Use this URL to cite or link to this record in EThOS: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.728811
Title: Retrieval of trace gases using IASI
Author: Vincent, Robert Anthony
ISNI:       0000 0004 6496 4642
Awarding Body: University of Oxford
Current Institution: University of Oxford
Date of Award: 2016
Availability of Full Text:
Access from EThOS:
Full text unavailable from EThOS. Restricted access.
Access from Institution:
Abstract:
Trace gases comprise less than 1% of Earth's atmosphere, yet they dominate the radiation budget and drive atmospheric chemistry. The Infrared Atmospheric Sounding Interferometer (IASI) is a set of two nadir viewing Fourier transform spectrometers aboard the MetOp-A and B satellites in polar orbit that measure spectral radiances in the thermal infrared and, therefore, observe the spectroscopic signatures of numerous trace gases. This thesis focuses on the practicalities of fully utilizing the vast amounts of data recorded via satellite with limited computational resources and prior knowledge of the target gas. First, a method for selecting a vertical retrieval grid is presented that minimizes the amount of prior knowledge appearing in the estimate due to applied constraints. This selection method uses the information content of the retrieval to optimally determine the number and spacings of vertical levels, while accounting for correlations between the various layers. Failing to select an appropriate vertical grid for water vapour and temperature retrievals, such as using levels equally spaced in pressure, was found to reduce the information content by as much as 30% on average. Overwhelmingly, the slowest part of an iterative retrieval is evaluating the forward model. As a result, the computation may fail to keep pace with data acquisition. Procedures to parametrise the Reference Forward Model (RFM) by creating a set of pre-tabulated look-up tables (LUTs) of absorption cross-sections for individual gases were developed. Since such LUTs can be unruly in size, a simple compression scheme based on linear interpolation was analysed that reduced the total LUT size to just 4.4% of the original. Additionally, the fine spectral grid for radiative transfer was also reduced using similar methods. Model comparisons to a global atmospheric ensemble showed negligible increases in error compared to IASI instrument noise and a factor of 130 increase in computational speed. These LUTs are openly available for use by the scientific community, whether using the RFM or to be incorporated into any forward model. Finally, a linear retrieval scheme was developed to estimate total column amounts of carbonyl sulphide (OCS) at a rate roughly 104 times faster than a typical iterative retrieval. This scheme incorporates two concepts not utilised in previously published linear estimates. First, all physical parameters affecting the signal are included in the state vector and accounted for jointly, rather than treated as effective noise. Second, the initialisation point is determined from an ensemble of atmospheres based on comparing the model spectra to the observations, thus improving the linearity of the problem. The entirety of IASI data from 2014 was analysed and showed spatial features of OCS total columns including depletions over tropical rainforests, seasonal enhancements over the oceans, and distinct OCS features over land. Error due to assuming linearity was found to be on the order of 11% globally for OCS. Comparisons to surface VMR in situ samples taken by NOAA show seasonal correlations greater than 0.7 for five out of seven sites across the globe. Furthermore, this linear scheme was applied to OCS, but may also be used as a rapid estimator of any detectable trace gas using IASI or similar nadir-viewing instruments.
Supervisor: Dudhia, Anu Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.728811  DOI: Not available
Share: