtering etalon fringes and provides superior performance to conventional least squares minimization techniques. This novel method can improve the accuracy of atmospheric retrievals and minimize biases.

Trace gases and their isotopic ratios in planetary atmospheres offer important but subtle clues as to the origins of a planet's atmosphere, hydrology, geology, and potential for biology. An orbiting laser remote sensing instrument capable of measuring trace gases on global scales with unprecedented accuracy and high spatial resolution can provide valuable measurements for planetary and earth science missions.

Frequency modulation (FM) spectroscopy was first demonstrated by Bjorklund1 in the visible region using a ring dye laser. The basic scheme of FM spectroscopy calls for phase modulation of the laser beam either by an external electrooptic modulator or in the case of diode lasers by modulating directly the injection current. The requirement of modulating the laser at high frequencies (several gigahertz), so that the spectral feature of interest can be probed by a single Fourier component of the laser power spectrum, introduces the need for high-speed detectors which are not always available commercially and have small and damage-sensitive surfaces.