Lange, Anne Caroline ORCID: 0000-0001-8027-5933 (2018). Observability of Sudden Aerosol Injections by Ensemble-Based Four-Dimensional Assimilation of Remote Sensing Data. PhD thesis, Universität zu Köln.
|
PDF
AC_Lange_Dissertation_2018.pdf - Published Version Download (47MB) | Preview |
Abstract
For sudden aerosol injections, uncertainties of emission source parameters impose the characterizing impediment for skillful numerical simulations. Large amounts of accidentally emitted aerosols can infer serious impacts on health, climate, environment, and economy. This highlights the societal need for reliable forecasts of released particles. Spatiotemporal assimilation techniques combine atmospheric dynamics as knowledge provided by the model with observations and induce constraints with potentially advantageous effects on the simulations. Ensemble-based analyses provide valuable information about the skill of forecast results. However, predictions remain uncertain in regions, where observational information is restricted. Observability investigates the impact of utilized observations, thus focusing on observation network optimization and information quantity specification. Taking volcanic eruptions as prototype for sudden aerosol injections, the research described in this thesis develops new methodologies to assess the impact of observations on the analysis. The emphasis is placed on assimilation-based analyses applying initial value and emission factor optimization for volcanic ash dispersion predictions of the Eyjafjallajökull eruption in April 2010. As observational input, two satellite-borne remote sensing principles are exploited: SEVIRI volcanic ash column mass loadings and CALIOP particle extinction coefficient profiles. For the assimilation within EURAD-IM, appropriate observation operators and their adjoints are constructed. The theoretical principles of observability in case of volcanic ash column mass loading observations are deduced from the viewpoint of the Kolmogorov-Sinai entropy. Ensemble versions of the 4D-var data assimilation technique and the particle smoother approach are implemented and processed, able to identify regions of high and low uncertainty in the dispersion simulation results. The analyses reveal a considerable constraining impact of SEVIRI retrievals to the ash dispersion, while CALIOP retrievals append information only on a very local scale. The variable degree of reliability is shown as a consequence of cloud cover dependent observability from space for both quasi-continuous SEVIRI data and sparse CALIOP overpasses.
Item Type: | Thesis (PhD thesis) | ||||||||
Translated title: |
|
||||||||
Creators: |
|
||||||||
URN: | urn:nbn:de:hbz:38-91485 | ||||||||
Date: | 14 December 2018 | ||||||||
Language: | English | ||||||||
Faculty: | Faculty of Mathematics and Natural Sciences | ||||||||
Divisions: | Faculty of Mathematics and Natural Sciences > Department of Geosciences > Institute for Geophysics and Meteorology | ||||||||
Subjects: | Natural sciences and mathematics Earth sciences |
||||||||
Uncontrolled Keywords: |
|
||||||||
Date of oral exam: | 2 March 2018 | ||||||||
Referee: |
|
||||||||
Refereed: | Yes | ||||||||
URI: | http://kups.ub.uni-koeln.de/id/eprint/9148 |
Downloads
Downloads per month over past year
Export
Actions (login required)
View Item |