Download files
Abstract
Changes to the present atmosphere and their effects on natural and human systems are challenges requiring mitigation strategies. These are contributed by anthropogenic emission of certain gases of which carbon dioxide (CO2) is the most significant. Our incomplete understanding of how the carbon is sequestered adds uncertainty in future climate prediction. A number of techniques are used to measure the CO2 content of the atmosphere and its spatial distribution. Satellites promise excellent spatial coverage but it is extremely challenging to achieve high accuracy. The in-situ methods are accurate but are sensitive to local distortions. Ground-based techniques are accurate but the network is too sparse. Accuracy of 0.1% is being achieved from measurement of CO2 lines in the spectrum of the Sun as used in the TCCON stations. These stations use Bruker Fourier Transform Spectrometers (FTS) and are relatively expensive and complex.
In this project we use an instrument that makes similar measurements to TCCON, in a compact, low-cost and low-power package that could potentially be deployed in remote locations. The instrument is developed from commercially available components using a Micron Optics fibre Fabry-Perot tunable filter, for simultaneous measurement of solar path CO2 absorption at 1.58 microns and O2 absorption at 1.27 microns. The principal innovation is to use fibre-optic devices developed for telecommunications, as miniature spectrometers. The O2 measurements are utilised as a proxy to infer the dry air number of molecules so that the CO2 columns concentrations can be corrected for common scatter via ratioing of the columns.
I have developed atmospheric models for the forward modelling and inversion of relevant physical quantities from the observations. As a test of the forward model I have used it to reproduce spectra observed with the TCCON spectrometer at the University of Wollongong (UoW), the agreement is excellent. This thesis is based on Fabry-Perot and FTS measurements recorded in a side-by-side monitoring at the TCCON observatory in the UoW. The Fabry-Perot measurements analysed correlate reasonably with the FTS measurements analysed by the TCCON benchmark. Our Fabry-Perot error budget analysis indicates that the a priori model is the most significant error-source.